HARVARD UNIVERSITY Library of the Museum of Comparative Zoology H E GREAT BASIN MURALIST VOLUME 54 NO 1 — JANUARY 1994 BRIGHAM YOUNG UNIVERSITY GREAT BASIN NATURALIST Editor Richard W. Baumann 290MLBM PO Box 20200 Brighaiii Young University Provo, UT 84602-0200 801-378-5053 FAX 801-378-3733 Assistant Editor Nathan M. Smith 190 MLBM PO Box 26879 Brigham Young University Provo, UT 84602-6879 801-378-6688 E-mail: NMS@HBLL1.BYU.EDU Associate Editors Michael A. Bovvers Blandy Experimental Farm, University of Virginia, Box 175, Boyce, VA 22620 J. R. Callahan Museum of Southwestern Biology, University of New Mexico, Albuquerque, NM Mailing address: Box 3140, Hemet, CA 92546 Jeffrey J. Johansen Department of Biology, John Carroll University University Heights, OH 44118 Boris C. Kondratieff Department of Entomology, Colorado State University, Fort Collins, CO 80523 Paul C. Marsh Center for Environmental Studies, Arizona State University, Tempe, AZ 85287 Stanley D. Smith Department of Biology University of Nevada-Las Vegas Las Vegas, NV 89154-4004 Paul T. Tueller Department of Environmental Resource Sciences University of Nevada-Reno, 1000 Valley Road Reno, NV 89512 Robert C. Whitmore Division of Forestry, Box 6125, West Virginia University, Morgantown, WV 26506-6125 Editorial Board. Jerran T. Flinders, Chairman, Botany and Range Science; Duke S. Rogers, Zoology; William Hess, Botany and Range Science; H. Duane Smith, Zoology. All are at Brigham Young University. Ex Officio Editorial Board members include Steven L. Taylor, College of Biology and Agriculture; Stanley L. Welsh, Director, Monte L. Bean Life Science Museum; Richaid W Baumann, Editor, Great Basin Naturalist. The Great Basin Naturalist, founded in 1939, is published quarterly by Brigham Young University. Unpublished manuscripts that further our biological understanding of the Great Basin and surrounding areas in western North America are accepted for publication. Subscriptions. Annual subscriptions to the Great Basin Naturalist for 1994 are $25 for individual sub- scribers ($30 outside the United States) and $50 for institutions. The price of single issues is $12. All back issues are in print and available for sale. All matters pertaining to subscriptions, back issues, or other busi- ness should be directed to the Editor, Great Basin Naturalist, 290 MLBM, PO Box 20200, Brigham Young University, Provo, UT 84602-0200. Scholarly Exchanges. Libraries or other organizations interested in obtaining the Great Basin Naturalist through a continuing exchange of scholarly publications should contact the Exchange Librarian, 6385 HBLL, PO Box 26889, Brigham Young University, Provo, UT 84602-6889. Editorial Production Staff JoAnne Abel Technical Editor Jan Spencer Assistant to the Editor Copyright © 1994 by Brigham Young University Official pubhcation date: 25 February 1994 ISSN 0017-3614 2-94 750 8744 The Great Basin Naturalist Pl BiJsiiKi) AT Phono, Utah, by Brk.iiam Young University ISSN 0017-3614 Volume 54 31 January 1994 No. 1 Great Basin Naturalist 54(1). © 1994, pp. 1-xx BIRDS OF NORTHERN BLACK MESA, NAVAJO COUNTX ARIZONA Charles T. LaRue^ You should see the sical isolation of this region has made detailed and e.xtensive biological sui-veys difficult to conduct; in some areas no such surveys have been done. Infor- mation on bird life is particularly scant. Phillips et. al (1964) refer to a "severe lack of data" for the birds of northeastern Arizona. Maps of the most recent annotated checklist of Arizona birds (Monson and Phillips 1981) show a large blank spot centered on Black Mesa. The casual traveler driving High\va\' 160 through Kayenta or visiting Monument Valley might notice the paucity of life in the area but would be unaware of the diverse environment behind the nearby scaip of Black Mesa. Hidden in the rugged complex of canyons and \alleys along the highest northern portions of Black Mesa is a diverse assemblage of vegetative 'Bin 1S1I2. k.iv.iit.i. Arizona 86033. Cheat Basin Natuhaijst [\blunie 54 ? =) .an O O r— O J3 > o o Chinle Ganado SAN FRANCISCO^ MT. \ Flagstaff Winslow 5 o m -0 > C N t m >< o o Fi^. 1. Northern Arizona and the Black Mesa communities. These communities, some of which are typical of the liigher mountains of the Colorado Plateau, support a surprising; numl)er of bird species. In a single 26()-ha (640-ac) block near Lolomai Point, 45 species are known to breed. If one searched further in this same block or doubled its size, an addi- tional 21 species coidd probably be found breeding. A total of 241 bird species (includ- ing those in the archaeological record) are known from northern Black Mesa. Another 40 species (all migrants) would be included in this list if the boimdar\' of the stiuK' area cov- ered in this report were to be moved just a few kilometers north to Laguna Creek at Kaventa (Jacobs 19(S6, personal records). The purpose of this report is to present what is currently known of the birds of north- ern Black Mesa and to discuss some ecologi- cal factors affecting them. This includes reporting the status of all bird species record- ed on northern Black Mesa in the late 1970s and the 1980s, reporting the avian assem- blages associated with each habitat found in this area, discussing possible causes and effects of environmental changes that impact 1994] Birds of Northern Black Mesa, Arizona bird coninuinities of the area, and providing a sonnd base for future work on the l)ird hfe of the mesa. Methods A review of historical accounts and other pubhcations was conducted to determine the extent of previous explorations and studies on northern Black Mesa. Results of this search and discussions with Gale Monson revealed the poor state of knowledge of the area s avi- fauna prior to the 1970s. Variable width transects (Emlen 1971, 1977) were used to derive bird densities in different habitats on the Peabody Coal Com- pany leasehold as part of the federally required baseline wildlife siu^veys. Each tran- sect was usually about 1500 m long and was traversed three times each season. Spot-mapping grids were established in three pinyon-juniper stands to census breed- ing bird densities in 1983 and 1984. Spot-map censuses were also conducted in mixed-shrub habitat from 1984 to 1986 and in three reclaimed mine spoil sites in 1985. These results are reported as number of pairs/40 ha. When presenting population density data from the literature, I have converted all values to number of individuals or number of pairs per 40 ha if this was not done by the authors. Waterfowl and shorebirds were coimted at ponds during migratory periods in 1982 and 1983. Each spring from 1982 until present, a sui-vey for nesting raptors within the Peabody lease was conducted. Nest sites located in any previous year were checked for use, and searches for new nesting sites were conduct- ed. The area north of the Peabody Coal Com- pany lease and the area below the rim of the mesa are unaffected by the Surface Mining Control and Reclamation Act (SMCRA) and, therefore, were not baseline surveyed. I visit- ed these areas on over 250 occasions from December 1981 through 1993, covering over 1770 km (1100 miles) by foot. On these visits I tried to determine the presence and status of those bird species that utilize this area. Many observations from throughout the area are incidental. The faunal resemblance index (see Tiible 12) used to compare the breeding bird com- position of each habitat is that applied by Hoffmeister (1986) to Arizona s mammalian fauna. Because of the variety of techniques used to determine bird densities in the lease area and the lack of density data for habitats outside the lease area, the usual techniques employed to determine similarity values for bird data were not used. These resemblance factors are based on the number of species shared in common between habitats and pro- vide adequate comparisons of the degree of similarity bet\\'een the breeding bird compo- sition of the habitats across the entire study area. Classification, sequence, and common and Latin names follow the American Ornitholo- gists Union (1983, 1985, 1987, 1989). Latin names of birds are presented only in the species accounts. Common and Latin names of plants, in general, follow McDougall (1973). Common names of birds are capitalized fol- lowing the opinion of Potter (1984). Geo- graphic place names and spellings are from U.S. Geological Sune}' 7.5-minute topograph- ic quadrangles. Additional names are my own creation, e.g., "east," "middle," and "west" forks of Coal Mine Wash (Fig. 2). The "upper," "mid- dle," and "lower" portions of washes refer to the upper third, middle third, and lower third, respectively, of those portions of each wash witliin the study area. Other names are those used b\' Peabody Coal Company. Figure 2 presents locality names used in this report. The periods of the month are as follows: early (Ist-lOth), mid (llth-20th), late (21st-end). Plumage terminology follows that of Humphrey and Parkes (1959). Terms of subjective abundance follow Monson and Phillips (1981): abundant — in large numbers; common — always to be seen, but not in large numbers; fairly common — very small num- bers or not always seen; uncommon — seldom seen, but not a surprise; sparse — always a sur- prise, but not out of normal range; casual — out of usual range, to be expected eveiy 20-50 years; and accidental — far from normal range and not to be expected again. The term area refers to the study area of this report. The term region refers to all of Black Mesa, Kayenta Valley, eastern Shonto Plateau, and Tsegi canyons. Lease means the Peabody Coal Company leasehold. Proof or confirmation of breeding consists of the following: nests with young and/or eggs, fledglings with or without adults, and remains of fledglings. Breeding is Ghk.m Basin Natihalist [Volume 54 FiK- 2. Localit\' nanu's of tlu' ,stiid\ area used in text. suspected when territorial or singing adults have heen found ire(]uently in the proper habitat in two or more breeding seasons. The seasons are defined as winter — December through February; spring — March through May; summer — June through August; and lall — September through November. Early, mid, and late seasonal references refer to the first, second, and third months of each season, respectively. A permanent resident is a species that is present in the study area in all four seasons. A summer resident is a species that is present in the study area primariK in summer. Certain summer residents arrive, however, in winter and leave in the fall. Like- wise, some summer residents may be more numerous as fall or spring migrants. Winter residents are species that spend the winter period and/or portions of fall and spring, but do not remain to breed. Migrants are species that pass through the area but do not over- winter or breed. Transient is generally 1994] Birds of Northern Black Mesa, Arizona synonymous with migrant but refers to species or individuals that wander infrequent- ly through the study area. Subspecies are dealt with primarily in ref- erence to species known or suspected to breed in the study area. The only specimens collected on Black Mesa (or nearby) consist of few individuals of few species. These were collected in the 1930s during the Rainbow Bridge-Monimient Valley Expedition (Wood- burA' and Russell 1945). To my knowledge, no recent specimens haxe been collected in the area. Subspecies that breed in the area are detemiined from Monson and Phillips (1981), Behle et al. (1985), Woodbury and Russell (1945), and Behle (1985). In several species the race(s) present on Black Mesa (or north- eastern Arizona) are unknown, questionable, or transitional. Salvage collecting these species in areas to be mined would help resolve these questions. History of Black Mesa Studies Northern Black Mesa remained biological- ly unexplored until the 1930s. Spanish, Mexi- can, and United States penetrations into this area were militaiy operations undertaken with extreme difficulty (McNitt 1962, 1972). Possi- bly the first white man to see northern Black Mesa was Colonel Don Francisco Salazar, who crossed "the almost impassible wilderness of Black Mesa" in August 1823 (McNitt 1972). Captain John C. Walker probably passed directly through the stud)' area of this report when he crossed the "high broken ground of Black Mesa" in September 1859 (McNitt 1972). Lieutenant H. R Kingsbuiy of Troop K, 6th Cavalry, crossed 117 km (73 miles) of "viciously rough wooded terrain" of the "unin- habited wastes of Black Mesa" on 19-20 August 1884 (McNitt 1962). Apparently, no reports dealing with the natural history of Black Mesa were made following these expeditions. The series of United States government surveys that traversed the Little Colorado Ri\'er Vallex' or skirted the southern fringe of Black Mesa in the late 1800s probably never reached northern Black Mesa (Woodbury and Russell 1945). Since no major trading posts were established on northern Black Mesa, sci- entists who frequently visited such outposts also never reached the area (Woodburv and Russell 1945, McNitt 1962). The earliest mention of specific birds observed on Black Mesa is that of Theodore Roosevelt (Roosevelt 1913), when he ascend- ed the mesa after leaving John Wetherills trading post at Kayenta on 17 August 1912. Roosevelt s route undoubtedly took him through Rock Gap at the head of Moenkopi Wash. He wrote: Our first day s march took us up this [the northern escarpment of Black Mesa]. We led the saddle horses and drove the pack animals up a ven' rough Navajo trail which zigzagged to the top through a partial break in the continuous rock wall. . . . On the summit we were once more among pines and we saw again the beautiful wild flowers and birds we had left onBuckskin Mountain [Kaibab Plateau]. ... I saw a Louisiana Tanager [Western Tanager]; the pinyon jays were everywhere; ra\ ens, true birds of the wilderness, croaked hoarse- ly. . . . From the cliff crest we traveled south through a wild and picturesque pass. The table land was rugged and mountainous; but it sloped gradually to the south, and the mountains changed to rounded hills. Extensive coal deposits helped to end the isolation of Black Mesa. Long before modern exploitation, prehistoric peoples utilized this coal for ornaments, firing pottery, and heating. The Hopi made extensive use of coal on the southern edge of Black Mesa (Hack 1942). During at least A.D. 600-1050, the Anasazi used coal on northern Black Mesa (Gummer- man 1984). The first automobile road built over the northern rim of Black Mesa was apparently constructed for the development of coal mines on Yellow Water Wash in the early 1900s (Johnston 1932). It was this road that allowed access by biologists to the high country of the mesa to record the only system- atic bird observations prior to the 1970s. Most of these records, concerning less than 40 species, were from a series of trips by H. N. Russell and A. M. Woodbury', participants in the Rainbow Bridge-Monument V^alley Expe- dition in June and July 1938 and in August 1935 and 1937 (Woodbuiy and Russell 1945). Approximately 24 specimens of 11 species were collected on trips made on 11 and 16 July 1938 and 17 August 1939. Brotherson et al. (1981) studied the bird community compo- sition in Betatakin Canyon, an area similar to the rim region of Black Mesa, at nearby Nava- jo National Monument. Bradfield (1974) reported on the birds of southern Black Mesa in the vicinity of Oraibi. Jacobs (1986) recently 6 Cheat Basin Natur.\list [Volume 54 compiled an annotated list ol birds Irom the Navajo and Hopi resenations that includes ohsenations from and around Black Mesa. Peahody Coal Company studies from 1981 to tlie present and records that form the core of this report were initiated to meet surface mining regulations. Mining activities were undenva\ on Black Mesa by 1971, but it was not until 1977 with the passage of the Federal Surface Mining Control and Reclamation Act (SMCRA) that detailed studies began. SMCRA. requires baseline wildlife censusing as part of the federal mine-peniiitting process. Peabody Coal Company hired Espey, Huston, and Associates to conduct the initial surveys in the western portion of the 25,900-ha (100- sc^-mile) lease in 1979-80. The results were presented in 1981. Peabody Coal Company biologists censused the remainder of the lease area in two units in 1981-82 and 1982-83. Results were presented in 1982 and 1983, respectively. Waterfowl, raptor, and small bird censuses have been conducted from 1982 to the present. I have made several hundred trips north of the lease area into the highest and more remote portions of Black Mesa from 1982 through 1993 recording the birds found there. Therefore, virtually all bird censusing from 1982 to 1993 was conducted by me. Black Mesa Environment Physiography and Geology Black Mesa encompasses a large expanse of uplands on the Navajo and Hopi Indian reser- vations in northeastern Arizona (Fig. 1). The mesa is roughly circular and 88-113 km (55-70 miles) across and covers nearlv 518,000 ha (2000 sq miles). The mesa itself is in the center of a large structural basin (Repenning and Page 1956); as a residt, the interior is lower than the rim on the east, north, and west sides. A network of five large, parallel flowing washes drains Black Mesa to the southwest and empties into the Little Col- orado River. These washes, previously called the Tusayan Washes, are (east to west) Jeddito, Polacca, Oraibi, Dinnebito, and Moenkopi. Pinyon and Forrest Lake are the only relative- ly large communities in the interior of Black Mesa. The larger communities of the region are situated in the valleys surrounding the mesa. Most of Black Mesa consists of low mesas, rolling hills, and shallow valleys. Elevations rise gradualK to the rim, which breaks off in a stair-step escaipment 150-610 m (500-2000 ft) high. Where the large washes exit the mesa, elevations may be near 1680 m (5500 ft), and the rim rises to above 2440 m (8000 ft). Most of the mesa is approximately 1830-2070 m (6000-6800 ft) in elevation. Black Mesa is underlain by a regionally downwarped series of Jurassic, Triassic, and older sedimentary formations. The topograph- ically elevated portion of the mesa itself is composed of a series of sedimentary beds deposited during a period of transgressing and regressing seas in the late Cretaceous (Repenning and Page 1956). The Jurassic Cow Springs and Morrison formations and the Dakota Sandstone are the oldest deposits to outcrop on the mesa flanks at the northern base. The Mancos shale forms broad slopes at the foot of the mesa and imderlies the broad valleys where the major washes leave the mesa. Most of the rocks exposed on Black Mesa are of the Toreva and Wepo formations. The former is composed of cliff-forming pale sandstones that cap small mesas and canyons. The Wepo is a complex of sand, silt, and mud- stones that contain significant coal deposits. The rolling baked-clinker hills of the Black Mesa interior are Wepo. The highest portion of Black Mesa along the impressive northern scarp is capped by the Yale Point Formation. This pale, crossbedded sandstone is dissected into a series of short, deep canyons. Recent alluvium is deposited in virtualK' every valley floor throughout Black Mesa. Aeolian deposits are more extensive on southern Black Mesa than in the northern portions (Coolev et al. 1969). The area covered in this report is the northernmost portion of Black Mesa south of the town of Kaventa (Fig. 1). This area lies between 36°42'30" and 36°22'30"N latitude and 110°10' and 110°30'W longitude. Includ- ed here is the upper 16-24 km (10-15 miles) of Moenkopi and Dinnebito washes and their tributaries. It includes the mesa rim from near Black Mesa Trading Post northeast to Lolomai Point and southeast to the head of Assayii Wash. It covers the mesa foot from upper Long House Valley to upper Owl Spring Valley and the upper basin of Assayii Wash (Fig. 2). The 25,900-ha (64,000-ac) Peabody Coal Com- pany lease is located in the southern portion of the study area. The study area encompasses 1994] Birds of Northern Black Mesa, Arizona approximatel)- 106,200 ha (232,573 ac). A few bird records reported herein come hom out- side this specific study area. Elevations range from 1768 m (5800 ft) at the mesa foot to 2490 m (8168 ft) at the high- est point on the rim. Where Moenkopi Wash leaves the study area, the elevation is about 1829 m (6000 ft). The southern portion is typi- fied by pinyon-juniper woodland-covered hills and shrub-filled valleys. To the northeast the land rises gradually and becomes incised by canyons up to 152 m (500 ft) deep. From the rim the mesa breaks off in a spectacular escarpment of cliffs and slopes to grasslands 610 m (2000 ft) below. Climate The climate of the study site is semiarid and, as is characteristic of the region, typified bv dailv and seasonal temperature extremes. Temperatures of -34°C (-30°F) and 40°C (104°F) have been recorded on southern Black Mesa (Thornthwaite et al. 1942). Excel- lent detailed descriptions of the regional cli- matic patterns directly applicable to Black Mesa are those of Hack (1942), Thornthwaite et al. (1942), Lowe (1964), and Sellers and Hill (1974). The mean annual temperature of sur- rounding sites ranges from 11.6°C (53°F) at Kayenta (elevation 1737 m [5700 ft]) to 9.8°C (49.7°F) at Navajo National Monument (eleva- tion 2220 m [7285 ft]). Mean annual precipita- tion in northeastern Arizona is relatively low for the moderate elevations typical of the region (Sellers and Hill 1974), and for this rea- son the area has been called a "rainfall sink" (Brown 1982). Mean annual precipitation at Kayenta is a scant 198 mm (7.78 in), and at Navajo National Monument it is 291 mm (11.46 in). Mean annual precipitation of seven sites (mean elevation 2054 m [6740 ft]) on the Peabody lease is 260 mm (10.22 in). This pre- cipitation is bimodal, with nearly 46% falling as convectional showers in July, August, and September. The remainder falls throughout the year as cyclonic rain and snow. January and June are the driest months. The diy cli- mate allows only intermittent perennial stream flow. Ephemeral flows may result from thunderstorms and melting snow. Prevailing winds are southwesterly. May and June fre- quently have warm, dry winds that add to the approximately 1015 mm (40 in) annual evapo- ration rate. Habitats and Associated Birds Fourteen habitat types have been identi- fied for this report and are mapped in Figure 3. Designation of a particular habitat is based on one or more of the following: perennial plant species composition, vegetative physiog- nomy, topographic uniqueness, and distinc- tiveness of the associated breeding bird species composition. The sequence of the pre- sentation of the habitats follows increasing breeding bird species richness, which is con- sidered here to be an approximation of breed- ing bird species diversity (see Ralph 1985, Brotherson et al. 1981, Wiens and Rotenberiy 1981). Tables are also arranged by this sequence. Habitats in the study area may be sharply demarcated as in ponds, riparian areas, or cliff scarps. As a lade, however, there is much interspersion between them. Breeding bird communities of northern Black Mesa comprise a complex of species derived principally from the Sonoran and boreal fliunal areas, utilizing a series of plant comminiities of Great Basin origins. (Faunal derivations of native terrestrial breeding birds of northern Black Mesa follow Behle [1985]). The upland nature of the study site may explain the high number of boreal species (44%), while the geographic position of the area in the southwestern United States proba- bly accounts for the similar proportion of Sonoran derivatives. Predominant habitats on northern Black Mesa are pinyon-juniper woodland and sagebrush, saltbush, and greasewood shrublands (Fig. 3), which are pri- marily Great Basin-derived plant communi- ties (Brown 1982). However, only 9 birds of Great Basin origins are present on Black Mesa because the Great Basin is characterized by few endemic bird species (Behle 1985). Near- ly 38% of boreal species and 57% of Sonoran species breed in these Great Basin-type habi- tats. Half of the 10 breeding species designat- ed as typical of the pinyon-juniper woodland are derived from the Great Basin. Montane Scrub The montane scrub on Black Mesa is best developed on slopes of the outer mesa escarp- ment where conifers have not yet invaded rockfalls, slides, and slumps (Fig. 4). Because of this, montane scrub habitat may often be a serai stage of succession. Its establishment in an old burn of several hectares in mixed- Great Basin Naturalist [Volume 54 I I = Pinyon-Juniper Woodland ^B = Chained Pinyon-Juniper Woodland [■:-:":-3 = Mixed Conifer |.;.;| = Reclaimed Mine Spoil & Active Mining Areas ^3 = Grasslands & Juniper Savanna liij = Shrublands (incl. sage, salt, grassweed & mixed shrubs) mU = Tamarisk-filled Channels ^^ = Major Cliff Scarps • = Larger Aspen Stands SCALE 5 kilonneters Fig. 3. Vegetative comiminities of northern Black Mesa. conifer woodland in upper Moenkopi Wash indicates such a condition. It occurs from 2133 to 2470 m (7000 to 8100 ft) and is distributed extensively east of the study site along the northern face of Black iMesa. The dominant species of this habitat include Gambel oak {Quercus gambelii), snowberry (Sijmphoricar- pos sp.), cliff fendler bush (Fendlera rupicola), wax cinrant {Ribcs ccreum), Utah senicebeny (Atnelanchier titaJicnsis), wild-rose {Rosa Birds of Northern Black Mesa, Arizona Fig. 4. Montane scrub below Loloniai Point, August 1987. arizoiuca). and skunkbiish sumac (Rhus trilo- hata). A deciduous shrubland has dexeloped extensively in chained pinyon-jiniiper wood- land on Lolomai Point. Five bird species, dominated by ground- nesting and foliage-gleaning forms, are known to breed in montane scrub in the study area (Tables 1, 2): Scrub Jay, Orange-crowned War- bler, Virginia's Warbler, MacGillivray's War- bler, and Rufous-sided Towhee. No breeding densities have been determined in this habitat on Black Mesa. In Gambel oak-mountain mahogany and scrub oak-mountain mahogany scrub in Colorado, the number of species per plot varied from 9 to 18, with breeding densi- ties of 53-116 individuals/40 ha (American Birds 1982, 1983). Rufous-sided Towhees and Scrub Jays were present in both habitats in both years. Brown (1982) considers the Vir- ginia's Warbler and Rufous-sided Towhee to be characteristic of this habitat. Sagebrush Shrubland The sagebrush shrubland is dominated by big sage {Artemisia fridentata) to densities of approximately 10,870 shrubs/ha (4400/ac). Understory associates include blue grama {Boiitcloiia gracilis) and squirrel-tail iSitanion Jujstrix) and a variety of forbs (Fig. 5). Pure sagebrush covers the deeper alluvial soils of the large valleys such as White-house Valley (Fig. 5), Reed Valley, and Coal Mine Wash, and canvons and basins at elevations of 1920-2377 m (6300-7800 ft). Wintering and migrant bird species are characteristically sparse. All breeding species are ground-feeders, most of which are sum- mer residents (Tables 1, 2). Bushtits forage frequently in sagebrush in winter, and Bewick's Wrens move into this habitat in late summer. Breeding species are the Horned Lark, Sage Thrasher, Green-tailed Towhee (restricted to Lolomai Point), Brewer's Spar- row, and Sage Sparrow, all of which are t>'pi- cal of sagebrush shrublands (Rotenbeny and Wiens 1980, Wiens and Rotenberry 1981). Overall spring densities are about 75 individ- uals/40 ha. Late-sininner densities may reach 100 individuals/40 ha (Table 3). Rotenberry and Wiens (1980) report breeding densities in Arf^'//H'.sJfl-dominated shrubsteppe of 56-182.3 individuals/4() ha. Three to five species were found breeding on each plot. Two sagebrush plots in California supported six breeding 10 Great Basin N vrufiAUsr [Volume 54 Table 1. Bivfcliiie; l)ircl foimnunity stnictuic iiicst site lial)itats) on nortlii'iii lilack Misa. At Nnnihei- ol breeding species l^roportion oi permanent residents Nesting site (projiortion of spc substrate ■cies utilizing) Habitat (ironnd Foliage Ca\it\ Li'dge Di\ersit\' Montane scrul) 5 ().4() 0.80 0.20 — — 0.501 Saiii'hrnsli 5 0.20 0.40 0.60 — — 0.673 Salthusli 5 0.20 0.40 0.60 — — 0.673 Crcasewood 6 0.17 0.17 0.83 — — 0.456 Reclaimed mine spoil - 0.29 0.57 0.43 — — 0.683 Juniper sa\anna 8 0.25 0.25 0.75 — — 0.347 Ril^arian Iial)itats 9 0.22 0.22 0.78 — — 0.527 Ponds 9 0.11 0.89 0.11 — — 0.563 Chained pin\()n-j\inii)er 11 0.45 ()..36 0.28 0.36 — 1.092 Mixed-shnil) 12 0.25 0.25 0.58 — 0.17 0.964 Aspen gn)\es 14 0.36 0.36 0.36 0.21 0.07 1.238 Cliffs, talus, hanks 20 0.50 0.05 — 0.401' 0.55 0.845 Pinxon-jnniper 42 0.55 0.07 0.49 0.34 0.10 1.133 Mixed-conifer 59 0.58 0.12 0.45 0.29 0.14 1.247 •'Di\irsit\ \,ilui-s ' \alnes deri\L-tl nsin«4 proportions ol speeics in each foiaiiii ,a M.uArlliur 19(ili Species at total densities of 47 and 17 individ- uals/40 ha (American Birds 1979). Smith et al. (1984) reported hreeding-hird densities of 61.2 and 65.2 individuals/40 ha in sagel)rush in Idaho over a 2-year period. Saltbush Shrubland Saltbush shrubland habitat occms as pure stands of fourwing saltbush {Atriplex canescens) in valle\'s on alluvial terraces at 1980-2070 m (6500-6800 ft). These stands are usually upstream from stands of greasewood {Sarcohatiis vennicuhitiis). Cheatgrass {Brofnus fectoniiii) and sticksecd {Lappida rcdowski) IrequentK dominate the understoiA. Saltbush stands are well developed in the middle por- tion of Yellow Water Wash, in Reed Valley, and in portions of Dinnebito Wash (Fig. 6). Open stands grow in places at the foot of the mesa. Like the sagebrush shrubland, the saltbush breeding bird assend)lage is comj:)osed of ground-feeding, summer-resident species (Tables 1, 2). Bird densities in saltbush are 1994] Birds of Northern Black Mesa, Arizona 11 Fig. 5. Sagebrush shriihlaiid, W'liite House Valley, June 1987. higher than those in sagebrush (Table 4). Mid- April 1982 densities in Reed Valle\' and Din- nebito Wash were 121 and 138 individnals/40 ha. The dominant breeding species is the Brewer's Sparrow. Four remaining breeding species are the Horned Lark, Sage Thrasher, Vesper Sparrow, and Sage Sparrow, although the latter species is sparse in saltbush as a breeding species. Of two saltbush plots (pre- sumably Atriplcx canescens) in California, one supported no breeding birds, the other six species at 21 individuals/40 ha (American Birds 1979). A third plot in shadscale scrub (A. confertifoUa) supported fi\'e species at 30 individuals/40 ha (American Birds 1979). Extensive grazing pressure in saltbush maintains an open understory that probably allows use of this habitat by Horned Larks. In Table 3. Bird densities in two stands of sagebrush shrubland in the J-19 and J-21 mining areas'' Bird density (no./40 ha) S pring Summer Fa 11 Species J-19 J-21 J-19 J-21 J- 19 J-21 Bushtit 10.6 — Bewick s Wren 6.1 4,7 11,5 .32.5 — — Blue-grav Gnatcatcher — — 9,4 5.3 — — Western Bluebird — 5..3 — — — — Mountain Bluebird 5.6 13.3 8.2 4.5 3.5 — Yellow-rimiped Warbler — — — — 4.6 4.2 Brewer's Sparrow 27.5 — .38.2 8.9 — 8.0 Sage Sparrow 41.2 37.5 29.3 33.3 12,4 — House Finch — — — 15.4 — — Winter 1-19 1-21 Total 80.4 71.4 96.6 99.9 20.5 12,2 4.4 4,4 ^Study conducted In PealiodN ( 19.S2-S3j 12 C,HEAT Basin Naturalist [\blunie 54 ^•*1>. \W- r*%?^^J'- r<5> \-jv' I Ht .-'^-^. '^Mk ?-."♦•. '■ -•jffi wm^ •. * -«^~...^?%j^ - -•^-..^ -t^^' Fi^. (i. Salthush shrubland, Diiuifhito \\'asli, June 1987. f^fejI^S •.^A«iU< ^- if ■* ■•^ A- .^^■/ivlf'r V.^ V."^* •y-v'- winter, Horned Larks are the only species that can be considered common in salthnsh. Greasewood Shrubland Greasewood shrubland dominates terraces of the larger and lower wash valleys of Moenkopi, Coal Mine, Yellow Water, Red Peak, and Yucca Flat washes (Fig. 7). It is typi- cally found below 1920 m (6300 ft). Shadscale {Atriplex coiifertifoUa), alkali sacaton {Sporo- bolus airoides), and cheatgrass are connnon imderstory associates. The breeding bird species connnunity of greasewood is dominated by summer-resi- dent, foliage-nesting, and ground-feeding forms (Tables 1, 2). The Northern Mocking- bird, Bendire's Thrasher, Loggerhead Shrike, Lark Sparrow, Brewer s Sparrow, and Black- throated Sparrow are known breeding species of greasewood. The greasewood stand cen- sused in Table 5 shows significant numbers of Sage Sparrows and House Finches, but it is not known whether they nested in this partic- ular higher-elevation (2042 m [6700 ft]) stand. The relativeK' large number of species record- ed for this stand may be due to the presence of surface water in the arro\ o that bisected it. Extensive greasewood stands on lower Moen- kopi Wash have not been censused. Results of such work would probably be typical of shrubsteppe vegetation, with breeding bird densities around 60-180 individuals/40 ha comprising 3-5 species (Rotenberry and Wiens 1980), particularly those noted above. Bradfield (1974) found the Northern Mock- ingbird, Bendire's Thrasher, and Black-throat- ed Sparrow to be characteristic of greasewood on Oraibi Wash at the southern edge of Black Mesa. I have found Sage Thrashers breeding in greasewood stands at Kayenta just outside the stud\ area. Reclaimed Surface Mine Spoil Initiation of the large-scale surface coal mining operation by Peabodx Coal Company in the 1960s saw extensive tracts of land undergo what is essentially a type conversion. By agreement with the Navajo and Hopi tribes, the postmining land use of the areas proposed for coal extraction was designated as livestock raising. Therefore, a grassland vege- tation type best suited for effective livestock production is the objective sought in postmin- ing reclamation efforts. 1994] Birds of Northern Black Mesa, Arizona 13 Table 4. Bird ck-nsities in two stands ol'saltljusli shnihland in Reed Xalley (RV) and Dinnehito Wash (DW)-> Bird density (no./4() Iia) Spring Sumniei Fall Species RV DW R\' DW^ Sa\'s Phoehe 24.0 8.8 5.7 1 lorned Lark — 8.8 — 50.3 Rock Wren 9.8 3.5 9.6 7.3 Bewick s Wren 5.9 — 5.7 1.8 Sage Thrasher 8.6 19.5 — — Monntain Bluebird 5..3 16.9 2.9 4,5 Vellow-rumped Warbler — — — — Brewers Sparrow 2.3.1 10.0 84.7 163.2 V'esper Sparrow — 27.5 — — Sage Sparrow 7.7 22.9 — — White-crowned Sparrow — — — — House Finch 37.3 20.8 7.6 13.3 Tdi \i 12L7 138.7 110.5 246.1 R\' DW \Vinter RV DW 134.8 134.8 49.8 7.8 26.6 14.8 8.9 .57.4 5.3 170.6 41.4 41.4 •'Stmlirs ccimluctril 1)\ P.mIkkK (1982-83). To date about 4300 ha (10,625 ac) of lie above 2134 m (7000 ft) and about 8900 ha regraded mine spoil has been seeded with (22,000 ac) will be reclaimed. The mix of grasses and other range plants (Fig. 8). Land species planted for each area varies, but typi- that has been reclaimed lies between 1980 cal species used are crested wheatgrass, west- and 2072 m (6500 and 6800 ft); however, by ern wheatgrass, intermediate wheatgrass the time mining is completed, some areas may [Agropyron iniennedium), various other ^.•'■^.^•iti^ -A. 1*C^ ^ .-'jf: tgfi^ .-*?■* *^:i^ ;,/■ Fig. 7. Greasewood slnubland, Moenkopi/Red Peak \'alle\ Wash connuence. June 1987. 14 Great Basin Nail iulist [Volume 54 'I'ahi.Ko. IJird c Ic'iisitics in ^reasewood-salthiisli sliiiil) laud in the J-2S niininii area-'. Bird density {no./40 liaj Species Spring Suinnier Fall Winter Rufous Iluniniinsihird Northern Flicker Gray FKcatclu'r Sax's Flioehe Asli- throated I'Kiatcher Horned Lark Scrni) |a\ Mountain Clhickadee Bushtit Rock \\ ren Bew ick s Wren Mountain Bluebird Solitan' X'iri'o W'arhler Chippinsi Sparrow Bri'wer s Sparrow \'esper Sparrow Lark Sparrow Black-throated Sparrow Sas^e Sparrou White-crowned Sparrow Dark-e\'ed J unco Sparrow sp. Meadowlark sp. Cassins Finch House Finch Pine Siskin Unidentified sp. ToiAL 8.9 6.7 ILl 17.8 2.2 L5.6 6.7 L5.6 8.9 22.2 3.6 8.9 15.6 1.8 37.8 1S3.4 2.2 1.8 8.0 2.7 0.9 1.8 0.6 — 0.9 2.7 26.7 — 3.6 5.3 4.4 — 2.7 7.1 — 5.3 — 20.0 — _ 39.1 3.6 3.6 0.9 — 18.7 6.2 — 33.8 18.8 — 0.9 55.2 30.2 0.9 248.9 126.2 95.2 337.9 ■'Studies tcrnliatril In PialiuiK il9Sl-S2l. wheatgrasses, Ru.ssiaii wildrye {Elyiuus junceus), smooth hrome {Bro)niis inennis), saltbush, alfalfa {Medicago sativa), sweet clover [Melilntus ojficinalis), some blue grama {Boiiteloua gracilis), and Indian ricegrass {Orijzopis hyinenoides). Areas reclaimed after mining contain the best-developed grasslands in the study area. The breeding bird conmiimity is composed entirely of ground-feeding forms, most of which are summer residents (Tables 1, 2). Horned Larks are the typical birds throughout the year in such areas (Table 6). Western Meadowlarks are common breeding birds within the study area only on reclaimed areas. Where saltbush is well established. Mourning Doves, Sage Thrashers, and Brewer s and Vesper Sparrows have been found nesting in reclaimed areas (Table 6). Maximum breeding densities under such conditions have been found as high as 33 pairs/40 ha. Cody (1966) reported 3-4 breeding species to be t\ pical of grasslands worldwide. Roten- berry and Wiens (1980) report 2-6 (average 3.8) in shortgrass prairie and 3-5 (average 3.8) in mixed-grass prairie. They found breeding densities in these types to be 81.6-132.8 and 40-126.4 individiials/40 ha, respectively. Therefore, the reclaimed areas on Black Mesa support breeding birds at densities compara- ble to natural grasslands elsewhere in North America. However, breeding bird species richness is lower than most grasslands. This may be related to the relative structiu-al and floristic simplicity of the reclaimed areas (MacArthur 1964, Karr and Roth 1971, Tomoff 1974, Willson 1974, Roth 1976). Juniper Savanna Jiuiiper savanna is a predominantly over- grazed grassland terrain at the lowest edge of the pinyon-juniper woodland (Figs. 1, 10). It is located as a band at the foot of the mesa from 1770 to 1860 m (5800 to 6100 ft). Juniper inxasion into grasslands has occurred in the southwestern United States in the past 130 1994] Birds of Northern Black Mesa, Arizona 15 Fig. 8. Reclainu'd mine spoil, N-2 reclaimed area, June 198"; \'ears (West et al. 1975); trees in this habitat, appearing relatively young, support this. Vegetation is dominated by snakeweed, galleta, bkie grama, squirreltail, Momion tea {Ephedra viridis), narrowleaf yucca {Yucca an(i,ustis.si)na), Indian ricegrass, sand dropseed {Sporohohts cryptandrus), and scattered Utah and one-seed junipers {Jiiniperus osteospenna and/ ]nonospenna). The niajoritx' of breeding species are simi- mer-resident, foHage-nesting, and ground- feeding forms (Tables 1, 2). No censusing has Table 6. Spot-mapped breeding bird densities in tliree reclaimed areas in 1985 (no. pairs/40 ha). Rec lainied area Species \-l J-l/N-6 J- Mourning Do\ e .3. .3 Horned Lark 19.2 17.7 16.3 Sage Thrasher — — 3.3 Brewer's Sparrow — — 3.3 \'espei- Sparrow — -1- 3.3 Lark Sparrow — + — Western Meadowlark — 2.2 3.3 Tf)T.\L 19.2 19.9 .32.8 + Pi->-s.iit Imt (k'iisit\ uiKlrlinimird, been done in juniper savanna in the study area, but studies in similar habitats (Grue 1977, Beatty 197S [both cited in Balda and Masters 1980]) indicated 11-23 breeding species/plot at densities of 35-179 pairs/40 ha. Since pinyon-juniper woodland on Black Mesa supports 68.5-106.7 pairs/40 ha (see Table 12), it seems reasonable to assume that breeding bird densities would be below 70 pairs/40 ha. Eight bird species are known to breed in juniper savanna in the study area: the Horned Lark, Northern Mockingbird, Bendire's Thrasher, Loggerhead Shrike, Chip- ping Sparrow, Lark Sparrow, Western Mead- owlark, and Scott's Oriole. Riparian Habitats Riparian habitats are the most restricted habitats in the study area. They are dominated by tamarisk {Tamarix chinesis) thickets that are usualh' less than 4.6 m (15 ft) tall. Fremont cottonwoods {Popidus fremontii), although present as a few young individuals, are essen- tially absent from the study area. Russian olives {Eleagnus angustifolia) are actively advancing up Moenkopi Wash and are currently well established at about 1829 m (6000 ft). Dense 16 Great Basin Naturalist [Volume 54 "^ti^F^r'-.-rv; 'i* '"-'i^i- -^:^*^' -V^ Fig. 9. Juniper saxanna and the northern escarpment, Owl Spring N'alle) , June 1987. tamarisk thickets grow on Coal Mine and Moenkopi washes (Fig. 9) below 1920 m (6300 ft). There is also a thicket at the confluence of Moenkopi \Viish and Reed Valley at 1981 m (6500 ft). Tamarisk-filled wash channels are considered disclimax strand communities by Brown (1982). During the 1980s, tamarisk continued to establish and spread conspicu- ously throughout the study area. This perhaps indicates that its invasion has not yet ended. Bird use of riparian habitats on Black Mesa, like that of ponds, is t> pified by heavy migrant use and few breeding species. Most breeding species are summer-resident, foliage-nesting, and ground-feeding forms (Tables 1, 2). The only species found nesting in tamarisk are the Killdeer, Send) Ja\, Bushtit, Blue Grosbeak, Lazuli Bunting, Indi- go Bunting, and Brewer s Blackbird. The Black-chinned Hunnningbird, House Finch, and Lesser Goldfinch are suspected breeders. Migrant densitites, usually much higher in fall than in spring (Table 7), averaged 666 indi\id- uals/40 ha from September to mid-October. Common fall migrants include the House Wren, Kuby-crowned Kinglet, Orange- crowned Warbler, Yellow-rumped Warbler, MacCillivray's Warbler, Wilson's Warbler, Green-tailed Towhee, Brewer's Sparrow, Chipping Sparrow, and White-crowned Spar- row. Wintering bird densities are dominated by Dark-e\'ed Juncos and White-crowned Sparrow. It is interesting that such numbers of insectivorous, foliage-gleaning migrants (near- ly 68% of the total densit\') utilize tamarisk in the fall, while virtually no foliage-gleaning forms nest in it. An increase in arthropod den- sities may occur in late summer and fall, attracting migrants. However, Hunter et al. (1988) believe that insects do not limit insecti- vore use of tamarisk on the Colorado River in southwestern Arizona. Species from adjoining habitats, such as the Northern Mockingbird, Loggerhead Shrike, and Black-throated Spar- row, frequently forage in or near tamarisk. The Blue Grosbeak is a characteristic breed- ing species in tamarisk throughout the South- west (Bradfield 1974, Jacobs 1986, Hunter et al. 1988). Dark-eyed Juncos and White- crowned Sparrows are common in tamarisk in winter in the southern edge of Black Mesa (Bradfield 1974). 1994] Birds of Northern Black Mesa, Arizona 17 9*^-- -■,, ■<*^^--^Kf"'-'->tt%^-*;.'-----^-: ..-..-rr- Jk'#^.'--^^-'«^ Fig. 12. Chained piinon-jiiniper woodland invaded liy Gamhel oak, Loloniai Point, August 1987. t\pe are Mancos Shale slopes at the mesa foot and in the area of the Moenkopi Wash-Coal Mine Wash confluence that are covered with shadscale and Russian thistle [Sahola iherica). Elevations in this area are 1829-1980 m (6000-6500 ft). Tomoff (1974), Wiens and Rotenberry (1981), and Smith et al. (1984) noted that cer- tain desert scrub and shrubland bird species are associated with specific shrub species. Tomoff (1974) states that "plant species com- position is highly significant in regulating breeding bird communities in desert scrub. ' Rotenberry (1985), in a study of grassland habitats, found that "over half (55%) of the variation in bird community composition was associated with floristic variation. Desert scrub bird communities in the study area sup- port these statements, with several shrubland breeding birds being associated with particu- lar shrub species. Sage Sparrows are most numerous in sagebrush, Sage Thrashers and Brewer's Sparrows prefer saltbush, and Black- throated Sparrows are typical of greasewood and shadscale. Because the mixed shrubland is a composite habitat, the breeding bird com- munity composition at any particular site reflects the shrub species composition repre- sented in it. For these reasons, an\' particular area of mixed shrubland t\'picall\ supports more species of breeding birds than a mono- typic shnibland (Tables 8, 9). Shadscale occasionally occurs as nearly monotypic stands on exposures of Mancos Shale (or on the outwash fans of material eroded from the same). Such situations appear to support the lowest breeding bird densities and species richness of any of the native habi- tats on Black Mesa. Fautin (1946), Smith et al. (1984), and Medin (1986, 1990) examined bird communities in shadscale-dominated desert scrub in the Great Basin. Each study found about three species breeding at densities of 39.2-64.8 individuals/40 ha. Horned Larks were the most numerous species in each stud>', topically constituting the majority of total bird densit)'. The mechanism(s) of the relationship noted above are not clearly known. The tendency of some birds to nest in certain plant species (Tomoff 1974, Petersen and Best 1985) helps explain part of the relationship. Species-spe- cific exploitation of arthropod faunas distinct to each shrub species may be important 1994] Birds of Northern Black Mesa, Arizona 21 --.---,-- 1^'- -►•- •*--#- .-T'-- : •fi-T' ^9ii?V' rf I'Sflk.- .i*.**" r; ^^^^^iiilfe 4.\- '.am X* 3l ^^^ -''■ -9ir» .V'- ^ wT- *^*, - -..•><\v. Fig. 13. Shadscale-dominated mixcil slinihlaiid, Moenkopi Wash, June 19S7 (Wiens and Rotenberry 1981). Rotenberry (1985) postulates that within similar habitat types "bird species/plant taxa associations . . . are mediated by the specific food resources that different plant taxa provide." Bird response to habitat physiognomy plays a key role in habitat selection and utilization (Mac-Arthur 1964, Karr and Roth 1971, Wiens 1973, Willson 1974, Roth 1976). Shrublands on Black Mesa represent a single physiog- nomic type, with the shrub species compris- ing them all having Great Basin affinities (Brown 1982). It may be that the distinct bird species associated with various Great Basin desert scrub communities evolved through resource partitioning brought about by com- petitive interactions (Cody 1985). Aspen Groves Aspen groves are found in cool, moist heads of box canyons (Fig. 14) or sheltered ravines at elevations above 2195 m (7200 ft). These groves are dominated by quaking aspen {Popuhis tremuloides), box-elder maple {Acer negundo), Gambel oak, chokecherry {Pruniis virginiana), and red osier dogwood {Corniis stolonifera). One grove contains several large narrowleaf cottonwoods {Populus angustifolia). Aspens, known from 39 sites within the study area, vary in stand size from a few trees to stands up to 1.86 ha (4.6 ac). Each major wash, except Dinnebito, contains aspen groves, but most are located in drainages of Yellow Water Ganyon in the vicinity of Lolo- mai Point. Understory vegetation includes meadow rue {Thalictrum fendleri) and Oregon grape {Berheris repens). Poison ivy {Rhus radi- cans) and bracken fern {Pteridium oqiulinum) are found in some areas. Gattle grazing has widely opened the shrub stratum in several groves, while in protected groves the under- story may be nearly impenetrable. Most of the 14 known or suspected breed- ing species in aspens are foliage-gleaning and ground- and foliage-nesting summer residents (Tables 1, 2). House Wrens and Warbling Vire- os breed only in this habitat. No extensive censusing has been conducted in aspens on Black Mesa. The results in Table 10 suggest that rather high densities may occur. An aspen stand with scattered conifers in Colorado sup- ported 30 species at 184 individuals/40 ha (American Birds 1979). 22 Great Basin Natuhalist [Volume 54 Table 8. Spot-mai^pccI l)r('ccliiiu I)ircl dcTisitit's in mixed shriihlaiul near tlif |-7 iiiininii area. 1984-80) (no. pairs/4() lia). Spc'cifS 1984 1985 1986 M()nrnin<4 Dovr — — + Sun's Plioelx- L5 — 2.1 Horned Lark S.9 + 5.9 Nortlicrn Raven — — + Rock Wren L5 L5 1.2 Mountain Rlnel)ird — — + Northern M<)ekingl)ird 0.8 + 0.6 Sajii- Tliraslier + — 3.0 Lojigerliead Shrike 0.8 + — Brewer's Sparrow 3.0 8.9 7.4 Vesper Spairow + + — Lark Sparrow — — 1.2 Black-thioated Sparrow 1L9 5.9 8.9 Saiie Sparrow 7.4 3.0 1.8 Brown-headetl C'ou bird — — + Total 35.8 19.3 32.1 Breedin(; species richness 8 4 9 Cliffs, Talus Slopes, and Wash Banks Cliffs, talus slopes, and eroded wash banks are found at all elevations throughout the study site. On the outer mesa escaipment, the geologic strata that form cliffs are Dakota Sandstone, the Toreva Formation, and Yale Point Sandstone (Figs. 10, 15). Cliff heights in this area may be up to 107 m (350 ft), but are usualK imdcr 46 m (150 ft). Yale Point Sand- stone is the cliff-former in canyons near the mesa rim. In the vicinity of the mine lease, the Wepo Formation rarely foniis cliffs. Toreva Sandstone cliffs flank the valley where Moenkopi Wash exits the study area. Eroded wash banks are present throughout the study site where wash channels have dissected allu- vial fill. Twenty bird species are known or suspect- ed to use ledges, or holes, in cliffs or wash banks as nesting sites. Several of these are restricted to such nesting sites in the study area and include the Prairie Falcon, White- throated Swift, Cordilleran Flycatcher, Violet- green Swallow, Northern Rough-winged Swallow, Cliff Swallow, Rock Wren, Canyon Wren, and Townsend s Solitaire. The majority are permanent residents, and aerial feeders and predators predominate (Tables 1, 2). Some require special conditions around suit- able nesting sites; e.g., Cordilleran FKcatch- ers must also have mixed-conifer present and Northern Rough-winged Swallows have been found nesting only in holes in wash banks. Piiu'on- Juniper Woodland Pinyon-jimiper woodland (Fig. 16) is one of the most widespread communities in the southwestern United States, occiuring where Table 9. Bird densities in sagehrusli i-nii.\ed shruhland near the J-7 mining area". Bird densitx (no. 40/ha) Species Spring Summer Fall Winter Cooper s I lawk Black-eliinned 1 lumniiny;l)ird Ilumminnhird sp. Say s Phoebe Ash-throated M\ catcher Horned Lark Rock Wren Western Bluebird Bluebird sp. Northern Mockingbird Sage Thrashei- Brewer's Sparrow Vesper Sparrow Black-throated Sparrow Sage Si)arr()\v Sparrow sp. Western Meadowlark Brewer's Blackbird House Finch ToiAL 0.9 17.8 0.9 2.7 0.9 0.6 0.4 24.2 0.9 0.9 0.4 10.1 4.7 0.4 0.6 1.8 4.4 24.2 109.3 0.9 0.4 0.9 2.7 1.8 0.9 116.9 1.4 186.7 188.1 ulutt.dln EH&A (1979-8(11 1994] Birds of Northern Black Mesa, Arizona 23 •r.-f'm^^l^'j' >Jl • "^i'^":-i I •'tf ^terV ;f:»-\ 'Ni^; Fig. 14. Aspen grove, Moenkopi Wash, October 1986. mean annual precipitation is 250-500 mni (9.8-19.7 in) (Brown 1982). It is the dominant plant communit\' of Black Mesa (Fig. 3). The dark aspect this woodland imparts to the mesa when seen from a distance is said to account for the name "Black Mesa." Throughout Black Mesa, the woodland begins appearing at about 1830 m (6000 ft) and is found from this elevation to the mesa's highest reaches at over 2470 m (8100 ft). Colorado pinyon {Piniis edulis) and Utah jimiper are the principal trees of this wood- land (Fig. 16). Junipers dominate at lower ele- vations, and as elevation increases, pinyons become dominant, total tree density increases, and trees become larger in stature (Table 11). Tree densities may exceed 400/ha near the mesa s nortliem rim. Understoiy is usually open but variable. In some places it is nearly bare, while in others big sage may be quite dense. Cliffrose {Cowanki mexicana) and Gambel oak are frequently found in the woodland. Above 2200 m (7200 ft), cliff fendlerbush and ante- lope bush {Purshia tridentata) are common understory associates. Silverleaf buffaloberry {Shepherdia rotiindifolia) is a common under- story' associate on the rocky mesa scarp and in canyons near the mesa rim. Manzanita {Arc- tostaphijlus piingens) is present in the wood- land in a few places near the mesa rim. The assemblage of bird species at an\' point in the woodland is dictated by stand charac- teristics, tree density, tree species composi- tion, and abiotic factors such as soil, slope, and exposure. No single area of woodland will support all of the 41 known or suspected breeding species. Several species breed in higher-elevation stands while others nest in the lower, open stands. Scott's Orioles, Gray Vireos, and House Finches are typical low- stand species. The Haiiy Woodpecker, Moun- tain Chickadee, White-breasted Nuthatch, Solitary Vireo, Black-throated Gray Warbler, and Rufous-sided Towhee are most common in higher-elevation stands. Eleven species are widespread throughout the woodland and can be considered typical of it: Gray Flycatcher, Ash-throated Flycatcher, Pinyon Jay, Moun- tain Chickadee, Plain Titmouse, Bushtit, Bewick's Wren, Mountain Bluebird, Solitary Vireo, Black-throated Gray Warbler, and Chip- ping SpaiTow. These species (excluding the jay) accounted for 80.1-92.8% of the total breed- ing bird density in the three stands censused 24 Great Basin Naturalist [Volume 54 Table 10. Ck'nsus results in aspens on nortliein lilaek Mesa-'. Species No. present Turke\ Vulture Shaip-shinned Hawk Red-tailed Hawk Broad-tailed I hnnniinuhird White-throated Swiit Downy Woodpecker Hairy Woodpecker Northern Flicker Dusky Flycatcher Cordilleran Flycatcher Violet-green Swallow S teller's Jay 2 Scrub Jay 2 Mountain Chickadee 4 Pygmv' Nuthatch 1 Brown Creeper 1 House Wren 1 American Robin 2 Solitary Vireo 1 Warbling Vireo 4 Orange-crowned Warbler 3 Virginia's Warbler 1 Wilson's Warbler 1 Black-headed Grosbeak 4 Northern Oriole 1 Total small forms 44 5 utilize nocturnal roost 2 pair with nest 1 noted o\ erhead 4 — foraging above canopy 1 4 3 2 9 foraging above canop>' 'KDensus conducted in die middle fork of Coal Mine Wash on 25, 2(i Nhi\ 1986. Six traverses counting all indi\idiials detected. ApiiroxiniateK 2 li.i censused. in 1983 and 1984 (Uble 12). Half of the breed- ing species are foliage nesters, but cavity nesters are also well represented (Tables 1, 2). Ten percent of the species nest on the ground (Table 1). In the three stands censused above, foliage-nesting pairs constitute 42.9% of the total nesting pairs in the highest tree-density stand. Foliage nesting pairs make up nearly 50% in each of the other two stands. Nearly 87% of foliage nests and 87% of all cavity nests located in Black Mesa woodland have been in Utah junipers. Balda and Masters (1980) found that breed- ing bird density increases with increasing tree density. On Black Mesa an average breeding density of 66.7 pairs/40 ha (1983 and 1984) used a stand containing 150 trees/ha (Tables 11, 12). In a stand of 283 trees/ha, an average of 93.4 pairs/40 ha was found for the same years. A density of 105 pairs/40 ha was deter- mined for a stand of 380 trees/ha. Total breed- ing bird density showed a strong positive cor- relation with total tree density (r = .99). Addi- tionally, pinyon density was positively corre- lated with densities of Gray Flycatchers, Mountain Chickadees, and Black-throated Grav Warblers. These results reflect those Fig. 1.5. Cliffs at the northern rim, September 19- cover (%): pinyon 31.90 49.15 69.07 Relative canopy cover (%): juniper 68.10 50.85 30.93 Mean tree height 4.02 in 3.66 m 5.00 m Mean height; pinyon 3.9 m 3.52 m 4.93 m Mean height; juniper 4.04 m 3.85 m 5.16 m elevations to the south and west where rainfall is greater on the Mogollon, Kaibab, and Coco- nino plateaus are well-developed ponderosa pine or ponderosa pine-Douglas fir forest (Brown 1982). Mixed-conifer habitats occur on northern Black Mesa in specific areas where local abiotic conditions favor its devel- opment. These conditions all increase avail- able moisture and are (1) deep, sheltered, and shaded canyons, (2) north-facing slopes, (3) joint traces and cracks where runoff is con- centrated by bare exposures of Yale point sandstone, and (4) small, shallow drainages above 2255 m (7400 ft) where runoff is also concentrated. The pinyon-juniper woodland is still the dominant vegetation type on the broad mesa top (see Fig. 15) at elevations above 2440 m (8000 ft), indicating that the mesa s highest elevations probably receive no more than about 356 mm (15 in) mean annual rainfall. This is less than the 470-540 mm (18.5-21.3 in) reported for the above-men- tioned uplands (Brown 1982) and approxi- mates the isohyet estimate for northern Black Mesa by Cooley et al. (1969). The mixed-conifer association supports the greatest richness of breeding bird species of all habitats present in the study area. Fifty- nine species are known or suspected to breed in mixed-conifer habitats (Table 1). Of these, nearly half are permanent residents. Foliage- and cavity-nesting forms predominate, although a variety of other nesting substrates are utilized by the remaining species (Table 1). Common characteristic breeding species include the Broad-tailed Hummingbird, Dusky and Cordilleran Flycatchers, Steller's Jay, Pygmy Nuthatch, Hermit Thrush, Grace's Warbler, and Dark-eyed Junco. Townsend s Warbler, along with MacGillivray's, Wilson s, and Orange-crowned Warblers, are common migrants. The Golden-crowned Kinglet is a typical winter resident. That Red-breasted Nuthatches may be common some winters is probably related to conifer cone crop irrup- tions (Widerlechner and Dragula 1984). No bird density data have been collected in mixed-conifer woodland on Black Mesa. Breeding bird densities no doubt greatly exceed that of the pinyon-jimiper woodland. A ponderosa pine stand on San Francisco Moun- tain, Arizona, supported 23 species at 232 pairs/40 ha, and a nearby mixed-conifer stand supported 27 species at 253 pairs/40 ha (Haldeman et al. 1973). A spnice-fir stand in the White Mountains, Arizona, supported 16 species at 169.7 pairs/40 ha and 17 species at 186.5 pairs/40 ha in two different vears (Carothers et al. 1973). Franzreb (1977) reported mid- to late breeding season bird densities of 632.9 and 865.9 individuals/40 ha (31 and 40 species, respectively) in two con- secutive years in unlogged White Mountain mixed-conifer forest. A logged stand support- ed 32 and 41 species at 544.0 and 758.0 indi- viduals/40 ha, respectively, in the same two years. Brotherson et al. (1981) reported an early July density of 280 individuals/40 ha (26 species) in an aspen-mi.xed-conifer woodland in nearby Betatakin Canyon. The mixed-conifer association and its bird species make northern Black Mesa unique. Most of the Navajo and Hopi reservations are dominated by arid deserts and semiarid grass- lands, shrublands, and pinyon-juniper wood- land. There are few montane "islands " rising 1994] Birds of Northern Black Mesa, Arizona 27 Tahlk 12. Breeding bird densities-' of tliree pin\'on-juniper woodland stands on the Blaek Mesa leasehold (ni pairs/40 ha). .[- 10 .1- 20 .1-: 30 Species 19S:3 1984 1983 1984 1983 1984 American Kestrel -1- -1- — — — — Mourning Do\'e + 3.8 + + — — Common Nighthawk + — + — — — Black-chinned I himmingliird + + + -1- -1- + Ilaiiy Woodpecker + + + 1.9 -1- 3.8 Gray FKcatcher 7.6 7.6 7.6 11.5 11.5 9.5 Ash-throated FKcatcher 7.6 5.7 3.8 3.8 3.8 3.8 Scrub Jay + — 3.8 1.9 3.8 1.9 Pinyon Jav' -1- + + -1- -f- -t- Mountain Chickadee 3.8 3.8 7.6 7.6 11.5 9.5 Plain Titmouse 7.6 11.5 7.6 7.6 7.6 11.5 Bushtit 3.8 5.7 — 3.8 -1- 3.8 White-hreasted Nuthatch — + 7.6 3.8 11.5 5.7 Rock Wren + — — — — — Bewick's Wren 11.5 11.5 19.1 15.3 11.5 19.1 Blue-gra\' Gnatcatcher — -1- 3.8 — 3.8 3.8 Western Bluebird — — -1- — 3.8 3.8 Mountain Bluebird 3.8 1.9 3.8 7.6 3.8 — Hermit Thrush — — — — — -1- Solitan' Vireo 3.8 1.9 7.6 3.8 3.8 3.8 Black-throated Gray Warbler 7.6 -1- 11.5 11.5 15.3 15.3 Rufous-sided Towhee — — — — — 3.8 Chipping Sparrow 7.6 7.6 7.6 11.5 11.5 7.6 Black-throated Sparrow + — + — — — Brown-headed Cowbird + -1- + + -1- + House Finch 3.8 3.8 + 3.8 + + Total densitt 68.5 64.8 91.4 95.4 103.2 106.7 Total number of species 21 19 21 18 19 20 ■'Spot-mappccl. 1983-84. "•"Pifsi'Tit Init ck'iisit\ uiick'ttniiincd Ix'c :)l insufficKiit (l.ila. out of the dry lowlands. In the vast region bordered by the Little Colorado, Colorado, and San Juan rivers and the New Mexico state line, only Defiance Plateau, Navajo Mountain, and the Carrizo-Lukachukai-Chuska Moun- tain chain rise high enough to support exten- sive montane coniferous forests. The mixed- conifer woodland on Black Mesa is restricted and isolated but is sufficient to allow such montane species as Clark's Nutcrackers, Brown Creepers, Hennit Thmshes, and Yellow- rumped Warblers to breed within a few kilo- meters of the arid flats of Monument Valley. Additional Discussion The highest portions of northern Black Mesa support a rich association of montane breeding birds compared to other insular montane habitats in the Great Basin and Col- orado Plateau. The mixed-conifer woodland alone supports 59 species, 32 of which are pennanent residents. Johnson (1975) reported a mean of 37 species from a series of 31 mon- tane islands in the Great Basin. The mean number of permanent resident species was 7.39. Behle (1978), in a similar analysis of 14 montane islands in Utah, had a mean of 42.7 species per island (17.0 permanent residents). All of those designated as "widespread species" (occurring on all islands) in both studies are known or suspected breeding species on Black Mesa. Both authors found the number of breeding species to be positively and signif- icantly correlated with habitat diversity. The remarkable breeding bird species rich- ness on northern Black Mesa is related to the number and types of habitats occurring in the area. A well-documented concept in ecology states that increasing structural diversity of a plant community allows bird species diversity to increase (MacArthur and MacArthur 1961, Mac-Arthur 1964, Karr and Roth 1971, Willson 1974, Roth 1976). If it is allowed that breed- ing bird species richness of a given habitat 28 Great Basin Naturalist [Volume 54 ■ \» V ^ ^. ^^r '-r;i, Fig. 17. Mixt'd-conifer vvoodlaiid, Loloinai Point, August 1987 corresponds to the breeding bird species diversity of that habitat, then a pronounced habitat complexity-l^reeding bird species diversity gradient is well illustrated on Black Mesa (Tables 1, 2). The structural simplicity of the scrub habitats allows utilization of them by fewer than 12 breeding bird species, with a given site generally supporting 3-4 species. The pinyon-juniper woodland bird diversity is far greater than the next nearest habitats. Mixed-conifer woodland supports the greatest diversity of breeding species for several rea- sons: greater stature and complexity inclusion of the pinyon-juniper woodland component, and the relatively pristine decadent old- growth nature of portions of the habitat. Reclaimed mine spoil, juniper savanna, chained pinyon-juniper, and mixed-shrub habitats have higher species richness values than monotypic shrublands because each includes characteristics and/or plants species of other habitats. Species richness of breeding bird associa- tions of several habitats is suiprisingly depau- perate. Riparian habitats and ponds that are typically used by many bird species support only 7 and 9 breeding species, respectively. Factors responsible ior this include the young age of the ponds, small size, isolated nature of both types, heavy livestock grazing in both types, and apparent inability of riparian breeding bird species to utilize the principal riparian plant species in the area, the exotic tamarisk. In pure montane scrub only 5 species are known to breed. In the relatively complex aspen groves, 14 species are known to breed. The small extent and isolated nature of the latter habitat probably prevent utiliza- tion of more species. The montane scrub, although appearing more complex than the Great Basin shrublands, is essentially a shrub- land of restricted regional occurrence; again, isolation may prevent use by more species. Faunal resemblance factors (Table 13) show limited overlap between breeding bird associ- ations in the study area. Half of the pairs of habitats in Table 13 show no faunal resem- blance (e.g., share no species). Approximately half of those pairs sharing species have factors below 0.21. The greatest resemblance factor is between sagebrush and saltbush shrublands (0.80), which probably reflects bird response to the very similar growth form of the two shrubs. Greasewood shrubland and juniper 1994] Birds of Northern Black Mesa, Arizona 29 Tablk 13. Faiinal resemblance factors of the breeding bird communities of northern Black Mesa, Arizona" Montane scrul) 5 — — — — — .12 Sagebrush — 5 .SO .IS ..50 .31 — Saltbush — 4 5 .IS .67 .15 — Gicasevvood — 1 1 6 .31 .57 — Reclaimed spoil — 3 4 2 7 .40 — Juniper savanna — 2 1 4 3 8 — Riparian 1 — — — — — 9 Ponds — — — — — — 2 Chained pinson- juniper 3 2 2 1 2 1 1 Mixed shrub — 4 4 6 5 6 — Aspen gro\es 3 — — — — — — Cliffs, talus, banks — — — — — — — l^in\()n-juniper 2 — — — 1 2 4 Mixed-conifer 4 — — — 1 1 2 .13 .25 .47 — — — — .25 .47 — — — — .12 .67 — — — — .22 .53 — — .04 .03 .11 .60 — — .OS .03 .10 — — — .16 .06 11 .IS .OS .13 .19 .14 1 12 — .13 .04 — 1 — 14 .06 .11 ..30 2 2 1 20 .29 .20 5 1 3 9 42 .61 5 — 11 10 31 59 ■'Ik'semblancc factor Inght ot diagonal) follows 1 lofTnicistc-r ! 19.S(i) and is dern ed b\ douliling tlif luinihcr of species in coninion (left of diagonal I and iln idnig !n flic total number of species in each community (diagonal). savanna have a resemblance factor of 0.57. This rehitively high vahie is also probably related to bird response to structurally similar habitats. Principal plants in these habitats are larger, densely foliaged shrubs and small trees, respectively. The relatively high faunal resemblance factors of the juniper savanna and mixed-shrub (0.60), saltbush and reclaimed mine spoil (0.67), greasewood and mixed shrubland (0.67), and pinyon-juniper and mixed-conifer woodlands (0.61) are all related to similar habitat structure and/or floristic composition. A few low faunal resemblance factors and their possible explanations are noteworthy. Pinyon-juniper and chained pinyon-juniper woodland have a resemblance factor of 0.19, reflecting the strong response of the breeding bird species composition to a pronounced phv'siognomic change. Furthermore, the value is this high because of four species that char- acterize the woodland edge and, therefore, utilize the openings created by chaining. Montane scrub shows no faunal resemblance with any of the Great Basin desert shrublands (sagebrush, saltbush, greasewood, and mixed shrublands). Although structinally similar, the lack of any faunal resemblance between these two shrubland types may be due to differences in their foliage growth form and geographical origins. The montane scrub is a high-elevation, cold-climate derivative (Brown 1982) of broad-leaved deciduous shrubs, while the Great Basin desert scrub communities are lower-elevation, desert-derived associations of sclerophyllous, often evergreen shrubs (Brown 1982). These differences are no doubt associated with the different avifaunas associ- ated with each. Indeed, foliage gleaners com- prise 80% of the breeding species in the mon- tane scrub, and ground-feeding species com- prise 78-100% of the species in the desert scrub habitats (Table 2). Ponds and riparian areas show a faunal resemblance only to each other, indicating a regional uniqueness of the associated breeding birds. With increasing habitat diversity and accompanying breeding bird species diversity, there is also an increase in the number of dif- ferent nesting site substrate and foraging sub- strate/mode guilds (Tables 1, 2). Ground- and foliage-nesting species are found in nearly all habitats. Cavity nesters, which utilize tree cavities, appear in tree-dominated habitats. Ledge nesters, logically, are dominant in cliffs, talus slopes, and wash banks. Ground-feeding species are found in all habitat t\pes and com- prise the principal guild in open, low-statured 30 Great Basin Naturalist [Volume 54 habitats. Foliage j^leauers are prominent in the habitats eomposed of trees. ENVT HON MENTAL CHANGES ON Black Mesa En\ironmental ehant^es during the past 2000 years on the Colorado Plateau, and on Blaek Mesa in particular, have been well stud- ied and documented. These changes have undoubtedly affected the avifauna of Black Mesa in many ways. The exact effect of many such changes on the bird life of the mesa, however, must remain conjectiual. A brief dis- cussion of several broad categories of change is presented below, along with comments on the effects such changes may have had on the birds of Black Mesa. Climatic Changes Climatic changes, specidated to be related primarily to changes in rainfall amounts, modes, or distribution, have been studied extensively in the Southwest (Euler et al. 1979). However, distribution of plant commu- nities in the Black Mesa region has been mini- mally affected by such changes in the past 2500 years (Dean 1989). The principal effect of precipitation change has been in levels of alluvial water tables (Karlstrom 1983). Euler et al. (1979) postulate that a 550-year cycle of rainfall changes and the accompanying water table fluctuations are responsible for repeated aggradation and degradation of alluvial deposits along wash courses throughout the southwestern United States. They believe degradation, or arroyo cutting and gullying, follows lowered water tables during periods ot low relative precipitation. Channels then refill with sediment as water tables rise during wet- ter periods. Dean (1989) states that significant arroyo cutting episodes occurred on Black Mesa around A. D. 225-250, 750-775, and 1275-1300. Cutting of the present widespread gully net- work began around 1880-90 (Thornthwaite et al. 1942) and by 1915 had reached current conditions in the Polacca Wash drainage ot eastern Black Mesa. Laguna Creek in neigh- boring Tsegi Canyon had gullied by at least 1918 (J. Wetherill unpublished letter to Tiilbot Hyde on file at American Museum of Natural History). Thus, in a period of only 25-35 )'ears water tables lowered and alluvial valleys had been extensively dissected. Any riparian vege- tation, particularly old cottonwood and willow growth, was probably drastically altered, if not eliminated, during this time. The pre\ ious occurrence of native riparian habitats in the study area is conjectural. Three cottonwood posts identified from a large Anasazi site on Moenkopi Wash are associated with a structure dating to about A.D. 869-876 (Sink et al. 1983). Virtually no cottonwoods are present in the upper 39 km (24 miles) of Moenkopi Wash today. It is possible that a cottonwood-willow association was present in the vicinity ol the Coal Mine and Moenkopi washes confluence where there are short reaches of perennial stream flow and dense, well-developed tamarisk thickets. An\' cotton- wood snags that may have existed after the cutting of the arroyo have disappeared. Karl- strom (1983) presents evidence that a series of small ponds occin^red in Yellow Water Canyon at about 2073 m (6800 ft) elevation. Small ponds may have existed elsewhere, again, par- ticularly along Coal Mine and Moenkopi washes. In 1896, Richard Wetherill reported ponds in the Tsegi Canyons similar to what may have occurred on a smaller scale in Yel- low Water Canyon. The canyon had "two lagoons . . . from cliff to cliff about one mile apart and each one a mile long and about 300 yards wide. Ducks are plent}' on these lakes" (R. Wetherill unpublished report on file at the American Museum of Natural Histoiy). Bird species affected by the reduction or elimination of riparian habitats would include all species closely associated with these habi- tats throughout northeastern Arizona (see Woodbiu)' and Russell 1945, Monson and Phillips 1981, and Hunter et al. 1987). Among these are the Western Kingbird, Yellow War- bler, Blue Grosbeak, Lazuli Bunting, Yellow- breasted Chat, and Northern Oriole. The dis- appearance of ponds would also have elimi- nated shorebirds and waterfowl and affected species characteristic of emergent vegetation such as the Marsh Wren, Common Yel- lowthroat, and Red-winged Blackbird. Brad- field (1974) discussed severe impacts of gully- ing on shorebirds and waterfowl on southern Black Mesa at Oraibi Wash. It is possible that the Black-billed Magpie, which disappeared from most of northeastern Arizona in the late 1800s (Woodbur)' and Russell 1945), was elim- inated by alteration of riparian areas due to 1994] Birds of Northern Black Mesa, Arizona 31 the current arroyo-cutting episode that began abont the time of the birds disappearance. Where this species does currently occur in extreme northeastern Arizona, it is associated with well-developed riparian growth (Jacobs 1986), indicating a dependence on such areas. Several bird species typical of rock talus, cliffs, and ledges have moved to the exposed dirt banks along arroyos which exhibit atten- dant holes, caves, and crevices. Say's Phoebes, Northern Rough-winged Swallows, and Rock Wrens are characteristic of wash banks. Amer- ican Kestrels, Great Horned Owls, Northern Flickers, Violet-green Swallows, and Moun- tain Bluebirds have been found to a lesser extent nesting in wash banks. Exploitation by Prehistoric Man Bird species identified from remains exca- vated from Anasazi archaeological sites are presented in Table 14. Excavation of these sites was conducted during the Black Mesa Archaeological Project (1968-83) as part of clearance procedures preliminary to mining operations by Peabody Coal Company. Six (39%) of the species identified were not found during this study (Table 14). It is striking that, as a group, Galliformes are so well represent- ed. The Wild Turkey was obviously extensive- ly utilized for a long period, and overuse by the Anasazi may have eliminated it from the area if it was not brought in as a domesticated species. The Scaled Quail may have been recently extiipated from the area by overgraz- ing (see below), but there are apparently no recent records of it for the region. Pat Ryan (personal communication), however, states that Scaled Quail are still present south of Black Mesa in the northern Hopi Buttes area. The Bobwhite may be a misidentification. Raptors as a group are also well represent- ed. The symbolism accorded raptors and their use by the Anasazi were probably similar to those by both present-day Hopi and Navajo. Rea (n.d.) indicates the broken and healed left radius and ulna of an immature Prairie Falcon from site AZ:D:7:98 are strong evidence that the bird was taken as a nestling and held in captivity. Several raptor species are currently taken from nests in the region by the Hopi (personal obsenation). Raptors have probabK' been used by Native Americans in this region nearly continuously for at least the last 1900 years. Periods of especially intensive use may Tabi,!': 14. Bird remains from Anasazi sites excavated during the Black Mesa Archaeological Project, 1968-83'' Specii's No. sites Approximate age(s) Cooper s Hawk 2 A.D. 100-300, SOO-1090 Red-tailed Hawk 3 A.D. 800-1090, 1100, 850-1150'' Biiteu sp. 4 A.D. 600-1100, 850-975, 1100 Eagle sp. (Aqiiila or Ualiaeetus) 3 A.D. 600-1100, 800-1030, 1070-1150 American Kestrel 2 A.D. 800-10.30, 850-975 Prairie Falcon 1 A.D. 850-975 \Vild Turke>-^ 20 A.D. 100-300, 800-1150 Northern Bohwhite^' 1 A.D. 1100 Scaled QuaiH 1 A.D. 1100-1150 Gambel's QiiaiU 1 A.D. 800-1090 Quail sp.*^ 1 A.D. 800-1030 Sandhill Crane'^^ 1 A.D. 800-1090 Mourning Dove 1 A.D. 800-900 Screech Owl sp. 2 A.D. 100-300, 1100 Otus sp. 1 A.D. 800-1090 Great Horned Owl 3 A.D. 600-1100, 800-10.30, 1100 BiuTowing Owl^ 1 A.D. 800-1030 Northern Flicker 3 A.D. 600-1100, 1100, one undated Horned Lark 1 A.D. 100-300 Scrub Jay 2 A.D. 100-.300, 600-1100 Pin\()n Jay 5 A.D. 600-750, 600-1100, 850-975, 850-1150, 600-1100 American Crow 1 A.D. 850-975 Clark's Nutcracker 1 A.D. 600-1100 Common Raven 2 A.D. 850-975, 1100 "Based on tlie following reports: Olsen 1972, Rea n.d.. Beezlcy 1974. Seme 1980, Seme son and Pern,- 198.5, and Leonard 1989. ''Anasazi occupation of northern Black Mesa ceased .\.D. 1150 (Gummerman 1984). 'Not found on Black Mesa during this stud\', 1979-9.3. 1982, SinilcN ct al. 1983. Nichols and Smiles 1984, Christen- 32 Great Basin Naturalist [Volume 54 have lowered the densities of some desirai>le or readily available species (e.g., Red-tailed Hawk and (iolden Eagle). The Ferruginous Hawk was probably eliminated from the region by human exploitation (Hall et al. 1988). Lixestock Grazing With their anival in 1540, the Spanish intro- duced livestock to the Southwest. However, extensive use of livestock did not develop in the Black Mesa area until after 1868 (Thornth- waite 1942). Excessive overgrazing had, by the 1930s, become such a severe problem on the Navajo Reservation that a livestock reduc- tion program was initiated by Collier in 1934-37 (Philp 1977). Overgrazing has brought about widespread changes in the bird life of Arizona (Phillips et al. 1964). Thorn- waite et al. (1942) have cited it as a cause of excessive erosion and gidlying in northeastern Arizona. While Euler et al. (1979) believe arroyo cutting is a cyclic natmal phenomenon associ- ated with precipitation cycles, Thornthwaite et al. (1942) may be correct in concluding that overgrazing initiated the current arroyo-ciit- ting episode. Under the influence of contin- ued, heavy, year-long grazing and the succes- sive replacement and supplanting of the rela- tively well-developed root systems of perenni- als with the shallow systems of weedy annu- als, regional erosion will continue. As long as this situation remains, the cycle of alluvial aggradation and degradation may likely be eclipsed in its current state. Grazing can have a significant impact on the plant community in which it occurs and in turn can affect the community's birds (Wiens 1973, Bock and Webb 1984). Wiens (1973) found that across a series of grassland types grazing caused "a uniform directional change towards dominance by plant species charac- teristic of drier climates. " In addition to floris- tic changes in the community, grazing may produce marked physiognomic changes (Wiens 1973), such as reducing shrubs in riparian habitats (Taylor 1986) and increasing shrubs in grasslands (Phillips et al. 1964). Grazing can cause changes in the arthropod fauna of grasslands (Smith 1940), which could in turn cause a change in bird species compo- sition (Wiens and Rotenbeny 1981, Rotenber- ry 1985). Grazing has contributed to the spread of pinyon-juniper woodlands into grasslands during the past centuiy (West et al. 1975). Grazing may cause changes in bird species composition with little change in over- all bird density (Wiens 1973, Medin 1986), reduce bird species richness and density (Monson 1941, Wiens 1973, Taylor 1986), or even cause increases in density (Bock et al. 1984). Monson (1941) found that elimination of grazing and initiation of revegetation efforts in a grassland/shrubland site caused bird den- sity to nearl) double. Horned Lark densities increased on grazed grassland and shmbland sites (Wiens 1973, Bock and Webb 1984, Medin 1986). Western Meadowlark densities are reduced by grazing (Monson 1941, Wiens 1973). Bock and Webb (1984) and Bock et al. (1984) found Mourning Doves, Horned Larks, Mockingbirds, and Lark Sparrows to be sig- nificantly more numerous on a grazed grass- land site than on an ungrazed site. The contin- uous, year-long grazing (by up to five species of livestock) typically practiced in the Black Mesa region has no doidit reduced prey species populations to the detriment of sever- al species of raptors (Kochert et al. 1988). Introduction of Exotic Vegetation Plant species not native to the southwest- ern United States (or North America) are pres- ent in the Black Mesa region. Some arrived accidentally, while others were introduced for a variety of reasons. Species, both plant and animal, when introduced into a new region frequently increase rapidK' in the absence of controlling factors with which they evolved in their native regions. Great disruptions in the numbers and composition of native flora and fauna can result where exotic species achieve dominance. The most conspicuous exotics within the study site are tamarisk {Tainarix chinensis), Russian thistle {Salsola iberica), and cheatgrass {Bromus tectorum). Other species include Russian olive {Eleagnus angustifoUa), Siberian elm {Uhniis piimila), filaree {Erodium ciciifarium), chorispora {Cho- rispora tenella), and summer cypress {Kochia scoparia). The greatest impact of exotics on the bird life of the mesa probably occurs in riparian habitats and in grasslands. It is in such habi- tats that exotics are most conspicuous and dominant. Tamarisk is widely established along nearly all major washes in the region. Most stands have spread and grown up 1994] Birds of Northern Black Mesa, Arizona 33 through natural dispersal, hut in some places the species was planted hy man, as at Keams Canyon (G. Monson personal communica- tion). Preferential grazing of young cotton- woods and willows by livestock helps con- tribute to the monotypic tamarisk stands typi- cal of the study area. On Black Mesa it is well developed along lower reaches of the major washes (see above). In this region few breeding birds of native (cottonwood-willow) riparian stands breed in tamarisk. Migrant species forage extensively in tamarisk, and a few species, especially Dark-eyed Juncos, winter in it. Russian olive, also planted extensively in nearby areas, is dispersing rapidly in riparian habitats in the Intermountain West (Knopf and Olson 1984). Russian olives are currently well established and advancing up Moenkopi Wash below the confluence with Coal Mine Wash. Breeding bird use of mature Russian olive stands is lim- ited, but winter use of fruits is extensive (B. Jacobs personal communication concerning stands in the Chinle Valley and personal observation at Keams Canyon; unpublished Black Mesa data). Tamarisk-dominated ripari- an strands in the study site support few ripari- an breeding obligates (e.g.. Western King- birds, Yellow-breasted Chats, orioles, and buntings) but are intensively used by insectiv- orous migrants. As the Russian olive grove on Moenkopi Wash matured during the 1980s, numerous species appeared during winter, relying on the fruit as a food source. These species included Downy Woodpeckers, flick- ers, ravens, Mountain Bluebirds, robins, star- lings. Cedar Waxwings, Yellow-rumped War- blers, White-crowned Sparrows, and Evening Grosbeaks. Grassland habitats and disturbed sites in the study area are frequently composed of several e.xotic plant species. Effects on bird life in the study area are poorly known. It can be assumed, however, that there haxe been alter- ations in numbers, kinds, and distributions of grassland birds as a result. Cheatgrass is prevalent in juniper savanna at the mesa foot, and especially in saltbush stands. Russian thistle occurs abundantly on disturbed sites and is frequent on Mancos Shale slopes. Sum- mer cypress is abundant early in the reclama- tion process. Horned Larks and Dark-eyed Juncos may feed extensively on Russian this- tle and summer cypress during the winter. particularly following storms when these plants are the only abimdant food source pro- tniding above the snow. Pinyon-Juniper Type Conversions The purpose, methods, results, and areas of pinyon-juniper control on the study site were discussed previously. In any type conversion, the effect on wildlife must be considered radi- cal. Impacts are especially pronounced when a woodland physiognomy is converted to that of a grassland or simple shrubland. In the case of birds, a nearly complete change in the com- position of species occurs. In any particular converted site most woodland obligates are eliminated, and grassland and shmbland species invade the newly created openings. A few species absent or sparse in surrounding habi- tats may find favorable conditions in convert- ed areas and experience population increases. Bird species typically eliminated from woodland stands that have been converted on the study site include the Haiiy Woodpecker, Gray Flycatcher, Mountain Chickadee, Plain Titmouse, White-breasted Nuthatch, Bewick s Wren, Blue-gray Gnatcatcher, Solitary Vireo, and Black-throated Gray Warbler. The majori- ty of the above species are inhabitants of high- er-elevation woodland stands where most conversion projects were conducted. Virtually all cavit\'-nesting species are eliminated. The effect of chaining on the Spotted Owl may be especially severe, particularly in the Lolomai Point area where the woodland was eliminat- ed on the mesa tops adjacent to numerous small, mixed-conifer-filled canyons. Several species characteristic of grasslands and shniblands and of woodland edge invade or increase as a result of woodland conver- sions. Chipping and Vesper Sparrows have invaded several chained sites in the study area. Mountain Bluebirds are common in chained sites, especially on Lolomai and Kayenta points. This open-area species proba- bly increased in these areas while Western Bluebirds, a species typical of higher, dense stands, was reduced in numbers. In older chained areas where Gambel oak is estab- lished (see Fig. 12), Rufous-sided Towhees have increased. The Green-tailed Towhee breeds in the study area primarily where invading oaks are established on Lolomai Point. This area has also been colonized by Virginia's Warblers, which are found naturally 34 Great Basin Natur.\mst [\blunie 54 in moiitune scrub. Mountain paiklike areas with scattered ponderosa pines are the only areas where Acorn Woodpeckers have l)een found and where Lewis Woodpeckers breed in the study area. These species may have been present before the clearings were made but have certainly increased since. These parklike areas are also readib' hunted b\ Red- tailed Hawks, which are infre(|iient in the neighboring dense woodland. The Rock Wren, suii^risingly, is frequently observed in cleared woodland areas where dead trees are left scattered. The ph\ siognomy of chained areas must bear enough resemblance to the open talus slopes this wren inhabits to allow it to occasionally utilize them. The Rock Wren has not, however, been found nesting in cleared woodland on Black Mesa. Ridgway obsei'ved this species in timber slash piles in California and Nevada in the late 1800s (in Ryser 19S5). Sedgwick and Ryder (1987) report Rock Wrens breeding in chained woodland in Colorado. Sedgwick and Ryder (1987) quantified the impacts on birds caused by chaining pinyon- juniper woodland in northwestern Colorado. Their results support my qualitative evalua- tion discussed above and can probably be applied to Black Mesa. Their study revealed that chaining caused an alteration in bird species composition and declines in overall bird use, density, species richness, and species diversity. They noted that species which were cavity or foliage nesters, foliage and bark gleaners, and aerial feeders underwent declines on a chained woodland plot. Specific species that declined in abundance due to chaining included the Hairy Woodpecker, Gray Flycatcher, Moimtain Chickadee, Plain Titmouse, White-breasted Nuthatch, Solitary Vireo, and Black-throated Gray Warbler. The Mountain Bluebirds and Chipping Sparrows that utilized the chained plot accounted "for 48 percent of the avifaima. They found the Rock Wren breeding on the chained plot and said it was a "common" breeder even in chain- ings larger than 500 ha (1236 ac). In addition, species that "foraged and/or nested on the ground were less affected by the chaining process" than other groups. Surface Mining Activities Extensive deposits of subsurface coal made northern Black Mesa attractive for coal extrac- tion. The opening of the first road o\er the northern rim of the mesa was associated with development of small underground mines (Johnston 1932). Peabody Coal Company began actual surface mining operations on its 25,900-ha (64,000-ac) lease in 1971 after sev- eral years of legal negotiations and prepara- tion. Currently, about 243 ha (600 ac) is mined annually. Reclamation practices, like woodland elim- ination, function ecologically as a habitat type conversion. Areas where native woodlands and shrublands stood before luining are plant- ed to grasslands. This grassland-dominated range is developed to meet the primaiy desig- nated postmining land use of livestock raising. Areas reclaiiued as such are structmalK* sim- ple, homogeneous communities and therefore do not support avifaunas as rich as the neigh- boring woodlands and shrublands. Such dras- tic geological (Hall 1983) and ecological changes in any surface-mined tract will markedly alter faunal assemblages found in them. Pinyon-juniper woodland, sagelirush shrubland, and saltbush shrubland are the pri- mary vegetation communities affected by mining. Leasewide, the areas mined or to be mined are covered by about 65% pinyon- juniper woodland, 30% sagebrush shrubland, and 5% saltbush shrubland. The total area reclaimed to grassland and grass-shrubland will be somewhat less than the total overall distiu-bance acreage since some roads, ponds, and facilities will be retained after mining. At the end of mining in about 2011, approximate- Iv 9771 ha (24,144 ac) will have been dis- turbed, of which about 8931 ha (22,069 ac) will be reclaimed. The total area of each habi- tat disturbed will be about 6238 ha (15,414 ac) of pinyon-juniper woodland, 3241 ha (8009 ac) of sagebrush shrubland, and 292 ha (721 ac) of saltbush shrubland. Until 1986, about 4295 ha (10,613 ac) had been disturbed. Bird species richness and density' generally decrease from native habitats to reclaimed areas. Howe\'er, where shrubs are reestab- lished to sufficient densities at about 2720/ha (1100/ac) or higher, breeding bird richness and densit> ma> approach that of the shrub- land habitats, particularly the mixed shrub- land (see Tables 1-7). Several studies have dealt with axian com- munities on reclaimed mine spoil. Apparently, 1994] Birds of Northern Black Mesa, Arizona 35 few dealing with western sites have been piib- hshed. Wray et al. (1982) found that sparrow productivity on a reclaimed site in West Vir- ginia was insufficient to maintain their popu- lations. Karr (1968) reported that the presence of water and diverse topography (in the form of ungraded spoil banks) greatly increased avian diversity on an abandoned mine site in Illinois. Krementz and Sauer (1982) compared a reclaimed site to an undisturbed desert scrub site in Wyoming and found that bird diversity was lower on the reclaimed site. In all, 12 species in 6 foraging guilds used the reclaimed site, while 37 species in 11 guilds occurred on the native site. Ground-gleaning guilds were predominant on the former. Horned Larks dominated the reclaimed site and were the only species breeding in it. Their reclaimed area was dominated by halogeton {Halogeton glomeratus), a weedy annual. Differences in avian commimities on their sites were attributed to habitat structure. Hickey and Mikol (1979) surveyed breeding birds on mine spoil in Montana and Wyoming and compared them to native grasslands and shrublands. Their reclaimed area bird densi- ties were lower than native sites with one exception and supported 4.2 species per site as compared to 4.0 and 8.0 species per site in grassland and sagebrush, respectively. A major change associated with mining activities on northern Black Mesa is the con- struction of numerous (over 150 currently) water impoundment and sedimentation struc- tures. Virtually all observations of shorebirds and waterfowl in the study area are a result of these man-made water impoundments. Red- winged Blackbirds nest only at these ponds. Only a few obsei^vations of Great Blue Heron, Black-crowned Night-heron, Mallard, Ginna- mon Teal, Killdeer, Solitary Sandpiper, and Common Snipe have been recorded away from the ponds in perennial reaches of Goal Mine and Moenkopi washes. The largest impoundment, created in 1973, is J-7 pond on Red Peak Valley Wash. Two other large dams are located in Reed Valley and in Wild Ram Valley. Very few impoundments are over 15 years old. Thus, only recently have waterfowl and shorebirds become frequent visitors on northern Black Mesa, although in numbers much lower than at ponds in the neighboring Klethla, Kayenta, and Ghinle valleys (personal obsei^vation). The Gommon Raven, European Starling, and House Sparrow have all increased on northern Black Mesa as a result of surface mining activities. The raven exploits garbage and may feed extensively on seed heads of grasses in reclaimed areas. The latter two species are present at mine shops and other support facilities. The Future As the previous discussion noted, both the status and distribution of birds on Black Mesa are constantly changing. The last two decades were a period of vmprecedented changes, the most pronounced of which were associated with habitat alterations residting from mining and type conversions. The increase in water- fowl and shorebirds utilizing the lease as a result of pond construction, while not great compared to major wetlands, may be greater than at any time in the past several thousand years. Gommon Ravens are certainly more abundant now than ever before. Starlings and House Sparrows, most likely arriving in num- bers only since 1970, have joined the Brown- headed Gowbird as recent emigrants. The Great-tailed Grackle may join these as well. For unknown reasons, the Gliff Swallow has apparently been recently lost as a breeding species. The chaining of nearly 3600 ha (10,000 ac) and mining and reclaiming of sev- eral thousand more hectares of woodland have allowed open-countn,' species to increase and woodland species to be locally eliminated. Tamarisk, whether one likes it or not, con- tinues to spread and is now the single most important riparian vegetation type in the region. It is here to sta\'. Roads have penetrat- ed farther into the upper canyons, and the clearing of pinyon-juniper woodland and sagebrush continues. Russian olive groves on lower Moenkopi Wash matured in a 7-year period to the point where a variety of winter- ing birds are now feeding on the abundant fruit and buntings were noted breeding in 1992. These recent trends will continue. The rate of increase in the number of ponds will become slower and eventually stop. As mining comes to an end, the number will begin to decrease as ponds are reclaimed. Gonsequently, numbers of ducks and shorebirds will decline. The land reclaimed at the end of mining will 36 Great Basin Natuiulist [Violunie 54 form a complex ot large grasslands. Open- country raptors such as Northern Harriers, Ferruginous Hawks, Rough-legged Hawks, and Merlins, as well as Northern Shrikes, will likely increase as migrants and/or winter resi- dents, utilizing the reclaimed landscape. With the continued spread of tamarisk and Russian olive, some species such as Northern Orioles may begin breeding along the washes. Robins, Yellow-rumped Warblers, and White-crowned Sparrows will continue to increase as winter residents in these areas. The number of Kill- deer breeding on the larger washes may decline if tamarisk continues to choke main channel beds. As grazing pressure mounts in the upper canyons, the understory of aspen groves and mixed-conifer woodlands will degrade further, with possible declines in scrub- and ground-dwelling species (e.g., MacGillivray's Warbler). Many species such as the Northern Goshawk, Spotted Owl, and other owls may be adversely affected if the last large, remote tracts of pinyon-juniper interspersed among the canyons of upper Coal Mine Wash are cleared. Many other changes to the bird life on Black Mesa will occiu; Although the majority of species in the study area will exist in peipetuity (with some declining and others increasing), the likely trend, as human environmental pressures rise, is a decline in overall abundance and diversity and the loss of some species as breeding resi- dents. I hope this report will help serve as a benchmark from which to gauge these coming changes both on Black Mesa and in the region as a whole. Species Accounts The 241 bird species identified from north- ern Black Mesa are treated individually in this section. Six species are known only from archaeological remains. Ninety-seven species are confirmed as breeding in the area. An additional 12 are suspected breeders, 10 of which almost certainly do nest in the area. Fort\-two families in 17 orders are represent- ed. Status, period of occurrence, and habitat preferences are discussed. Subspecies desig- nations are from Monson and Phillips (1981) and Behle (1985). Where these authors are in agreement, no citation is given. Where there is disagreement, differences are noted. Ranges of densities presented for smaller forms were extracted from the series of Peabody reports listed in the Literature Cited section. Order Gaviiformes Family Gamidae C]ommon Loon {Gavia immer). A fairly coninion migrant that has been observed only at J-7 pond. Most observations have been of single birds in alternate plimiage from early April to early May. A single basic-plnmaged bird was present at J-7 pond (S June-5 October 1984. Four were seen on 3 April 1986 and 12 on 14 April 1987. Fall records include one on 6, 9, and 12 November 1986 and one on 30 October 1987. Order Podicipediformes FaM I lA P( )DIC: I FED I DAE Pied-billed Grebe {Podilymbiis podiceps). A common migrant at J-7 pond and other larger impoundments. Spring migrants appear in mid- March and are mostly gone by mid-May. It is com- mon during the fall and into early winter until the ponds freeze. An adult with two young was observed on IS June 1990 at Pond N14-F Horned Grebe {Podiceps auritus). A sparse migrant. One was observed on 4 April 1982, anoth- er on 13 December 1982, both at J-7 pond. Eared Grebe (Podiceps nigricoUis). A common to abundant migrant at J-7 pond. Spring migrants appear from mid-March to late May with numbers peaking in mid-April. Small numbers have been seen in late summer. Fall birds appear in mid-Sep- tember and ]ia\e been recorded imtil mid-Decem- ber Western Grebe (Aechmophorits occidentalis). A fairh common migrant obserxed onl\ at J-7 pond. Spring migration records are from April only. Fall migration is from July to November and peaks in October No Clark's Grebes (A. chirkii) ha\'e been obseiA'ed. Order Pelecaniformes Family PELEf:AMDAE American White Pelican {Pelecanits erijth- rorhynchos). A sparse migrant. Four individuals were obsei\'ed at a pond in J-3 reclaimed area on 18 October 1984. Flocks observed at J-7 pond include 30 on 15 April 1986 (b\' B. Moreo), 1 on 8 October 1986, 215 on 13 October 1986, 27 on 5 October 1989, and 7 on 6 August 1990. FaM ILY PlIALACROCORACIDAE Double-crested Cormorant {Phalacrocorax auritus). A sparse transient. An immature individ- ual was observed at a freshwater storage pond 17-24 August 1990. 1994] Birds of Northern Black Mesa, Arizona 37 Order Ciconiiformes Family Ardeidae Great Blue Heron {Ardea herodias). A common migrant at J-7 pond and other impoundments throughout the lease. Rarely observed along the major washes. Spring migrants are seen from mid- Xhirch through \la\. Fall birds begin appearing in earl\' Jul\ and are seen until the ponds freeze in December. One wintered at the old Kaxenta Mine freshwater pond in 1984-85. Great Egret [Casmerodiiis albus). A sparse transient. One was seen at J-7 pond 25 Septem- ber-! October 1990. Snowy Egret [Egretta thula). A common migrant at J-7 and other ponds throughout the lease. Most are observed in April and May, August and September. Spring migrants have been seen imtil mid-June, and fall birds have been obsei-ved as early as late July. Cattle Egret {Bubulcus ibis). A sparse migrant. One was reported by EH&A Consultants on 1 May 1980, one was seen at J-7 pond on 19 October 1984, three were seen at N-2 reclaimed area on 28 April 1988, two were at Pond CW-A on 22 August 1990, one was at N-1 reclaimed area on 23 April 1991, and three were at J-7 pond on 24 April 1992. Black-crowned Night-heron {Nijciicorax nycti- corax). A faii^b' common migrant most often seen at J-7 pond, although an adult was seen in tamarisk on Moenkopi Wash 9 June and 27 August 1987. A dead immature in pre-basic molt was found on Coal Mine Wash at Navajo Rt 41 on 7 May 1985. Most records fall in May and from late August to mid-September. A lone immatme was seen in N-2 reclaimed area on 24 July 1992. Family Threskiormtiiidae White-faced Ibis {Plegadis chihi). A common migrant at ponds throughout the area. Most are observed in April, May, August, and September (with smaller numbers noted in June and July) as singles and in small flocks. However, 60 were at J-7 pond on 29 August 1986, 183 were at N-1 reclaimed area on 3 May 1989, 100+ were seen soaring over N-1 reclaimed area on 16 April 1991, and 70 were at J-7 pond on 22 April 1992. Order Anseriformes Family Axatidae Snow Goose {Chen caerulescens). A sparse migrant. One was present at the reclamation barn on 17 and 18 November 1982, a second bird was seen in N-6 reclaimed area on 18 November 1982, and a "blue" morph individual was seen in N-14 reclaimed area on 24 February' 1992 by C. Salt and S. Begay. Ross' Goose [Chen rossii). A sparse transient. An adult was seen at Pond N14-D on 9 November 1989, and two were at J-7 pond on 10 April 1992. Canada Goose (Branta canadensis). Primarily a fairly common migrant, l)ut 15 wintered at J-7 pond in 1983-84. The largest group recorded was a northbound flock of 60 birds on 15 February 1991. Most records are from November and December. Small flocks have been seen feeding on wheatgrass- es in reclaimed areas and resting on nearby ponds. Most birds seen are presumably B. c. inoffitti, but a much smaller form(s) is frequently seen with the t\'pical size birds. Wood Duck {Aix sponsa). A sparse migrant. A female was seen at J-7 pond on 6 December 1985. Green-winged Teal {Anas crecca). The most common migrant duck at ponds throughout the lease. Flocks of over 100 birds have been observed at J-7 pond. Most spring migrants pass through from mid-February to mid-April, but some are seen until mid-June. Males are most numerous early in this period and especially in late February and early March. Fall migration is primarily in August and September and is more drawn out than spring. A flock of near!) 100 birds was seen at J-7 pond on 6 December 1985 when males were in alternate phmiage. Mallard {Anas platyrhynchos). A common per- manent resident at ponds throughout the lease. Spring numbers peak in mid-March. Twenty nest- ings at 12 ponds were noted from 1986 through 1992. Male Mallards were at the Coal Mine- Moenkopi Wash confluence on 20 December 1988 and 15 November 1989. The Mallard's occurrence on the leasehold increased during the 1980s, prob- abh' due to increasing niunbers of sediment ponds. Northern Pintail {Anas acuta). An uncommon migrant. An unsuccessful nesting occurred at N-1 reclaimed area in May and June 1985. Pintails are more numerous as basic-plumaged, late-summer and early fall migrants. In the spring they are quite sparse but most frequent during mid-March. On 11 October 1985, 15 alternate-plumaged males were seen at J-7 pond. Blue-winged Teal {Anas discors). A fairly com- mon spring migrant from mid-March to earl\' May. Its fall status is imknown but it ma\' be more fre- quent then than in spring (Jacobs 1986). A pair of adults was seen in N-1 reclaimed area on 17 June 1992. Cinnamon Teal {Anas cyanoptera). A common spring migrant from mid-February to early May. Numbers peak in mid-April. A common fall migrant from July to October. Five birds were seen on Moenkopi Wish on 10 April 1992. Occasionally, small numbers may summer, but breeding has not been documented. Northern Shoveler {Anas clypeata). A common migrant throughout the lease lingering into December at J-7 pond. Spring numbers peak in 38 Great Basin Naturalist [Volume 54 niid-April. Eleven were at Heed \alle\ Dam on 1 June 1988. Gadwall {Anas strepera). A eonunon nii,u;rant in small numbers at ponds throughout the lease. Seen from mid-February to early May and from late July to mid-December. Spring observations peak markedly in early April. On 19 June 1989, 13 were seen at Reed Valley Dam. Two individuals wintered in 1990-91. Eurasian Wigeon [Anas penelope). Possibly a sparse migrant. Alternate-plumaged males in mi.xed Anas sp. flocks have been seen at J-7 pond (5 March 1982 and 17 Nhuch 1992), Reed Valle>- Dam (19 October 1988), and at ponds in the N-1 and N-2 reclaimed areas (10-27 November 1988). Another male, possibK' the same individual seen November 1988, was observed at several ponds leasewide 20 October-5 December 1989. American Wigeon {Anas americana). A fairly common spring and fall migrant, generally in small numbers, throughout the lease. Most fall occur- rences are from late Octolier to early December. Two individuals wintered in 1990-91. Canvasback {Aythya valisineria). An uncom- mon migrant as singles or pairs at J-7 pond and at the old Ka\enta Mine freshwater pond. It has been obser\ed from October through April (no Februar\ records). One male was seen .5 August 1984. Redhead [Aythya americana). A common migrant at J-7 and other larger ponds. Most birds pass through in March and April. Infrequent in the fall when most are seen in November Single males were seen on 26 June 1986 and 6 Jul\ 1987. Ring-necked Duck {Aythya coUaris). A com- mon migrant in small mnnbers seen most fretjuent- ly at J-7 pond. Most records are from mid-March to early May. Notable concentrations include 60 at J-7 pond on 2 December 1987, up to 40 that wintered in 1989-90, and 60 that wintered in 1990-91. Greater Scaup {Aythya marila). A sparse migrant. A pair was seen at J-7 pond on 2 April 1982. Lesser Scaup {Aythya affinis). A common migrant seen most frequentb' at J-7 pond. This is the most numerous of the diving ducks found on the lease. Most spring migrants pass through from early March to early May. It is less numerous in the fall when most records are from October and November. White-winged Scoter {Melanitta fusca). A sparse fall transient. A female was seen at J-7 pond on 30 October 1989. Common Goldeneye {Bucephala clangiihi). A sparse spring migrant recorded primariK during Vlarch. There are five records: two females on 1 March 1984, three males and two females on 19 March 1984, single females on 8 and 20 March 1989, and a female seen at a pond in N-2 reclaimed area 4 March-4 April 1991. Bufflehead {Bucephala albeola). A fairK com- mon spring and fall migrant. Seen uncommonly ihiough winter. Most birds pass througli as singles or in groups of two to si.\ indi\ iduals from March to mid-April. Most fall records are from late October and early Noxcmber. Hooded Merganser {Mergus cucullatus). A sparse migrant and winter resident. A male was seen on 17 November 1982 and a female on 3 November 1986, both at J-7 pond. (Jn 20 March 1989, one male and three females were at J-7 pond. A female wintered at a freshwater storage pond December 1989-February 1990. Another female was at J-7 pond on 7 November 1990. Common Merganser {Mergus merganser). A fairh common spring migrant at J-7 pond from mid-February to late April. Fall records include four females on 1.5 November 1982, seven females on 3 November 1986, and a single female 28 November-1 December 1988. Lone females were at Reed Vallex pond on 19 Jime 1989 and J-7 pond on 9 June 1993. Red-breasted Merganser {Mergus serrator). An uncommon spring migrant at J-7 pond with nearly all records from April. Seen twice in fall: a single female at J-7 pond on 17 November 1986 and six females there on 29 October 1987. Ruddy Duck (Oxyura jamaicensis). A fairly connnon to connnon spring migrant from February to early May. It is less numerous in the fall and early winter. Most observations consist of five or fewer individuals. Seen occasionally at J-7 pond during the summer. Order Falconiformes Family Cathartiedae Turkey Vulture {Cathartes aura). A common summer resident seen throughout the area. How- ever, numbers seem to have declined dining the 1980s. This species appears regularly in the first week of April (the earliest being 23 March 1993), and the last are seen in early October. The latest was one seen on Dinnebito Wash on 8 October 1982. Breeding is unconfirmed, but this species probably uses numerous caves in the upper canyons as nest sites. An Anasazi granary in VIoenkopi Wash appears to have been used as a nest site prior to 1987. In early Jul\' 1985 evidence was found of a small roost in an aspen grove in Coal Mine Wash. This roost was in use in late Ma>' 1986. Although occasionally seen feeding on small road- killed animals, most feeding observed has been on dead livestock, which is probably the primaiy food source of Turkey Vultures throughout the region. The breeding form in northeastern Arizona is not well known (see Monson and Phillips 1981). Family Accipitridal Osprey {Pandion haliaetus). A fairly common migrant observed widely throughout the area. 1994] Birds of Northern Black Mesa, Arizona 39 Ahout half of all records are from J-7 pond where thev have been observed feeding on green simfish {Lepoinis cyauelhis) and largeniouth bass {Microptenis salmoides). Seen three times on the mesa rim: Kayenta Point on 2 May 1985, Lolomai Point on 15 April 1987 and 1 April 1989. Most spring records are from April, with extreme dates of 25 March (1993) and 18 May (1988). Most fall occurrences are from September, but records range from 9 August (1990) to 15 October (1992). Bald Eagle {Haliaeetus leiicocephahis). A sparse earlx' winter transient. A single adult was observed in Coal Mine Wash on 16 December 1982, an immatme bird was seen over lower Yellow Water Canyon on 4 December 1984, and an imma- ture was seen over Lower Moenkopi Wash on 20 December 1988. B. Moreo saw two adults at J-7 pond in Januaiy of 1985. One was seen over Din- nebito Wash on 16 March 1993. Leonard (1989) reports Aqiiila I Haliaeetus remains from Anasazi sites e.xcaxated on the lease (Table 13). Northern Harrier (Circus cyaneus). A fairK common migrant and uncommon winter resident in open grasslands and shrublands throughout the area. Records range from 15 August (1991) to 1 May (1990). On 6 January 1982 one was observed feeding on a junco in the J-16 Mine Plan area. Sharp-shinned Hawk (Accipiter striatus). A fairly common permanent resident. As a breeding bird it is found in mixed-conifer and aspen-oak habitats in the upper canyons (seven (jf nine nests). Ellis (personal communication) found a nest in pinyon-juniper woodland in 1982. This nest held three eggs on 4 June and two young ready to fledge on 23 July. Two active nests were found in late May 1986: one in a Gambel oak fringing an aspen grove and one with three eggs on 26 May in Rocky Mountain juniper in dense mixed-conifer. Three other \'ocal birds were found in earlx' Jime 1986. It may be observed an\'\vhere throughout the rest of the year as a migrant, but particularly in Septem- ber and April. Late-spring migrants have been seen to mid-May. Bluebirds {Sialia sp.) accounted for 37% of the prey remains found below two perch sites in 1985 and 1986. Eight other prey species were identified from these remains. The breeding form is A. s. velox. Cooper's Hawk {Accipiter cooperi). A fairly common to common permanent resident through- out the area, but quite sparse in winter. Migrant individuals may be seen anywhere. Breeding pairs are widespread and restricted to higher-ele\'ation pinyon-juniper (where junipers appear to be select- ed for nest sites) and mixed-conifer woodland where it is common but inconspicuous. Several winters have no records. The first spring migrants may appear as early as late January and the latest pass through in early May. However, the majority arrive in mid-March and are frequently seen into April. Resident birds arrive during this period. with territorial individuals being noted from mid- April to mid-Ma\. Aerial courtship displays have been seen from late April to June. Eggs are appar- entl\' laid in mid- to late May and hatch in mid- to late June. Fledging takes place in mid- to late July, and possibly into early August for late clutches. Fledglings are conspicuous in or near breeding sites in August. Migrants are common from late August through October. Remains are reported from Anasazi sites excavated on the lease (Table 14). Northern Goshawk [Accipiter gentilis). A fairly common winter resident in wooded sites through- out the area. More numerous in the 1983-84 win- ter. Two nests were found in mixed-conifer wood- land: one with two young on 5 July 1987 and anoth- er on 20 June 1989. Nested in dense pinyon- juniper woodland in White House Valley in 1993. The breeding form is A. ^. atricapillus. Swainson s Hawk (Buteo swainsoni). A sparse migrant with foiu" records: one seen at J-7 pond on 16 April 1984, one over the mesa rim on 19 June 1985, an immature over N-2 reclaimed area on 30 October 1990, and one over Reed Valley on 23 Sep- tember 1991. Red-tailed Hawk {Buteo jamaicensis). A com- mon permanent resident throughout the area. Cliffs are preferred for nest sites where they are available. However, pinyons are preferentially selected over jimipers in cliffless areas of the pin- yon-jimiper woodland. Larger, taller trees such as ponderosa pine and Douglas fir that occur in mixed-conifer/pinyon-juniper woodland ecotone are selected for nest sites in areas where suitable cliffs are unavailable. An estimated 12-18 pairs nested in or near the lease from 1982 to 1984 and again in 1992. Perhaps 75 pairs may nest in the study area during peak prey years. Egg laying occurs from late March to early May and, in one exceptional case, early March. Cottontails {Sylvila- gus sp.) are the principal prey species, comprising 63% of prey taken (based on nest remains and observed kills). A nearly complete lack of nesting attempts in 1985 and the simultaneous reduction in number of adults observed from 1985 through 1988 (particularh' in winter) may be related to very low cottontail numbers preceding and during this peri- od. All breeding attempts during this period appar- entl>- failed. Low pre>' densities may have forced an adult to feed repeatedly on carrion (a road-killed dog) 7-9 February 1987. A "Harlan's" form was ob- served in White House Valley on 31 March 1983. The majority of breeding birds are light morphs. Only two dark morphs have been observed as breeding birds. Dark birds are more frequently obsei^ved as migrants. Three adults were obsen'ed at a nest containing young in late June 1984, indi- cating a possible helper role for one of them (see Santana et al. 1986). Collecting of young by Native Americans accoimts for some nesting mortality in 40 Crrat Basin Naturalist [Volume 54 the art'a. Remains have been identified from Anasa/.i sites in the lease (Table 14). B. J. raliini.s is the breedin.t; form. Ferruginous Hawk {Buteo regalis). An uncom- mon migrant. One was seen in upper Reed Valley on 30 March 1983. Single birds seen 27-29 Sep- tember 1983 at Dugout Valley, J-3 reclaimed area, and over lower Coal Mine Wash may have been the same individual. One was seen over Coal Mine Wash on 19 April 1985. Immatures were seen at J-27 reclaimed area on 10 September 1987 and at N-1 reclaimed area on 12 September 1988. Only one summer record: an immature seen over Moenkopi Wish at Navajo Rt 41 on 2 August 1988. A dark morph adult was seen hunting over N-l/N-2 reclaimed areas several times from 23 October 1988 to 3 January 1989. An immature was seen at N-1 reclaimed area on 2 May 1989. Possibly more common previousK'. Rough-legged Hawk (Buteo lagopits). Usually a sparse winter resident and spring transient. One was observed by C. Kling of Mariah, Inc. Consul- tants at N-1 reclaimed area on 4 December 1984. One was seen over N-1 reclaimed area on 30 March 1987. A male wintered at the N-1 and N-2 reclaimed areas in 1987-88. Perhaps as many as 12 different individuals, mostly immatures, wintered at the N-7/N-8, N-1, N-2, N-14, J-l/N-6, J-16, N-14, and J-21 reclaimed areas in 1988-89. The first birds were seen on 23 October and the last on 7 April 1989. During November one was observed feeding on small rodents disturbed and exposed by a bulldozer and reclamation seeding drill operating in N-7/N-8 reclaimed area. Golden Eagle (Aqiiila chrysaetos). A sparse permanent resident. Informants leport current col- lecting of this species in the region by Hopis. Prob- ably reduced in numbers from former times due to considerable use for religious and other practices by Native Americans. However, reduction of prey populations related to overgrazing may also play a lole in low eagle densities in the region (Kochert et al. 1988). Family Falconidae American Kestrel (Falco sparverius). A com- mon summer resident throughout the area. The uncommon November to mid-March wintering population is composed almost exclusively of males. This species may be abundant during peak migration periods in early April and late Septem- ber-early October. Nests have been found in cavi- ties of junipers, Douglas firs, and ponderosa pines, and in holes in alluvial banks and cliffs. Eggs and downy young have been seen in mid-Jime. Fledg- lings have been seen from late jime to early August with most seen in mid-July. During the fall migra- tion, kestrels feed extensively on grasshoppers in reclaimed areas. Horned Larks and Dark-eyed J un- cos appear to be imi)()rtant winter prey species. A female caught a Violet-green Swallow in flight over J-7 pond on 18 September 1985. Another adult female was seen feeding on a Horned Lark at N-1 reclaimed area on 2 August 1989. Other verte- brates known to have been taken include a fledg- ling Sage Sparrow, several Mexican voles {Microtus mexicanus), horned lizards {Phrtjnosona doiiglassi), and Sceloporiis sp. lizards. F. s. sparverius is the breeding form. Merlin (Falco columbarius). A fairly common winter resident in open habitats throughout the area. The 31 records to date (most of which are from reclaimed areas) span from 24 October (1984) to 26 April (1991). Horned Larks probably form a staple of the w inter diet, and hunting flights direct- ed at this abundant winter reclaimed-area species have occasionally been seen. Merlins have been observed feeding on Horned Larks near J-3 reclaimed area (7 January 1983) and J-16 reclaimed area (2 January 1989). Another was feeding on a Spizella sparrow at N-2 reclaimed area on 21 April 1988. VIost birds appear to be F. c. richardsoni. Peregrine Falcon (Falco peregrinus). A sparse transient on the leasehold. Two were seen hunting Horned Larks at N-1 reclaimed area on 20 June 1984. Another pursued two Bairds Sandpipers at N-2 on 3 August 1987. An immature female was seen chasing a Mourning Dove in Long House Val- ley on 17 September 1985. Ellis (1982) reports a substantial breeding population from throughout Arizona. Prairie Falcon (Falco mexicanus). An uncom- mon migrant and winter resident. Reported from a single Anasazi site (Tible 14) by Rea (n.d.). Sparse as a breeding species with onI\ two nesting sites known from the stud\' area. Order Galliformes Family Piiasianidae Chukar (Alectoris chukar). A sparse introduced species. Birds released at J-7 reclaimed area in the fall of 1982 (three seen 22 September) have appar- ently not survived. Chukars were heard and tracks were found in talus at the mesa foot south of Owl Spring Valley on 13, 19, 27 October 1985 and 3 Ma\ 1989. Three birds were seen at the same area on 5 September 1989. The\' are probabb' descen- dants of transplants made near Chilchinbito by F Taber in 1958 (Ryan personal conuiiunication). Wild Turkey (Meleagris gallopavo). Reports of introductions near the mesa rim in the 1981 mine peiniit application are apparently unsubstantiated. No recent evidence of their occurrence in the study area has been found. This is the most fre- (juentK- reported bird from excavated Anasazi sites that date almost continuoush' from A.D. 100 to 1150 (Table 14), suggesting its previous occurrence in the area, probably in the mi.xed-conifer habitats 1994] Birds of Northern Black Mesa, Arizona 41 near the rim. Hargrave (1970) and McKusick (1986) think that feral domesticated turkeys established populations in the southwestern United States. Northern Bobwhite {Coliniis virginianus). Known from a single 12th-century Anasazi site (Table 14). If not a misidentification, this may have been C. v. ridgway (Masked Bobwhite) perhaps brought in by prehistoric trading activities. Scaled Quail {CaUipepla sqiiamata). Known from a single 12th-century Anasazi site (Table 14). Gambel's Quail {CaUipepla gambelii). Known from a single Anasazi site. A CaUipepla/Loi)hurtyx (sic) determination is also reported (Table 14). Order Gruiformes Family Rallidae American Coot {Fulica americana). Common to abundant migrant at J-7 pond and other larger ponds from mid-March to early May and from late September to November. Smaller numbers are pres- ent in summer. It has nested at J-7 pond and spo- radicall)' at several others. Infrequent in winter Sora (Porzana Carolina). A sparse migrant with seven records from late July to late September: one seen in tamarisk at J-7 pond on 24 and 27 August 1986, an immature flushed from weeds at a pond in N-2 reclaimed area on 12 September 1988, single adults at a pond in N-2 reclaimed area and in tamarisk below J-7 pond on 12 September 1991, one heard at a pond in N-10 reclaimed area, one in N-1 reclaimed area on 30 September 1991, and another there on 24 JuK- 1992. Family Gruidae Sandhill Crane [Griis canadensis). Known from a single Anasazi site (Table 14). Order Charadriiformes Family Charadhiidae Semipalmated Plover {Charadrius semipalma- tus). A sparse spring migrant. Two were seen at J-7 pond on 9 Ma\' 1983. On 26 April 1984 five were seen at J-7 pond and 14 at a pond near the N-5 pit. Killdeer (Charadrius vociferus). A common permanent resident. It winters in small numbers and may leave during cold periods. Migrant birds pass through from early March to late April and in late September. Breeds at ponds and along the lower washes where densities of 1/620 m of wash bed were recorded over perennial reaches of Coal Mine and Moenkopi washes in late May 1986. Small yoimg have been seen from mid-May to mid- July. The breeding form is C. v. vociferus. Family Recurvirostridae Black-necked Stilt (Himantopus mexicanus). An uncommon migrant. Two were seen at J-7 pond on 15 April 1983, five were seen there on 6 May, and three more on 20 May 1983. On 27 June 1983 three were seen at a pond in the N-1 reclaimed area. A flock of 23 was seen at J-7 pond on 16 May 1986. American Avocet (Recurvirostra americana). An uncommon migrant. Six of the 10 records are from April, two from August, and one each from September and May. Usually seen in groups of five or less, but 19 were seen on 9 April 1982 at J-3 reclaimed area by B. R Dimfee. Has been seen only once since 1985. Family Scolopacidae Greater Yellowlegs (Tringa melanoleuca). A fairly common migrant as singles or pairs at ponds throughout the lease from late March to mid-April. A flock of 18 was seen at Kelly Pond on 14 April 1983. Seen in fall from late June to late November. Lesser Yellowlegs (Tringa flavipes). A common migrant more frequently seen than the preceding species. Most records are from late April and Sep- tember, but it has been observed in the region in all months between these periods. Solitary Sandpiper (Tringa solitaria). A fairly common migrant from early July to late September. Usually seen as single individuals at ponds throughout the area. Two were seen at the conflu- ence of Moenkopi and Yucca Flat washes on 13 August 1990. Two spring records: 23 April 1982 (two seen) and one on 19 April 1985. Willet (Catoptrophonis semipaJmatus). A fairly common migrant from mid-April to early May, and from early July through September. The largest flock seen was 34 birds at J-7 pond on 24 April 1992. Spotted Sandpiper (Actitis macularia). A com- mon migrant at ponds from mid-April to early June and mid-July to early October. Twenty seen at J-7 pond on 3 May 1991 were the largest recorded group. This species may breed at J-7 pond. Long-billed Curlew (Numenius americanus). A sparse migrant. A single bird was seen at a pond in N-1 reclaimed area on 27 August 1985. Three were seen on 26 April 1982. One was at N-14 reclaimed area on 22 May 1988 and one at N-1 reclaimed area on 27 June 1989. Marbled Godwit (Limosa fedoa). A fairK com- mon spring migrant. All observations are restricted to the brief period of 9 April (1982) to 27 April (1990). Western Sandpiper (Calidris mauri). A com- mon migrant at ponds throughout the area. Most are seen in late April and early May and from early July to September. Least Sandpiper (Calidris minutilla). A com- mon migrant at ponds in late April and early May and from earl\' July to early September. This species is considerably more frequent and numerous 42 Ghfat Basin Natur\list [Volume 54 than tlic prc'ct'dintz;. Lar.utT flocks arc seen occa- sioiialK : IN on IS April 19(S2 and 55 on 27 April 1984. Baird's Sandpiper {Calidris bairdii). A fairly coninion fall nii' pond on 2 November 1989. Ring-billed Gull (Lanis delaivarensis). C'om- mon spring migrant at ponds throughout the area from early March through mid-May. Adults are most numerous early in this period, and progres- sively younger birds appear as the season passes. Considerably less numerous and frequent in the fall; three at J-7 pond on 28 November 1988 were late. Seen as singles to flocks of 30 or more, although 107 were at J-7 pond on 28 March 1988. An earK fall individual was in Long House Valley on 3 August 1989. An immature seen at J-7 pond on 20 June 1983 was probabK- a late-spring migrant. California Gull (Larits californiciis) . An uncom- mon spring migrant with most records in March and April. Four fall records: one on 17 October 1983, two at J-7 pond on 25 October 1985, three at J-7 pond on 27 October 1986, one there on 29 October 1987. Usually seen as singles or pairs with flocks of Ring-billed Gulls, but seven were seen on 27 March 1984. Herring Gull (Larus argentatus). A sparse migrant. One was seen at J-7 pond on 14 April 1982. Common Tern (Sterna hiriindo). A sparse migrant. An adult and an immature were at J-7 pond on 9 September 1988. Three were seen at J-7 pond on 4 October 1983, si.x on 28 March 1984, and one on 20 April 1984. Two adults and an imma- ture were seen at Reed Valley (Pond J28-G) on 25 August 1989. Another was at Pond J16-A on 21 September 1990. Forster's Tern (Sterna forsteri). A fairly com- mon spring migrant at J-7 pond, with most records restricted to the brief period of 26-30 April, plus a single adult on 18 May 1992. Three fall records: one on 17 August 1983, one 9-12 September 1986, and a single immature at Reed Valle}' Dam on 9 September 1988. Black Tern (Chlidonias niger). A sparse late- summer migrant w ith four records: two at J-7 pond on 13 August 1984, two diffeient individuals at J-7 pond on 13-17 August 1990, and one at pond NIO- A on 10 September 1991. Order Columbiformes Family Columbidae Rock Dove (Columba livia). A sparse transient. Remains were found on the mesa rim on 1 August 1983. Three sight records: one at the mesa rim 15 miles east of the study area on 1 September 1984, another feeding along Navajo Rt 41 and Hwy 160 on 5-9 September 1989, and another at the Coal Mine Wash crossing of Na\'ajo Rt 41 on 15 March 1991. Rock Doves are present in Ka\enta and may stray into the stud\ area. Band-tailed Pigeon (Columba fasciata). Proba- bK a sparse transient. One seen on Lolomai Point on 22 June 1985. This species may increase in numbers as the oaks invading Lolomai Point and other chained areas fiuther increase in density and stature. Mourning Dove (Zenaida macroitra). A com- mon summer resident throughout the area. Winters sporadicalK' in small numbers at lower elevations. It usually arrives in early April and, except for stragglers, is gone by late September. A nest with 1994] Birds of Northern Black Mesa, Arizona 43 two eggs was found in J-7 reclainied area on 9 Ma\' 1985. Feathers of a fledgling were found on bellow Water Wash on 25 June 19S2. A breeding density- of 3.8 pairs/40 ha was found in pinyon-juniper at J- 10 mine plan area. The breeding form is Z. in. inar- ginella. Order Cuculiformes Family Clculid.^e Greater Roadrunner (Geococcijx californi- aniis). Probably a sparse permanent resident in brush}' valleys throughout the region. Nearly all Black Mesa records fall from August to late December. One was reported by B. Hector from lower Yucca Flat Wash (no date). J. Gilbert saw one at Black Mesa Junction 19 December 1984. One was seen in tamarisk on Moenkopi Wash on 27 August 1986. B. Clutter reported one in Dugout Valle\' on 8 August 1988. M. Koffler reported one on Moenkopi Wash on 15 September 1988. Anoth- er was seen in pinyon-juniper woodland in the J-1 mine plan area on 29 November 1988. One was seen on Navajo Rt 41 at Coal Mine Wash on 3 August 1989. C. Salt reported one on Navajo Rt 41 at Moenkopi Wash on 30 August 1989. Another was found road-killed in Long House Valley on 26 October 1989. The only spring record is one seen on 11 March 1993. The preponderance of late- summer through early winter observations from the study area and other areas of Black Mesa (per- sonal observation) suggests that rather long-dis- tance, post-breeding dispersal from primaiy breed- ing areas may be occurring. The Little Colorado River (Phillips et al. 1964) and the lower reaches of the Tusa\an washes (e.g., Moenkopi Wash) max be such breeding areas. Order Strigiformes Family Strigidae Flammulated Owl {Otiis flammeolus). An uncommon summer resident. One was seen in Upper Reed Valley on 16 October 1982. A nest with young was reported from Long House Valley on 16 June 1936 (Woodbury and Russell 1945). Calling birds have been heard in the east fork of Coal Mine Wash in a mi.\ed-conifer/oak/aspen habitat on 18 May 1986, on Lolomai Point on 31 Ma\ 1988, and in upper Moenkopi Wash on 12 May 1992. Another was seen in a cavity in an oak in the west fork on 9 June 1986. The breeding form is apparentlv undescribed (see Monson and Phillips 1981). Western Screech-Owl (Otits kennicottii). An uncommon permanent resident in pinyon jimiper woodland. Encountered primarily as road-kills: in Long House Valley 3 January 1983, at the foot of the mesa 4 miles west of Kayenta on 3 March 1984 and 3 December 1984, a molting adult in the J-21 mine plan area on 9 September 1988, and an adult female on Navajo Rt 41 near Yellow Water Wash on 15 July 1992. Remains were found in mixed-conifer on Coal Mine Wash on 1 July 1983 and one was seen in dense pinyon-juniper in Coal Mine Wash on 23 October 1984. Both the J-21 bird and the Navajo Rt 41 bird contained arthropods: six Jerusalem crickets [Stenopelmatus sp.) in the first and a large centipede in the latter The wing chords of these individuals were 171 mm for the J-21 bird (indicating a female) and 176 mm for the Navajo Rt 41 female, making both larger than average for birds reported from Arizona (Phillips et al. 1964) and larger than any known from Utah (Behle 1985). Reported from Anasazi sites as well as an Otus sp. determination (Table 14). The breeding form is O. k. aikeni. Great Horned Owl {Bubo virginianits). A fairly common permanent resident throughout the area. Nests have been found in old Common Raven and Red-tailed Hawk nests in trees and on ledges and potholes in cliffs in pinyon-juniper woodland on Yucca Flat Wash, Dinnebito Wash, and Moenkopi Wash. Nests on cliffs have been found in mixed- conifer in Coal Mine Wash (middle fork) and in Yel- low Water Can\ on. Eggs were seen on 5 April 1983 (J. Ohlman personal communication); half-grown \oung ha\'e been seen on 1 and 2 May 1983. Two nests foimd in 1989 contained three young, each on 18 May. Black Mesa birds show marked color varia- tion typical of the local race, B. v. pallescens. Wing chord and/or weights of seven individuals found road-killed over 10 years are as follows: male, 21 September 1984, 330 mm, 875 g; male, 25 June 1990, 356 mm, no weight; male, 25 July 1990, 359 mm, no weight; female, 24 August 1985, 380 mm, 1300 g; female, 23 January 1987, 390 mm, 1284 g; female, 29 June 1989, 378 mm, 1063 g; female, 11 October 1991, 380 mm, 1432 g. These measure- ments, especially those of females, indicate that Black Mesa birds (if the above individuals are not migrants) are larger than B. v. pallescens from near- by southern Utah (Behle 1985). Northern Pygmy- Owl {Glaucidium gnoma). An uncommon permanent resident. An increase in observations in fall indicates a possible influx of migrants. Most records are from mixed-conifer habitats. One was heard in Coal Mine Wash on 19 June 1982; another was seen on 20 June 1989 in the east fork of Coal Mine Wash. Two birds were seen and heard calling in Yellow Water Canyon on 8 August 1987, and another was seen on 30 Sep- tember 1987. Remains were found in Coal Mine Wash from a probable Accipiter kill on 12 May 1986. G. g. pinicola is the breeding form. Burrowing Owl (Speotyto cunicularia). Known only from a single Anasazi site (Table 14). I have seen this species just northeast of the study area, so it may be a sparse transient through it. 44 (iREAT Basin Naturalist [Volume 54 Spotted Owl iStrix occidentalis). At least a fairly common summer resident in shad\' mixed-conifer canyons and ravines. There are no winter records, hut it is prohahly a permanent resident (Ganey and Balda 1989). Molted feathers have heen found vir- tually across the extent of suitable habitat in the study area. All fi\e known nest sites are in caves in cliffs adjacent to mixed-conifer-filled canyon floors. La\inu; takes place from late March to early Ma>-. W'oodrats {Ncotoiiw sp.) are the principal prey species identified from pellets (Ganey 1992). S. o. lucida is the race on Black Mesa. Long-eared Owl [Asio otus). Sparse in pinyon- juniper and mixed-conifer. The only records are of one seen 29 March 1982 in J-28 mine plan area, one seen in White House Valley on 19 April 1983, remains of an adult found in a Great Horned Owl nest on 2 May 1983, a molted rectrix found in N-2 reclaimed area on 2 August 1989, and a fledgling and adult observed on 22 Jinie 1990 at an aban- doned Cooper's Hawk nest in pinyon-juniper in \\ hite House Valley. A single bird was in mixed- conifer on 18 June 1993. Northern Saw-whet Owl {Aegoliiis acadicus). Prior to 1993, known only from a single feather (identified by J. T. Marshall) found in mi.xed-conifer at the head of Yellow Water Canyon on 23 June 1985. Marshall (personal communication) says this species may move into an area for a few years, nest, and then disappear In the spring and sinnmer of 1993, 10 were foimd in mixed-conifer. Order Caprimulgiformes Family Caprimulgidae Common Nighthawk {Chordeiles minor). A common summer resident from early June to Sep- tember throughout the area. A nest with two eggs was found by C. Salt and S. Begay on 15 June 1989. A nest with one egg was found in pinyon-juniper on 1 August 1983 in the J- 10 mine plan area. Two downy young were foimd near J-7 pond on 10 July 1985. The subspecies C. in. hesperis and C. m. hen- ryi are reported from the area (Woodbury and Rus- sell 1945), with the latter breeding. One late record from Yellow Watei- Wash on 7 October 1986. Common Poorwill (Phalaenoptihis niittaUii). An uncommon sununer resident in pinyon-juniper. This species appears to be uncharacteristically sparse on northern Black Mesa compared with the southern region of the Hopi Reservation (personal observation). Most records come from the pinyon- juniper-covered benches on the outer mesa scarp and consist of the following; four to six heard call- ing northwest of Tees Spa Spring on 29 April 1985, two seen below Tees Spa Spring on 5 July 1987, three heard there on 7 July 1987, one heard near Black Mesa junction on 3 June 1991, and four heard below Rock Gap on 4 June 1991. The only records from the interior of the mesa are of one seen in middle Coal Mine Wash on 26 April 1983, one heard calling in mi.xed-conifer in Coal Mine Wash on 5 June 1986, two heard in upper Moenkopi Wash on 11 May 1992, and one flushed from tamarisk on lower Moenkopi Wash on 23 July 1992. The nominate race is the breeding form. Ordei- Apodiformes Family Apooidae White-throated Swift {Aeronaiites saxatalis). An abundant migrant and summer resident throughout the area from late March through mid- October. This species breeds in cliffs along the mesa rim and in the upper canyons. A steady stream of migrants was observed for several hours on 29 August 1984 at the rim where Navajo Rt 41 crosses. Thousands of migrants were seen over the lease area 21-25 September 1988. A. s. s(ix(it(dis is the lireeding form. Family Trociiilidae Black-chinned Hummingbird (Archilochus cdexandri). A conuuon summer resident in pinyon- juniper, along the major washes, and in mixed- conifer from late April to September A female was seen on a nest in a pinyon near the head of the middle fork of Coal Mine Wish on 19 June 1982. S. Hamilton found a nest on a greasewood root along an alluvial bank on Moenkopi Wash in August 1983. Calliope Hummingbird iSteUida calliope). A sparse migrant. One was seen among a swarm of Rufous Hummingbirds at a stand of bee plant {Cleome serndata) on Yellow Water Wash on 22 August 1989. An adult male was seen near the Yel- low Water Wash crossing of Navajo Rt 41 on 16 Jul) 1990. A small female hummingbird, which ma\- have been this species, was seen in the east fork of Coal Mine Wash on 20 June 1989. This date suggests this species may breed on Black Mesa. Broad-tailed Hummingbird (Selasphonis platycercus). A common migrant throughout the area and a common breeding resident in mixed- conifer. Records range from earK April to mid- Sep- tember, with the last northbound males noted to late May. Two females on nests with eggs were found in Douglas firs in Coal Mine Wash on 23 and 27 Jime 1983. A nest with >oung was found by M. Williams in a ponderosa pine on 17 June 1985 at Lolomai Point. Fall migrants appear in mid-July. Rufous Hummingbird (Selasphonis rufus). A common late-summer migrant. Usually first seen in mid-July, but may show up as early as late June. This species is often abundant at stands of Indian paintbrush (Castillcja linariaefoUa) and Rocky Mountain bee plant. 1994] Birds of Northern Black Mesa, Arizona 45 Order Coraciiformes Family Alcedinidae Belted Kingfisher {Ceryle alcyon). A fairly com- mon migrant from mid-April to early May and from mid-August to early October. Most records are from J-7 pond. Order Piciformes F.WIILV PlC:iD,\E Lewis' Woodpecker {Melanerpes lewis). A fairly common summer resident in chained woodland at Lolomai Point. Seen once in a chained area on Kayenta Point on 25 September 1984. Young being fed in a ponderosa pine snag on Lolomai Point were seen on 17 June 19S3 and 25 Jime 1986. An uncommon migrant elsewhere: three in pinyon- juniper at Dinnebito Wash on 8 April 1987, one above Black Mesa junction on 18 September 1987, one in pinyon-juniper in White House Valley on 13 September 1988, and one foraging with Pinyon Jays near J-3 reclaimed area on 19 September 1988. One was seen in pinyon-juniper on 17 December 1992. Acorn Woodpecker {Melanerpes formicivorits). Fairly common in chained woodland areas. Breed- ing is unconfirmed. There are no winter records. Seen most frequenth' in the Kaventa and Lolomai points chained areas. M. f. aculcatus is the form present in this area. Red-naped Sapsucker (Sphyrapicus nuchalis). A common migrant in wooded areas, including tamarisk on the larger washes. All records are from April, September, and October, except for one on 6 December 1991 seen feeding on Russian olive fruit. Small numbers were noted in tamarisk in Moenkopi Wash during September-October 1986. Williamson's Sapsucker {Sphyrapicus thyroi- deiis). A fairly common migrant in mixed-conifer Apparently resident in very small numbers. A female was seen in Coal Mine Wash on 23 Febru- ary 1983 and another female in pinyon-juniper in White House Valley on 19 January 1984. One pair was seen in pinyon-juniper near Lolomai Point on 27 February 1986; another pair was seen feeding young in a Douglas fir snag in dense mi.xed-conifer in the east fork of Coal Mine Wash on 23 June 1983. A vocal male was seen in Yellow Water Canyon on 23 June 1985. One record from tamarisk: a female seen on Moenkopi Wash on 10 October 1991. Downy Woodpecker {Picoides pubescens). A sparse permanent resident more frequent in fall and winter Seen in oaks in Yellow Water Canyon on 4 March 1984, in tamarisk on Lower Moenkopi Wash on 6 January 1986, and in aspens on Lolomai Point on 14 October 1986. Another was in the Russian olive grove on Moenkopi Wash on 12 December 1989. A single bird, probably breeding. was seen in aspens in Coal Mine Wash on 26 Mav 1986. Hairy Woodpecker {Picoides villosus). A com- mon permanent resident in pinyon-juniper and mi.xed-conifer woodlands. Nests with young (all in the third week of June) have been found in pon- derosas, pinyons, and aspens. Densities of 1.8-4.4 individuals/4() ha have been found using Emlen transects. Breeding densities of 1.9 and 3.8 pairs/4() ha were determined with spot-mapping. This species was not found breeding in a woodland stand of 150 trees/ha. Black Mesa birds may be intermediates between P. v. orius, the breeding form according to Monson and Phillips (1981), and P. V. leticothorectis from nearby Navajo Mountain (Behle 1985). Northern Flicker {Colaptes auratus). A com- mon permanent resident in pinyon-juniper and mi.xed-conifer A nest with young was found in a Gambel oak on 15 June 1984 in Coal Mine Wash. A nest with \'oung was found in an aspen on 22 Jime 1985 in Yellow Water Canyon. Also, nests were found in holes in alluvial banks on lower Coal Mine and Moenkopi washes, where a nest contained five eggs on 8 Ma>' 1987, two nests held small nestlings on 6 June 1990, and another had yoimg read) to fledge on 27 June 1990. Transect densities of 0.4-3.5 birds/40 ha have been observed. This species is frequent along the major washes in win- ter. Reported from a 12th-century Anasazi site (Table 14). The nomenclature of the breeding form from this region is presently under debate (see Monson and Phillips 1981 and Behle 1985). Order Passeriformes F,\MILY TyRANNIDAE Olive-sided Flycatcher {Contopus horealis). An uncommon fall migrant in August and late May-early Jvme in pinyon-juniper and in mixed- conifer One was found singing on the mesa rim on 17 June 1983. Western Wood Pewee {Contopus sordidulus). A common migrant and sparse summer resident. Breeding in mixed-conifer is unconfirmed. Migrants are seen in pinyon-juniper and in tamarisk. A transect density of 8.6 birds/40 ha was found on Moenkopi Wash in mid-September 1985. Two were seen in tamarisk at the Reed Valley-Moenkopi Wash confluence on 10 June 1986. This species seems uncharacteristically sparse as a breeding bird considering the available habitat. Willow Flycatcher {Empidonax trailli). A sparse migrant in riparian areas. Two were seen at the confluence of Moenkopi and Yucca Flat washes on 13 August 1990, another was there on 11 Sep- tember 1989, one was seen at J-7 pond on 6 Sep- tember 1989, and one was seen on lower Moenkopi on 13 August 1992. Seemingly more common as a 46 Great Basin Naturalist [Volume 54 transient in tlie larger lower valii'\ areas throniili- out the region (Jaeohs 1986, personal oi)ser\ati()n). Hammond's Flycatcher (Empidonax liam- mondii). A sparse migrant. One was seen in tamarisk on Moenkopi Wash on 7 October 1986 and another was secMi 7 May 1987. Perhaps more common than noted. Diisk\ FKcatcher (Empidonax oberholseri). A common snmmer resident in open mixed-conifer associations with decidnous scnih nnderstory. Nest hnilding was obsei-ved on 30 Ma\ 1990. A nest with lour et^gs was found on 17 June 1985 in a thicket of Primus below aspens in Yellow Water Canyon. Another nest with three young was found in an oak-covered slope in the same canyon on 23 June 1985, and \'oung were being fed in a nest in a Doug- las fir on 28 June 1992. Singing males have been seen on 30 April. Birds were seen in breeding habi- tat as well as in tamarisk on Moenkopi Wash in late August 1986. Gray Flycatcher [Empidonax wrightii). A com- mon smnmer resident in all but the lowest pinyon- juniper from mid-April to September. A nest with \'(nmg was foimd in a juniper in White House Val- le\' by B. Ebbers on 22 June 1983. Woodbiny and Russell (1945) report two nests also in junipers. Adults with full-sized \oung have been seen from mid-June to mid-jul\'. Transect densities range from 1.8 to 19.0 indi\'iduals/40 ha. Spot-map densi- ties of 6.7-11.5 pairs/40 ha have been noted. Seen as late as 24 September (1986). Cordilleran Flycatcher [Empidonax occiden- talis). A common summer resident in mi.xed-conifer with deciduous scrub nnderstory. This species is less numerous than the Dusky Flycatcher and occupies more shaded areas containing cliffs and banks for nesting. Nests with eggs were found on ledges in Coal Mine Wish on 24 June 1983 and 1 July 1985. Another with four eggs was found on 15 June 1986. Five short-tailed young were seen in Coal Mine Wash on 5 August 1985. E. a. hcUmayri is the breeding form. Black Phoebe [Sayornis nigricans). A sparse transient. One was seen on Moenkopi Wash near the confluence with Yucca Flat Wash on 29 June 1990. Say's Phoebe [Sayornis saya). A common migrant and simimer resident from late February to early November; one late bird seen on 6 December 1991. This species is distributed along larger open wash courses, rock outcrops, and near houses. Nests with eggs have been found as early as 20 April (1989). Nests with large voung have been found on 28 May 1992 and 11 and 14 June (1985 and 1984, respectiveK). Fledglings have been seen on 29 June. Transect densities range from 0.3 to 24.0 individuals/40 ha. Summer densities are lower than spring, indicating a rather heavy influx of migrants. One was seen on the mesa summit in chained pinyon-juniper on 9 August 1986. Another was seen over dense pinyon-juniper on the mesa rim on 6 September 1986. The nominate race is the breeding form. Ash-throated Flycatcher [Myiarchus cineras- cens). A common summer resident in pinyon- juniper woodland throughout the area (absent from die highest, densest stands) but present in mixed- conifer habitats adjacent to shrub-filled canyon floors. Records range from 24 April to an abrupt mid-August departure. Spot-map densities of 3.8-7.6 pairs/4() ha have been recorded. Transect densities range from 0.6 to 12.4 individnals/40 ha. The nominate form breeds. Cassin's Kingbird [Tyrannus vociferans). A common summer resident in open pinyon-juniper. Most numerous in the small canyons at the mesa foot. On 19 Jiuie 1986 nests with \'oung were found on power poles at the Black Mesa Mine office and below Rock Gap. It has been seen by 27 April (1982); most leave by early September. Fledglings were seen at the mesa foot on 7 July 1985. T. v. vociferans is the breeding form. Western Kingbird [Tyrannus verticalis). A fairly common migrant and a sparse summer resident. B. Ebbers found a nest near die Black Mesa Archaeo- logical Project camp on 15 June 1983. Post-breed- ing dispersal from breeding habitats begins in mid- JuK; and migrants become common in early Sep- tember. Scissor-tailed Flycatcher [Tyrannus forficatus). A sparse transient. One was seen in juniper savan- na in Long House Valley on 24 June 1993. Eastern Kingbird [Tyrannus tyrannus). A sparse transient. One was seen at J-3 reclaimed area on 29 August 1988. Family Alaldidae Horned Lark [Eremophila alpestris). An abim- dant permanent resident in grasslands, open shrub- lands, and reclaimed areas. Flocks of up to 300 have been observed in reclaimed areas during mid- winter when there may be an influx of northern birds. These winter flocks begin breaking up and males become territorial in late February. Nests with eggs have been foimd from mid-April to mid- May. Fledglings and new xoung have been found in late May. Mixed-age flocks begin forming by early Jul\'. Reported from a single early Anasazi site (Table 14). E. a. occidentalis is the breeding form. Family Hiiu ndimdae Purple Martin [Progne subis). A sparse migrant. A female was seen over Moenkopi Wash on 11 September 1986, and another was in N-1 reclaimed area on 24 August 1992. Tree Swallow [Tachycineta hicolor). A common migrant near ponds from late March through May and from JuK to earh' October. 1994] Birds of Northern Black Mesa, Arizona 47 Violet-green Swallow {Tachycineta thalassina). A common migrant throughout the area. It is a common summer resident nesting in cliffs in the upper canyons and along the mesa rim and in holes in dirt banks along washes in areas of pinyon- juniper Seen from 23 March (1990) to late Septem- ber T. t. lepida is the breeding form. Northern Rough-winged Swallow (Stelgi- dopteryx serripennis). A common siuumer resident antl migrant. Tliis swallow breeds as widely scat- tered pairs in holes in alluvial banks along larger dissected wash courses. Records range from mid- April to mid-September. The form breeding in northeastern Arizona is not clear, but may be inter- mediate between S. s. psammochrous and S. s. ser- ripennis (see Behle 1985 and Monson and Phillips 1981). Bank Swallow [Riparia riparia). A common migrant. Seen most frequently at J-7 pond in August and September Cliff Swallow (Hiritndo pyrrhonota). A com- mon migrant throughout the area. Seen from 8 April (1982) to 13 September (1982). Old nest site foundations were found on a cliff on lower Moenkopi Wash on 1 July 1986 and below J-7 pond in 1987, but no active colonies were found from 1979 to 1990. Five vocal birds were seen at the Moenkopi Wash site on 1 July 1986, and an active colony appeared there in 1993. This species was imusuallv numerous throughout the studv area in June 1989. Barn Swallow {Hiriindo riistica). A common migrant throughout the area from mid-April to early Jinie and mid-July to mid-October Breeding is known from a single nesting at the N-8 shop in 1992. Family Cory i da e Steller's Jay (Cyanocitta stelleri). A common permanent resident in mixed-conifer habitats. A few wander in the fall to lower areas of pinyon- juniper woodland, as at Owl Spring Valley on 13 October 1985. A nest with three young was found in Coal Mine Wash on 25 May 1986. Feathers of predated immatures were foimd on 27 May (1984) and on 4 JuK (1985). FamiK' groups have been seen on 9 June (1984) and 16 June (1982). From Monson and Phillips (1981) and Behle (1985), it is unclear what the breeding form in this region is. It may be intermediate between C. s. diodemata and C. s. )n(icr()Iopl\a. Scrub Jay {Aphelocoma coendescens). A com- mon permanent resident throughout the area in pinxon-juniper and mixed-conifer woodland, mon- tane scrub, and tamarisk. FamiK' groups in tamarisk on 8 June 1992 and 27 June 1990 indicate breed- ing. A nest containing four week-old young was found on 24 May 1990. Adults with fledglings have been seen from mid-Ma\ to late June. Transect densities in woodland range from 0.4 to 3.5 individ- uals/40 ha. Spot-map densities of 1.4-3.8 pairs/40 ha have been found in pinyon-juniper woodland. Remains are reported from a single Anasazi site Cnible 14). Monson and Phillips (1981) call the birds in this region A. c. siittoni. Behle (1985) calls them A. c. woodhousei. Pinyon Jay {Gymnorhinus cyanocephcdus) . A common permanent resident throughout the area. Adults building nests have been seen on 8 March (1982). Nests with young and/or eggs have been seen in early May. Young fledglings have been seen from 13 April to the sin-prisingly late date of 12 Jul\' (1982). Flocks of Pinyon Jays are occasionally seen feeding on wheatgrasses in reclaimed areas. Dur- ing the fall Pinyon Jays form the largest bird flocks in the area. On 10 September 1983, 285 were seen at the mesa foot. On 23 September 1983, over 470 were seen at Reed Valle\'. A loose aggregation of several flocks totaling nearly 600 indi\ iduals was seen at Dinnebito Wash on 19 Januar\' 1984. Reported from three Anasazi sites (Talile 14). Clark's Nutcracker {Nucifraga columbiana). A fairh' common to common permanent resident in the upper canyons. It occasionally wanders widely in late summer and fall to the pinyon-jiniiper below the mesa rim. Fledglings have been seen in the upper canyons from 24 May (1985) to 22 June (1983). Remains of a young fledgling were found in Moenkopi Wash on 13 April 1989. These dates indicate that laying takes place from early March to mid-May. A mixed-age flock of 35 indi\'iduals was seen in Coal Mine Wash on 30 Ma>- 1990. This species apparentK' left the mesa rim region during most of the 1985-86 winter American Crow {Corvus brachyrhynchos). A sparse migrant. One was seen near J-3 reclaimed area on 16 November 1987, two at the highway crossing of Coal Mine Wash on 28 March 1988, and one at N-14 on 13 November 1991. Another was seen on 6 November 1989. Known from a single Anasazi site (Table 14). Common Raven [Corvus corax). An abundant permanent resident areawide. The population in the lease is probably artificially high because of readily available food in garbage, waste grain in straw mulch, and other sources. Adults frequenting nest sites and building nests have been seen on 15 January (1985). Nests with eggs have been found by early May. Young of all ages have been found from late May to early July, but most fledge by mid-June. Flocks of over 100 birds have been seen feeding on cured seed heads in reclaimed areas on several occasions. Approximately 100 birds were obsei^ved using Russian olives for a nocturnal roost on 23 January 1992. Remains are reported from Anasazi sites (Table 14). 48 Great Basin Naturalist [Volume 54 Famii,\ Pa hi dak Mountain Chickadee (Panis ^amheli). A com- mon permanent resident in hi^her-elexation pin- yon-juniper and mixed-conifer thronghout the area. Small numbers occupy tamarisk of the larger wash- es in fall and winter. Song activity is frequent by early March. Nests with young and eggs have been found during the first three weeks of June, and fledglings have been seen from mid-June to mid- Jul\'. Transect densities in pinyon-juniper range from 1.8 to 32.2 individuals/4() ha. Spot-map densi- ties range from .3.8 to 11.5 pairs/4() ha. Monson and Phillips (1981) call birds in this region P. <,'. wasatchensis, but Behle (1985) would assign them to at\'pical P. ^. ^(nnbcli. Plain Titmouse {Parus inornatus). A common permanent resident in pinyon-juniper but general- ly absent in the highest-elevation stands. Fall tran- sients have been seen twice in tamarisk. A nest with yoimg ready to fledge was found on 13 June 1982. Family groups have been seen from mid- June to mid-July. Transect densities in pinyon- juniper range from 1.8 to 30.2 individuals/40 ha. Spot-map densities range from 7.6 to 11.5 pairs/40 ha. The breeding form is P. i. ridgwayi. Family Aegitiialidae Bushtit {Psahripanis minimus). A common per- manent resident in pinyon-juniper and mi.xed- conifer throughout the area. A nest was found in tamarisk on Moenkopi Wash on 25 May 1993. Flocks wander into shiublands and tamarisk during fall and winter, begin to break up into pairs by late March, and are forming again by late June. B. Ebbers found an active nest in Jime 1983 that had fledged young by the 15th. Family groups were noted in 1992 on 29 May and 5 June. Pinyon- juniper transect densities range from 1.8 to 31.3 individuals/40 ha. Spot-map densities of 3.8-5.7 pairs/4() ha have been recorded. The breeding form is P. 1)1. plwnhcHS. Family Sittidae Red-breasted Nuthatch {Sitta canadensis). An irregularly common winter resident in mixed- conifer. A sparse summer resident in the same areas. An adult with an immature seen at Lolomai Point on 7 August 1990 may not represent local breeding. However, family groups seen on 28 July 1992 and 16 June 1993 confirm local nesting. Two were seen in pinyon-juniper in White House Valley on 10 August 1984. This species was common in the upper canyons in the 1984-85 winter when a heavy, widespread Douglas fir cone crop ripened. It was absent the following v\inter when no cones were produced. White-breasted Nuthatch (Sitta carolinensis). A common permanent resident in pinyon-juniper woodland and in mixed-conifer. This species does not breed in lower-elevation pinyon-juniper where it was absent as a breeding species in a stand of 150 trees/ha. Spot-map breeding densities in pinyon- juniper range from 3.8 to 11.5 pairs/40 ha. Transect densities range from 1.4 to 14.8 individuals/40 ha. Nests with yoimg have been foimd from late May to mid-June. Family groups have been seen from mid-June to mid-July. The breeding form is S. c. nclsoni. Pygmy Nuthatch (Sitta pijgmaea). A common permanent resident in flocks in ponderosa pine-dominated mixed-conifer association. Strays to adjacent pinyon-juniper in late summer and fall. Fledglings have been seen from mid-June to late July. The breeding form is S. p. melanotic. Family Certiiiidae Brown Creeper (Certhia americana). A fairly common permanent resident in mixed-conifer. Seen once in pinyon-juniper in winter: 5 January 1983 in White House Valley. Young were seen being fed in a nest on a Douglas fir in Yellow Water Canyon on 22 June 1985, and a family group was seen in Coal Mine Wash on 2 July 1992. The breeding form is C. a. inoutdna. Family Troglodytidae Rock Wren {Salpinctes obsoletus). A common summer resident of talus slopes, rock outcrops, alluvial banks, and riprap of dam spillways and road slopes. Birds are present on the lease from early March to early December, but numbers decrease sharply in early October. One midwinter record was 20 January 1992. Numerous family groups were noted 11-12 June 1992. On 20 June 1983 a nest with six eggs was found at Tees Yah Toh Spring. On 19 July 1984 a family group was seen below J-7 dam. It is frequently seen in chained pinyon-juniper on the mesa summit. The nominate race breeds. Canyon Wren (Catherpes mexicanus). A com- mon permanent resident in the cliffs of the upper canyons, along the Dakota Sandstone, and in other cliff and talus areas. During fall and winter it is seen along alluvial banks and occasional!}' at small rock outcrops. Family groups have been seen from mid- June to early July. A nest with a fledgling was found on 27 July 1987. C. ;/;. consperus is the breeding form. Bewick's Wren (Thyromanes hewickii). A com- mon summer resident in pinyon-juniper and mixed-conifer. Fairly common in winter when it wanders into shrublands and tamarisk. Males are acti\'el\' singing b\ late February. Nests, both in junipers, with eggs were found on 2 and 11 June 1982. Family groups have been seen by 20 June and 2 July. Transect densities in pinyon-juniper 1994] Birds of Northern Bijvck Mesa, Arizona 49 range from 1.8 to 28.8 inclividiials/4() ha. Densities of 11.5-32.5 inclivicluals/40 ha have been recorded in shruhland areas in late snnimer. Spot-map breeding densities range from 11.5 to 19.1 pairs/40 ha in pinyon-juniper. Uncommon as a breeding bird in the densest pinyon-juniper on the mesa summit. The breeding form is T. h. oberholser. House Wren {Troglodytes aedon). A common migrant along washes in tamarisk. A common simi- mer resident in aspen groves. Migrants are usualh' seen in spring in late April and early May, but one on 11 April 1989 was atypically early. Most fall migrants pass through from late August to mid- October, with one on 31 October 1989 being late. A nest with young in an aspen was found on 15 June 1984. Young fledglings were seen in the east fork of Coal Mine Wash on 24 June 1984. Several individ- uals were observed in slash piles in the Kayenta Point chaining on 14 June 1986. Sedgwick and Ryder (1987) report this species in pinyon-juniper slash in Colorado. Post-breeding dispersal away from aspen groves into adjoining mi.xed-conifer woodland begins in mid- to late Jime. T. a. parh- inani is the breeding form. Winter Wren {Troglodytes troglodytes). A sparse migrant. One was seen in an aspen grove in the west fork of Coal Mine Wash on 6 November 1983. Another was seen in the same grove on 15 November 1986. Marsh Wren {Cistothorus palustris). A fairly common migrant at weedy or brushy pond edges. One was seen by EH&A Consultants on 5 October 1979. Seen at a pond in N-2 reclaimed area and in tamarisk at J-7 pond several times in early Septem- ber 1986. One was seen at Pond N14-G on 12 March 1989 and at N-2 reclaimed area on 12 and 28 April 1989. Family Muscicapidae Golden-crowned Kinglet {Regtdus satrapa). A fairly common fall, winter, and early spring resi- dent in mixed-conifer of the upper canyons. Records range from 12 October (1988) to 4 March (1984). Ruby-crowned Kinglet {Regulus calendula). A common migrant in wooded areas throughout the area. A fairly common summer resident in mi.xed- conifer in the upper canyons. Breeding is uncon- firmed. One was seen in Russian olives on 9 December 1992. Singing males have been record- ed throughout June in all of the upper canyons including five on 16 June 1982 in the middle fork of Coal Mine Wash. Transect densities of migrants range from 1.8 to 10.6 individuals/40 ha. Monson and Phillips (1981) call Arizona breeding birds R. c. calendula, but Behle would assign birds of this area to R. c. cineraceu.s. Blue-gray Gnatcatcher {Polioptila caerulea). A common summer resident in higher-elevation pinyon-juniper and in areas of deciduous scrub. A nest with young near fledging was found in a pin- yon near Tees Spa Spring on 5 July 1987. An old nest was found in a small pinyon in Reed Valley in December 1988. Wanders into sagebrush in late summer. Seen from mid-April to late October. Migrants are frequent in tamarisk. Transect densi- ties range from 2.3 to 9.4 individuals/40 ha. All spot-map densities are 3.8 pairs/40 ha. The breed- ing form is P. c. ohscura. Western Bluebird {Sialia mexicana). A common permanent resident throughout the area. It breeds in higher-elevation pinyon-juniper and mixed- conifer woodland. Pinyon-juniper stands of 150 and 182 trees/ha contained no nesting individuals. It may be absent during periods of cold and deep snow in midwinter. Nests with young have been seen from mid-June to late July. Pinyon-juniper transect densities range from 1.8 to 8.9 individu- als/40 ha. A spot-map density of 3.8 pairs/40 ha has been foiuid. S. ;/;. occidentali.s is the breeding form. Mountain Bluebird {Sialia currucoides). A common permanent resident of pinyon juniper throughout the area. It may be absent during extreme cold and heavy snow in midwinter but was seen several times in winter at the Russian olive grove on Moenkopi Wash. It consistently appears by late January and is common by late February. The largest flock recorded was 250 in the Russian olive grove on 29 January 1992. Mountain Blue- birds occupy more open areas such as sage clear- ings and chained pinyon-juniper. It frequently nests in holes in alluvial banks (nestlings being fed in such situations were seen 4 June 1991 and 12 May 1992) and cliffs (nest with eggs and young seen 28 May 1992). Nests with young have been seen in mid- and late June. Pinyon-juniper transect densities range from 0.9 to 16.0 individuals/40 ha. Spot-map densities range from 1.8 to 7.6 pairs/40 ha. Townsend's Solitaire {Myadestes townsendi). A fairly common to common permanent resident in mixed-conifer of the upper canyons. It has been seen occasionally in pinyon-juniper elsewhere. Fledglings have been seen from early July to early August. An old nest was found in a hole in a boul- der in Coal Mine Wash on 6 September 1986. M. t. townsendi is the breeding form. Hermit Thrush {Catharus guttatus). A fairly common migrant throughout the area. A common summer resident in mixed-conifer of the upper canyons. Singing has been heard in late April. Stubby-tailed fledglings have been seen on 27 June (1983,' 1985) in Coal Mine Wash and Yellow Water Canyon, respectively. The breeding form is C. g. auduboni. Occasionally heard singing in summer in dense pinvon-juniper as at White House Valley on 17 June 1983 and 20 June 1984. American Robin {Tardus migratorius). A com- mon migrant throughout the area and a fairlv 50 Great Basin Naturalist [Volume 54 common summer resident in mixed-conifer ol the upper canyons. Nested in pinyon-juniper in 1993 (adults with fledgUng seen 26 June). Uncommon and irregularly seen in winter until about 1990 when up to 70 began wintering in Russian olives. An incubating adult was seen in oaks fringing an aspen grove in Yellow Water Canyon on 15 June 1986. Woodbur>' and Russell (1945) report breed- ing from Black Mesa. T. in. propinquiis is the breeding form. Varied Thrush (Ixoreus naevius). A sparse fall transient. One record: an adult male was seen with robins in Russian olives on Moenkopi Wash on 12 November 1992. Family Mimidae Northern Mockingbird (Mimus polyglottos). A common summer resident in juniper savanna at the mesa foot and in greasewood and tamarisk on lower washes and in smaller drainages in mixed shrub- lands. One seen 14 March 1989 was exceptionally early by six weeks. Seen in oaks in the Lolomai Point chaining on 14 July 1984. A nest in grease- wood contained four eggs on 6 June 1990. A fledg- ling was seen in juniper savanna on 3 July 1985. A transect density from Moenkopi Wash is 5.3 indi- viduals/40 ha. The nominate race is the breeding form. Sage Thrasher {Oreoscoptes montanus). A com- mon summer resident in greasewood and saltbush in the middle to lower reaches of the major washes. It is less numerous in sagebrush. The first spring appearance is usually in early April. A nest in salt- bush in a mixed-shrub habitat contained three young on 19 June 1986. Fledglings were seen in Reed Valley on 21 July 1982. Sage Thrashers are common in J-7 reclaimed area where saltbush is well developed (3.3 pairs/40 ha). Transect densities range from 0.9 to 9.5 pairs/40 ha. Bendire's Thrasher {Toxostoma bendirei). A fairly common summer resident in open grease- wood in lower washes and in juniper savanna at the mesa foot. Small numbers are found in open salt- bush in mixed shrubland. Records span from mid- April to August. A nest with three small young was found in a juniper at the mesa foot on 18 June 1986. A transect density of 1.8 individuals/40 ha was recorded on Moenkopi Wash in 1980 by EH6fA Consultants. numbers in October. Usually seen in small groups, but a flock of 41 was seen on 12 April 1989 in N-1 reclaimed area, and over 200 were seen there on 26 April 1991. Family Bombycillidae Bohemian Waxwing (Bomhycilla garrulus). Two records: a flock of 29 was seen in pinyon-juniper in White House Valley on 13 April 1982 and a single bird in Russian olives on 12 November 1992. Cedar Waxwing [Bomhycilla cedrorum). A sparse fall migrant with records as follows: six seen in Dinnebito Wash on 16 November 1982, seven in mixed-conifer in Coal Mine Wash on 22 October 1983, one over Moenkopi Wash on 10 October 1986, two near "Kelly Pond" on 12 September 1991, one in Russian olives on 15 November 1990, and eight there on 29 January 1992. Family Lamidae Northern Shrike {Lanius excubitor). An uncom- mon to fairly common winter resident in open pinyon-juniper from mid-October to late March. A large invasion occurred in 1988-89, with more birds seen during this period than in the previous seven winters combined. Loggerhead Shrike [Lanius ludovicianus). A fairly common to common permanent resident in juniper savanna and greasewood in the major washes. Seen less frequently in other open shrub- land. The 1984 spot-map density in mixed-shrub habitat was 0.8 pairs/40 ha. Fledglings and family groups have been noted from early June to early July. The breeding form is L. /. excubitorides (see Monson and Phillips 1981). Family Sturnidae European Starling [Sturnus vulgaris). A com- mon permanent resident that is local in the study area. Starlings frequent the reclamation barn and the area near the mine facilities. This species was probably absent from the area prior to the mine development in the late 1960s. The largest group seen was a flock of 350 in Russian olives on Moenkopi Wash on 15 November 1990. This species appears to be widely dispersed throughout the region during winter. Fa.mily Motacillidae American Pipit [Anthus rubescens). A common migrant at ponds and reclaimed areas throughout the area. Sparse in winter at unfrozen ponds. Spring migrants peak in April and have been seen from mid-February to early May, with a single bird seen 3 June 1991 being late. Fall migrants are seen from mid-August to early December with peak Family Vireonidae Gray Vireo [Vireo vicinior). A fairly common summer resident fiom 23 April (1992) to 10 Septem- ber (1983). Most observations are from open pin- yon-juniper-covered slopes and small canyons that support a scattered growth of Utah serviceberry. A nest with two small young was found in a small pinyon above Tees Yah Toh Spring on 20 June 1986. 1994] Birds of Northern Black Mesa, Arizona 51 Solitary Vireo (Vireo solitariiis). A common migrant and summer resident in pin\ on-juniper woodland and in mixed-conifer of the upper canyons. Records span from 22 April (1983) to 7 October (1982). The migrant form, V. s. cassinii, has been seen in the fall. A nest with young was found on 21 June 1983 in lower Reed Valley and in White House Valley. An adult with a full-sized fledgling was seen in White House Vallex' on 27 Jul\- 1982. A flock of 10 was seen on the rim of \ellow Water Canyon on 11 Ma\' 1984. Transect densities range from 0.4 to 12.4 individuals/40 ha. Spot-map densi- ties range from 1.9 to 7.6 pairs/40 ha. The breeding form is V. s. phnnbeus. Warbling Vireo {Vireo gilvus). A common sum- mer resident of aspen groves in the upper canyons. A common migrant in tamarisk from early August to mid-October. The breeding form is V. g. brew- ■stcri. Family Emberizidae Tennessee Warbler {Vermivora peregrina). A sparse transient. A single bird was seen in tamarisk on Moenkopi Witsh on 7 October 1986. Orange-crowned Warbler {Vermivora celata). A common migrant in deciduous scrub and tamarisk throughout the area. Transect densities as high as 177 individuals/40 ha have been recorded in tamarisk (1 October 1986). A fairly common summer resident in mixed-conifer in the upper canyons. Breeding habitat is Gambel oak-covered slopes usually adjacent to aspen groves and with a grassy, leaf-littered understory. A pair of adults was found with fledglings in the west fork of Coal Mine Wash on 27 June 1983. Records span from 1 Ma\ (1983) to 22 October (1986). The breeding form is V. c. orestera. Nashville Warbler {Vermivora riificapilla). A common to abimdant fall migrant in montane scrub and tamarisk from early August to early October. Virginia's Warbler {Vermivora virginiae). A common to abundant migrant areawide. A common summer resident in montane scrub of the upper canyons, upper mesa slopes, and Lolomai Point chained area. Records span from mid-April to earl\' October. Several pairs with \'oung fleglings were found on 15 June 1986. Fall migrants appear in tamarisk along the lower washes beginning in late June and become progressively more common, peaking in late August. The breeding form is V. v. virginiae. Lucy's Warbler {Vermivora luciae). A sparse transient: a singing male was seen in tamarisk on Moenkopi Wash (elevation 1768 m) on 11 April 1992. Yellow Warbler {Dendroica petechia). A com- mon fall migrant in tamarisk along the major wash- es from early August to early October Apparently sparse in spring: one in aspens in Yellow Water Canyon on 5 May 1987, one in pinyon-juniper on 6 May 1987, one at J-28 settling ponds on 20 May 1990, and a singing male in Russian olives on 8 June 1992. A nonbreeding, singing male was seen in tamarisk on Moenkopi Wash on 26 June 1990. See Monson and Phillips (1981) concerning the complex of races migrating through the region. Yellow-rumped Warbler {Dendroica coronata). A common to abundant migrant throughout the area. Up to 713 per 40 ha were recorded in tamarisk on Moenkopi Wash (26 September 1986). A fairly common summer resident in mixed-conifer of the upper canyons. Spring migration peaks in late April-early May, and fall migration peaks in late September-mid-October. (Records of migrants span from 13 April to 8 December) By 1990 this species was wintering in small numbers in Russian olives on Moenkopi Wash. Two adults were seen feeding \'oimg in a ponderosa on 16 June 1984 on Kayenta Point. A fledgling was found on 14 June 1986. An adult male in pre-basic molt was seen in upper Moenkopi Wash on 30 July 1985. A male "myrtle form was seen in J-27 reclaimed area on 27 April 1982. Monson and Phillips (1981) call the breeding form D. c. memorahlis, but Behle (1985) assigns it to D. c. audiihoni. Black-throated Gray Warbler (Dendroica nigrescens). A conunon summer resident of pinyon- juniper and mixed-conifer woodland. Records span "from 8 April (1991) to 21 September (1982). The males, initially seen in small groups, are singing and actively defending territories by late April. Females have been observed nest building from mid- to late May. A nest on 25 May contained four eggs. Young in nests have been found on 8 and 17 June. Fledglings have been seen 25 June-27 July. Adults were seen feeding young Brown-headed Cowbirds on 19 July and 4 August 1982. Transect densities range from 0.9 to 26.6 individuals/40 ha. Spot-map densities range from 7.6 to 15.3 pairs/40 ha. Monson and Phillips (1981) recognize the sub- specific rank of D. n. halsei as the breeding form in the region. Breeding densities correlate positively with pinyon density, and this species appears to forage preferentiallv' in pinyons. Further study may prove it to be a pinyon specialist in the Black Mesa area. Townsend's Warbler (Dendroica townsendi). A common fall migrant in mixed-conifer of the upper canyons. Fairly common in aspens, pinyon-juniper, and tamarisk. Records range from 21 August (1992) to 22 October (1986). Hermit Warbler {Dendroica occidentalis). A sparse fall migrant. An immature female was seen in tamarisk on Moenkopi Wash on 29 September 1989. Grace's Warbler (Dendroica graciae). A com- mon simimer resident in ponderosa pine-dominat- ed mi.xed-conifer woodland. Adults with fledglings were seen on 16 Jime and 13 Jul\' 1984 in the west 52 C,REAT Basin Naturalist [Volume 54 fork of (^oal Mine VVasli. Woodhiiry and Knsscll (1945) report collecting two adults and an innna- tiire in July 1938. Records rantie from IS A])ril (1989) to 6 September (1986). American Redstart {Setophaga rulicilla). Remains were found near the mesa rim by D. Ellis (personal communication) in 19S3. Probably a sparse migrant. Northern Waterthrush {Seiiiriis noveboracen- sis). A sparse miurant. One was seen at a pond in |- 27 reclaimed area on 13 May 1982, one in tamarisk on Moenkopi Wash on 12 May 1987, and another in mixed-conifer on 7 Ma\ 1992. One was seen in tamarisk on 16 September 1992. Kentucky Warbler [Oporornis formosus). A sparse transient. An adult male was seen in mon- tane scrub at the head of the east fork of (Joal Mine Wash on 7 May 1983. Another adult was seen in tamarisk on Moenkopi Wash on 17 September 1986. MacGillivray's Warbler (Oporornis tolmiei). A common migrant in montane scrub and tamarisk. It is most numerous during fall migration, which peaks in late August to mid-September, but it is seen until mid-October. An uncommon breeding resident of montane scrub at three localities in the upper canyons. An adult male in Yellow Water Canyon was seen feeding a fledgling in a dense tangle of oaks, aspen, chokecherry, clematis, and dogwood on 23 June 1986. Six adults (three carry- ing food) were seen in a similar habitat on 20 June 1986 in the east fork of Coal Mine Wash. The breeding form is O. t. nionticola. Grazing may threaten this species as a breeding bird on Black Mesa. Common Yellowthroat (Geothlypis trichas). A fairly connnon migrant in earl\ Ma\ and from late August to early October in tamarisk on Moenkopi Wash and at weedy ponds. A singing, nonbreeding adult male was seen on Moenkopi Wash on 27 June 1990. Wilson's Warbler (Wihonia pusilla). A common migrant in montane scrub and tamarisk. Fall migra- tion peaks in late August to early September but lasts until late October A singing male was seen in chokecherry and aspens in the middle fork of Coal Mine Wash On 26 .\hi\ 1986. Yellow-breasted Chat [Icteria virens). A sparse transient. One was seen in tamarisk on Moenkopi Wash on 27 August 1990, and another in tamarisk 25 May-8 June 1993. Western Tanager {Piranga hidoviciana). A com- mon migrant throughout the area in pinyon-juniper and tamarisk. A common summer resident in mixed-conifer of the upper canyons. Records range from 2 May (1985) to 25 September (1986). A pair was seen attending a nest in Yellow Water Canyon on 15 June 1986. Another pair was feeding nestlings in Coal Mine Wash on 27 June 1989. The last spring migrants are seen in early June. Fall migrants are seen awa\ ironi the breeding habitat beginning in mid-JuK. Black-headed Grosbeak {Pheiicticus melanoce- phalus). A common migrant and summer resident in aspen groves and mixed-conifer of the upper canyons. Migrants are occasionalK' seen in tamarisk. A pair was seen building a nest in the east fork of Coal .Mine Wash on 24 June 1984, and three lamiK groups were seen on 2 July 1992. Woodbury and Russell (1945) report collecting a male. The breeding form is P. in. inelanocc})hali(s. Blue Grosbeak {Guiraca caeriilea). A common summer resident in tamarisk on Moenkopi. Coal Mine, and Red Peak Valley washes. Two fledglings were seen below J-7 dam on 29 August 1985. Another was seen on 30 August 1989. This is the most numerous breeding bird of the tamarisk thick- ets along the lower washes. A transect breeding density of about 11 pairs/40 ha was determined in late Jime 1986 on Moenkopi Wash. G. c. i)iteijusa is the breeding form. Lazuli Bunting [Passerina amoena). A common fall migrant from mid-August to early October in tamarisk of the large washes. Less common else- where. On 12 September 1985, 19.5 individuals/40 ha were counted on Moenkopi Wash. Up to five males and two females were present in Russian olives on Moenkopi Wash throughout the summer of 1992. A sixth singing male was at the mouth of Yucca Flat \Vash on 15 June 1992. An adult female with a fledgling was seen in the olives on 14 lulv 1992. Indigo Bunting [Passerina cijanea). A sparse transient. A singing (but nonbreeding) adult male was seen in tamarisk on Moenkopi Wash near the Yucca Flat Wash confluence on 26-27 June 1990. A male paired with a female Lazuli Bunting bred suc- cessfulK' (fledgling seen 29 July) in Russian olives on Moenkopi Wash in 1992. Green-tailed Towhee (Pipilo chlorurus). A com- mon migrant in tamarisk and brush along the major washes and less commonly elsewhere. A common summer resident in Gambel oaks in chained pin- yon-juniper on Lolomai and Kayenta points. Also found breeding on Lolomai Point in an undisturbed basin of big sage and wax currant bordered by (iambel oak and chokecherry. Spring migrants appear in mid-April. Most are gone in fall by late September. Nearly 80 individuals/40 ha were counted on Moenkopi Wash on 12 September 1985. Rufous-sided Towhee [Pipilo erythrophthal- mus). A common summer resident in montane scrub in mixed-conifer of the upper canyons. A fairly common simimer resident in high-elevation pinxon-juniper with a hea\'\' big sage imderstory or in areas with Gambel oak. Small numbers winter in Russian olives on Moenkopi Wiish. Migrates widely throughout the area. A nest with four eggs was found at the head of the west fork of Coal Mine 1994] Birds of Northern Black Mesa, Arizona 53 Wash on 23 Ma\- 1983. Small fled,<:;lings have l)een seen in late June. An individual in first pre-hasic molt was seen on 2 September 1986. Towhees are actively singing in the upper canyons in early March. A spot-map density from dense pinyon- juniper in White House Valley indicated 3.8 paiis/40 ha. The breeding form is P. e. niontanKs. American Tree Sparrow {Spizella arhorea). A sparse winter resident. One record: a single bird was seen on Moenkopi Wash on 1 December 1992. Chipping Sparrow (Spizella passerina). A com- mon sinnmer resident in pinyon-juniper and open mixed-conifer areawide. Records span from 1 April (1982) to 2 November (1990). Nest building was observed on 31 May (1983), small young (5-7 days) were found on 19 Jivme 1984 and 21 June 1983, and fledglings or full-sized juveniles have been seen from 19 June (1983) to 30 JuK- (1985). Flocks begin forming in mid-Jul>' and ha\ e been seen until mid- October. Pinyon-juniper transect densities range from 2.6 to 42.7 indi\ iduals/40 ha. Spot-map breed- ing densities in pinyon-juniper range from 7.6 to 11.5 pairs/40 ha. The breeding form is S. p. arizonae. Brewer's Sparrow [Spizella breweri). A com- mon sunmier resident in sagebrush and saltbush. Observed from 10 April (1984) to 17 October (1985). Individuals have been noted singing on 22 April (1983), nest building has been seen on 4 May (1982), nests with eggs have been found on 26 Ma\' (1983) and 27 May (1985), nestlings were found on 4 June 1984, and fledglings have been seen 4-27 July. Flocks begin forming in late July and have been seen until mid-October. Brewer's Sparrows are abundant in tamarisk in September on the larg- er washes. Transect densities range from 0.4 to over 500 indi\ iduals/40 ha. Spot-map breeding densities are 3.0-8.9 pairs/40 ha in mi.xed-shrub shrubland and 3.3 pairs/40 ha in a reclaimed area with well- developed saltbush (4700 shrubs/ha). The nominate form is the breeding bird. Clay-colored Sparrow [Spizella pallida). A sparse transient. One was seen on Moenkopi Wash on 4 September 1992. Vesper Sparrow [Pooecetes gramineus). A com- mon migrant in open terrain throughout the area. Less numerous as a breeding resident. Records span from 9 March (1983) to mid-November. Breeding is docimiented by nests with eggs found in heavily grazed saltbush in upper Reed Valley on 6 May 1982 and in N-2 reclaimed area on 25 May 1992. It may breed in other open shrublands and in chained pinyon-juniper on the mesa summit. Tran- sect densities range from 0.9 to 27.5 indiv iduals/40 ha. Monson and Phillips (1981) assign the breeding form to P. <;. altiis, but Behle (1985) calls it P. g. con- finis. Lark Sparrow [Chondestes grammacus). A common summer resident in greasewood, open mixed-shrub, juniper savanna, and less numerously in open pinyon-juniper. Lark Sparrows seen in the Lolomai Point chaining on 14 Jul) 1984 and 27 June 1985 may not breed there. Adults feeding two \oung still in the nest were seen on 24 June 1988. A fledgling was seen at the mesa foot on 19 June 1986. Flocks begin forming in late July. Records range from mid-April to early September. The breeding form is C. g. strigatus. Black-throated Sparrow [Amphispiza bilinea- ta). A common summer resident of the lower wash- es in greasewood and adjacent shadscale-covered terraces and in mixed-shrub habitats containing greasewood and/or shadscale. Transient individuals have been seen in pinyon-juniper and reclaimed areas. A nest with three eggs was found in a shad- scale bush on lower Moenkopi Wash on 19 June 1989. Records span from early April to early Sep- tember. Transect densities range from 1.8 to 12.4 indi\ iduals/40 ha. Spot-map densities in mixed- shrub for 1984 and 1985 are 11.9 and 5.9 pairs/40 ha, respectixely. Perhaps strong competitive pres- sure from large numbers of Chipping, Brewer's and Vesper Sparrows in late August and early Sep- tember accounts for the abrupt late -summer depar- ture of this species from the area. The breeding form is A. h. deserticohi. Sage Sparrow [Amphispiza belli). A common late-winter to earl\ fall resident of sagebrush, salt- bush, and mixed-shrub terrains. Small numbers winter in saltbush at the mesa foot and in lower Moenkopi Wash where 12 were seen on 13 December 1989. Usually first seen on the lease in late February when singing is frequent. Nest build- ing was noted on 6 April 1982. Nests with eggs have been found from late May to late June. Fledg- lings were noted on 19 June and 23 July 1982. Shrubland transect densities range from 2.7 to 22.9 indi\iduals/40 ha. Spot-map densities for two years in mixed-shrub are 7.4 and 3.0 pairs/40 ha. A. b. nevadensis is the breeding form. Lark Bunting [Calamospiza melanocorys). A sparse spring migrant. A male was seen at J-3 reclaimed area on 19 May 1983. Savannah Sparrow [Passerculus sandwichen- sis). A common migrant in weed)' pond and stream edges. Spring observations range from mid-March (27 February 1989 being early) through April; one seen 25 May 1992 in N-2 reclaimed area was exceptionally late. Fall records are from mid-July to mid- October. Song Sparrow [Melospiza melodia). A fairly common migrant in tamarisk along the major wash- es and at weedy pond edges. It winters in small numbers in the same areas. Lincoln's Sparrow [Melospiza lincolnii). A fairly common fall migrant at weed)- pond edges and in tamarisk. \'ery few have been seen in spring. Swamp Sparrow [Melospiza georgiana). A sparse migrant. A single bird in alternate plumage was seen in weeds at a pond in the N-2 reclaimed area 25-27 April 1990. 54 Ghi:at Basin Naturalist [Xxjlunie 54 White-throated Sparrow {Zonotricliia alhicol- lis). An adult male was seen with a Hock of juncos in an oak thicket in chained pin\'on-jiniiper wood- huul on L()h)niai Point h\ Ciale Monson on 4 November 1990. Another adnlt was on Moenko])i Wash on 24 April 1992. White-crowned Sparrow (Zonotricliia leu- cophrys). A conunon migrant in brush and taniaiisk throughout the length of the major washes. Also common in brushy chained pinyon-juniper on Lolomai Point. This species became markedly more numerous in the 1980s as a winter resident in tamarisk and adjacent greasewood thickets along the lower washes. Indicative of this increase was a flock of 195 seen in Russian olives on 9 January 1992. One was seen in chained pinxon-juniper on Lolomai Point on 17 June 1983. Transect densities range from 3.5 to 134.8 indi\iduals/40 ha. Harris' Sparrow {Zonotrichia guerilla). A sparse winter resident. Single indi\iduals were seen in the Russian olive grove on 29 January 1992 and 24 April 1992. Dark-eyed Junco ijtinco hyemalis). An abun- dant winter resident throughout the area but pri- marily distributed in small flocks along washes. A common sinnmer resident in dense mixed-conifer of the upper canyons. The forms that winter in the area are present from late September to early May. The arrival of fall birds appears to displace the flocks of Chipping and Brewer's Sparrows in tamarisk of the lower washes. The local breeding form apparently winters in the upper canyons where they were segregated from wintering forms on 4 March and 25 September 1984, 26 October 1985, and 16 December 1989. Nests with eggs have been found from 22 May (1983) to 17 June (1985). Fledglings have been seen from 16 June (1982) to 13 July (1983). Full-sized immatures were seen on 30 Jul\' 1985, and a bird in pre-basic molt was seen on 14 August 1984. Birds seen in mid- September were all in basic plumage. The breed- ing form is intermediate between /. h. dorsalis and J. h. caniceps, which is typical for all juncos that breed in northeastern Arizona (see \\^oodbin-\' and Russell 1945 and Phillips et al. 1964). Chestnut-collared Longspur (Calcariiis orna- fiis). PrimariK a sparse fall migrant recorded onK from reclaimed areas. Records include one seen by EH&A Consultants on 9 October 1979, one in J-1/ N-6 reclaimed area on 2 October 1985, one at N-2 reclaimed area on 16 October 1989, one at J-7 reclaimed area on 1 October 1990, one at N-2 reclaimed area on 30 October 1990, and up to 10 in N-1 reclaimed area 14-17 October 1991. The onK spring occurrence is two males at N-2 reclaimed area on 30 March 1990. Bobolink {Dolichonyx oryzivorus). A sparse migrant. A male in alternate phmiage was seen in N-1 reclaimed area on 17 Mav 1989. Red-winged Blackbird [Agelaiits phoeniceus). A common migrant at ponds throughout the area. Small nimibers nest at several ponds in the lease area. A flock of about 50 wintered on lower Coal Mine and Moenkopi washes in 1988-89. The breeding race is A. p.fortis. Eastern Meadowlark (Stiirnella magna). A sin- gle bird identified 1)\ call, song, and throat color pattern was seen in J-7 reclaimed area 3-4 June 1991. Western Meadowlark (Sturnella neglecta). A common summer resident in reclaimed areas, the highway right-of-way in Long House Valley, the valley of lower Moenkopi Wash, and the J-8 mine plan area. Two nests found in reclaimed areas on 16 May 1989 contained four and five eggs, respective- ly. Another found on 30 May 1989 held five eggs. Broods typically leave the nest in late May-early June and again in late July-early August. A com- mon migrant in reclaimed areas and occasionally in shrublands. A sparse winter resident in reclaimed areas, not returning in numbers until early March. From 2.2 to 3.3 pairs/40 ha have been found in reclaimed areas. The nominate race breeds. Yellow-headed Blackbird (Xanthocephahis xan- thoceplialiis). A common migrant, usualK near ponds, throughout the area. Most records are from April and from mid-JuK' to October A single male was noted at N-1 reclaimed area on 24 Jime 1992. Brewer's Blackbird [Euphagiis cyanocephalus). A common migrant at ponds throughout the area. Small numbers breed at J-7 and other ponds. Also breeds in small numbers in moist tamarisk sites on Moenkopi Wash. A female feeding fledglings was seen at J-7 pond on 5 JiiK' 1985 and at a pond in N-2 reclaimed area on 19 Jul>' 1985. A pair feeding small \'oimg was seen on 10 June 1986 at a pond in N-2 reclaimed area. Great-tailed Crackle (Quiscalus mexicaniis). A sparse migrant. Three males and a female were seen in J-27 reclaimed area on 23 April 1984, a sin- gle male was seen at a pond on 5 May 1989, and another lone male was seen on 18 June 1990. Brown-headed Cowbird (Molothriis ater). A common summer resident in small numbers throughout the area, the largest group being a migrant flock of 147 birds seen on 24 April 1992. Records span from late April to mid- September Black-throated Cray Warblers are the only host species noted from the area. Transect densities range from 0.4 to 8.9 individuals/40 ha. M. a. artcinisiae and M. a. obscunis overlap in northeast- ern Arizona. Probably sparse or absent in the study area before the 1930s (Wootlburx and Russell 1945). Northern Oriole [Icterus galhula). A connnon fall migrant in tamarisk from early August to mid- September. Less numerous in the spring. A male was seen in pinyon-juniper woodland on 11 May 1982. Two were seen in tamarisk on lower Coal 1994] Birds of Northern Black Mesa, Arizona 55 Mine Wash on 27 Ma\ 1986. Another was seen in the same area on 23 lune 19(S6. A female was seen in aspens in the middle fork of Coal Mine Wash on 26 Ma>' 1986. Scott's Oriole {Icterus parisorum). A common summer resident of jnniper sa\anna and open pin- yon-jnniper. Records range from 23 April (1993) to 13 Angust (1986 and 1992). A pair with t^vo small fledglings was seen in the J-S mine plan area on 18 Jnne 1986. This species possesses large territories; fonr closely observed pairs (including the above) had territories averaging 51.6 ha/pair (127 ac/pair) in June 1986. Family Fhi\c;illidae Rosy Finch (Leucosticte arctoa). A sparse win- ter resident. On 3 Januar\' 1984 fom^ indi\'iduals of a "gray-crowned form were seen feeding on Rus- sian thistle seeds on a shadscale-covered Mancos Shale slope on lower Moenkopi Wash. Pine Grosbeak {Pinicola enucleator). A sparse winter resident. Seven were seen in mixed-conifer on 1 Noxember 1992. Two females were seen feed- ing on quaking aspen buds at the head of Yellow Water Canyon on 12 April 1985. On 4 January 1986, 22 females were seen in pin\on-jimiper and Douglas fir on the mesa rim at Ka\'enta Point. B. Mellberg reported a pair in the east fork of Coal Mine Wash on 17 December 1988 in mixed- conifer. Cassin's Finch (Carpodaciis cassinii). A fairly common to common winter resident in wooded terrain throughout the area. It is absent some years. Singing and displaying males on Lolomai Point on 12 April 1985, singing males on 14 June 1984 and 25 June 1986, and a male foraging and carrying food on 23 June 1988 may indicate breed- ing near the rim in mixed-conifer habitats. House Finch {Carpodaciis mexicanus). A com- mon permanent resident throughout the area. Retreats in winter to lower elevations. Fledglings have been seen on 22 July and 11 August 1982. G. Swan found a new nest under the cowling of his Cessna on 3 May 1985, suggesting that the nest causing an exciting engine failure in J. Gibbs' Cess- na four years earlier was also from this species. Transect densities range from 0.4 to 53.3 individu- als/40 ha. Spot-map densities in pinyon-juniper are all 3.8 pairs/40 ha. C m. frontalis is the breeding form. Flocks of up to 300 winter on lower Moen- kopi Wash. Red Crossbill {Loxia curvirostra). An irregular- 1\' common permanent resident in mixed-conifer in the upper canyons. Frequently seen in dense pin- \ on-juniper woodland away from the rim area. Very \'oung fledglings were seen in mi.xed-conifer in Coal Mine Wash on 10 March 1989. Several family groups were noted in upper Moenkopi Wash on 2 May 1989. Another family group was seen in Yel- low Water Can>'on on 5 Ma\' 1992. For the complex of races in the region, see Monson and Phillips (1981). Pine Siskin {Carduelis pinus). A common win- ter resident throughout the area, more numerous in the spring. Oxer 300 were seen feeding on dande- lion seeds {Taraxacum officinale) in a side canyon of Moenkopi Wash on 24 May 1985. Breeding by birds seen fairly commonly in summer in mixed- conifer is unconfirmed. Seemingly more numerous in 1992, at which time it wandered widely during the summer throughout the area, with birds noted in tamarisk on 23 and 29 July, 5 August; 30 were seen in thistles {Cirsium vulgare) at Pond J28-G on 18 August. If breeding occurs, Monson and Philips (1981) would assign them to the nominate form. Behle (1985) assigns birds in the region to C. p. vagrcni. Lesser Goldfinch {Carduelis psaltria). A com- mon summer resident throughout the area. Most are seen from late April to early December. During October the\' feed extensively on seeds of threadleaf groundsel {Senecio longilohus), as a flock of 35+ were obsei-ved doing on 4-5 October 1989. Fledglings were seen on 13 July 1986. The nomi- nate race is the breeding form, and both black- backed and green-backed males are frequently observed. American Goldfinch {Carduelis tristis). Irregu- larly common from early September to mid-May. It frequents weedy roadsides throughout the area. These birds are most mmierous from December to mid-Ma\'. Males usually undergo pre-alternate molt in April and early Ma\ before disappearing. Evening Grosbeak {Coccothraustes vesperti- nus). An uncommon fall migrant with most records being from early October into January. Spring records include two on 14 April 1985, 27 on 29 April 1991, and many widespread flocks in March- April 1993. Nearly 50 were seen feeding on Rus- sian olive fruit on lower Moenkopi Wash on 13 December 1989 and four were seen there on 3 Jan- uar\- 1991. Family Passeridae House Sparrow {Passer domesticus). A common permanent resident at mine shops and facilities. This species was probably absent from the area prior to mine development. Acknowledgments I especially thank Bonnie Dillon, who typed many versions of the manuscript, and the management at Peabodv' who allowed me the time to pursue this project. Reviews by Gale Monson, Steve Carothers, David Ellis, Art Phillips, Brian Maurer, John Spence, and 56 Great Basin Naturalist [Volume 54 Richard Jeiikinson improved the manuscript immeasurably. Peregrine Smith Books gave permission to reprint the Everett Ruess quote from VV. L. Rusho's Everett Ruess: A ta^ahond for beauty. Judy Colemann drew and provid- ed the drawing of the bird petroglyph for reprint covers. Mr. Pat Ryan provided notes and records for several species. Bob Leonard provided a summary of the faunal remains from the archaeological record. After failing several times to find an outlet for this paper, 1 must also thank Dr. James R. Barnes, Dr. Clayton White, Dr. Richard Baumann, and the Editorial Board of the Great Basin Natu- ralist for accepting such a long manuscript and for providing clerical assistance late in the game. My wife, Anne, provided many reviews, criticisms, good discussions, and good company as she always does. Many other people contributed in many ways, and I am grateful to all of them. This paper is dedicated to my parents who, many years ago, stopped for an oriole. Literature Cited American Birds. 1979. Forty-second breeding bird cen- sus. American Birds 34: 84-95. . 19S2. Forty-fifth breeding bird census. American Birds 36: 86-95. . 1983. Forty-sixth breeding bird census. American Birds 37: 84-93. American Ornitik)LOGI.sts' Union. 1983. Check-Hst of North American birds. 6th ed. American Oniitliolo- gists' Union, Washington, D.C. . 1985. Thirty-fifth supplement to the American Ornithologists' Union Check-list of North American birds. Auk 102: 680-686. . 1987. Thirty-sixth supplement to tlu' American Ornithologists' Union Check-list of North American birds. Auk 104: 591-596. . 1989. Thirty-seventh supplement to the American Ornithologists' Union Check-list of North American birds. Auk 106: 532-,538. Balua, R. p., and N. L. Masters. 1980. Avian communi- ties in the pinyon-juniper woodland: a descriptive analysis. Pages 146-169 in Proceedings of a work- shop for managing nongame birds of the Rocky Mountains. USES Technical Report INT-86. Beezley, J. 1974. Personal communication on faunal remains. In: S. Powell, ed.. Excavations on Black Mesa, 1971-1976. A descriptive report. Center for Archaeological Investigations Research Paper No. 48. Southern Illinois University at Carbondale. (Published 1984.) Beiile, W. H. 1978. Avian biogeograpin of the Croat Basin and Intermountain Region. Pages 55-80 in Intermountain biogeography: a symposium. Great Basin Naturalist Memoir 2. Brigham Young Univer- sity, Provo, Utah. . 1985. L tall birtls; geographic distribution and sys- tematics. Utah .Museum of Natural History Occa- sional Publication No. 5. University of Utah, Salt Lake Cit>-. Beiile, \V. IL, E. D. Sorensen, and C. .M. White. 1985. Utah birds: a revised checklist. Utah .Museum of Natural History Occasional Publication No. 4. Uni- versity of Utah, Salt Lake City. Bock, C. E., and B. Webb. 1984. Birds as grazing indica- tor species in southeastern Arizona. Journal of Wildlife .Management 48: 104.5-1049. Bock. C. E., J. H. Bock, W. R. Kenney, and V. M. Hawthorne. 1984. Responses of birds, rodents, and vegetation to livestock exclosure in a semidesert grassland site. Journal of Range Management 37: 239-242. Bradfield, M. 1974. Birds of the Hopi Region, their Hopi names and notes on their ecology. .Museum of Northern Arizona Bidletin No. 48. The Northern Arizona Society of Science and Art, Inc., Flagstaff. Brotherson, J. D.. L. A. Szyska, and W. E. Evenson. 1981. Bird community composition in relation to habitat and season in Betatakin Canyon, Navajo National Monument, Arizona. Creat Basin Natural- ist 41; 298-309. Brown, D. E., ed. 1982. Biotic communities of the Amer- ican Southwest — United States and Mexico. Desert Plants. Vols. 1-4. Carothers, S. W., R. p. Balda, and J. R. Haldeman. 1973. Habitat selection and density of breeding birds of a coniferous forest in the White Mountains, Arizona. In: S. W. Carothers, J. R. Haldeman, and R. P. Balda, eds.. Breeding birds of the San Francis- co Mountain area and the White Mountains, Ari- zona. Museum of Northern Arizona Technical Series No. 12. Northern Arizona Society of Science and Art, Inc., Flagstaff. Christenson, A. L., and W. J. Perry, eds. 1985. Appen- dix L: Faunal remains and bone artifacts. Pages 645-653 in Excavations on Black Mesa, 1983. A descriptive report. Center for Archaeological Inves- tigations Research Report No. 46. Southern Illinois University at Carbondale. 714 pp. Cody, M. L. 1966. The consistency of intra- and intercon- tinental grassland bird species counts. American Naturalist 100: 371-376. . 1985. An introduction to habitat selection in birds. Pages 3-56 in M. L. Cody, ed., Habitat selection in birds. Academic Press, Inc. Cooley, M. E., J. W. Harshbarger, J. P. Akers, and F. W. Hardt. 1969. Regional hydrogeolog>' of the Navajo and Hopi Indian resenations, Arizona, New Mexi- co, and Utah. Geological Sunex' Professional Paper ,521-A. Dean, J. 1989. Paleoenvironment. Chapter 4 in S. Powell, and G. J. Gimimerman, eds., 10,000 years on Black Mesa, Arizona: prehistoric culture change on the Colorado Plateau. Center for Archaeological Investi- gations, Southern Illinois University, Carbondale. Docimient submitted In' Peabody Coal Compan> to the Office of Surface Mining, Denver, Colorado. Ellis, D. H. 1982. The Peregrine Falcon in Arizona: habitat utilization and management recommenda- tions. Institute for Raptor Studies Report No. 1. Emlen, J. T. 1971. Population densities of birds derived from transect counts. Auk 88: 323-342. 1994] Birds of Northern Black Mesa, Arizona 57 191 Estimating hit'eding bird densities tioni transect counts. Auk 94: 455-468. EuLKH, R. C, G. J. GuMMERMAN, T. N. V. Karlstrom, J. S. Dean, and R. H. Hevlv. 1979. Tlie Colorado Pla- teaus: cultural dynamics and paleoenvironment. Sci- ence 205: 1089-1101. Fautin, R. W. 1946. Biotic communities of the northern desert shrub biome in western Utah. Ecological Monographs 16: 251-310. Floyd, M. E. 1982. The interaction of pinyon pine and Gambel oak in plant succession near Dolores, Gol- orado. Southwestern Naturalist 27: 143-147. Franzreb, K. E. 1977. Bird population changes after tim- ber hanesting of a mi.xed conifer forest in Arizona. USDA Forest Ser\'ice Research Paper RM-184. Ganey, J. L. 1992. Food habits of Mexican Spotted Owls in Arizona. Wilson Bulletin 104: 321-326. Ganey, J. L., and R. P. Balda. 1989. Home-range charac- teristics of Spotted Owls in northern Arizona. Jour- nal of Wildlife Management 53: 1159-1165. GuM.MERMAN, G. J. 1984. A view from Black Mesa. Uni- versity of Arizona Press, Tucson. Hack, J. T. 1942. The changing physical environment of the Hopi Indians of Arizona. Report No. 1. Reports of the Awatovi Expedition, Peabody Museum, Har- vard University, Cambridge, .Massachusetts. Haldeman, J. R., R. P. Balda, and S. W. Carothers. 1973. Breeding birds of a ponderosa pine forest and a fir, pine, aspen forest in the San Francisco Moun- tain area, Arizona. In: S. W. Carothers, J. R. Halde- man, and R. P. Balda, eds., Breeding birds of the San Francisco Mountain area and the White Moun- tains, Arizona. Museum of Northern Arizona Tech- nical Series No. 12. Northern Arizona Society ot Science and Art, Inc., Flagstatt. Hall, G. A. 1983. Western Virginia birds. Carnegie Museum of Natural History Special Publication No. 7. Pittsburgh, Pennsylvania. Hall, R. S., R. L. Glinskl D. H. Ellis, J. M. R-wl^kk.a, and D. L. B.\se. 1988. Ferruginous Hawk. Pages 111-118 in R. L. Glinski et al., eds.. Proceedings of the Southwest raptor management symposium and workshop. National Wildlife Federation Scientific and Technical Series No. 11. National Wildlife Fed- eration, Washington, D.C. Hargran E, L. L. 1970. Feathers from Sand Dune Cave: a Basketmaker cave near Navajo Mountain, Utah. Museum of Northern Arizona Technical Series No. 9. HiCKEY, J. J., AND S. A. Mikol. 1979. Estimating breed- ing bird densities on coal lands in Montana and Wyoming. FWSAVELUT-79/03. Fort Collins, Col- orado. HoFFMEiSTER, D. F. 1986. Mammals of Arizona. Univer- sity of Arizona Press, Tucson. 602 pp. Humphrey, P. S., and K. C. Parkes. 1959. An approach to the study of molts and plumages. Auk 76: 1-31. Hunter, W. C, R. D. Ohmart, and B. W. Anderson. 1987. Status of breeding riparian-obligate birds in southwestern riverine systems. Western Birds 18: 10-18. . 1988. Use of exotic saltcedar {Tamarix chinensis) by birds in arid riparian systems. Condor 90; 113-123. Jacobs, B. 1986. Birding on the Navajo and Hopi reserva- tions. Jacobs Publishing Co., Sycamore, Missouri. Johnson, N. K. 1975. Controls ot number of bird species on montane islands in the Great Basin. Evolution 29; 545-567. Johnston, P. 1932. Black Mesa phantoms. Touring Top- ics 24(10): 10-15,48-59. Karr, J. R. 1968. Habitat and avian diversity on strip- mined land in east-central Illinois. Condor 70: 348-357. K.-\RR, J. R., .\nd R. R. RcniL 1971. Vegetation structure and avian diversity in several New World areas. American Naturalist 105: 423-435. Karlstrom, E. 1983. Soils and geomoiphology of north- ern Black Mesa. Chapter 5 in Preliminan report of the 1981 archaeological field season. Southern Illi- nois University, Carbondale. Knopf, F. L., and T. E. Olson. 1984. Naturalization of Russian olive: implications to Rocky Mountain wildlife. Wildlife Society Bulletin 12: 289-298. KocHERT, M. N., B. A. MiLLSAP, AND K. Steenhof. 1988. Effects of livestock grazing on raptors with empha- sis on the southwestern U.S. Pages 325-340 in R. L. Glinski et al., eds., Proceedings of the Southwest raptor management symposium and workshop. National Wildlife Federation Scientific and Techni- cal Series No. 11. National Wildlife Federation, Washington, D.C. Krementz, D. G., AND J. R. Sauer. 1982. Avian commu- nities on partially reclaimed mine spoils in south central Wyoming. Journal of Wildlife Management 46: 761-765. Lanner, R. M. 1981. The pinon pine: a natural and cul- tural history. University' of Nevada Press, Reno. Leonard, R. D. 1989. Anasazi faunal exploitation: prehis- toric subsistence on northern Black Mesa, Arizona. Center for Archaeological Investigations Occasional Paper No. 13. Southern Illinois University at Car- bondale. Lowe, C. H. 1964. Arizona's natural environment. Uni- versity of Arizona Press, Tucson. MacArthi'R, R. H. 1964. Environmental factors affecting bird species diversity. American Naturalist 98; 387-397. MacArthur, R. H., and J. W. MacArthur. 1961. On bird species diversity. Ecology 42: 594-598. McDougall, W. B. 1973. Seed plants of northern Ari- zona. Museum of Northern Arizona, Flagstaff. McKusiCK, C. R. 1986. Southwest Indian turkeys: prehis- tory and comparative osteology. Southwest Bird Laboratoiy, Globe, Arizona. McNiTT, F. 1962. The Indian traders. University of Okla- homa Press, Norman. . 1972. Navajo wars; militan' campaigns, slave raids and reprisals. University of New Mexico Press, Albuquerque. Medin, D. E. 1986. Grazing and passerine breeding birds in a Great Basin low-shrub desert. Great Basin Naturalist 46; 567-572. . 1990. Birds of a shadscale {Atriplex confertifolia) habitat in east central Nevada. Great Basin Natural- ist 50; 295-298. MoNSON, G. 1941. The effect of revegetation on the small bird population in Arizona. Journal of Wildlife Man- agement 5; 395-397. MoNSON, G., .'\ND A. R. Phillips. 1981. Annotated check- list of the birds of Arizona. 2nd ed. University of Arizona Press, Tucson. 58 Cheat Basin Naturalist [Volume 54 Nichols, D. L., and F. B. Smiley, i:ns. 19(S4, Aiipcndix M: Faunal remains from exca\'atc'd sites, 1982. Pages 813-832 in Excavations on Black Mesa, 1982. A descriptive report. Center for Archaeological Investigations Research Report No. 39. Soutlierii Illinois Universit) at Carbondale. 856 pp. Olsen, S. J. 1972. Unpublished Black Mesa Arcliaeologi- cal Project faunal analysis report for the 1971 exca- vations. .Manuscript on file. Center for Archaeologi- cal Investigations, Southern Illinois l'ui\crNit\', Car- bondale. Peabody Coal Co.vipany. 1982. Vegetation and wildlife resources Black Mesa and Kayenta mines. 1981 report. Peabody Coal Company, Arizona Di\ision. (Baseline bird surveys of the J-16 and J-28 coal resonrce areas.) . 1983. Mine plan modification. Appendix I: Vege- tation and fish and wildlife resources information. (Baseline bird sun'eys in the J-19 through J-21 coal resource areas.) . 1984. Vegetation and vsildlife resources. Black Mesa and Kayenta mines. 1983 report. Peabody Coal Company, Arizona Division. (Waterfowl and shorebird census report.) . 1985. Vegetation and wildlife resources. Black Mesa and Kayenta mines. 1984 report. Peabody Coal Company, Arizona Division. (Raptor census report and pin\-on-juniper spot-map census report.) . 1985. Indian lands permit application package. Volume 8, Chapter 10: Wildlife resources. Peabody Coal Company, Arizona Di\ision. (Leasewide base- line wildlife presentation.) . 1986. Vegetation and wildlife resources. Black Mesa and Kayenta mines. 1985 report. Peabody Coal Company, Western Division. (Reclaimed mine spoil and mixed shrub spot-map census report and raptor nesting census report.) Peterse.n, K. L., and L. B. Best. 1985. Nest-site selec- tion by Sage Span-ows. Condor 87: 217-221. Phillips, A. R., J. T. Marsh.\ll, and G. Monson. 1964. The birds of Arizona. University of Arizona Press, Tucson. Philp, K. R. 1977. John Collier's crusade for Indian reform: 1920-1954. Uni\'ersit>' of Arizona Press, Tucson. Potter, E. F. 1984. On capitalization of vernacular names of species. Auk 101: 895-896. Ralph, C. J. 1985. Habitat association patterns of forest and steppe birds of northern Patagonia, Argentina. Condor 87: 471-i83. Rea, A. M. n.d. Unpublished avifaunal remains identifica- tion report for the Black Mesa Archaeological Pro- ject. Manuscript on file. Center for Archaeological Investigations, Southern Illinois University, Car- bondale. Repenning, C. A., and H. G. Pace. 1956. Late Creta- ceous stratigraphy of Black Mesa, Navajo and Ilopi Indian resen'ations, Arizona. American Association of Petroleum Geologists Bulletin 40: 255-294. Rotenberry, J, T. 1985. The role of habitat in avian com- munity composition: physiognomy or floristics? Oecologia 67: 213-217. Rotenberry, J. T., and J. A. Wiens. 1980. Habitat struc- ture, patchiness, and avian communities in North American steppe vegetation; a multivariate analysis. Ecology 61: 1228-1250. Rom. R. R. f97(). Spatial heterogeneit\ and birtl species di\frsit\ . Ecolog\' 57: 773—782. RoosEXEl.T, T. 1913. Across the Navajo Desert. Outlook Magazine 105:304-317. I^YSKR, F. A. 1985. Birds of the Great Basin: a natural his- tor>'. University of Nevada Press, Reno. Santana, E. C, R. L. Knight, and S. A. Te.mple. 1986. Parental care at a Red-tailed Hawk nest tended by three adults. Condor 88: 109-110. Sedc;\\TCK, J. A. 1987. A\ian hal)itat relationships in pin- yon-juniper woodland. Wilson Bulletin 99: 413—131. Sedgwick, J. A., and R. A. Ryder. 1987. Effects of chain- ing pinyon-juniper on non-game wildlife. In: Pro- ceedings, pinyon-juniper conference. USDA Forest Service Intermountain Research Station, General Technical Report INT-215. Sellers, W. D., and R. H. Hill. 1974. Arizona climate, 1931-1972. 2nd ed. University of Arizona Press, Tucson. Seme, M. 1980. Appendix VI: Faunal analysis of material from the 1979 field season. Pages 463-508 in S. Powell et al., eds.. Excavation on Black Mesa, 1979. A descriptive report. Center for Archaeological Investigations Research Paper No. 18. Southern Illi- nois Universit}' at Carbondale. 516 pp. Seme, M., and A. H. Harris. 1982. Appendix IX: 1980 faunal analysis. Pages 321-350 in P. P. Andrews et al., eds.. Excavations on Black Mesa, 1980. A descriptive report. Center for Archaeological Inves- tigations Research Paper No. 24. Southern Illinois University at Carbondale. 359 pp. Sink, C. W., M. C. Trachte, L. K. Miciialik, B. M. EsTES, T. Bastwick, L. M. Anderson, and D. Jes- sup. 1983. Arizona D:ll:2068. In F. E. Smiley, D. L. Nichols, and P. P. Andrews, eds., Excavations on Black Mesa, 1981. A descriptive report. Center for Archaeological Investigations Research Paper No. 36. Southern Illinois University- at Carbondale. 537 pp. S.MiLEY, F. E., D. L. Nichols, .^nd P. P. Andrews, eds. 1983. Appendix X: 1981 faunal recoveiy: frequency tables. Pages 489-498 in F. E. Smiley, D. L. Nichols, and P. P. Andrews, eds.. Excavations on Black Mesa, 1981. A descriptive report. Center for Archaeological Investigations Research Paper No. 36. Southern Illinois University at C^arbondale. 537 pp. Smith, C. C. 1940. The effect of overgrazing and erosion upon the biota of the mixed-grass prairie of Okla- homa. Ecology^ 21: 381-397. Smith, G. W., N. C. Nydegger, and D. L. Yensen. 1984. Passerine bird densities in shrubsteppe vegetation. Jom"nal of Field OrnithologN 55: 261-264. Taylor, D. M. 1986. Effects of cattle grazing on passerine birds nesting in riparian habitat. Journal of Range Management 39: 254-257. Thornthwaite, C. W., C. F. S. Sharp, .\nd E. F. Dosch. 1942. Climate and accelerated erosion in the arid and semi-arid Southwest, with special reference to the Polacca Wash drainage basin, Arizona. USDA Technical Bulletin 808. ToMOFF, C. S. 1974. Avian species diversity in desert scrub. Ecology 55: 396-304. West, N. E., Jr., K. H. Rea, and R. J. Tauscii. 1975. Basic synecological relationships in woodlands. Pages 41-53 in G. F. GifFord and F. E. Busby, eds., 1994] Birds of Northern Black Mesa, Arizona 59 The piiiNon-juniper ecosystem: a sNinposium. Ttah State University', Logan. WiDERLFXHNER, M. P., AND S. K. Diuca'LA. 1984. Rela- tion of cone-crop size to irruptions of four seed-eat- ing birds in California. American Birds 3(S; 840-846. WiENS, J. A. 1973. Patterns and process in grassland bird communities. Ecological Monographs 43; 237-270. WiENS, J. A., AND J. T. RoTENBEKRV. 1981. Habitat associ- ations and community structure of birds in shrub- steppe environments. Ecological Monographs 51: 21-41. WiLLSON, M. F. 1974. Avian community organization and habitat structure. Ecology 55: 1017-1029. Woodbury, A. M., and U. N. Russell, Jr. 1945. Birds of the Navajo country. Bulletin of the University ot Utah Vol. 35, No. 41 (Biol. Series Vol. IX, No. 1).' Wrav, T., II, K. A. Strait, and R. C. Whitmore. 1982. Reproductive success of grassland sparrows on a reclaimed surface mine in West Virginia. Auk 99: 157-164. Received 4 September 1992 Accepted 13 July 1993 Appendix 1. Bird species recorded on northern Black Mesa, Arizona. Relati\ e se; isonal abundance'' Common name and Seasonal Habitat breeding status'' status'' preference'^' Spring Sunmier Fall Winter Common Loon M P F S S — Pied-billed Grebe* M P C U c C Horned Grebe M P S — — s Eared Grebe M P c u c C Western Grebe M P F u F — American White Pelican M P S s s — Double-crested Cormorant M P — s — — Great Blue Heron M P c c c u Great Egret M P — — s — Snow7 Egret M P c c c — Cattle Egret M P S s s — Black-crowned Night-Heron M P F F F — White- Faced Ibis M F c F c — Snow Goose M P — — s s Ross' Goose M P S — S — Canada Goose M,W P, Re U s u F Wood Duck M P — — — s Green-winged Teal M P A u c C Mallard* P P C c c C Northern Pintail* M,S P u c F F Blue-winged Teal M P u s — — Cinnamon Teal?* M P C c c c Northern Shoveler M P C s c u Gadwall M P c c c u Eurasian Wigeon M P s — s s American Wigeon M P F — F F Canvasback M P u s u s Redhead M P c s c c Ring-necked Duck M,W P c — c c Greater Scaup M P s — — — Lesser Scaup M P c — c c White-winged Scoter M P — — s — Common Goldeneye M P s — — — Bufflehead M P F s c u Hooded Merganser M P s — s s Common Merganser M P F s F s Red-breasted Merganser M P U — s — Ruddy Duck M P c c c F Turkey Vulture?* S Ct c c c — Osprey M P,To F s F — 60 Great Basin Naturalist [Volume 54 Appendix 1. C^ontimu'd. Seasonal Habitat l{elati\e seasonal abundance'' Coninion name and breeding status'' status'' preference'^ Spring SuHimer Fall Winter Bald Eagle M McPJ S — — S Northern Harrier W To F U F U Sharp-shinned Hawk* P A,Mc,PJ C F C u Cooper's Hawk* P Mc,PJ,A C C C u Northern Goshawk* P Mc,PJ F S F F Swainson's Hawk M Ms, Pc S s S — Red-tailed Hawk* P Ct,PJ,Mc c c C c Fernigiiious Hawk M Ms u s u u Rough-legged Hawk W Re s — s s Golden Eagle* P Ct s s s s American Kestrel* P Pc,Ct,PJ c c c u Merlin W Js,Ms s — F F Peregrine Falcon* M Rc.PJ s s s — Prairie Falcon* P Ct u u u u Chukar?* P Ct s s s s Wild Turkey AR — — — — — Northern Bohwhite AR — — — — — Scaled Quail AR — — — — — Gambel s Quail AR — — — — — American Coot* M,S P c c c F Sora M P s s s — Sandhill Crane AR — — — — — Semipalmated Pkner M P s — — — Killdeer* P P,Ri c c c F Black-necked Stilt M P u u — — American Avocet M P u u u — Greater Yellowlegs M P F u F — Lesser Yellowlegs M P c F c — Solitary Sandpiper M P u F F — Willet' M P F u F — Spotted Sandpiper?* M,S P C F c — Long-billed Curlew- M P s S — — Marbled Godwit M P F — — — Western Sandpiper M P C c c — Least Sandpiper M P C c c — Baird's Sandpiper M P — s F — Pectoral Sandpiper M P — s S — Long-billed Dowitcher M P F u F — Common Snipe M P,Ri F s F S Wilson's Phalarope M F C s S — Red-necked Phalarope M P U s s — Franklin s Gull M P U — — — Bonaparte's Ciull M P F — s — Ring-billed Gull M P C F F — California Gidl M P U — U — Herring Gull M P s — — — Common Tern M P s s s — Forster's Tern M P F s s — Black Tern M P — s s — Rock Dove T PJ — s s — Band-tailed Pigeon T Pc — s — — Mourning Dove* S RcPJ c c c s Greater Roadnmner P G,Ri s s s s Flammulated Owl* M PJ,Mc u u s — Western Screech-Owl?* P PJ,Mc u u u u Great Horned Owl* P Ct,Mc F F F F Northern Pygmy-(Jwl?* P Me u U F F Burrowing Owl Ar — — — — — Spotted Owl* P? McCt F F F F Long-eared Owl* P PJ,Mc S S — — Northern Saw-whet Owl S Mc s S — S 1994] Birds of Northern Black Mesa, Arizona 61 Appendix 1. Continued. Rflativf seasonal abundance'' Common name and Seasonal Habitat breeding status'' status'' preference'^ Spring Sunnner Fall Wintei C F — U U — C C — C C — S — — C — — C C — F F — F F S F F — — C S U F U S S S c c c c c c u — — S C — S S — — S — C — — C C — C F — S — — CCS C — — cu- re— S — — S — — AAA s s — c c — c c — c c — c c — u c — u c — c c c c c c c c c c c c — s — AAA c c c c c c c c c u c c c c c c c c F F F C F S C C F C C C — s — c c — — F — — F F CCS C C — ecu C C F Common Nighthawk* S PJ — Common Pooi-will?* S PJ U White-throated Swift* S Ct c Black-chinned Hummingbird* S PJ,Mc c Calliope I hunmingbird M PJ — Broad-tailed Hummingbird* S Mc c Rufous I hunmingbird M PJ — Belted Kingfisher M P F Lewis Woodpecker* M Pc F Acorn Woodpecker?* S Pc U Red-naped Sapsucker M Mc F Williamson's Sapsucker* M Mc F Downy Woodpecker* P A,Ri S Hairy Woodpecker* P A,PJ,Mc c Northern Flicker* P Ct,PJ,Mc c Olive-sided Flycatcher M PJ u Western Wood-Pewee?* M,S Ri,Mc c Willow Flycatcher M Ri — Hannnond's Flycatcher M Ri S Dusk\ Flycatcher* S Mc c Gray Fl\ catcher* S PJ,Mc c Cordilleran FKcatcher* s Ct,Mc c Black Phoebe T Ri Say s Phoebe* s Ct,Ms c Ash-throated Flycatcher* s PJ,Mc c Cassins Kingbird* s PJ c Western Kingbird* s PJ F Eastern Kingbird M Re — Scissor-tailed Flycatcher T Js — Horned Lark* P Sg,St,RcJs,Ms A Purple Martin M Ms — Tree S\\ allow M P c Violet-green Swallow* S Ct,PJ,Mc c Northern Rough-winged Swallow* S Ct c Bank S\\ allow M P c Cliff Suallow* M Ct c Barn Swallow* M P c Steller's Jay* P Mc c Scrub Jay* P S,PJ,Mc c Pinyon Jay* P PJ,Mc c Clark's Nutcracker* P Mc c American Crow T,AR Ms s Conmion Raven* P Ct,PJ,Mc A Mountain Chickadee* P PJ.Mc c Plain Titmouse* P PJ,Mc c Bushtit* P PJ,Mc,Ri c Red-breasted Nuthatch* W,S Mc c White-breasted Nuthatch* P PJ,Mc c Pygmy Nuthatch* P Mc c Brown Creeper* P Mc F Rock Wren* S Ms,Ct,PJ C Canyon Wren* P Ct,PJ,Mc C Bewick's Wren* P PJ,Mc c Winter Wren T A — House Wren* M,S A c Marsh Wren M P F Golden-crowned Kinglet W Mc U Ruby-crowned Kinglet?* M Mc C Blue-gray Gnatcatcher* S PJ,Mc C Western Bluebird* P PJ,Mc c Mountain Bluebird* P PcPJ.Mc c 62 Cheat Basin Naturalist [Volume 54 Api'KNDIX I. Coiitiiiiifd. Rclativi' seasonal ahundaiici''' Common name and breeding status' Seasonal status'' Ilahitat preference" Spriuy; Summer Fal Winter Townsend s Solitaire* Hermit Thrush* Varied 'I'hrush American Robin* Northern Mockinj^bird* Sage Thrasher* Bendire's Thrasher* American Pipit Bohemian Waxwing CAxUir W'axwiiig Northern Shrike Loggerhead Shrike* European Starling* Gray Vireo* Solitarv' Vireo* Warbling Vireo* Tennessee Warbler Orange-crowned Warbler* Nashville Warbler Virginia's Warbler* Lucy s Warbler Yellow Warbler Yellow-nimped Warbler* Black-throated Gray Warbler* Townsend s Warbler Hermit Warbler Grace's Warbler* American Redstart Northern Waterthrush Kentucky Warbler MacGilIivra\'s Warbler* Common Yellowthroat Wilson's Warbler Yellow-breasted Chat Western Tanager* Black-headed Grosbeak* Blue Grosbeak* Lazuli Bunting* Indigo Bunting* Green-tailed Towhee* Rufous-sided Towhee* American Tree Sparrow Chipping Sparrow* Brewer's Sparrow* Clay-colored Sparrow Vesper Sparrow* Lark Sparrow* Black-throated Sparrow* Sage Sparrow* Lark Bunting Savannah Sparrow Song Sparrow Lincoln s Sparrow- Swamp Sparrow White-throated Sparrow White-crowned Sparrow Harris' Sparrow Dark-eyed Junco* Chestnut-collared Longspur Bobolink Red-winged Blackliird p A,Ct,Mc C s A,Mc C T Ri — P A.Mc C S G,Js,Ms C s Sg,St,Rc.Mx c s GJs,Ms F w P C T PJ S M PJ s W PJ,Sa s P GJs,Ms c P Ms c s PJ F s PJ,Mc C M,S A C M Ri — M,S S,A,Mc c M Ri — S,M S,Pc,A,Mc c T Ri s M Ri.A s M,S Mc c S PJ,Mc c M McRi — M Ri — S Mc c T PJ — M P s T S,Ri s M S,A c M P,Ri s M Ri c M Ri s S,M Mc c M,S A,Mc c S Ri c M,S Ri F S Ri s M,S Pc.Sg c P S,Pc,PJ,Mc c W Ri — s Js,Pc,P|,Mc c S,M Sg,St,Rc,Pc.G c T Ri — M,S St,Mc,Pc c S G,RcJs,Ms c S G,Ms c P Sg,St,Ms c M Re s M P c M,W Ri F M Ri F M Ri s M Pc s M Ri,S c W Ri s w,s A,Mc c M Re s M Re s M P c 1994] Birds of Northern Black Mesa, Arizona 63 Appendix 1. Continued. Relative seasonal ahiindanee^' Common name and breeding status' Seasonal Habitat status'' preferenee' M? Re P RcJs,Ms M P S P,Ri M Re S PJ M Ri,PJ S Js,PJ w Ms w A,Mc \v,s Mc p Ri,PJ p Mc p Mc s Ri.PI w PJ M Me.PJ P Ms Spring Sununer Fall Wintei Eastern Meadowlark Western Meadowlark* Yellow-headed Blackbird Brewer's Blackbird* Great-tailed Crackle Brown-headed Cowbird* Northern Oriole Scott's Oriole* Rosy Finch Pine Grosbeak Cassin s Finch?* House Finch* Red Crossbill* Pine Siskin?* Lesser Goldfinch* American Goldtinch Evening Grosbeak House Sparrow* tlu- L'tJlKT ■ md spflliiii; iif tlic Anu-ntaii Oniillioldiiisls' Lr "ConiiiiDii nanus toll pected. "M = migrant S = summer resident W = winter resident P = permanent resident AR = only in archaeological record T = transient If two seasonal status listings are given, the principal status is gi\en first. ^'S = montane senib P = ponds Sg = sagehnish Pc = chained pinNon-juniper St = saltbush .Ms = mixcd-shruli G = greasewood A = aspen groves Re = reclaimed mine spoil Ct = cliffs, talus slopes, wash banks Js = juniper savanna PJ = pinyon-juniper Ri = riparian habitats Mc = mi.xed-conifer If a species is found in two or more habitat t\pes, the principal one is given. II it is Ibund tl only the breeding habitat(s) are those listed. "A = abundant S = sparse C = common Cs = casual F = fairb common Ac = accidental U = uncommon 1983, 19S.5. 19S7, 1989). * = bi ding conhi brrrding sus- lighout, "To" is designated. If breeding is confinued or suspected. Great Basin Naturalist 54(1), © 1994, pp. 64-70 EFFECTS OF COBBLE EMBEDDEDNESS ON THE MICRODISTRIBUTION OF THE SCULPIN COTTUS BELDINGI AND ITS STONEFLY PREY Roger J. Haro'-2 and Merlyn A. Brusvenl-^ Abstiuct. — Lahoraton- experiments were undertaken to assess the effeets of three levels of eohhle eniheddedness on the niicrodistrihution of the sculpin Cotttis heldin^i and its stonefly prey, Skivala americuna. Experiments were con- dueted separateK and totjether as predator and prey in temperature- and flow-eontrolled artiiieial streams. When tested either separate!} or tojiether, both the predator seulpin and its stonefly prey oceurred in siy;nifieantK' greater numbers on substrata ha\ing unembedded eobbles than substrata having half- or eompletely embedded eobbles. Stonefly densi- ties were greater in substrata having unembedded cobbles even though predator densities within the more embedded cobble patches were significantly lower. These findings support the hypothesis that higher predator densities influence prey densities less than the structural habitat (jualitN of unembedded-cobble patches. Keij words: pndator-prcy. cobble cuibt'ddedness. nunliihal effects, stoneflics, scidpins. nonpoiiit source sedimenta- tion. CJottus beldingi. Reduced summer flows and increased sedi- mentation in many western North American streams may significantly diminish the size and avaihibility of adequate microhabitat patches for benthic fish and insects. Sedimen- tation from agricultural sources has been linked to pronounced changes in the trophic structure of lotic fish assemblages (Berkman and Rabeni 1987) and may affect macroinver- tebrate conmumit\' structure, further altering trophic relations within the lotic food web. Such trophic changes, in part, may result from alterations in prey refugia brought about by the embeddedness of cobble substrata. Brusven and Rose (19(S1) found that cobble embeddedness significantly influenced the vulnerability of two insect predators, Hesper- operla pacifica (Plecoptera: Perlidae) and Rhy- acophila vaccua (Trichoptera: Rhyacophili- dae), to predation b\' Cottits rhotJicus. They suggested high sculpin predation success in the embedded substrata was due to the loss of macroin vertebrate refugia under cobbles. Microhabitat shifts by macroinvertebrate prey in response to vertebrate and macroin- vertebrate predators have been reported b\ several workers (Stein and Magnuson 1976, Stein 1977, Peckarsky and Dodson 1980, Peckarsky 1983). Feltmate et al. (1986) found that, under laboratoiy conditions, Para' and ecological overview. Pages 203-224 in C. W. Kerfoot and A. Sih, eds., Predation — direct and in- direct impacts on acjuatic communities. University Press of New England, Hanover, New Hampshire. Sill, A., L. B. ICvrs, and R. D. Moore. 1992. Effects of predation of sunfish on density, drift, and refuge use of stream salamander larvae. Ecology 73: 1418-1430. Stein, R. A. 1977. Selective predation. optimal foraging, and the predator-prey interaction between fish and crayfish. Ecology 58: 1237-1253. Stein, R. A., and J. J. Magnuson. 1976. Behavioral response of crayfish to a fish predator. Ecology 57: 571-561. Received 23 November 1991 Accepted 24 Augiifit 1993 Great Basin Naturalist 54(1), © 1994, pp. 71-78 PERSISTENT POLLEN AS A TRACER FOR HIBERNATING BUTTERFLIES: THE CASE OF HESPERIAJUBA (LEPIDOPTERA: HESPERIIDAE) Amy E. Berkhousen^- and Arthur \I. Shapiro^-^ Abstract. — Pollen grains of plants with well-defined flowering seasons may persist on insects through episodes of dormancy, such as hibernation. When readily recognizable and possibly confounding ta.\a can be excluded, these pollen grains can sene as direct evidence of life-histon' phenomena that are notoriousK' difficult to verify in the field. Pollen of the autumn-flowering composite Chrysothainniis naiiseosus was used to demonstrate that the common montane skip- per, Hesperia jtilxi. hibernates as an adult in the Sierra Nevada. This is the first demonstration ot adult ovenvintering in a temperate-zone hesperiid and may represi'nt the smallest butterfly known to overwinter in a cold climate. Key uorcis: pollen carryover, hibernation, phenology, body size, Hesperia juba, Chrysothamnus nauscosus, Nymphalis antiopa. Many life-history phenomena of insects, including seasonal dormancy and migration, can be inferred from phenological data but are notoriousK difficult to dononstrate directly in the field; the animals are too few and/or diffi- cult to find. Many Holarctic nymphaline but- terflies have been suspected of hibernation since the 18th century, but there is still no direct evidence in many of them. Direct evi- dence requires either the post-hibernation recovery of individuals marked the previous season (or otherwise identifiable), or the dis- covery of hibernating individuals and the demonstration that they are able to survive winter in situ and enter the reproductive pool in spring. Despite the rarity of such evidence, the hypothesis of adult hibernation is repeat- ed uncritically in most modern accounts of nymphaline biology. It is much more difficidt to persuade anyone that a butterfly belonging to a lineage hitherto unsuspected of hiberna- tion in fact ovei'winters as an adult. Kettlewell (1961) and Ketdewell and Heard (1961) were able to use a radioactive particle originating from an atmospheric nuclear test in the Sahara Desert to trace the origin of a migrant moth {Nomophila noctiiella Schiff., Pyralidae) collected in Britain. Pollen grains are a more prosaic surface contaminant of insect specimens which under favorable cir- cumstances can serve the same purpose. Demonstration of even a single adherent grain of pollen of an autumn-flowering plant on a spring-collected insect could document the ovenvinter survival of that individual. There seems to be no prior docimientation of such long-term pollen persistence. We applied the concept to the Californian skipper, Hesperia juba Scudder (Hesperiidae), which has been suspected of adult hibernation, using as our index pollen species the composite shrid) Chrysothamnus nauseosus (Pall.) Britton, with encouraging results. The difficulty of verifying hibernation is illustrated b}' Shapiro's previous studies of H. juba (Shapiro 1981). He marked 104 individu- als at Donner Pass in the Sierra Nevada in September 1979. In June 1980 he captured 18 individuals in the same area, but none was marked. This was an unusually large-scale attempt at a direct demonstration of overwin- tering; hibernating nymphalids seldom occur at similar densities. Shapiro noted that the negative result was uninterpretable; the only meaningfid result would have been the cap- ture of one or more marked specimens. Natural History of Hesperia juba Hesperia juba is a conuiion and widespread montane skipper in California and adjacent states. Despite its commonness, its life histoiy remains very poorly known. MacNeill (1964) described the early stages based on laboratoiy 'Section of Zoolog\- and Center for Population Biology, University of California, Davis. California 9,5616. ^Present address: 808 Comet Dr. #104. Foster City, California 94404. ■'Antlior to ulidui reprint requests sliouki \iv addres.sed. 71 72 Great Basin Natuhalist [Volume 54 rearing hut did not understand its plienoloi^). Misled 1)\ pooling data from a \er\ heteroge- neous mix ol localities, he wrote that the adults are present from April through October, with some \ariation according to localitv'; evidently emergence is rather continuous and there are no distinct seasonal broods. Emmel and Enimel (1973) recorded two dis- tinct tlights in southern C>alifornia (April-Jmie and August-Septemher); and all suhsequent local or regional data, from Oregon (Dornfeld 1980) to Baja California (Brown et al. 1992), have been similar. Shapiro (1979) was the first to point out that the phenology of H. jiiha was not only well defined but very unusual for a skipper. He presented eight years of Donner Pass data (1972-79) resulting from a biweekly sampling program. This study is now in its 21st year, and the pattern evident in 1979 has continued with great consistency (Table 1). Hesperia juba occurs at 5 of the 10 stations on Shapiro's northern California transect and is definitely resident at 4 of these; at all 4 it is spring/autumn bivoltine with slight variation in phenology, over an altitudinal range of 1500-2750 m. No other hesperiine skipper is bixoltine at Donner Pass (2100 m) or higher, and no other skipper Hies either so early in spring or so late in autiunn. The only other bivoltine hesperi- ine in the region is Polites sabuleti tecumseh (Grinnell) at the lower limits of its range (1500 m on both the Sierran east and west slopes, but only irregularly bivoltine on the west). It is only about 60% the size o{U. juba, but even so it emerges later in spring (much later at Donner Pass, where it is univoltine). At Don- ner, //. juba is often one of the first species to fly after snowmelt, along with the presumably adult-hibernating nymphalids and pupal- hibernating Celastrina argiolus echo (W.H. Edwards)(Lycaenidae) and Pontia occidentalis (Reakirt)(Pieridae). These and other circum- stances described in Shapiro (1979) led to the suggestion that H. juba hibernates as an adult. Nonetheless, Scott (1986) ignored this sugges- tion, proposing instead that //. juba ovenvin- tered as a larva. An overwintering larva would not be expected to feed and grow under a snowpack that normally persists for 6-7 months. Thus, growth would be limited to the periods of good weather between autumn flight and onset of snow, and between snowmelt and some (pupation) time before the spring flight — in all, a few weeks. Skippers grow slow- ly e\'en under seemingly optimal conditions. Table 1. Phenolog\' of Hesperia juba at Donner Pass (2100 m) in the Sierra Nevada of California, based on rouglily biweekly sampling, 1972-92. Year First thght Second (light \.l' 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 198.5 1986 1987 1988 1989 1990 1991 1992 \'.24-\i.7 not seen vi.9 vi.ll v.l4-vii.l vi.4 vi.l4-vii.l vi.l-vii.l2 vi.7-vii.5 vi.7-vi.21 vi.23— vii.9 vi.5-vii.l2 v.27-vi.20 v.23-vi.l9 v.l7-vi.l5 vi.2 iv.26-v.l4 v.20-vi.l8 v.l2-vi.2 vi.2-vii.5 iv.28-vi.8 viii.lO-x.4 ix.7-.\.5 viii.24-ix.27 ix2.-ix.30 vii.20-x.8 ix.2-ix.23 viii.15-x.23 ix.4— ix.30 ix.4-ix.27 viii.6-ix.6^' ix.l-ix.20 viii.30-x.26 ix.6-ix.22 viii.20-x.5 ix.4 \'iii.ll-ix.l7 viii.19-ix.22 viii.28-x.8 ix.5-ix.l8 ix.4-x.l2 viii.22-ix.26 64 75 82 49 90 45 54 61 36 54 48 78 63 81 70 97 72 95 61 79 78 75 82 96 90 61 96 89 50 70 85 102 90 110 70 115 101 116 94 116 "Number ol'clays from last observation of first tliabt lo first oliscnation of sc-coiul flijiht. "Number of days from first oliservation of first fli^bt (o first olisei-vation of seeoncf flight. 'No late-September sample taken (investigator out of eountry). 1994] Persistent Pollen on Hesperiajvba 73 While attempting to carry various stages of H. jiiba overwinter in the hihoratory, we reared a single larva honi egg to (male) adult in 83 days at a continuous temperature of 26 °C under outdoor photoperiod on growing Poa sp. (Gramineae). This indi\idual devel- oped without interruption (except for molts). We estimate the time available for circum- hibernal larval development in an average year to be on the order of 75-85 da\'s, nearly all with night temperatures below freezing and with afternoons reaching 20 °C for per- haps 2-3 hr/da\'! Most Holarctic Hesperia are univoltine (MacNeill 1964) although phenologically diverse even in a given location. In the north- eastern United States, H. metea Scudder flies in early spring, H. sossocus Harris in late spring, H. attains (W.H. Edwards) in summer, and H. leonardus Harris at the end of summer into autumn. In northern California there are distinctive populations of the H. comma L. complex that are univoltine from June to October in different localities (more than one species may be involved); H. nevada (Scud- der) flies in earK' to midsimimer at high eleva- tion; H. Imdseyi (Holland) is univoltine in late spring-early summer, slightly earlier than sympatric populations of the comma complex; only H. cohmihia (Scudder) is bivoltine, flying in foothill habitats in midspring (oxerlapping H. lindseiji a little) and again in early autumn (overlapping late "comma"). All Nearctic Hes- peria appear to feed on perennial bunchgrass- es, and all except H. jiiba have seasonal phe- nologies consistent with larval overwintering (albeit in different instars). One potential alternative explanation of the H. juba phenol- ogy, apparently falsified by this study, is that the nominal species juba consists of two whol- ly allochronic (spring and fall) univoltine pop- ulations, indistinguishable phenotypically. (A situation of this sort occurs in the comma complex on the western slopes of the Sierra Nevada, apparently involving sibling species that are, however, weakly phenotypically dis- tinguishable.) The apparently universal sym- patry of the putative populations argues against this hypothesis; if they were truly independent, surely there would be places where one occurs without the other. The second flight of H. juba at Donner Pass coincides with peak flowering by the composite shrub loibber rabbitbrush, Chnjso- tliatunus nauseosus. It and the polygonaceous subshriib Erio374 grains examined (<7%). Chrysotliamnus pollen was unevenly distributed among individuals: most grains were found on specimens bearing tew pollen grains overall, and none or only one or two other pollen species. Specimens with heavy and diverse pollen loads tended to have no detectable Chnjsothamnus. We interpret this as indicating that individuals just out of hiber- nation, which have had little chance to feed, are most likely to have cany over grains from the previous autumn. Subsequent feeding would either dislodge such grains or bury them in new pollen of other species, making them difficult to see. (Because fewer plants are in flower in early spring at the colder, drier east end of Donner Pass than in the west, spring specimens collected in the east are usually taken from bare soil and have veiy light or no detectable pollen loads in compari- son to those from the west. Unfortunately, specimens taken before 1992 are merely labeled "Donner Pass. ") There are several composite genera pre- senting more or less similar spherical, tricol- porate, echinate pollens that occur at or near Donner Pass. Several of these bloom at mid- Table 2. Occurrence of Chrysotliainnii.s pollen grains on spring-collected Hesperiajuba from Donner Pass. Date of capture Total pollen grains Ch, ■ysothainnits (all species) Pt )llen grains vi.6.88 2 2 vi.6.88 0 0 V. 12.90 10 10 V. 12.90 10 O'' vi.1.90 10 10 v.8.92 15 1 v.8.92 >50 0 v.8.92 >100 0 v.8.92 7 0 v.8.92 2 0 v.8.92 4 0 v.8.92 2 1 v.8.92 10 0 v.8.92 20 0 v.8.92 7 0 V. 16.92 25 1 V. 16.92 >100 0 17 >;374 2,5^' **Onc' amliigiioiis grain on this thaimuis. .12.90 spccinun not countt-d as Chnj>i()- summer, in between the two flights oi H.jiiba, and are not at issue. Aster and SoU(la 1.50 0 >100 0 13 6 0 0 7 7 14 8 >293 30 •'Elevation 1501) m, ChnisDilitiiiuiii.s i)rc-srnt. ''Elevation < 10(1 iik Chnisollnniiniis ahseiit. 1994] Persistent Pollen on Hesperiajuba 77 this is its exclusive mode of ovenvintering. At lower elevations the time constraint on fall/spring larval development is less severe, and such a phenology becomes at least plausi- ble. It would rarely be possible at Donner, but in rugged relief many microclimates occur, some of which might allow lai^val ovei^winter- ing at least in some years. Some autumn females from Donner will lay fertile eggs, though they usually must be confined for at least a week before they do. Of 9 Donner females confined in autumn 1992, 3 laid a few fertile eggs and contained spermatophores; 2 contained matiue or nearly matine eggs but no spermatophore; and 4 contained neither eggs nor a spermatophore at death. In 1992, 1 1 females were collected at Donner; of these 2 had mature eggs and none had sper- matophores (Table 4). In addition, 3 were taken at Carson Pass, Alpine Co. (2700 m), 27 September 1992; of these 2 had neither eggs nor a spermatophore and the thiid had man>' eggs but no speniiatophore. InterestingK', this last was being courted by a male when taken, the only courtship observed in autumn 1992. Perhaps females do not become attractive to males imtil they carry mature ova. Ford (1975) states categorically that nymphalids hibernate as virgins; clearly that need not be true for H. jiiha. (We attempted a small-scale experiment at carr\ing eggs ovenvinter 1991-92 at Don- ner, but all disappeared. We have also failed to carry third-instar larvae overwinter in refrigerators or freezers.) Spring females are almost always in breeding condition; 8 of 9 spring 1992 females dissected had both mature eggs and a spermatophore. H. juha is the largest hesperiine skipper in the Sierra and one of the two largest in Cali- fornia, but it is also the smallest butterfl\' known to hibernate as an adult in a climate with severe winters. A variety of relationships between bod\' size and thermal biolog\' has been postulated in Lepidoptera (Douglas 1986, Miller 1991). Naively, one might sup- pose that the heat-loss properties implied by surface/volume ratios might impose a lower limit of body size on butterfly hibernators. No hibernating insect can keep itself warm over winter by metabolic thermogenesis alone. Ability to hibernate must be related to the abilitv' to get into a wanii microclimate, lower- ing the freezing point b\ biochemical mecha- nisms, insulation, or some combination of Table 4. Repi"()ducti\e condition of 14 //. jiiha feiiuiles collected in tlie Sierra Nevada in antumn 1992. Date Local it\ N imiher of Spc ■nnatophore 111 atiire ova present? ix.2 D< )nner P; iss 0 no ix.2 " 0 no ix.2 " 0 no ix.18 " 0 no ix.18 " 0 no ix.18 " 3 no ix.18 " 0 no ix.18 " 13 no ix.18 " 0 no ix.18 " 0 no ix.26 " 0 no ix.27 C; iir.son Pa LSS >25 no ix.27 " 0 no ix.27 " 0 no these. Bod\' size comes into play primarily in spring or autimin when the insect is active for part of the da\- but at risk for sudden, critical decrease of ambient temperature when the sun is obscured, or at dusk. A disadvantageous surface-to-volume ratio, uncompensated by insulation, could keep the insect from reach- ing shelter before it was immobilized by cold. We have watched nymphaline butterflies return to lava jumbles in late afternoon in the Sierra and quickK' crawl to shelter, but we do not know where H. jiiba go. Many insects much smaller than H. juha hibernate successfully in cold climates, but none is a butterfly. Those butterflies generally supposed to hibernate in the Holarctic (Nymphalidae: Nymphalini and Vanessini; Pieridae: Gonepterijx in the Palearctic and the Californian Zerene eunjdice Bdv.) are remark- ably imiform in size, though many butterflies of similar size do not hibernate. This is an interesting point, but we know too little of the energetics and physiology of butterfl\ hiber- nation to assess its significance. We merely note that H. juha is smaller than other known hibernators but is exceptionalK' large for a hesperiine skipper. Acknowledgments This study was made possible by a Univer- sity of California President's Undergraduate Fellowship awarded to AEB. We thank Rick Harris and Robert J. Munn for help with SEM; Robbin Thoip, Grad\' Webster, and Jim Doyle for discussion and suggestions; Owen 78 Gkkai- Basin Natukaust [\blunH' 54 K. Da\is and John Brooks ior checking their SEM files lor ns; and C. Don MacNeill for conipariny notes on H. jiiha biology and for acknowledging (by letter of 2 Jnne 1992) that "I am now a believer in yonr snggestion that at least some ll.jiiba oveiAvinter as adnlts!' Literature Cited Brown, J. W., H. G. Rkal, and D. K. Faulkner. 1992. Buttcrflifs of Baja California: faiinal sui"\'t'y, natural histoiy, consei-vation biologv'. Lepidoptera Researcli Foundation, Beverly Hill.s, California. 129 pp. CoiRTNEY, S. P., C. J. Hill, .\nd A. Westernl\n. 19S2. Pollen carried for long periods by butterflies. Oikos 38: 260-263. Dornfeld, E. 1980. The butterflies of ()reii;on. Timber Press, Forest Grove, Oregon. 276 pp. Douglas, M. M. 1986. The lives of butterflies. University of Michigan Press, Ann Arlior. 241 pp. Em.mel, T. C, and J. F. Emmel. 1973. The butterflies of Southern California. Natural History Museum of Los Angeles County, California. 148 pp. Ford, E. B. 1975. Butterflies. Fontana New Naturalist, Collins, London. 368 pp. Handel, S. N. 1983. Pollination ecology, plant popula- tion structiu-e, and gene flow. Pages 163-211 in L. Real, ed.. Pollination biolog\'. Academic Press, Orlando, Florida. Hendrlx, W. H., IIL T. F. Mueller, J. R. Phillips, .^nd O. K. Dams. 1987. Pollen as an indicator of long- distance movement of Heliothis zea (Lepidoptera: Noctuidae). Environmental Entomology 16: 1148-1151. Ke'ITLEWELL, H. B. D. 1961. A report on the in\estiga- tion into a radioactive migrant in 1961. Entomolo- gist 94: 49-52. Kehlewell, H. B. D., and M. J. Heard. 1961. Acciden- tal radioacti\'e labelling of a migratoiy moth. Nature (London) 189: 676-677. M.\cNeill, C. D. 1964. The skippers of the genus llespe- ria in western North America (Lepidoptera: Hes- periidae). University of California Pul)Iications in I'-ntomology 35: 67-77. MiKKOLA, K. 1971. Pollen anaKsis as a means of stud\ing the nugrations of Lepidoptera. Annales Entomoiogi- ci Fennici .37: 1.36-1.39. Miller, W. E. 1991. Positive relation between body size and altitudi' of capture site in tortricid moths (Tort- ricidae). Journal of the Lejiidopterists Societ\' 4.5: 6fS-67. Sfxrrr, J. A. 1986. The butterflies of North America. Stan- ford University Press, Stanford, California. .583 pp. SiiAi'iRO, A. XL 1979. Does Uespcha juha (Hesperiidae) hibernate as an adult? Journal of the Lepidopterists' Society 33: 258-260. . 1981. How to do a Beld experiment. News of the Lepidopterists' Society 1981(3): 41-42. . 1986. Seasonal phenology and possible migration of the mourning cloak, Nyinpluilis anfiopa (Lepi- doptera: Nymphalidae) in California. Great Basin Naturalist 46: 112-116. Solomon, A. M., J. E. King, P. S. Martin, and J. TllONi\S. 197.3. Further scanning electron photomi- crographs of southwestern pollen grains. Journal of the Arizona Acadenu' of Sciences S: 13.5-157. Tepedino, V. J. 1983. Pollen carried for long periods by butterflies: some comments. Oikos 41: 144-145. Venarles, B. a. B., and E. M. Barrows. 1985. Skippers: pollinators or nectar thieves? Journal of the Lepi- dopterists" Society 39: 299-312. Waser, N. M. 1983. The adaptive nature of floral traits: ideas and evidence. Pages 241-285 in L. Real, ed., Pollination biology. Academic Press, Orlando, Flori- da. WiKLUND, C, T. Eriksson, and H. Lundrerg. 1979. The wood white butterfly, Leptidea sinapis, and its nectar plants: a case of mutualism or parasitism? Oikos 33: 358-362. . 1982. On the pollination efficiency of butterflies: a reply to Courtney et al. Oikos 38: 263. Received 19 Janiianj 1993 Accepted 4 August 1993 Great Basin Naturalist 54(1), © 1994, pp. 79-85 BREEDING ECOLOGY OF LONG-BILLED CURLEWS AT GREAT SALT LAKE, UTAH Peter W. C. Patoni and Jack Daltonl Abstr.\CT. — We quantified nest site cliaracteristies, breeding densities, and migraton clironolog>- of Long-billed Curlews at Great Salt Lake, Utah. Tbe speeies is apparenth deelining in Utah, and little is know about their breeding eeologN in the eastern Great Basin Desert. This study was designed to provide wildlife biologists with baseline data useful for their suecessful management. Curlews arrived in northern Utah in late March and generally departed by mid-August. Nest densities at Great Salt Lake ranged from 0.64 to 2.36 males/km^. The habitat at curlew nest sites con- sisted of significantly shorter \egetation than nearby random locations (.v = 5.7 versus 9.0 cm, respectively; P < .01). Nests tended to be located in small patches of vegetation near barren ground. Maintenance of relatively short vegeta- tion appears to be important in managing curlew habitat. In addition, only 2 of 10 nests we monitored in 1992 were successful, with most lost to mannnalian predators. Fmther research is needed to determine the impact of mammalian predators on curlew populations. Ket/ words: Long-hiUed CtirJcic. Numenius americanus, nest site charaeteristics. initiation chronologij, Utah. Long-billed Curlews {Niiinenius ameri- canus) historically were a common species in the grasslands of North America (Pampiish 1980). Although quantitative population trend data are limited, it appears that habitat alter- ations and hunting dramaticalK reduced pop- ulations throughout their breeding range (Allen 1980, Pampush 1980). In Utah, Long- billed Curlews are presently being considered for listing as a sensitive species due to declin- ing populations in the northern part of the state (Frank Howe, Utah Division of Wildlife Resomces, Salt Lake City, personal communi- cation). However, reasons for this decline are unknown. Therefore, wildlife managers in Utah require quantitative information on their breeding ecology in the eastern Great Basin Desert to successfully manage this species. Two variables that wildlife biologists can manage to some extent are vegetation and predators. Previous studies in Idaho (Bicak et al. 1982, Redmond and Jenni 1986), Oregon (Pampush 1980), and Wyoming (Cochran and Anderson 1987) suggest that Long-billed Curlews select nest sites in grasslands with relatively short vegetation. Changes in vegeta- tion height due to field fertilization, grazing, and precipitation can significantly affect curlew nest success (Bicak et al. 1982, Red- mond and Jenni 1986, Cochran and Anderson 1987). In addition, predators can have a major impact on a curlew population because Long- billed Curlews initiate only one clutch per year and do not re-nest once a nest has been depredated (Redmond and Jenni 1986). Little quantitative information has been published on the breeding ecology of Long- billed Curlews in Utah. Wolfe (1931) provided qualitative information on their habitat char- acteristics, and Forsythe (1972) described four nests found near Great Salt Lake. Our objec- tive is to provide quantitative estimates of curlew migration chronology, current distribu- tion and breeding densities, nest success, and nest site habitat characteristics at Great Salt Lake so that biologists managing this shore- bird in northern Utah will have baseline infor- mation. Study Area Our principal study areas were three state- owned wildlife refuges located along the east- ern shores of Great Salt Lake (Paton and Edwards 1990): Howard Slough Waterfowl Management Area (WMA), 311 ha surveyed (41° 10' N, 112° 10' E); West Layton marsh, 339 ha (4rO'N,112°0'E), and West Warren WMA, ami Wildlife, I'tali CoopiTath f Fisli cli Unit, L'tali State University. Logan, Utah 84.322. 79 80 Great Basin Natl halist [\'()luiiie 54 400 ha (4r20'N,112°05'E). Satellite study sites included North Ogden Ba\' WMA, 500 ha (4ri5'N,lI2°10'E); Harold Crane WMA, 1300 ha (4r20'N,112°05'E); Locomotive Springs WMA, 1000 ha (41°40'N,112°55'); and northeast of Saltair Beach, 600 ha, (40°45'N,112°10'E). All sites receive approxi- mately 25-38 cm of precipitation annualK (Greer 1981) and are located at an elevation of 1283-1289 m. Marsh vegetation in these areas is dominated b\ bulrush iScirjXis spp.) and cattail (Typlui spp.). Upland habitats are domi- nated by greasewood {Sacrobatus vennicula- tiis) and several species of Chenopodiaceae (Salicornia europaea, Bassia hyssopifolia, KocJud scoparUi, Suacda calceolifonnis). Methods Fieldwork was conducted 14 April-11 August 1990, 27 March-19 September 1991, and 19 March-10 September 1992. To deter- mine curlew migratory chronologx', distribu- tion, and breeding densities, we surveyed principal study areas one day per week 1 April-31 August, except 1990, using spot- mapping techniques (Redmond et al. 1981). Surveys were started at sunrise and continued for 3-5 hoins per day. Curlews were censused using onl\ spot-mapping techniques at West Warren in 1992. Satellite study sites were vis- ited 1-3 times per month all 3 years, with curlews counted from road transects dining shorebird sui-veys (Paton et al. 1992). Weekly census data at Howard Slough and West Lay- ton from 1991 and 1992 were compared using a paired t test. To determine their breeding chronology in Utah, we actively searched for curlew nests in 1992 following methods outlined in Redmond (1986). Egg-laying dates for active clutches were determined using egg-floating tech- niques (Hays and LeCroy 1971). Obsei-\'ations of juveniles in 1990 and 1991 were used to supplement chronology data gathered in 1992. Clutch initiation dates for juveniles observed in the field were calculated b\' esti- mating age of the chick and then back-dating based on a 28-day incubation period (Red- mond and Jenni 1986) and 6-day egg-la\ing period (Cochran and Anderson 1987). We determined cmlew nest site character- istics based on nests found in 1992. Vegetation was quantified using a line-intercept tech- nique (Ha>'s et al. 1981:40). To minimize the probability of attracting predators to active nests, measurements were made <1 week after nests either hatched or failed. Nest site habitat characteristics were quantified along four 15-m transects initiated at the rim of each nest scrape, with transects arranged in the four cardinal directions. To quantify the curlew habitat patch use patterns versus the available landscape, each nest had a paired set of transects, centered on a point located 50 m in a random direction (hereafter referred to as random sites). Random sites were located only in areas with potentially suitable habitat (i.e., dry, upland vegetation < 15 cm tall; Pam- push 1980, Redmond and Jenni 1986, Cochran and Anderson 1987). Vegetation height was measured at 0.5-m increments along the transect, starting at the nest rim (that is, 31 points per transect). The height of the tallest plant within 5 cm of the transect was measured. Plant species composition at nests and random sites was determined b\ measuring to the nearest 1 cm each plant species that touched the transect tape. For vegetation coverage analyses, we classified each 1-cm segment along transects as either live vegetation, dead vegetation, or barren groimd. We compared vegetation height at nests and the paired random sites with a paired t test to quantify vegetational differences between used and available patches. To quan- tify variation in vegetation height as a func- tion of the distance from plot center, at both nests and random sites, we categorized the data into five 3-m-long distance segments. These distance segments were then compared using analysis of variance (ANOVA) and Dim- can's Multiple Range Test to determine which segments differed using PROC ANOVA in SAS (SAS Institute 1988). Alpha values <.05 were considered statistically significant. We compared ground coverage between nests and random sites using a paired t test for three vegetation categories (live, dead, and barren ground). To determine if the three veg- etation categories differed as a function of the distance from plot center, we again classified the data into five 3-m-long segments. The dis- tance segments were then compared using ANOVA and Duncan's test, at both nests and random sites. 1994] Long-billed Curlew in Utah 81 Table 1. Maximum number of Long-Billed Curlews eounted during weekly censuses at two studv sites at Great Salt Lake. Howard Slough WM A W. Layton WMA Date 1990 1991 1992 1990 1991 1992 Apr 1-7 NC' 5 4 NC 1 1 Apr 8-15 NC 13 6 NC 12 4 Apr 16-22 NC 8 10 NC 5 4 Apr 23-30 NC 12 7 NC 4 10 May 1-7 NC 9 14 NC 3 15 May 8-15 NC 6 12 NC 4 23 Mav 16-22 NC 5 14 NC 6 15 .\ lav 23-31 NC 9 11 NC 3 15 Jun 1-7 4 5 11 NC 5 14 Jun 8-15 0 3 4 2 5 9 "|un 16-22 1 0 2 2 3 3 jun 23-30 0 1 6 3 6 3 |u] 1-7 0 2 3 3 9 5 (ul 8-15 1 1 10 1 2 10 [ul 16-22 0 0 10 3 5 8 Jul 23-31 0 0 20 1 8 5 Aug 1-7 0 0 0 0 4 8 Aug 8-15 0 0 1 0 1 3 Aug 16-22 0 0 0 0 0 0 Aug 23-31 1 0 0 0 0 0 •'No census data Results Migration and nesting chronology. — Our weekly surveys were generally initiated 1 week after Long-billed Curlews started to arrive in Utah. Cursory surveys from mid- to late March and observations by Forsythe (1970) indicated that Long-billed Curlews arrive in Utah during the last week of March (Paton et al. 1992; Table 1). No curlews were observed at Howard Slough on 27 and 30 March 1991, while one bird was seen on 31 March. In 1992 no curlews were seen on vis- its to Howard Slough or West Layton on 19 March and Harold Crane on 26 March, whereas three birds were seen on 30 March at Howard Slough and two curlews were at West Layton on 31 March. By mid-April most curlews that nested around Great Salt Lake appeared to have arrived and established territories. However, not all curlews seen during April were local breeding birds, as flocks of 2-20 birds were often seen flying north over the study sites during the second and third weeks of April. For example, a flock of 11 birds was migrating north on 11 April 1991 over Howard Slough. On 14 April, 20 curlews were foraging near Brigham City in a pasture not used by resi- dent curlews. By the end of April all three years, flocks that appeared to be migrating curlews were no longer observed. Long-billed Curlew nests were initiated from mid-April to mid-May in northern Utah, based on floating eggs and observations of juveniles. Analysis of egg-floating data from 1992 showed that four clutches were initiated in late April, three the first week of May, and three the second week of May. In addition, juveniles 3—4 da>'s old were found on 23 May 1990 and 2 June 1990 at Locomotive Springs, and five broods (all >1 week old) were seen on the east side of Antelope Island on 23 May 1992. Based on back-dating, their nests were all started about the third to fourth week of April. Fall migration was relatively early for most curlews at Great Salt Lake compared with other shorebirds (Paton et al. 1992). The num- ber of curlews seen on our two principal study sites declined dramatically after the first week of June. We saw no obvious evidence to suggest that Long-billed Curlews attempted to re-nest after nests were depredated. In fact, most adults remained on territory for onl\' 2-3 weeks after nests were depredated and then vacated the study areas. There was an influx of birds at West Layton and Howard Slough from mid- July to late July (Table 1). These flocks were probably 82 Ghp:at Basin Naturalist [Volume 54 migrants, either from other areas around Great Salt Lake or possibly farther north. These floeks often had one or two adults (both sexes) and 2-4 juveniles, suggesting the possi- bilit)' they were sometimes migrating family groups, although this has not been previously reported. The largest late-summer migratoiy floek we obsen^ed during 3 years of fieldwork was 38 birds on 25 July l'99() at Salt Well Flats, located at the northwestern corner of Promontory Point. It was extremely rare to see any curlews during surveys at our study areas after 15 August (Table 1). Our latest Utah record was one bird on 27 August at Lay ton. Density estimates. — In 1990, surveys were initiated too late to estimate the nimiber of breeding adults at the principal study areas. Survey data from the two principal study areas suggested more curlews were sighted in 1992 than in 1991 (Howard Slough: / = -2.51, df = 19, F < .02; Layton: t = -2.4, df = 19, P < .03; Table 1). Howard Slough had 2 nest- ing pairs of curlew in 1991 (0.64 pairs/km-) and 6 pairs in 1992 (1.92 pairs/km-), while West Layton had 2 nesting pairs in 1991 (0.59 pairs/ km-) and 8 pairs in 1992 (2.36 pairs/km^). At West Warren, we estimated 9 breeding pairs in 1992 (2.25 pairs/km-). Although the data are limited, nearest-neigh- bor nest distances averaged 480.4 m (range = 351-1158 m, a = 6). Surveys at satellite study areas found no evidence to suggest that Long-billed Curlews nested at Harold Crane or north of Saltair during any year of the study. No curlews nest- ed at North Ogden Bay in either 1990 or 1991, and 1-2 pairs nested there in 1992. Locomotive Springs was surveyed most thor- oughly in 1990, when we estimated a mini- mum of 6 pairs nesting in the area. One of the largest nesting concentrations of curlews we obserx^ed at Great Salt Lake was in late May 1992 on the east side of Antelope Island, where at least 8 pairs were seen along 2 km of road approximatcK' 1.5 km northwest of Sea- gull Point. Interestingly, other ground-nesting species (Short-eared Owls [Asia flatnineus] and Northern Harriers [Circus cyaneii.s]} were also relatively common on the east side of Antelope Island, compared to other areas at Great Salt Lake (R Paton personal observation). Clutch size and nest success. — All nests in which we were able to determine final clutch size had four eggs (n = 9). Only 2 of the 10 nests we found in 1992 were success- ful. Seven nests were depredated by mam- malian predators. Red fox {Vulpes viilpes) was the primary nest predator for Snow\' Plovers on the east side of the lake (Paton and Edwards 1990) and probably depredated most curlew nests. In addition, one nest at Layton was possibly depredated by another curlew, based on the diameter of puncture holes found in the egg shells (Redmond and Jenni 1986). We were imable to determine fledging success. Nest site characteristics. — Ten Long- billed Curlew nests were found in 1992. Cinlews appeared to select nest sites in habi- tats with relatively short vegetation, often near barren patches of ground. Vegetation within 15 m of nest sites was significantly shorter than vegetation at random sites (paired / = -10.7, df = 1239, P < .0001; Table 2). Vegetation near nest sites (<6 m) was sig- nificantly taller than that far (> 6) from nests (Table 2). In contrast, there was no significant variation in vegetation height at random sites as a function of distance from plot center (Table 2). Curlews selected nest sites in small clumps of live/dead vegetation, and near the nest there was relatively little barren ground (Table 3). In fact, the amount of barren groimd near nest sites was the only vegetation vari- able that showed any significant variation as a Rmction of distance from plot center (Table 3). Therefore, it appears that Long-billed Cinlews did not select habitat patches based on the Table 2. Mean (±SE) vegetation fieiulit (cin) of Long-Billed Curlew nest and random sites at Great Salt Lake, Utah (n = 10). Means lacking similar letters are significantly different (ANOVA, F < .0.5, Dimcans Multiple Range Test). Distance from pl( :)t center (m) F 0-2.9 .3.0-5.9 6.0-8.9 9.0-11.9 12.()-L5.() 0- 15.0 P Nest sites Random sites H..5 ± ().:3A S.l ±0.5A B.O ± 0.4AB 9.5 ± O.SA 3.3 ± 0,;3B 9.:3 ± O.SA 4.9 ± ().4B 9.S ± ().7A 5.5 ± 0.5B S.(i ± O.TA 5.6 9.0 ± 0.2 ± 0..3 2.9.S 1.11 .OlS .35 1994] Long-billed Curlew in Utah 83 Table 3. Mean (±SE) vegetation coverage of Long-Billed Cnrlew nest and random sites at Great Salt Lake, Utah. Means lacking similar letters are significantly different (ANOVA, P < .05, Duncan s Multiple Range Test). Dis tance from pi ot center (m) F 0-2.9 3.0-5.9 6.0-8.9 9.0-11.9 12.0-15.0 0- -15.0 P Nest sites % live vegetation 56 ±4.6A 45 ±5.6A 43 ± 6.1 A 40 ± 6.3A 46 ±6.6A 46 ±4.2 1.1 .357 % dead vegetation 26 ±4.2A 27 ± 5.9A 19 ± 5.4A 21 ±5.7A 15 ±4.5A 22 ± 4.3 1.0 .418 % barren ground 18 ± 3.9A 28 ± 5.2AB 38 ± 6.3B 39 ± 6.7B 39 ± 6.7B 32 ±4.9 2.6 .037 Random sites % live vegetation 39 ± 6.3A 37 ± 6.0A 32 ± 5.5A 28 ±5.9A .36 ± 6.4A 34 ±4.6 0.5 .735 % dead vegetation 28 ± 6.1 A 28 ± 5.6A 33 ± 6.9A 36 ± 6.9A 30 ± 6.2A 31 ±5.2 0.3 .894 % barren ground 33 ± 6.4A 35 ± 6. 7 A 35 ± 6. 7 A 36 ± 6.4A 34 ± 6.6A 35 ±5.9 0.1 .997 proportions of dead and live vegetation avail- able, but rather vegetation height seemed to be the key variable. There was a weak tenden- cy for curlew nest sites to be located in areas with slightly more live vegetation than in ran- dom transects {t = 1.81, df = 18, P = .07), whereas nests and random sites did not differ in the amount of dead vegetation {t = -1.3, df = 18, P = .19) or barren ground {t = -0.3, df = 18, P = .74; Table 3). The most common plant species, with common defined as averaging >3% total cov- erage, within 15 m of the 10 nests were Sal- icornia europaea {x = 13.2% live, 7.7% dead), Bassia hyssopifoUa (14.7% live, 3.2% dead), Suac'da calceolifonnis (11.5% live, 6.1% dead), Distichlis spicata (4.3% live), and Chenopodi- iim album (0.3% live, 3.2% dead). Discussion Nesting densities in northern Utah found during this study were intermediate relative to estimates for other regions of western North America. Sadler and Maher (1976) reported relatively low densities (0.14-0.17 pairs per km^) at the northern limits of their range in Saskatchewan, which would be expected. Densities similar to those in our study were found in southeastern Washington (0.58-1.45 pairs per km^; Allen 1980) and north central Oregon (up to 3.6 per km-; Pam- push 1980), which would be expected given that both sites were at latitudes similar to those in northern Utah. An area with consis- tently high densities is the shortgrass range- lands of western Idaho (6.4 males and 5.3 females per km-; Redmond et al. 1981). The exact reasons for this variation in population densities across the species' range are unclear. yet should be studied further to assess factors regulating their populations. For example, lit- tle is known about prey and predator densi- ties in various parts of the curlew range. Other aspects of Long-billed Curlew breed- ing ecology at Great Salt Lake were similar to results reported from other parts of their range. Four eggs is the typical clutch size for the species (Pampush 1980, Redmond 1986). Somewhat surprisingly, the migratory chronology of Utah birds was different from that of southeastern Washington (Allen 1980), with birds in northern Utah arriving later and remaining longer. However, although south- eastern Washington is farther north than Utah, it is also lower in elevation (ca. 225 m) and has a milder climate than Great Salt Lake, which probably explains why curlews arrive earlier in Washington. As with the migratoiy chronology, clutch initiation dates vary with climate. Clutches in northern Utah were start- ed from mid-April to mid-May during our study, which was 2 weeks later than in west- ern Idaho (Redmond 1986), southeastern Washington (Allen 1980), and north central Oregon (Pampush 1980). However, in central Wyoming, clutches were initiated 1-2 weeks later than at Great Salt Lake (Cochran and Anderson 1987). Vegetation height seems to be one of the fundamental habitat characteristics used by Long-billed Curlews to select breeding areas. Curlews tend to nest in areas with vegetation <10 cm tall (Allen 1980, Pampush 1980, Bicak et al. 1982, Cochran and Anderson 1987, this study). Structural characteristics of their nesting habitat at Great Salt Lake are relatively similar to those in other regions of western North America, although specific plants were different. As in this study, Pampush 84 Great Basin Naturalist [\bluiiie 54 (1980) touiid that curk-ws in north central Oregon selected nest sites with generally lower vertical profile and lower vertical densi- t\ than the surrounding hahitat. Bicak et al. (1982) found a negative correlation between Long-billed Curlew abundance and vegeta- tion height, with more birds using areas with short vegetation. Since curlews use areas with relatively short vegetation, Bicak et al. (1982) suggested that livestock grazing prior to the onset of the breeding season could increase use of an area by nesting curlews. Redmond (1986) reported that relatively tall vegetation (40 cm tall) affected their foraging activities, and that an increase in plant height in nesting habitat (>12 cm tall) due to the previous year's growth delayed egg laying the subse- quent year. Therefore, all studies in western North America indicate that relativeK' short vegetation is among the key habitat variables that wildlife managers must be concerned with to maintain curlew nesting habitat. Nesting Long-billed Curlews at Great Salt Lake seem to prefer areas that provide good visibility of the surrounding habitat during incubation. This conclusion was similar to habitat studies from other parts of its range (Allen 1980, Cochran and Anderson 1987). At Great Salt Lake the ground is relatively level and curlews prefer to nest near the edges of barren alkali flats. Wolfe (1931) also reported that curlews nested near barren areas at Great Salt Lake. Interestingly, Cochran and Ander- son (1987) reported that Long-billed Curlews avoided fields with extensive barren groimd, although they did not determine if curlews had a threshold value for barren ground. Again, these data suggest that relatively short vegetation is preferred b\' nesting curlews. Finally, more must be learned about the impact of nest predators on curlew popula- tions in western North America. Red fox were first sighted at Great Salt Lake in the late 1960s, with fox numbers dramaticalK increas- ing during the recent Great Salt Lake Hood years (1983-90; Val Bachman, Ogden Bay WMA, personal communication). Currently, red fox are commonly sighted on the eastern shores of Great Salt Lake (personal observa- tion), whereas during 3 years of fieldwork on the eastern shores of the lake, we sighted only one coyote (Caniis hitrans) on one occasion. Interestingly, one area at Great Salt Lake where Long-billed Curlews are still relatively conniion, Antelope Island, also has coyotes. The interaction between coyotes and red fox requires further study. Impacts of nest preda- tors on Long-billed Curlew populations could be dexastating because Long-billed Curlews apparently do not re-nest after their eggs are depredated (Redmond and Jenni 1986). Therefore, additional work may be required of wildlife management to minimize depreda- tion rates and thus maintain curlew popula- tions in certain parts of their range. Acknowledgments We thank T C. Edwards, Jr., Frank Howe, Roland Redmond, Dennis Forsythe, and one anonymous reviewer for comments on earlier drafts of the manuscript. V. Bachman, R. Berger, J. Dolling, S. Manes, D. Paul, and R. Radant helped with access to state lands. Financial support was provided by the Utah Division of Wildlife Resources, Nongame Section (Contract No. 90-2028) Literature Cited Allen, J. N. 1980. The ecology and behavior of the Long- hilled Curlew in southeastern Washington. The Wildlife Society, Wildlife Monographs No. 73. 65 pp. BK:Afv, T. K., R. L. Redmond, and D. A. Jennl 1982. Effects of grazing on Long-billed Curlew {Numenius (unericaniis) breeding behavior and ecology in southwestern Idaho. In: J. Peek and P. D. Dalke, eds.. Wildlife-livestock relationships symposium: proceedings 10. University of Idaho, Forest, Wildlife and Range Experiment Station, Moscow. CociiRXN, J. F., AND S. H. Anderson. 1987. Comparison of habitat attributes of stable and declining Long- billed Curlew populations. Great Basin Naturalist 47: 459-466. FoHSVTiiE, D. M. 1970. Vocalizations of the Long-billed Curlew. Condor 72: 213-224. . 1972. Observations on the nesting biology of the Long-billed Curlew. Great Basin Naturalist 32: 88-90. Greek, D. C. 1981. Atlas of Utah. Brigham Young Uni- versity Press, Provo, Utah. Hays, H., and M. LeCroy. 1971. Field criteria for deter- mining incubation stage in eggs of the Common Tern. Wilson Bulletin 83: 425-429. H.\YS, R. L., C. Simmers, and W. Seliz. 1981. Estimat- ing wildlife habitat variables. U.S. Fish and Wildlife Service FSW/OBS-81/47. Ill pp. Pampush, G. J. 1980. Breeding chronology, habitat uti- lization, and nest-site selection of the Long-billed Curlew in northcentral Oregon. Llnpublished thesis, Oregon State University, Coi-vallis. 49 pp. P.vroN, P. W. C, AND T. C. Edwards, Jr. 1990. Status and nesting ecology of the Snov\y Plover at Great Salt Lake— 1990. Utah Birds 6: 49-66. 1994] Long-billed Curlew in Utah 85 Paton, p. W. C, C. Kneedy, and E. Sorensen. 1992. Sadler, D. A. R., and W, J. Maher. 1976. Notes on the Chronology of shorebird and ibi.s use of selected Long-billed Curlew in Saskatchewan. Auk 93: marshes at Great Salt Lake. Utah Birds 8: 1-19. 382-.384. Redmond, R. L. 1986. Egg size and laying date of Long- SAS Institute. 1988. SAS/STAT user's guide. SAS Insti- billed Curlews, Nwnenius americamis: implications tute. Inc., Gary, North Carolina, for female reproductive tactics. Oikos 46: 330-338. WoLFE, L. R. 1931. The breeding Limicolae of Utah. Red.mond, R. L., T. K. Bicak, and D. A. Jenni. 1981. An Condor 33: 49-59. evaluation ot breeding season census techniques for Long-billed Curlews (Nwnenius (nnericanus). Stud- ies in Avian Biology 6: 197-201. Received 3 Fehnianj 1993 Redmond, R. L., and D. A. Jenni. 1986. Population ecol- Accepted 20 May 1993 ogy of the Long-billed Curlew, Nionenius ameri- camis, in western Idaho. Auk 103: 755-767. Great Basin Naturalist 54(1), © 1994, pp. S(i-9() HABITAT USE AND BEHAVIOR OF MALE MOUNTAIN SHEEP IN FORAGING ASSOCIATIONS WITH WILD HORSES Kevin F C.'oatesl- and Sanford I). Schemnitz^ Key words: mountain slwep, Ovis c. canack-nsis, uiUl horses. Juihitat use, licharior. Bifjunn Cauyott National Recre- ation Area. Montana, Wijoininfi. R()ck\ Mountain bighorn sheep {Ovi.s canadensis canadensis) maximize survival by foraging in secure habitats that afford high visibiht>' and have good interspersion of pre- ferred forage plants with escape cover (Risen- hoover and Bailey 1985). Good visibilit\ and precipitous escape cover are structiual habitat elements that provide mountain sheep with security from predators (Buechner 1960, Geist 1971, Wishart 197S, Risenhoover and Bailey 1985). Wild horses {Eqmis cabaUus) also maximize survival by foraging in secure habitats with good interspersion of preferred forage plants. However, different structural elements of the habitat provide security for mountain sheep and horses; mountain sheep select foraging areas near precipitous escape terrain while horses select foraging areas near open, flat ter- rain. This is due to basic differences in preda- tor escape tactics for the species: mountain sheep climb to avoid predation and horses iim. Although grasses dominate the diets of both horses and mountain sheep, each species' predator-avoidance strategy selects for structurally different habitats. However, when spatial distributions overlap, a competi- tive situation may occur, with mountain sheep being negatively impacted. In several instances such competition with feral equids has resulted in mountain sheep declines (McMichael 1964, Weaver 1973, Seegmiller and Ohmart 1981). A growing body of literature supports the hypothesis that horses and other exotics may, in some respects, facilitate the foraging effec- tiveness of some native ungidate species either bv habitat modification or increased protection from predators (Berger 1978, 1986, Festa-Bianchet 1991). The purpose of this note is to present unique observations which suggest that male mountain sheep may benefit from close foraging relationships with wild horses. Few data exist on resource competi- tion between mountain sheep and feral horses (Berger 1986), and though not statistically quantifiable, these limited observations sup- port Berger's (1986) hypotheses regarding for- age facilitation of native species by exotics. Study Area and Methods The study was conducted at Bighorn Canyon National Recreation Area (BICA), a 48,679-ha National Park Service unit that has as its focal point a 114-km-long reservoir in southeastern Montana and north central Wyoming. Moimtain sheep recolonized BICA in 1975 because of dispersal of 4-6 animals from a nearby transplant. By 1986 the popula- tion had increased to over 60 animals (Coates and Schemnitz 1986). Portions of BICA are federally designated as the Prvor Mountain Wild Horse Range (PMWHR). The 17,402-ha PMWHR supports approximately 120 wild horses and is located 80 km south of Billings, Montana (Bureau of Land \hmagement 1984). The area is characterized as a desert-shrub woodland (Lichvar et al. 1985), and dominants include a sparse overstory of curlleaf moun- tain mahogany {Cercocarpus ledifolius var. intercedens), Utah juniper ijiiniperus osteo- sperma), sagebrush {Artemisia spp.), and greasewood {Sarcobatus spp.), with a poorly developed understory of bunchgrasses (Lich- var et al. 1985). Annual precipitation averages 'DcpartilU'lit III I'-islici-N anil Wikllilc Sciences. N, ^Present address; Box 271. Tios. Montana .5993.5. Mc'xieo Slate l'niversit\, Las Graces, New Mexico 8800.3. 86 1994] Notes 87 15-20 cm. Soils present include limestone and sandstone in the precipitous canyonland and dolomite in the nonprecipitous areas (Knight et al. 1987). Elevations vary from a mean pool level of 1109 m at the reservoir to 2682 m at East Piyor Mountain. Gray limestone cliffs rise >250 m vertical- ly from the lakeshore. Cliff faces, ledges, and eroded limestone soils (karst topography) pro- vide abundant escape terrain for mountain sheep. Escape terrain predominates the entire study area, from East Pryor Mountain to the reservoir. Other than an alluvial fan located at the northern extreme of the study area, virtually all habitat is within 300 m of cliffs, ledges, or karst topography (Coates 1988). Three adult ewes (>18 months old) and a 6-year-old ram were captured and equipped with radio collars manufactured by Telonics (Mesa, Arizona). Systematic radio relocation of these animals provided the opportimit\' to locate and observe 328 groups of mountain sheep between June 1986 and November 1987. Group size and age/sex composition were recorded for each observation. Additionally, three habitat parameters were analyzed: horse use (Yes/No), distance to precipitous terrain, and vegetation type. A preference ratio (per- cent use/percent availability) was used to ana- lyze preference and/or avoidance of vegeta- tion types (Risenhoover and Bailey 1985). Because escape terrain was nearly continuous throughout the southern portion of the study area (distance rarely >300 m), all habitat tvpes were considered available to nioimtain sheep. The alluvial fan was considered available to mountain sheep, primarily to investigate dif- ferences in habitat selection between male and female cohorts of mountain sheep, and to analyze the influence of distance to escape ter- rain on foraging behavior The foraging behavior of adult mountain sheep was analyzed to determine the effects of habitat security on foraging efficiency (Risenhoover and Bailey 1985). Once a group of moimtain sheep was located, a focal animal was selected for analysis of foraging behavior. Recognition of focal animals was aided by identifying marks on pelage or scars. Foraging behavior was obsei'ved for five consecutive 3- min periods to detemiine the amoimt of time the focal animal devoted to three behavioral categories: foraging, social, alert (Risenhoover and Bailey 1985). An animal was engaged in foraging when it actively ingested forage and when it moved about with animals that were actively ingesting forage. An animal was engaged in social behavior for all intraspecific and interspecific interac- tions. Social interactions included looking at another animal, moving toward/away from another animal, and mother/young interac- tions. Alert behavior was recorded if the focal animal stopped foraging to look up in the typi- cal alert posture for mountain sheep (i.e., ears up and neck outstretched; Geist 1971), if it looked at a disturbance (e.g., a vehicle on the highway, or a person approaching on foot), or when it ran to avoid a disturbance (e.g., a per- son approaching on foot). Foraging efficiency was calculated as percentage of time devoted to foraging behavior during the 15-min peri- od. Percentage of time spent in alert or social interactions provided a measure of the rela- tive security of moimtain sheep in different habitats. Results and Discussion Four vegetation types (Knight et al. 1987) occur within the obseived range of mountain sheep: Utah juniper/mountain mahogany woodland (JU/CE), Utah juniper woodland (JUOS), mountain mahogany woodland (CELE), and Douglas fir woodland (PSME). Distribution of JUOS was limited to an allu- vial fan at the north end of the study area and narrow fingers interspersed within the JU/CE habitat type. Horse use was always "No" for karst topography and "Yes" for the alluvial fan at the northern extreme of the study area, based on the presence/absence of horse feces observed during fieldwork. Horse use was also obsei'ved along fingers of nonprecipitous habitat interspersed throughout the JU/CE type. Distribution of PSME was restricted to a deeply incised drainage present in the core use area occupied by rams. Overall, 85.7% of male mountain sheep observations involving mixed age/sex groups occurred in JU/CE woodland. JUOS, CELE, and PSME woodlands were used in 13.6, <1, and <1% of the observations, respectively (Table 1). The preference ratio for JU/CE is 4.5, indicating that mountain sheep foraging with conspecifics prefer this type (Risenhoover and Great Basin Naturalist [Volume 54 Table 1. Percent lial)itat utilization 1)\ male mountain sheep in foraging associations with conspecifics com- pared with associations with wild horses. Habitat prefer- ence ratios are expressed as + or - and are given in parentheses below each appropriate categor\'. Habitat T\pe JLOS CELK JU/CE PSME Male mountain sheep: With conspecifics 85. 'J Habitat preference ( + ) W itli w ild horsi's Habitat preference 16.7 13.6 83.3 ( + ) <1 <1 Bailey 1985). Preference for JU/CE habitat probably resulted more from the intersper- sion of escape terrain than from differences in visibility between habitats. Juniper was sparsely distributed throughout both JU/CE and JUOS types. Distinction between types was based on occurrence of curlleaf mountain mahogany rather than on increasing frequency of Utah juniper (Lichvar et al. 1985). Visibility obstruction was low in both JU/CE and JUOS habitats. Ewes never occupied the PSME type, even though it was located on rocky slopes, because visual obstruction was much higher than in JU/CE or JUOS. Male mountain sheep were observed for- aging with wild horses 22 times on 20 differ- ent days, and habitat parameters were record- ed for 12 observations. Foraging associations usually involved 2 specific male horse/harem groups with bachelor ram groups. Ram group size ranged from 3 to 7 animals, 3 to 10 years of age. Female mountain sheep were never ob- served in association with wild horses. Horse group size was dynamic, but association usu- ally involved 1 of 2 specific male horses accompanied by 5 to 8 mares and subadults. Of the 12 observations, 83.3% (n = 10) occurred in the JUOS vegetative type, and 16.7% (n = 2) occurred in JU/CE (Table 1). The preference ratio for JUOS is 1.2. The preference ratio for JUOS by male mountain sheep foraging with wild horses is noteworthy because habitat utilization patterns for JU/CE and JUOS were reversed when male moun- tain sheep associated witli wild horses (Table 1). These limited observations suggest that male mountain sheep foraging with con- specifics may prefer the JU/CE vegetation type, but male moimtain sheep foraging with wild horses may prefer JUOS. Conversely, male mountain sheep foraging with con- specifics avoided JUOS, but male mountain sheep foraging with wild horses avoided JU/CE. Grasses accounted for < 1% of the vegeta- tive cover in the JU/CE type but approxi- mately 6% of the JUOS type (Knight et al. 1987). Although grasses were present in low composition in both vegetation types, moun- tain sheep foraging in JUOS had a higher availability of grasses. Average distance to escape terrain was determined for male mountain sheep that for- aged with conspecifics and compared to the distance for male mountain sheep that foraged with wild horses (Table 2). Male mountain sheep foraging with conspecifics remained within an average of 47 m (SD 69.5 m) from escape terrain, partially because of the ewes' reluctance to venture farther than 50 m from secure habitat. However, male mountain sheep foraging with wild horses were an aver- age of 217 m (SD 310 m) from escape terrain. These limited data suggest that male moun- tain sheep foraged farther from escape terrain (in less secure habitat) when associated with wild horses than with conspecifics. Foraging efficiency of mountain sheep with wild horses was 100% for all 12 locations (no alert or social interactions). Male mountain sheep that foraged with wild horses ignored disturbance (e.g., they could be approached readily, and they rarely looked up to scan their surroundings even when horses were fighting in their vicinity). Group size ranged from 9 to 16 animals, including rams and horses. Foraging efficiency of male moimtain sheep with conspecifics was only 66% [n = 67) and was characterized by high levels of aggressive or social interaction (Table 3). Aggressive interactions were exhibited between rams when two or more followed a ewe, and when they established dominance rank in the male cohort. Social interactions between rams occurred when they attended ewes. Aggressive or social interactions were never observed when male mountain sheep foraged with wild horses. This may have been due to size-related dominance in mountain sheep (Geist 1971) and subordinate behavior of male mountain sheep in the presence of the relatively large wild horse (Berger 1986). 1994] Notes 89 Table 2. Average distance to escape terrain (m) of male nionntain sheep in association with conspecifics compared to distance when associated with wild horses. Table 3. Average foraging efficiency of male mountain sheep in foraging associations with conspecifics com- pared with associations with wild horses. stanaara aeviations sno\ vn m parentneses. Percentage of time de three activities while voted to Distance to escape terrain (ni) foraging Male mountain sheep; 47 (SD 69) 217 (SD 310) Foraging Social Alert With conspecifics With wild horses Male mountain sheep: With conspecifics With wild horses 66 100 32 0 1 0 The subordinant/dominant relationship between male mountain sheep and wild horses was suggested both by the sheep's lack of aggressive behaviors while foraging and by the behavior of male wild horses directed toward male mountain sheep. Male wild horses were observed herding, or driving, male mountain sheep [n = 3) in a manner similar to the typi- cal posture used when herding females (Feist 1975, Berger 1986). This typical herding pos- ture consisted of running toward a female horse, or in this case a male mountain sheep, with ears flattened against the head, neck out- stretched, and head held low to the ground. Another indication of the subordinant/ dominant relationship between mountain sheep and wild horses was extended penis behaviors that Feist (1975) described as a mechanism to establish dominance in wild horse groups. These extended penis behaviors were directed by a subordinant male wild horse (without harem) to a 9-year-old male mountain sheep with three other rams ages 3 to 7. There were no other horses in the vicini- ty. By exerting dominance over male moun- tain sheep or allowing rams to enter their harem, stallions can potentially elevate their own dominance rank and subsequent repro- ductive success by attracting additional females. In summary, we believe that, contrary to some literature (Mc Michael 1964, Weaver 1973, Seegmiller and Ohmart 1981), male mountain sheep and wild horses can have beneficial relationships. Habitat selection by mountain sheep is a complex function of sea- son, age, reproductive status, and sex of the animal (Smith 1992). This paper presents analyses suggesting that habitat selection and foraging efficiency may also be influenced by association with another species during forag- ing periods. These data support Berger's (1986) hypothesis that feral horses may per- haps serve either as competitor or as facilita- tor, depending on ecological conditions. In this case they served as competitor for a patchy supply of grasses, but possibly also as facilitator by increasing foraging efficiency in insecure habitat. Dominance rank of male horses may have increased as a result of the re- lationship. Sample sizes were small, but these unique observations suggest that male moun- tain sheep in association with wild horses for- aged farther from escape terrain, enabling them to use areas that supported higher com- position of grasses than areas used with con- specifics. Also, male mountain sheep did not exhibit aggressive behaviors while in associa- tion with wild horses and thus had higher for- aging efficiency than those with conspecifics. To the best of our knowledge, foraging associ- ations of this type have not been previously reported. Acknowledgments We thank the National Park Service (NFS), Bureau of Land Management, Agricultural E.xperiment Station, and New Mexico State University for financial support. Equipment, advice, and expertise were provided by the Montana Department of Fish, Wildlife and Parks (MDFWP) and Wyoming Game and Fish Department. We especially thank J. T Peters, NFS, and C. Eustace, xMDFWP, for their expert support and guidance throughout the project. We also thank T. S. Smith for edi- torial review. This is Journal Article 1610 of the New Mexico Agricultural Experiment Station. Literature Cited Berger, J. 1978. Group size, foraging, and antipredator ploys: an analysis of bighorn sheep decisions. Behavioral Ecology and Sociobiology 4: 91-99. 90 Great Basin Natuhaijsi f\'()lunie 54 . 1986. Wild horses of the Great Basin. Uii\t'rsit\ of Chicago Press, C>hieago and London. 326 pp. BUECll.NliH, II. K. 1960. The bighorn sheep in the United States: its past, jiresent and fnture. Wildhfe Mono- graphs 4. 174 pp. BURE.\U OF Land M.\n.\c:emem. 1984. Herd manage- ment area plan: Pr>'or Mountain Wild Horse Range. Publication BLM-MT-P-019-4321. 63 pp. CoATES, K. P. 1988. Habitat utilization, interspecific interactions and status of a recolonized population of bighorn sheep at a wild horse range. Unpublished master's thesis, New Mexico State University, Las Cruces. 59 pp. Co.vlES, K. P., AND S. D. SciiEMMTZ. 1986. Habitat uti- lization, interspecific interactions and status ot a recolonized population of bighorn sheep at a wild horse range. University of Wyoming-National Park Ser\ice Research Center, Research Proposal BICA- N-019. 24 pp. Feist, J. B. 1975. Behavior of feral horses at the Piyor Mountain Wild Horse Range. Unpublished master's thesis. University of Michigan, Ann Arbor. 130 pp. Fest.\-Bi.\NCHET, M. 1991. The social system of bighorn sheep: grouping patterns, kinship and female domi- nance rank. Animal Behavior 42: 71-82. Geist, V. 1971. Mountain sheep: a study in behavior and evolution. University of Chicago Press, Chicago and London. 371 pp. Knicmt, D. L., G. p. Jones, Y. Akashi, .\nd R. W. Myers. 1987. Vegetation map of Bighorn Canyon National Recreation Area. University of Wyoming, National Park Ser\'ice Research Center, Laramie. 114 pp. Lk;ii\ah, R. W., E. I. Collins, and D. L. Knioht. 1985. Checklist of xascular plants for the Bighorn Canyon National Recreation Area, Fort Smith, Montana. University of Wyoming, National Park Service Research Center, Laramie. 51 pp. M(;Ml(:iL\EL, T. J. 1964. Relationships between desert bighorn and feral burros in the Black Mountains of Mohave County. Desert Bighorn Council Transac- tions 8: 29-35. RiSENHOOVER, K. L., AND J. A. Bailev. 1985. Foraging ecology of mountain sheep: implications for habitat management. Journal of Wildlife Management 49: 707-804. Seecaiiller, R. F., and R. D. Oiinlkrt. 1981. Ecological relationships of feral burros and desert bigliorn sheep. Wildlife Monographs 78. 58 pp. Smith, T. S. 1992. The bighorn sheep of Bear Mountain: ecological in\'estigations and management recomen- dations. Unpublished doctoral dissertation, Brigham Young University, Provo, Utah. 425 pp. Wea\'ER, R. a. 1973. Burro versus bighorn. Desert Bighorn Council Transactions 17: 90-97. Wtsii ART, W. 1978. Bighorn sheep. Pages 161-172 in J. L. Schmidt and D. L. Gilbert, eds.. The big game of North America. Stackpole Books, Hanisburg, Penn- svlvania. Received 20 April 1992 Accepted 2 SeptcinJ)er 1993 Great Basin Naturalist 54(1), © 1994, pp. 91-95 FISH MORTALITi' RESULTING FROM DELAYED EFFECTS OF FIRE IN THE GREATER YELLOWSTONE ECOSYSTEM Michael A. Bozek^- and Michael K. Youngl Key uords: fish kill, fire. sii.si)etule(l sediment. Greater Yelloivstone Ecosystem, storm, debris torrent. Often public concern focuses on the imme- diate, terrestrial impacts of wildfire. Such was the case during the summer and tall ot 1988 when fires burned 562,000 ha in the Greater Yellowstone Ecosystem (GYE; Christensen et al. 1989, Schullery 1989). Besides the obvious loss of vegetation, less apparent, short-term consequences of these fires to terrestrial ecosystems included greater nutrient avail- ability, widespread soil modification, and direct and indirect mortality of wildlife (Christensen et al. 1989, Singer et al. 1989). But as a result of the linkages between streams and their valleys (Hynes 1975), fires also may affect the hydrology, water chem- istry, and geomorphology of aquatic ecosys- tems (Tiedemann et al. 1979, Schindler et al. 1980, Minshall et al. 1989). One consequence of fire is that bed- and suspended-sediment loads are often abnormally high in streams after storm events (Minshall and Brock 1991). Depending on the concentration and duration of exposure, suspended sediment can induce physiological stress, reduce growth, and cause direct mortality' in fish (Newcombe and Mac- Donald 1991). However, as terrestrial vegeta- tion recovers and soils stabilize, concentra- tions of suspended sediment in streams are expected to decline (Minshall et al. 1989). Unfortunately, little else is known about rela- tions between watershed recovery and aquatic ecosystems following fire. During the 1988 fires in the GYE, Minshall et al. (1989) observed fish kills in streams, but the extent and causes of mortality were not reported. While conducting other studies of watersheds in the GYE, we observed a fish kill in a burned watershed that occurred two years after the fires. In this paper we describe aspects of this fish kill and relate them to hydrologic conditions in this stream and those in a nearby stream with an unburned water- shed. Study Area We studied two tributaries of the North Fork Shoshone River in the North Absaroka Wilderness Area adjacent to the eastern bor- der of Yellowstone National Park in the Absaroka Mountains of northwestern Wyoming. Jones Creek drains a 6423-ha watershed that was almost completely biunied in 1988. Because of steep topography, drainage comprises numerous high-gradient tributaries and steep ephemeral chutes. The 4946-ha Crow Creek watershed is the next drainage south of Jones Creek watershed, has similar topographic relief and watershed ori- entation, and still supports extensive mixed- age stands of conifers. Both watersheds con- sist of geologically young and highK' erodible volcanic soils (Minshall and Brock 1991). Methods On 17 and 18 August 1990 we surveyed 1774 m of the stream channel and lower and upper banks of Jones Creek for dead fish fol- lowing storm flows that had been caused by rain that began at 1600 hoins on 16 August. Fish were identified, measured, and examined to determine possible causes of death. We examined the external anatomy of the fish, including skin, eyes, and gills, as well as stom- achs of several fish. Suspended sediment and discharge data from April to September 1990 were obtained 'Rock>' Mountain Forest and Range Experiment Station, 222 S. 22nd Street, Laramie, Wyoming 82070 USA. ^Present address; Centre for Northern Forest Ecosystem Research, Lakehead University Campus, 955 Oh\er Road, Thunder Bay, Ontario P7B 5E1 Canada. 91 92 Great Basin Naturalist [Volume 54 from continuous remote sampling stations operated by the U.S. Geological Survey (USGS 1990). Suspended sediment concen- trations are daily means l)ased on 2-4 samples per day; mean dail>' discharge was based on hourK ohserxations. On 17 August we collect- ed two grab samples from Jones Creek to quantitativeK assess the unusually high con- centration of suspended sediment observed during the storm. Samples were collected at mid-depth in 0.5 m of water We analyzed the grab samples for total suspended sediment by filtering them through Whatman grade 934AH fiberglass filters (1.5 ^tm effective pore size), oven-drying them for 1 week to constant weight, and then measuring them and averag- ing the results to determine the total concen- tration of suspended sediment (APHA 1989). Results and Discussion On 17 and 18 August 1990, 1 rainbow trout {Oncorhijnchus )nykiss), 4 Yellowstone cut- throat trout (O. clarki boiivieri), 11 brook trout {Salvelinus fontinalis), and 2 Yellowstone cut- throat trout X rainbow trout hybrids, all rang- ing from 190 to 410 mm total length, were found dead during surveys of Jones Creek. We found fish only in or near obstructions to flow (e.g., debris accumulations and boul- ders); thus, our survey probably overlooked dead fish that had been transported down- stream or buried in newly formed bars. We believe that fish collected on 17 August had died recently because rigor mortis had not set in. Fish collected the following day were rigid and had started to diy; we suspect these also had succumbed on 17 August. Surveys on seven other occasions on Jones Creek (includ- ing 16 August) and eight other occasions on Crow Creek failed to reveal anv moribund fish. Each fish we examined appeared to have been asphyxiated by sediment. Typically, sedi- ment completely embedded the gills of each fish, and individual lamellae often were diffi- cult to see (cf Cordone and Kelley 1961: 192). Eyes and skin appeared to be relatively nor- mal, and fish lacked contusions and lacera- tions. Stomachs we examined appeared nor- mal, and all contained recently consumed invertebrates. Timing of the fish kill coincided with storms that began on 16 August and contin- ued into the early morning of 17 August. Nearly 2.3 cm of rain was recorded on 16 August, followed by an additional 0.7 cm the next day. Though discharge in Jones Creek peaked on 16 August, concentration of sus- pended sediment did not appear to peak until 17 August (Fig. 1). And though suspended sediment concentrations on 17 August were the highest recorded bom April to September 1990 (uses 1990), concentration of suspend- ed sediment in grab samples (9680 mg/L) was more than an order of magnitude higher than the daily mean concentration (587 mg/L) recorded for that date. Because the continu- ous remote sampling station collects samples at intervals, it is likely that the automated sampling missed the instantaneous peak con- centration of suspended sediment. Likewise, it also is possible that our grab samples did not represent the instantaneous peak concen- tration. Suspended sediment is known to be lethal to salmonids, but usually at higher concentra- tions and/or for longer exposures (Redding et al. 1987, Newcombe and MacDonald 1991) than we obseiA'ed in Jones Creek. For exam- ple, Newcomb and Flagg (1983) calculated that a 36-h exposure to a suspended sediment concentration of 9400 mg/L would kill 50% of juvenile chinook salmon (O. tshawytscha) and sockeye salmon (O. uerka). However, lethal effects of suspended sediment ma\' be more pronounced in the field than in the laboratory'. In live-box tests in streams affected by ashfall from Mount St. Helens, concentrations of sus- pended sediment as low as 488 mg/L killed 50% of chinook salmon smolts after a 96-h exposure (Stober et al. 1981). But in compara- ble laboratory tests, a concentration of 19,364 mg/L was required to produce the same mor- tality rate (Stober et al. 1981). Clearly, sus- pended sediment concentrations in Jones Creek were stressful for trout. For the 24 h beginning at 1600 on 16 August, we estimated a minimum stress index of 11.3 mg'h«L~^ (Newcombe and MacDonald 1991), which is near the value associated with lethality in adult salmonids (12 mg-h-L-^; C. R New- combe personal commimication). Other factors may have contributed to the fish kill in Jones Creek. Newcombe and Mac- Donald (1991) and C. R Newcombe (personal comnumication) suggested that high or fluctu- ating temperatures may increase the sensitivity 1994] Notes 93 0) D 600 500 - 400 - 300 - 200 - 100 600 - 500 - 400 - 300 200 - 100 c o c CD U c o u "c E ID 0) 00 ■D (D X5 C (D Q_ m CO April 1 May 1 June 1 July 1 August 1 September 1 Date Fig. 1. Discharge and suspended sediment concentrations from April through September 1989 in Jones Creek (above) and Crow Creek (below), Wyoming (USGS 1990). The peak in suspended sediment in Jones Creek on 17 August coincides with the fish kiU. 94 Great Basin Naturalist [Volume 54 of trout to suspended sediment. In Jones Creek, water temperature varied from 10.1 to 17.3°C on 17 August (USGS 1990), but this fluctuation was equaled or e.xceeded on 70 of the 137 monitored days. Innthermore, these temperatures are largely within the range of those reported from other tests (e.g.. New- comb and Flagg 1983, 15-17°C; Redding et al. 1987, 12.5-13.5 °C). A reduction in dis- solved oxygen concurrent with peak suspend- ed sediment concentrations, or other changes in water chemistn; may also have contributed to mortality, but we did not measure these parameters. Tiedemann et al. (1979) indicated that landslide activity in steep drainages increases after wildfires. In Jones Creek a debris torrent down a tributary, apparently caused by heavy rainfall on unstable burned slopes, may have produced the high concentrations of suspend- ed sediment on 17 August. After surveying farther upstream on subsequent days, we found a fresh debris and mud jam in Jones Creek near the mouth of a severely eroded tributary. Possibly because the stream was downcutting through this material, concentra- tions of suspended sediment remained high for several days (Fig. 1). Once activated by rainfall, numerous other ephemeral channels also carried silt-laden water for several days, but in concentrations visibly less than the peak concentration observed in Jones Creek. The effects of fires on streams include increases in discharge and suspended sedi- ment (Tiedemann et al. 1979, Schindler et al. 1980), and these differences seem evident in the comparison between the burned Jones Creek and the unburned Crow Creek water- sheds. Discharge and suspended solid con- centrations in Jones Creek were relatively high and erratic throughout the 137-day sam- pling period (Fig. 1). During this time the daily mean concentration of suspended sedi- ment averaged 73.9 mg/L (USGS 1990). Though concentrations of total suspended sediment often increased with stream dis- charge (e.g., during spring runoff), pro- nounced episodic peaks in the concentration of suspended sediment also occurred during lower discharges, apparently associated with summer rainfall (e.g., as seen 16-21 August). In contrast, discharge and suspended solid concentrations in the unburned Crow Creek watershed were markedly lower and more sta- ble (Fig. 1). As in Jones Creek, suspended sediment increased during snowmelt and storm events, but changes appeared more proportional to increases in discharge. Daily mean concentrations of suspended sediment averaged 8.2 mg/L; the maximum daily mean concentration recorded (59 mg/L) was less than the Jones Creek average for the entire sampling period. Unfortunately, though the contrast between responses of both water- sheds was quite marked, the lack of pre-fire hydrologic data makes it difficult to conclude that this contrast was the result of fire. How- ever, during observations made on both streams for 4 weeks over 2 years, we did not witness similar concentrations of suspended sediment or a fish kill in either stream. At least circimistantially, the fish kill appears to be related to the unusualK' large hydrologic event associated with a rainstorm in the Jones Creek watershed. The fish kill that we obsei^ved was notable because it occmred 2 years after the fire and appeared to result from an acute exposure to sediment. The extent and frequency of lethally acute concentrations of suspended sediment, as well as their effect on entire fish populations, are unknown. Previous fires in the Yellowstone area in the 1700s were at least as intense as those that occurred in 1988 (Romme and Despain 1989) and may also have produced slope instability, high suspended sediment concentrations, and, consequently, fish kills. In both Jones and Crow creeks we found rem- nants of the toes of landslides and the termini of debris flows that may have resulted from past fires. Because fire is a natural distur- bance that will recur, further investigations are needed to gain a better understanding of the effects of fire and how watersheds, streams, and fish populations respond immediately and during the successional recovery of adjacent terrestrial vegetation. Acknowledgments We thank K. Benson, J. Hansen, and J. Blake for assistance in the field; G. Laidlaw for access to the U.S. Geological Survey data; and A. J. Cordone, G. W. Minshall, C. P New- combe, and R. Zubik for reviewing the manu- script. 1994] Notes 95 Literature Cited APHA (American Public Health Association). 1989. Standard method.s for the examination of water and wastewater. 17th ed. American Puhhc Health Asso- ciation, Washington, D.C. ClIRlSTENSEN, N. L., ET .-^L. 1989. Interpreting the fires of Yellowstone. BioScience 39: 678-685. CoRDONE, A. J., AND D. W. Kelley. 1961. The influences of inorganic sediment on the atjiiatic life of streams. California Fish and Game 47: 189-228. Hynes, II. B. N. 1975. The stream and its valley. Ver- handlungen Internationale Vereinigung fiir Theo- retische und Angewandte Limnologie 19: 1-15. MiNSllALL, G. \V., AND J. T. Brock. 1991. Ohsened and anticipated effects of forest fire on Yellowstone stream ecosystems. Pages 123-135 in R. B. Keiter and M. S. Boyce, eds., The Greater Yellowstone Ecosystem: redefining America's wilderness her- itage. Yale University Press, New Haven, Connecti- cut. MiNsiiALL, G. W., J. T. Brock, and J. D. Varley. 1989. Wildfires and Yellowstone's stream ecosystems. Bio- Science 39: 707-715. Newcomb, T. W., and T. A. Flagg. 1983. Some effects of Mt. St. Helens volcanic ash on juvenile salmon smolts. U.S. National Marine Fisheries Service Review 45: 8-12. Ne\vc;()MBE, C. p., and D. D. M.acDonald. 1991. Effects of suspended sediments on aquatic ecosystems. North American Journal of Fisheries Management 11:72-82. Redding, J. M., C. B. Schreck, and F. H. Everest. 1987. Ph\siologica! effects on coho salmon and steelhead of exposure to suspended solids. Transac- tions of the American Fisheries Societ\' 116: 737-744. RoMME, W. H., AND D. G. Despain. 1989. Historical per- specti\es on the Yellowstone fires of 1988. Bio- Science 39: 695-699. SciiiNDLER, D. W., R. W. Newbury, K. G. Beaty, J. Prokopowich, T. Ruscznski, and J. A. Dalton. 1980. Effects of a windstorm and forest fire on chemical losses from forested watersheds and on the quality of receiving streams. Canadian Journal of Fisheries and A(]uatic Sciences 37; 328-334. ScilULLERY, p. 1989. The fires and fire polic> . BioScience 39: 686-694. Singer, F. J., W. Sciireier, J. Oppenheim, and E. O. Carton. 1989. Drought, fires, and large mammals. BioScience 39: 716-722. Stober, Q. J., B. D. Ross, C. L. Melby, P. A. Dinnel, T. H. Jagielo, and E. O. Salo. 1981. Effects of sus- pended volcanic sediment on coho and chinook salmon in the Toutle and Cowlitz rivers. Fisheries Research Institute, University of Washington, Seat- tle. Technical Completion Report FRI-UW-S124. Tiedem.\nn, a. R., C. E. Conr.\d, J. H. Dietericii, J. W. HoRNBECK, W. F. Megaiian, L. a. Viereck, a.nd D. D. Wade. 1979. Effects of fire on water. USDA Forest Service General Technical Report WO- 10. Washington, D.C. uses (United St.vies Geological Survey). 1990. Water resoiuces data — Wyoming — water year 1990. U.S. Geological Survey, Washington, D.C. Water Data Report WY-90-1. Received 28Janiiani 1993 Accepted 25 August 1993 INFORMATION FOR AUTHORS The Great Basin Naiiiralist welcomes previousK unpublished manuscripts pertaining to the biologi- cal natural history of western North America. Pref- erence will be given to concise manuscripts of up to 12,000 words. SUBMIT MANUSCRIPTS to Richard W. Baumann, Editor, Great Basin Naturalist, 290 MLBM, PC) Bo.x 20200, Brigham Young University, Provo, UT 84602-0200. A cover letter accompanying the man- uscript must include phone number(s) of the author submitting the manuscript, and FAX number and E-mail address when applicable; the letter must also provide information describing the extent to which data, te.xt, or illustrations have been used in other papers or books that are published, in press, submitted, or soon to be submitted elsewhere. Authors should adhere to the following guidelines; manuscripts not so prepared may be returned for revision. MANUSCRIPT PREPARATION. Consult Vol. 51, No. 2 of this journal for specific instructions on style and format. These instructions, GUIDELINES FOR Manuscripts Submitted to the Great Basin Naturalist, supply greater detail than those pre- sented here in condensed form. The guidelines are printed at the back of the issue; additional copies are available from the editor upon request. Also, check the most recent issue of the Great Basin Nat- uralist for changes and refer to the CBE Style Man- ual, 5th edition (Council of Biology Editors, One Illinois Center, Suite 200, 111 East Wacker Drive, Chicago, IL 60601-4298 USA,' phone: (312) 616- 0800, F.w: (312) 616-0226; $24). TYPE AND DOUBLE SPACE all materials, including literature cited, table headings, and figure legends. Avoid hyphenated words at the right-hand margins. Underline words to be printed in italics. Use stan- dard bond (22x28 cm), leaving 2.5-cm margins on all sides. SUBMIT 4 COPIES of the manuscript and the origi- nal on a 5.25- or 3.5-inch disk utilizing WordPerfect 4.2 or above. Number all pages and assemble each copy separately: title page, abstract and key words, text, acknowledgments, literature cited, appen- dices, tables, figure legends, figures. TITLE PAGE includes an informative title no longer than 15 words, names and addresses of authors, a running head of fewer than 40 letters and spaces, footnotes to indicate change of address and author to whom correspondence should be addressed if other than the first author. ABSTRACT states the purpose, methods, results, and conclusions of the research. It is followed by 6-12 key words, listed in order of decreasing importance, to be used for indexing. TEXT has centered main headings printed in all capital letters; second-level headings are centered in upper- and lowercase letters; third-level head- ings begin paragraphs. VOUCHER SPECIMENS. Authors are encouraged to designate, properly prepare, label, and deposit high-quality voucher specimens and cultmes docu- menting their research in an established permanent collection, and to cite the repositoiy in publication. REFERENCES IN THE TEXT are cited by author and date: e.g., Martin (1989) or (Martin 1989). Multiple citations should be separated by commas and listed in chronological order. Use "et al. " after name of first author for citations having more than two authors. ACKNOWLEDGMENTS, under a centered main heading, include special publication numbers when appropriate. LITERATURE CITED, also under a centered main heading, lists references alphabetically in the fol- lowing formats: Mack, G. D., and L. D. Flake. 1980. Habitat rela- tionships of waterfowl broods on South Dakota stock ponds. Journal of Wildlife Management 44: 695-700. Sousa, W. R 1985. Disturbance and patch dynamics on rocky intertidal shores. Pages 101-124 in S. T. A. Pickett and P S. White, eds.. The ecolo- gy of natural disturbance and patch dynamics. Academic Press, New York. Coulson, R. N., and J. A. Witter. 1984. Forest ento- mology: ecology and management. John Wiley and Sons, Inc., New York. 669 pp. TABLES are double spaced on separate sheets and designed to fit the width of either a single column or a page. Use lowercase letters to indicate foot- notes. PHOTOCOPIES OF FIGURES are submitted initially with the manuscript; editors may suggest changes. Lettering on figures should be large enough to withstand reduction to one- or two-column width. Originals must be no larger than 22x28 cm. NOTES. If the manuscript would be more appro- priate as a short communication or note, follow the above instructions but do not include an abstract. A CHARGE of $50 per page is made for articles published; the rate for individual subscribers will be $35 per page. However, manuscripts with com- plex tables and/or numerous half-tones will be assessed an additional charge. Reprints may be pur- chased at the time of publication (an order form is sent with the proofs). FINAL CHECK: • Cover letter explaining any duplication of information and providing phone number(s), FAX numl)er, and E-mail address • 4 copies of the manuscript and WordPerfect diskette • Conformity with instructions • Photocopies of illustrations (ISSN 0017-3614) GREAT BASIN NATURALIST Vol 54. no 1. January 1994 CONTENTS Articles Birds of northern Black Mesa, Navajo County, Arizona. . . . Charles T. LaRue 1 Effects of cobble embeddedness on the microdistribution of the sculpin Cottus beldingi and its stonefly prey Roger J. Haro and Merlyn A. Brusven 64 Persistent pollen as a tracer for hibernating butterflies: the case of Hesperia juba (Lepidoptera: Hesperiidae) Amy Berkhousen and Arthur M. Shapiro 71 Breeding ecology of Long-billed Curlews at Great Salt Lake, Utah Peter W. C. Paton and Jack Dalton 79 Notes Habitat use and behavior of male mountain sheep in foraging associations with wild horses Kevin P. Coates and Sanford D. Schemnitz 86 Fish mortality resulting from delayed effects of fire in the Greater Yellowstone Ecosystem Michael A. Bozek and Michael K. Young 91 H E IBRARY UiMi tf'ui<3^iyV GREAT BASIN NMRAUST VOLUME 54 N2 2 — APRIL 1994 BRIGHAM YOUNG UNIVERSITY GREAT BASIN NATURALIST Editor Richard VV. Baumann 29()MLBM PO Box 20200 Brighani Yoiinj^ University Provo, UT 84602-0200 ' 801-378-5053 FAX 801-378-3733 Assistant Editor Nathan M. Smith 190 MLBM PO Box 26879 Brighani Yoimg University Provo, UT 84602-6879 ' 801-378-6688 E-mail: NMS@HBLL1.BYU.EDU Associate Editors MiciiAKi. A. Bowers Bland\' Experimental Farm, University of Virginia, Box 175, Boyce, VA 22620 J. R. Callahan Museum of Southwestern Biology, University of New Mexico, Albuquerque, NM Mailing address: Box 3140, Hemet, CA 92546 Jeffrey J. Joiiansen Department of Biologv, John Carroll University University Heights, OH 44118 BORLS C. KONDRATIEFF Department of Entomology, Colorado State University Fort Collins, CO 80523 PaulC. Marsh Center for Environmental Studies, Arizona State University, Tempe, AZ 85287 Stanley D. Snhth Department of Biology University of Nevada-Las Vegas Las Vegas, NV 89154-4004 Fault. Tueller Department of Environmental Resource Sciences University of Nevada-Reno, 1000 Valley Road Reno, NV 89512 Robert C. Whitmore Division of Forestry, Box 6125, West Virginia University, Morgantown, WV 26506-6125 Editorial Board. Jerran T Flinders, Chairman, Botany and Range Science; Duke S. Rogers, Zoology; William Hess, Botan\' and Range Science; H. Duane Smith, Zoology. All are at Brighani Young University. Ex Officio Editorial Board members include Steven L. Taylor College of Biolog\' and Agriculture; Stanle\ L. Welsh, Director, Monte L. Bean Life Science Museum; Richard W Baumann, Editor Great Basin Naturalist. The Great Basin Naturalist, founded in 1939, is published quarterly by Brighani Young University. Unpublished manuscripts that further our biological understanding of the Great Basin and surrounding areas in western North America are accepted for publication. Subscriptions. Annual subscriptions to the Great Basin Naturalist for 1994 are $25 for individual sub- scribers ($30 outside the United States) and $50 for institutions. The price of single issues is $12. All back issues are in print and available for sale. All matters pertaining to subscriptions, back issues, or other busi- ness should be directed to the Editor, Great Basin Naturalist, 290 MLBM, PO Box 20200, Brigham Young University, Provo, UT 84602-0200. Scholarly Exchanges. Libraries or other organizations interested in obtaining the Great Basin Naturalist through a continuing exchange of scholarly publications should contact the Exchange Librarian, 6385 HBLL, PO Box 26889, Brighani Young University, Provo, UT 84602-6889. Editorial Production Staff JoAnne Abel Technical Editor Jan Spencer Assistant to the Editor Copyright © 1994 by Brighain Young Uni\ersity Oflkial publication date: 29 April 1994 ISSN 0017-3614 4-94 750 355 The Great Basin Naturalist Published at Provo, Utah, by Brigham Young University ISSN 0017-3614 Volume 54 30 April 1994 No. 2 Great Basin Naturalist 54(2), ©1994, pp. 97-105 COLONY ISOLATION AND ISOZYME VARIABILITY OF THE WESTERN SEEP FRITILLARY, SPEYERIA NOKOMIS APACHEANA (NYMPHALIDAE), IN THE W^ESTERN GREAT BASIN Hugh B. Britten^, Peter E BrussarcU, Dennis D. Murphy^, and George T. Austin^ Abstract. — Thirteen Speijeria nokomis apacheana (Edwards) (Nymphalidae) populations from the western Great Basin were assayed for isozyme variability' using starch-gel electrophoresis. Eight of the 25 presumptive isozyme loci analvzed were found to be polymorphic. Collections made in 1991 and 1992 allowed for between-year comparisons of heterozygosity' and the estimation of effective population size for five of the sampled populations. Speijeria rwkomis apacheana populations exhibit lower mean heterozygosity levels than other nymphalids. This may be attributed to genetic drift in apparently isolated populations with small effective sizes. Key words: Lepidoptera, protein electrophoresis, population structure, heterozygosity. Great Basin, gene flow. The western seep fritillary, Speijeria numbers of populations. In this way extirpated nokomis apacheana (W. H. Edwards) (Nymph- populations are recolonized, and declining alidae), is confined to mesic areas in the Great populations are genetically and demographi- Basin where both its larval foodplant, Viola cally augmented. Protein electrophoresis is a nephrophijlla (Greene) (Violaceae), and the most useful tool for assessing population structure important adult nectar source, Cirsium (Mill.) and levels of genetic variability in species with (Asteraceae), co-occur Adults of the single brood this t\'pe of distribution (e.g., Vrijenhoek et al. are present from late July through mid-Sep- 1985, Waller et al. 1987, Dinerstein and tember and are rarely observed far from McCracken 1990). The two goals of this study colony sites. Population sizes are variable, were to ascertain population structure of Some colonies contain many hundreds of indi- Speijeria nokomis apacheana in the western viduals, while others can be quite small with Great Basin and to estimate levels of isozyme fewer than 10 adults observed over several days. variability within its populations. Small, isolated populations theoretically are exposed to a number of demographic, environ- Materials and Methods mental, and genetic threats to their persis- tence (Gilpin and Soule 1986, Shaffer 1987, Speijeria nokomis apacheana adults were Boyce 1992). Long-term persistence of such collected from 10 sites in Nevada and eastern populations usually requires dispersal among California in late August and September 1991 ^Nevada Biodiversity- Research Center, Department of Biology, University of Nevada-Reno, Reno, Nevada 895.57 ^Nevada State Museum and Historical Society, 700 Twin Lakes Drive. Las Vegas, Nevada 89107. 97 98 Great Basin Naturalist [Volume 54 \ Secret Pass • Ruby Valley • • Reese River ^ Reno 1 Reno L o J O • Kingston Canyon Tahoe \^ X_y^ Scossa Ranch \ ^^ * ^^ ^ Sweetwater North ^^ ^ '^y^ Canyon * CA Route 108# ^V • Sweetwater South * Devil's Gate^^^ Mono County Park A . _^V LeeVining^V^ ) ^^ Mono Lake ^^ \y^ Round Valley % ^^S^C^ Fig. 1. Map of the western Great Basin with Speyeria nokoinis apacheana collection sites denoted with closed circles (•). Sites sampled in 1991 and 1992 are indicated by *. and from 8 sites in late summer 1992 (Fig. 1). One or two collectors sampled each popula- tion, and collecting efforts required 1-3 h per site. Captured individuals were frozen in liquid nitrogen for transport and were subsequently stored in an ultra-cold freezer at -80°C. Allozyme variation was assayed at 25 pre- sumptive loci (Table 1); general methods and procedures followed Brussard et al. (1985). Genotype frequencies were obtained by direct count from phenotypes observed on the gels. The most common electromorph (allozyme) at each locus was designated as "C, " with relatively faster migrating allozymes scored as "B." Still faster migrating allozymes were scored as "A" alleles. Likewise, allozymes that migrated slower than the "C" alleles were designated as "D," and progressively later let- ters in the alphabet were assigned to still slower allozymes. Data from each year of sampling were analyzed separately. Estimates of polymor- phism level and heterozygosity and tests for conformance to Hardy-Weinberg expectations were made using BIOSYS-1 (Swofford and Selander 1981). A X- test for heterogeneity (Sokal and Rohlf 1981) was used to test the significance of allele frequency differences between populations at all polymorphic loci. Fixation indices (F-statistics) were estimated for a hierarchy with three levels: total sample, regional samples, and individual populations. Regions were delineated as (1) western, including nine sites in eastern California and 1994] Western Seep Fritillary Colony Isolation 99 Table 1. Enzymes assayed and buffer systems used in the protein electrophoretic analysis oi Speyeria nokomis apacheana populations in the Great Basin. Locus Enzyme Enzyme commission number Buffc 2.6.1.1 R" 2.7.4.3 4b L8.L4 R C<^ 5.3.L9 4 LLL49 4 LLL30 R 1.1. L14 R 1.1.1.42 C 1.1.1.37 4 1.1.1.40 C 5.3.1.8 R 3.4.-.- R 3.4.11.1 4 1.1.1.43 C 5.4.2.2 R 1.15.1.1 C AAT-1,2,3 Aspartate aminotransferase AK Adenylate kinase DIA NADH diaphorase GP-1,2,3 General (unidentified) protein GPI-1,2 Glucosephosphate isomerase G6PDH Glucose-6-phosphate dehydrogenase HBDH Hydrcxybuteric dehydrogenase IDDH L-iditol dehydrogenase IDH-1,2 Isocitrate dehydrogenase MDH-1,2 NAD Malate dehydrogenase MDHP NADP Malate dehydrogenase MPI Mannosephosphate isomerase PEP- A Peptidase (glycyl-leucine) PEP-E-1,2 Aminopeptidase (cytosol) PGD Phosphogluconate dehydrogenase PGM Phosphoglucomutase SOD Superoxide dismutase "From Ridgeway et al. (1970). ''From Selander et al. (1971). 'Electrode buffer, 0.04 mol/dm'' citric acid adjusted to pH fi.l with N-(3-amino-propyl)-morpholine; diluted 1:10 for gel buffer (Clayton and Tretiak 1972). western Nevada; (2) central, including Reese River and Kingston Canyon sites; and (3) east- ern, including Secret Pass and Ruby Valley sites (Fig. 1). Because F-statistics are hierar- chical "inbreeding coefficients " (Hartl and Clark 1989), they can be used to compare directly relative levels of gene flow among populations within regions and among popula- tions within the total sample. The simultane- ous test procedure (Sokal and Rohlf 1981) was used to test for homogeneity of genotype fre- quencies among samples in the western region to provide further insight into popula- tion structure along the eastern slope of the Sierra Nevada. The genetically effective num- ber of migrants per generation {Nm) within homogeneous groups of populations was esti- mated from the F-statistics following Slatkin (1987). Samples from sites collected in both years were pooled for the calculation of genet- ic distances. The UPGMA clustering algo- rithm was used to derive a phenogram based on genetic distances. Genetically effective population sizes {N^'s) were calculated for populations sampled in both 1991 and 1992 using the methods of Nei and Tajima (1981) and Pollack (1983). These methods calculate standardized variances in allele frequency change at polymorphic loci sampled at two or more different times. These variances provide an estimate of genetic drift which, in turn, is inversely related to N^. Results Allele Frequencies and Genetic Variability Six of the 25 presumptive allozyme loci assayed were polymorphic in at least one of the populations sampled in 1991 (Table 2), and all populations conformed to Hardy-Weinberg expectations at these loci. The 10 sampled populations were fixed for the same alleles at all other loci analyzed. Mean observed het- erozygosity estimates ranged from 0.014 in the Secret Pass population to 0.042 for the Nye Canyon population (Table 2). Seven of the 25 presumptive allozyme loci assayed in the 1992 samples were polymor- phic (Table 2), and all these samples con- formed to Hardy-Weinberg expectations at all polymorphic loci. As in 1991, all populations were fixed for the same alleles at the monomorphic loci. Mean observed heterozy- gosit\' estimates ranged from 0.020 in the Nye Canyon population to 0.044 in the Sweetwater North population (Table 2). Geographical Structure Among Colonies Sample sizes for 1991 and 1992 collection efforts are given in Table 2. Five sites were repeat-sampled in 1992 (Sweetwater North and South, Nye Canyon, Reese River, and Kingston Canyon), and yearly allele frequencies for 100 Great Basin Naturalist [Volume 54 Table 2. Sample sizes and allele frequencies at polymorphic loci assayed in Speyeria nokomis apacheana populations sampled in 1991 and 1992 with percent polymorphic loci (P) and direct count mean heterozygosity estimates (H). te^ Western Central Eas item Sample si SVVN SWS DG NC CI 08 ML SG RVC LV RR KG RV SP Sample size (1991) 21 42 18 20 10 20 0 0 0 58 34 17 14 Locus Allele AAT-2 B 0.00 0.00 0.00 0.00 0.00 0.00 _ _ _ 0.00 0.00 0.00 0.25 C 1.00 1.00 1.00 1.00 1.00 1.00 — — — 1.00 1.00 1.00 0.75 GPI-1 C 1.00 1.00 0.94 1.00 1.00 1.00 _ _ _ 1.00 1.00 1.00 1.00 D 0.00 0.00 0.06 0.00 0.00 0.00 — — — 0.00 0.00 0.00 0.00 GPI-2 B 0.00 0.00 0.00 0.00 0.00 0.00 _ _ _ 0.00 0.00 0.32 0.00 C 1.00 1.00 1.00 1.00 1.00 1.00 — — — 1.00 1.00 0.68 1.00 MPI C 0.62 0.88 0.94 0.62 1.00 0.97 _ _ 1.00 1.00 1.00 1.00 D 0.38 0.12 0.06 0.38 0.00 0.03 — — — 0.00 0.00 0.00 0.00 PGM B 0.00 0.02 0.00 0.00 0.00 0.00 _ _ 0.00 0.00 0.00 0.00 C 0.74 0.33 0.47 0.60 0.45 0.32 — — — 0.81 0.26 0.32 1.00 D 0.26 0.65 0.53 0.40 0.55 0.68 — — — 0.19 0.74 0.68 0.00 SOD C 1.00 1.00 1.00 1.00 1.00 1.00 _ 0.65 1.00 1.00 1.00 D 0.00 0.00 0.00 0.00 0.00 0.00 — — — 0.35 0.00 0.00 0.00 P 8 8 12 8 4 8 _ 8 4 8 4 H 0.036 0.024 0.027 0.042 0.020 0.016 — — — 0.032 0.016 0.028 0.014 Sample size (1992) 37 32 0 10 0 0 48 80 13 33 31 0 0 Locus Allele GPI-1 C 1.00 1.00 1.00 1.00 1.00 1.00 0.98 1.00 D 0.00 0.00 — 0.00 — — 0.00 0.00 0.00 0.02 0.00 — — GPl-2 B 0.00 0.08 0.00 0.00 0.00 0.00 0.00 0.00 C 1.00 0.92 — 1.00 — — 1.00 1.00 1.00 1.00 1.00 — — MDH-1 B 0.00 0.00 0.00 0.00 0.00 0.00 0.02 0.00 C 1.00 1.00 — 1.00 — — 1.00 1.00 1.00 0.98 1.00 — — MPI C 0.62 0.84 0.75 0.83 0.99 1.00 1.00 1.00 _ _ D 0.27 0.04 — 0.05 — — 0.17 0.01 0.00 0.00 0.00 — — E 0.11 0.12 — 0.20 — — 0.00 0.00 0.00 0.00 0.00 — — PGD B 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 _ _ C 1.00 1.00 — 1.00 — — 1.00 0.99 1.00 0.97 0.95 — — D 0.00 0.00 — 0.00 — — 0.00 0.00 0.00 0.03 0.05 — — PGM B 0.00 0.11 0.00 _ _ 0.00 0.08 0.12 0.00 0.00 C 0.69 0.33 — 0.90 — — 0.56 0.28 0.58 0.98 0.35 D 0.31 0.56 — 0.10 — — 0.44 0.64 0.30 0.02 0.65 — — SOD C 1.00 1.00 1.00 1.00 1.00 1.00 0.61 1.00 _ _ D 0.00 0.00 — 0.00 — — 0.00 0.00 0.00 0.39 0.00 — — P 8 8 8 _ _ 8 12 4 20 8 H 0.044 0.032 — 0.020 — — 0.034 0.023 0.022 0.021 0.022 — — "Sample sites: SWN = Sweehvater North, SWS = Sweetwater South. DG = Devils Gate, NC = Nye Canyon, C108 = Cal Rt 108, ML = Mono Lake, SC Scossa Ranch, RVC = Round Valley, CA, LV = Lee Vining, RR = Reese River, KC = Kingston Canyon, RV = Ruby Valley, SP = Secret Pass. 1994] Western Seep Fritillary Colony Isolation 101 Table 3. Unbiased genetic distances (Nei 1978) above diagonal and unbiased genetic identities (Nei 1978) below diagonal for 13 Great Basin populations of Speyeria nokomis apacheana sampled in 1991 and 1992. Population SWS SWN DC NC C108 ML SC RVC LV RV SP RR KC Sweetwater South ***** 0.007 0.000 0.006 0.000 0.000 0.002 0.000 0.003 0,004 0.020 0.017 0.001 Sweetwater North 0.993 ***** 0.005 0.000 0.007 0.010 0.002 0.011 0.005 0,015 0.011 0.012 0.011 Devil's Gate 1.000 0.995 ***** 0,004 0,000 0.000 0.000 0.001 0.001 0,005 0.014 0.012 0,001 Nye Canyon 0.994 1.000 0.996 ***** 0,006 0.008 0.001 0.009 0.004 0,013 0.009 0.010 0,010 CA Rt. 108 1.000 0.993 1.000 0,994 ***** 0.000 0,001 0.000 0,001 0,004 0.014 0.012 0,000 Mono Lake 1.000 0.990 1.000 0.992 1.000 ***** 0,003 0.000 0,004 0,004 0.021 0.018 0,000 Scossa Ranch 0.998 0.998 1.000 0.999 0.999 0.997 ***** 0,003 0,001 0.007 0.011 0.011 0.004 Round Valley 1.000 0.989 0.999 0.991 1.000 1.000 0.997 ***** 0,004 0.004 0.022 0.019 0.000 Lee Vining 0.997 0.995 0,999 0.996 0.999 0,996 0.999 0,996 ***** 0.008 0.008 0.008 0.004 Ruby Valley 0.996 0.985 0.995 0,987 0.996 0,997 0.993 0,996 0,992 ***** 0.025 0.022 0.004 Secret Pass 0.981 0.989 0,987 0.991 0.986 0.979 0.989 0,979 0,992 0.975 ***** 0.008 0.022 Reese River 0.983 0.988 0.988 0.990 0.988 0.982 0.989 0,982 0,992 0.978 0.992 ***** 0.019 Kingston Canyon 0.999 0.989 0.999 0.990 1.000 1.000 0.996 1,000 0,996 0.996 0.978 0.981 ***** these populations were pooled to represent an intergenerational mean that can be used to calculate genetic distances among the sampled populations. A UPGMA phenogram based on Nei's (1978) unbiased genetic distances (Table 3) did not represent the geographical relation- ships of the assayed populations particularly well (Fig. 2). For example, Secret Pass, an east- ern population, clustered on the UPGMA phenogram most closely with Reese River, a central population. Similarly, Kingston Canyon, a central population, clustered among a group of western populations in the phenogram. Because of the nonconcordance of the phenogram in Figure 2 with the geo- graphical dispersion of the populations (Fig. 1), we undertook a more detailed analysis of the genetic structure of the sampled popula- tions. Genetic Structure Among and Within Regions Significant heterogeneity in allele frequen- cies (X^ test, p < .05) at all polymorphic loci among all sampled populations in both years of sampling indicated substantial population structuring. When regions were considered independently, significant levels of hetero- geneity in allele frequencies were observed among the populations within each region. Hierarchical F-statistics also pointed to fine- scale population structuring within regions for both years. For example, the 1991 within- region fixation index was 0.227. Using the relationship F = 1/(1 + 4Nm) (Slatkin 1987), we arrive at a figure suggesting an average of only 0.851 individuals dispersed between colonies within each region in 1991. The six 1991 western samples were sub- jected to a simultaneous test procedure (Sokal and Rohlf 1981) to obtain further insight into the geographic structures among them. All populations were heterogeneous except two groups. The first, consisting of California Route 108, Devil's Gate, Mono Lake, and Sweetwater South, was found to be homoge- neous with respect to allele frequencies across two of the three (MPI and PGM) polymorphic loci in the samples. The greatest linear dis- tance between any pair of these populations is approximately 46 km (Fig. 1). The mean F^^ among the four populations was 0.022. The second homogeneous group consisted of Nye Canyon and Sweetwater North, two popula- tions separated by 2 km of apparently suitable habitat (Fig. 1). The mean F^^ estimate for these two populations was 0.010. The simultaneous test procedure revealed no homogeneous groups among the 1992 west- em populations. Nye Canyon and Sweetwater 102 Great Basin Naturalist [Volume 54 Similarity .98 .99 1.00 Region Sweetwater South Devil's gate — California Route 108 Mono Lake Round Valley Kingston Canyon — Scossa Ranch I - Lee Vining — Ruby Valley I Sweetwater North I Nye Canyon - Secret Pass - Reese River West West West West West Central West West East West West East Central •+ — .98 99 1.00 Fig. 2. UPGMA phenogram based on Neis (1978) unbiased genetic identities for 13 Great Basin populations o{ Spetj- eria nokomis apacheana. Regional designations indicate that topology of the phenogram does not correspond to the geo- graphic arrangement of tlie populations. Cophenetic correlation coefficient is 0.873. North formed a marginally heterogeneous group (total X^ for two loci = 8.45, df = 3, p = .04) with a mean F^( estimate of 0.05 {Nm = 4.7; Slatkin 1987). Estimates of Effective Population Size Five Speyeria nokomis apacheana popula- tions, Sweetwater North and South, Nye Canyon, Reese River, and Kingston Canyon, provided large enough samples in 1991 and 1992 (Table 2) to allow estimation of N^'s using the methods of Nei and Tajima (1981) and Pol- lack (1983). Estimates of iV^ derived from both methods for replicated samples are given in Table 4. Estimates were smaller than the num- ber of individuals sampled at each site. The large confidence intervals around these esti- mates (Table 4) result from the small niniiber of alleles used in their estimation and the small number of generations (n = 1) over which the study was conducted (Nei and Taji- ma 1981, Pollack 1983, VVaples 1989). Despite these limitations, the estimates are consistent with the high degree of structuring observed and indicate that the N^'s of Speyeria nokomis in the Great Basin are generally small. Discussion Mean population heterozygosity estimates for Speyeria nokomis apacheana are consis- tently lower than heterozygosity in other 1994] Western Seep Fritillary Colony Isolation 103 Table 4. Estimates of effective population sizes {N^'s) for five repeat-sampled Speyeria nokoinis apacheana pop- ulations collected in 1991 and 1992. Confidence intenals are in parentheses. Estimates ofN^ Population Nei and Tajima Pollack (1981) (1983) Sweetwater North 36 10 (4-oc) (0-37) Sweetwater South 9 6 (2-00) (0-17) Nye Canyon 2 1 (0—) (0-4) Reese River 4 3 (l-oo) (0-11) Kingston Canyon 73 70 (3_oo) (0-1118) nymphalid butterflies. For example, Brussard et al. (1989) estimated a range of mean het- erozygosities of 0.17-0.26 among western North American populations in a complex of semispecies within the highly variable species Euphydryas chalcedona. Britten et al. (1993) estimated mean heterozygosities of 0.041- 0.127 in Canadian Boloria itnproba, and 0.031 for the closely related, endangered, and nar- rowly endemic butterfly Boloria acrocnema (Britten et al. 1994); both values are near the high range of estimates for those of Speyeria nokomis in the Great Basin (Table 2). In addi- tion, Brittnacher et al. (1978) estimated that mean heterozygosity in a number of California Speyeria species and subspecies ranged from 0.141 in Speyeria coronis coronis to 0.067 in Speyeria atlantis. Brittnacher et al. (1978) esti- mated a mean heterozygosity of 0.034 for Speyeria nokomis apacheana at Round Valley, a figure somewhat higher than our estimated heterozygosity for that population (0.023, Table 2), but within the range of estimates made herein for other Speyeria nokomis apacheana populations (Table 2). A number of evolutionaiy forces could be responsible for the apparent lack of heterozy- gosity in the sampled Speyeria nokomis apacheana populations. Selection against het- erozygous individuals is theoretically capable of reducing heterozygosity. There is, however, little indication that selection acts frequently on allozvmes. Furthermore, selection is weak relative to genetic drift in small populations (Crow 1986, Hard and Clark 1989). Because genetic drift in small, isolated pop- ulations can erode heterozygosity over num- bers of generations (Hartl and Clark 1989), it is the most plausible explanation for the observed low levels of isozyme variability in the Speyeria nokomis populations included in this study. Drift is most effective in eroding heterozygosity and causing loss of neutral alle- les when populations are small and isolated. Speyeria nokomis populations in the Great Basin appear to meet both criteria. Although fairly large sample sizes were obtained at sev- eral sites, much smaller samples were taken at most sites with similar collecting efforts (Table 2). Sample size, therefore, is a rough indicator of relative population size, and most popula- tions appeared to consist of far fewer than 100 individuals on the days they were sampled. Estimates of N^ (Table 4) corroborate the evi- dence for small effective population sizes in this taxon. Several authors (e.g., Frankel and Soule 1981, Allendorf 1986, Hedrick and Miller 1992) have stressed the importance of drift in the loss of selectively neutral alleles from pop- ulations subjected to bottlenecks. Populations with chronically small sizes may be consid- ered analogous to populations that have suf- fered a series of bottlenecks. In that light, larger Speyeria nokomis populations would be expected to retain a larger complement of alle- les over longer periods of time than would smaller ones. Furthermore, allele frequencies are expect- ed to fluctuate less between generations when N^'s are consistendy large. This hypothesis is partially testable by comparing the results of our study with those of Brittnacher et al. (1978) for the Round Valley population. Britt- nacher et al. (1978) collected 58 Speyeria nokomis from this colony in 1974 and 1975. A single locus (PGM) was polymorphic with the following allele frequencies: 0.69 for "PGM97," 0.23 for "PGM 100," and 0.08 for "PGM 103" (Brittnacher et al. 1978). Based on relative migration rates of alleles within each study, PGM97, PGM 100, and PGM 103 are assumed to be homologous with PGMD, PGMC, and PGMB, respectively, of the present study. As shown in Table 2, allele frequencies at the PGM locus have changed litde in this large colony over the 18 years intervening 104 Great Basin Naturalist [Volume 54 between the two studies. In contrast, the Reese River sample, with an estimated N^ of 3-4 (Table 4), experienced some rather large changes in allele frequencies at several loci in the short interval between the 1991 and 1992 generations (Table 2). Dispersal among butterfly colonies is expected to ameliorate the erosive effects of genetic drift in individual populations. Het- erogeneity tests and F-statistics, however, sug- gest that Speyeria nokomis colonies are gener- ally isolated from each other and that even geographically proximate populations are like- ly to be drifting independently. Even though the 1991 data suggest that two sets of popula- tions were homogeneous and that there was substantial gene flow among them (Slatkin 1987), sample sizes from these populations were necessarily small and the resultant power of the G-test (Sokal and Rohlf 1981) to detect heterogeneity among allele frequencies was low (mean power for pairwise G-tests = 0.046). Thus, the apparent homogeneity and implied high rate of gene flow among these populations may not be real. Alternatively, dis- persal rates among colonies may change between years depending on weather or other environmental conditions. In any case, colonies are not static; the number of individ- uals dispersing among colonies apparently changes from generation to generation. This situation probably reflects environmentally mediated fluctuations in population sizes and resource availability. The results of this study further confirm that Speyeria nokomis populations are con- fined to mesic seep, spring, and riparian areas in the Great Basin. Such habitats are often separated by tens of kilometers of unsuitable habitat in which individual butterflies have never been observed. The isozyme data pre- sented here indicate very low levels of gene flow among the sampled populations and sug- gest that these populations may have lost genetic variabilit)' as a result of small effective sizes and genetic drift. Because a number of unique alleles were detected in several popu- lations (Table 2), conservation of individual colonies may be important to the evolutionary potential of this subspecies (Frankel and Soule 1981, Allendorf 1986). The apparent philo- patric nature of this butterfly results in geneti- cally unique colonies whose habitat should be preserved in order to achieve this goal. Acknowledgments This work was supported by the Nevada Biodiversity Research Center. We wish to thank two anonymous reviewers for their use- ful comments on the manuscript. Literature Gited Allendorf, F. VV. 1986. Genetic drift and the loss of alleles versus heterozygosity. Zoo Biolog>' 5; 181-190. BOYCE, M. S. 1992. Population viabilit>' analysis. Annual Review of Ecology and Systematics 23: 481-506. Britten, H. B., P. F. Brussard, and D. D. Murphy. 1994. The pending e.xtinction of the Uncompahgre fritillary butterfly. Conservation Biology: In press. Brittnacher, J. C, S. R. Sims, and F. J. Ayala. 1978. Genetic differentiation between species of the genus Speyeria (Lepidoptera: Nymphalidae). Evolution 32: 199-210. Brussard, P. F., J. F. Baughm.\n, D. D. Murphy, P. R. Ehrlich, and J. Wright. 1989. Complex popula- tion differentiation in checkerspot butterflies {Euphijdnjas spp.). Canadian Journal of Zoolog\' 67: 330-335. Bruss.\rd, P. F., P. R. Ehrlich, D. D. Murphy, B. A. Wilcox, and J. Wright. 1985. Genetic distances and the taxonomy of checkerspot butterflies (Nymphalidae: Nymphalinae). Journal of the Kansas Entomological Society 58: 403^12. Cla\ton, J. W., AND D. N. Tretiak. 1972. Aminocitrate buffers pH control in starch gel electrophoresis. Journal of Fisheries Research Board of Canada 29: 1169-1172. Crow, J. F. 1986. Basic concepts in population, quantita- tive, and evolutionaiy genetics. W. H. Freeman and Co., New York. 213 pp. DiNERSTEiN, E., AND G. F. McCracken. 1990. Endan- gered greater one-homed rhinoceros cany high lev- els of genetic variation. Conservation Biologv 4: 417-422. Frankel, O. H., and M. E. Soule. 1981. Conservation and evolution. Cambridge University Press, Cam- bridge. 327 pp. Gilpin, M. E., and M. E. Soule. 1986. Minimum viable populations: the process of species extinction. Pages 13-34 in M. E. Soule, ed., Conservation biolog\': the science of scarcity and diversit)'. Sinauer Associates, Inc., Sunderland, Massachusetts. Hartl, D. L., and a. G. Clark. 1989. Principles of popu- lation genetics. Sinauer Associates, Inc., Sunderland, Massachusetts. 682 pp. Hedrick, p. W., and P. S. Miller. 1992. Consenation genetics: techniques and fundamentals. Ecological Applications 2; 30—46. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583-590. Nei, M., and F. Tajima. 1981. Genetic drift and the esti- mation of effective population size. Genetics 98: 62.5-640. Pollack, E. 1983. A new method for estimating the effec- tive population size from allele frequency changes. Genetics 104: ,531-548. 1994] Western Seep Fritillary Colony Isolation 105 RiDGEWAY, G. J., S. W. Sherburne, and R. D. Lewis. Swofford, D. L., and R. B. Selander. 198L BIOSYS-L 1970. Polymorphism in the esterases of the Atlantic a FORTRAN program for the comprehensive analy- herring. Transactions of the American Fisheries sis of electrophoretic data in populations genetics Society 99: 147- L5L and evolution. Journal of Heredity 72: 281-283. Seij\nder, R. K., M. H. Smith, S. Y. Yang, W. E. John- Vrijenhoek, R. C, M. E. Douglas, and G. K. Meffe. SON, and J. B. Gentry. 1971. Biochemical polynior- 1985. Conservation genetics of endangered fish pop- phism and systematics in the genus Peromyscus. 1. ulations in Arizona. Science 229: 400^02. Variation in the old-field mouse Pero»i!/sci/,s /Jo//o/io- WALLER, D. M., D. M. O'Malley, and S. C. Gawler. tus. Studies in Genetics VI. University of Te.xas Pub- 1987. Genetic variation in the extreme endemic lication 7103: 49-90. Pedicuhiris furbishiae (Scrophulariaceae). Conserva- Shaffer, M. 1987. Minimum viable populations: coping tion Biology 1: 335-340. with uncertainty. Pages 69-86 in M. E. Soule, ed., Waples, R. 1989. A generalized approach for estimating Viable populations for conservation. Cambridge effective population size from temporal changes in University Press, Cambridge. allele frequency. Genetics 121: 379-391. Slatkin, M. 1987. Gene flow and the geographic struc- ture of natural populations. Science 236; 787-792. Sokal, R. R., and F. J. Rohlf. 1981. Biometry; the prin- Received 24 August 1993 ciples and practice of statistics in biological research. Accepted 29 September 1993 W. H. Freeman and Co., San Francisco. 805 pp. Great Basin Naturalist 54(2), ©1994, pp. 106-1 13 INFLUENCE OF FINE SEDIMENT ON MACROINVERTEBRATE COLONIZATION OF SURFACE AND HYPORHEIC STREAM SUBSTRATES Carl Kicluirtls^-^ and Kermit L. Bacon^ Abstract. — Colonization of macroinvertebrates was assessed in a stream impacted by inputs of fine sediments. Col- onization was e.xamined at the stream surface and within the hyporheos with Whitlock-Vibert (W-V) bo.xes. Hyporheic areas accumulated much greater amounts of all size categories of sediment. No significant difference was observed in the amounts of organic matter accumulated at either depth. Only 150-^(,m sediment had significant effects on macroin- vertebrate total numbers and number of titxa. Total numbers of invertebrates at 30 cm were only 20% of those at the sur- face. Canonical Correspondence Analysis indicated that the strongest influence on macroinvertebrates colonizing W-V boxes at the surface was stream size and secondarily fine sediments. Within the hyporheos, abundance of fine sediment was the dominant influence on macroinvertebrate assemblages. Impacts of sedimentation on hyporheic environments can have significant and persistent impacts on streams. Key words: stream ecology, hyporheos, sediment, organic matter, macroinvertebrate. The addition of fine substrates to streams can result in significant changes to stream macroinvertebrate assemblages. Substrate plays an important role in structuring stream macroinvertebrate assemblages. Numerous studies (see Minshall 1984) have demonstrat- ed the importance of both substrate type and size in determining distributions of specific taxa. In general, the number of taxa and pro- ductivity of substrates composed of small par- ticle sizes are less than those of larger, more heterogeneous substrates (Pennak and Van Gerpen 1947, Allan 1975, Ward 1975). Reduced invertebrate utilization and production from small substrates may be attributed to a variety of reasons, ranging from the need of some insects for large particles for attachment, to the need for interstitial pore space for movement among substrate particles. Macroinvertebrate responses to variation in substrate size and composition can result in distributi(^n patterns that are observed within streams longitudinal- ly (Allan 1975) and among several streams within a region (Richards et al. 1993). The addition of fine substrates to streams may also affect macroinvertebrate abundance and distribution in the hyporheos. Taxa within the hyporheic region of streams can be found as deep as 70 cm below the stream bottom (Williams and Hynes 1974). The benthic assemblage within the hyporheic region is associated with overall stream productivity and surface assemblage structure (Strommer and Smock 1989, Ward 1989). Because macroinvertebrates utilize the hyporheos dur- ing all seasons, this area can provide a refuge for new colonists following high flows or other disturbance events (Williams 1984, Palmer et al. 1992). Alterations to physical characteris- tics of the hyporheos could cause significant changes in the dynamics of macroinvertebrate populations that utilize these areas. This study was undertaken to determine whether fine sediment inputs from both point and nonpoint sources influenced macroinver- tebrate assemblages along the length of a stream in central Idaho. We hypothesized that assemblage structure could be related to the proportion of fines in surface and hyporheic substrates. Methods Study Area The study was conducted in Bear Valley Creek, a headwater tributary to the Middle Fork of the Salmon River watershed in central Idaho. The stream flows through subalpine meadows and lodgepole pine {Pinus contorta) 'Fisheries Department, Shoshone-Bannock Tribes, Box 306, Fort Hall, Idaho 8.3208. ^Present address: Natural Resources Research Institute, University of Minnesota-Duluth, .5013 Miller Trunk HiKlnvay, Duluth, Minnesota 55811. 106 1994] Macroinvertebrate Colonization of Substrates 107 forests on a granitic batholith. Alluvial deposits of erosive sandy soils typify the region. Historically, the stream had high sec- ondary productivity and supported large pop- ulations of anadromous salnionids. Since the 1950s, however, large volumes of fine, inor- ganic sediments have entered the stream through both point and nonpoint sources along the length of the stream (Konopacky et al. 1986). Consequently, stream substrates have high proportions of fine sediments in many areas, and fish production has declined partly as a result of sediment introduction. Experimental Method To examine whether sediment influences macroinvertebrates assemblages, we conduct- ed colonization studies at 19 sites along a 50- km section of Bear Valley Creek. Colonization studies are effective means of examining the dynamics of stream macroinvertebrate assem- blages and have been used extensively in many streams and geographic areas (Robinson et al. 1990, Mackay 1992). Study sites were located in riffle habitats at approximately even intervals along the length of the stream. Sites reflected the full range of substrate character- istics found in the stream, including low pro- portions of fine sediments and high propor- tions of sediments. At each site, stream width, gradient, and substrate composition were assessed. Sub- strate was assessed by determining the pro- portion of surface particles <4 mm in diame- ter. One hundred points were randomly selected along a transect that bisected the study riffles, and the closest substrate particle to each point was measured. This size class corresponds well to proportions of smaller- sized surface substrate particles in Bear Valley Creek (Konopacky et al. 1986). We used small basket samplers (Whitlock- Vibert [W-V] boxes; Wesche et al. 1989) for macroinvertebrate colonization. The polypro- pylene (14 X 6.4 X 8.9-cm-deep) boxes enclosed a known volume of standardized sub- strate that allowed comparisons among sites. These boxes are typically used to incubate fish eggs in stream gravels (Federation of Fly Fish- ermen personal communication). The sides, top, and bottom of the boxes are perforated with rectangular slots (3.5 X 13 mm) to allow water circulation. The bottom of each box was covered with duct tape to reduce sediment loss. Boxes were filled with 3/4-inch-grade clean gravel. This readily available substrate approximates the size of clean gravels in Bear Valley Creek. Two boxes were placed at each site: one approximately 30 cm below the sur- face of the substrate (using a small shovel) and the other flush with the surface directly above the below-surface box. Both boxes were locat- ed in the center of the stream channel. Boxes were placed in the substrate the last week of July 1988 and retrieved 10 weeks later Dur- ing this period little or no rainfall was received in the area, and the stream was in a baseflow condition. Colonization was examined in relation to variation in fine sediments that accumulated in the boxes. Fine sediment accumulation in W-V boxes has been shown to be correlated with the amount of fines in surrounding sub- strates in streams and laboratory channels (Wesche et al. 1989). W-V boxes were re- moved from the stream as carefully as possible so as to retain any fine substrate materials and macroinvertebrates in the boxes. While still under water, the W-V box was slipped into a plastic bag with minimal disturbance. The lower box was removed in the same way after excavating the substrate material between the upper and lower boxes. Material from the boxes was preserved in 10% formalin. In the lab the 3/4-inch-grade gravel was removed from the samples with a large sieve. Macroin- vertebrates were removed ft-om these samples under a dissection microscope, identified to family, and enumerated. The remaining mate- rial was divided into portions that collected on 150-/xm and 850-^Lm sieves. These portions were dried at 60°C and then ashed in a muffle furnace to obtain a weight for organic and in- organic (fine sediment) fractions. The 850-;Ltm sieve collected material < 3.5 mm in diameter. Sediment particle sizes <4 mm are frequently implicated in negative impacts on the abun- dance of stream invertebrates and productivi- ty (Nuttall 1972, Alexander and Hansen 1986). Smafler particle sizes (<850 /xm) include clay- sized particles that also decrease invertebrate abundance (Cederholm and Lestelle 1974) and clog interstitial spaces. Differences in sediment and organic accu- mulation between surface and below-surface boxes were examined by group comparisons, as were macroinvertebrate assemblage com- parisons (species richness, total numbers). 108 Great Basin Naturalist [Volume 54 Macroinvertebrate assemblage composition among the sites was examined with multivari- ate direct gradient analysis (Canonical Corre- spondence Analysis; ter Braak 1986, 1987). Macroinvertebrate data were log transformed prior to analysis. In CCA, axes are selected to be linear combinations of environmental vari- ables so that taxa are related directly to a set of these variables. This technique is particu- larly useful for examining the relative strength of various environmental variables on influ- encing assemblage composition (ter Braak and Prentice 1988, Richards et al. 1993). Environ- mental variables used in the analysis were sediment and organic accumulations in the boxes, proportion of 4-mm surface sediments in rifines, gradient, and stream width. The lat- ter two variables were included to account for some differences in stream size and channel moiphology among the sites. Results The width of the study sites ranged from 2.9 to 24 m (mean = 8.57). The proportion of sediments <4 mm in diameter in riffles varied from 0 to 56% (mean = 8.6); all sites had gra- dients <2% (mean = 0.47). A much larger amount of fine sediment accumulated in the below-surface boxes than in the surface boxes (/ test, p < .05; Table 1). This was true for both the 850-^tm and 150- fim size classes. There was no significant dif- ference in the amounts of organic material that accumulated between treatments for either size class. Twenty-two macroinvertebrate families were identified from the W-V samplers (Table 2). With the exception of Perlidae, Cerato- pogonidae, and Tabanidae, all taxa were found in both surface and below-surface locations. Significantly (p < .05) greater numbers of taxa and total numbers of individuals per box were found in the surface samples (Table 2). The most abundant taxa in below-surface samples were Heptageniidae, Leptophlebiidae, Chloro- perlidae, and Chironomidae. These taxa also had relatively higli abvmdance in surface sam- ples. Baetidae and Ephemerillidae had rela- tively high abundance in the surface samples but were not well represented in below-sur- face samples. No taxa were more abundant in below-surface samples than in surface samples. Pearson correlation coefficients were calcu- lated between each size class of sediment and taxa richness and total number of individuals to determine whether fine-sediment variables had relationships to macroinvertebrate assem- blage characteristics. Separate calculations were made for surface and below-surface samples. No significant correlations (p < .05) were found with surface samples. In below-surface samples a significant correlation {p < .05) was observed between the 0.15-mm sediment size class and both number of taxa and total num- bers of individuals (Fig. 1), but no significant correlations were found with the 0.85-mm size class. Results of the CCA analysis for surface samples indicated that sediment accounted for a relatively small proportion of the variance in assemblage composition among sites. The first axis, which described the greatest amount of variation in the ordination, was most strongly influenced by gradient and width (Table 3). This axis differentiated the taxa most abun- dant at sites with high stream gradient and narrow widths from those taxa most abundant at sites with low gradient and wide widths. These data suggest that longitudinal position of the station along the stream course played the greatest role in determining assemblage composition. Nematodes, Ceratopogonidae, Hydropsychidae, and Pteronarcyidae were found in narrow, high-gradient sites, and Rhy- acophilidae and Hydracarina were found at Table 1. Macroinvertebrate numbers and sediment and organic material accumulations in experimental colonization boxes. * denotes a significant difference (p < .05, t test) between surface and below-surface boxes. Surface Below -surf; ice Variable Mean Std. dev. Mean Std. dev. 850-;U.m sediment* (gr/liox) 150-^(,m sediment* (grA^ox) 850-/i,m organic (gr/box) 150-p(,m organic (gr/box) 11.01 11.44 0.435 0.643 11.22 17.98 0.460 0.679 101.01 79.76 0.572 0.706 51.04 50.74 0.359 0.438 1994] Macroinvertebrate Colonization of Substrates 109 Table 2. Macroinvertebrate taxa that colonized surface and below-surface W-V boxes. Surface Below-surface Taxa Mean Std. dev. Mean Std. dev. Baetidae (BAE) 38.02 50.13 3.84 8,51 Ephemeriliidae (EPH) 35.49 45.53 3.50 5.05 Heptageniidae (HEP) 54.81 60.59 10.96 27.68 Leptophlebiidae (LET) 36.05 37.96 23.80 25.70 Siphlonuridae (SIP) 7.89 20.03 2.96 6.78 Brachvcentridae (BRA) 5.81 15.96 0.11 0.48 Hvdropsvchidae (HYD) 1.75 6.66 0.22 0.66 Hvdroptilidae (HYP) 10.03 20.78 0.22 0.66 Lepidostomidae (LEP) 58.28 106.43 2.19 4.52 Limniphilidae (LIM) 12.58 18.95 1.53 3.00 Rhyacophilidae (RHY) 0.66 1.96 0.11 0.48 Chloroperlidae (CHL) 27.35 28.25 12.01 11.75 Perlidae (PED) 0.77 1.88 0.00 0.00 Perlodidae (PER) 5.47 8.27 0.77 1.98 Pteronarcyidae (PTE) 0.77 1.73 0.11 0.48 Ceratopogonidae (CER) 0.55 1.94 0.00 0.00 Chironomidae (CHI) 718.15 868.08 414.84 618.80 Tabanidae (TAB) 3.40 13.82 0.00 0.00 Tipulidae (TIP) 9.65 16.35 5.38 10.08 Elmidae (ELM) 9.75 23.07 2.74 5.24 Annelid (ANN) 10.44 28.29 1.33 2.54 Mollusca (MOL) 11.07 34.45 3.40 7.64 Number tax.\/box 11.3 3.24 7.0 2.62 Total number/box 1120.3 1087.5 492.8 641.33 14 12 10 ir 8 6 50 100 150 2000 E 1500 - 1000 B 500 -•«•%.' 50 100 150 150 um sediment (gr/box) 200 200 Fig. 1. Correlations between number of taxa (r = -.52, p < .05) and total numbers of individuals (r = -.48, p < .05) and the amount of fine sediment that accumulated in W-V boxes within the hyporheos. low-gradient, wide sites (Fig. 2). The second axis was most strongly influenced by total organic weight and width. No environmental variables had strong (r < .35) correlations with the third axis. Sediment was more important in defining differences in assemblage composition among below-surface samples. The 0.85-mm and 0.15-mm sediment volumes had highest corre- lations with the first CCA axis (Table 3). These variables acted in an opposite manner. The majority of taxa were associated with decreas- ing amounts of 0.15-mm sediment; however, Brachycentridae and Hydroptilidae were most abundant at sites with relatively high amounts of 0.15-mm sediment (Fig. 2). Gradient and width had the highest correlations with the second axis. There appeared to be little corre- spondence between taxa preferring high-gra- dient, narrow sites or low-gradient, wide sites in below-surface samples and above-surface samples (Fig. 2). Taxa preferring high-gradient sites were Mollusca, Perlodidae, and Tipuli- dae. Hydracarina, Elmidae, and Siphlonuri- dae preferred wide, low-gradient sites. As with surface samples, the third axis was diffi- cult to interpret. no Great Basin Naturalist [Volume 54 Table 3. Correlations between environmental variables and CCA axes. Percentages refer to the proportion of vari- ance in species data explained. Above surface Axis 1 Axis 2 Axis 3 (17.3%) (9.6%) (8.0%) Surface sediment 0.01 0.32 0.17 (<4 mm) Box sediment 0.14 0.29 -0.13 (850 /im) Box sediment -0.28 0.19 0.06 (150 Mm) Total organic weight 0.03 0.64 0.13 Gradient 0.51 -0.47 -0.28 Width -0.53 0.53 -0.32 Below surface Axis 1 Axis 2 Axis 3 (9.3%) (7.,5%) (6.0%) Surface sediment -0.10 0.17 0.30 (<4mm) Box sediment 0.40 0.34 -0.03 (850 Mm) Box sediment -0.60 -0.01 0.17 (150 Mm) Total organic weight 0.28 0.38 0.38 Gradient 0.02 0.60 -0.18 Width -0.10 -0.61 0.42 Discussion Colonization patterns on the stream surface were most strongly influenced by variation among sites with respect to stream size and gradient and not fine-sediment abundance. Several other studies within the Middle Fork of the Salmon River basin also found that macroinvertebrate assemblages exhibit pre- dictable changes with increasing stream size (Bruns et al. 1982, 1987, Bruns and Minshall 1985). In our study the available pool of colonists probably shifted within the study area and masked our ability to examine fine- sediment impacts. Within the 60-km study region on Bear Valley, the stream increases from a first- to a fourth-order stream and exhibits longitudinal changes in channel mor- phology and riparian characteristics along this gradient that influence macroinvertebrate assemblage composition (Bruns et al. 1982, 1987). Surface substrate characteristics played a secondary role to other stream features in influencing macroinvertebrate abundance. However, this does not mean that substrates do not influence macroinvertebrate distribu- tions on surface substrates. Other studies of sediment effects in the Bear Valley watershed found that biomass of macroinvertebrate drift from sand substrates was less than that from larger substrates (Konopacky 1984) and that macroinvertebrate densities were greater in riffles with low amounts of fines than riffles with higher proportions (Bjornn et al. 1977). Both studies were conducted within relatively small areas that did not encompass longitudi- nal variation in stream characteristics. Fine-sediment abundance did have distinct effects on macroinvertebrate colonization within the hyporheos. The greatest effect was with the smallest sediment size class (<1.50 mm). Sediment particles in this size range may have the most potential for clogging interstitial spaces within gravel. Although most sediment studies have not explicitly assessed impacts of sediments in this size range on macroinvertebrates, at least one study (Cederholm and Lestelle 1974) noted that particles <0.84 mm in diameter had strong negative correlations with total number of stream invertebrates. In addition, particles <1 mm in diameter are known to reduce availability of dissolved oxygen in stream grav- els (Tagart 1984), and clay-sized silt impairs periphyton production in riffles (Graham 1990). Our results suggest that macroinvertebrate habitat in Bear Valley Creek is impaired because of fine-sediment abundance in the hyporheos. Cobble and gravel bed streams without high proportions of sediment in the hyporheos typically exhibit much less differ- ence in macroinvertebrate composition and abundance between surface and hyporheic zones than we observed in this study. For example, Coleman and Hynes (1970) found little differentiation in macroinvertebrate numbers in the upper 30 cm of substrate. Williams and Hynes (1974) reported differ- ences among near-surface and below-surface macroinvertebrate assemblages; however, they found total numbers at 30-cm depth were typ- ically at least 50% of those near the surface. In Bear Valley, total numbers at 30 cm were only 22% of those near the surface. Bear Valley 1994] Macroinvertebrate Colonization of Substrates 111 < O O Surface 1.5 1 0.5 0 -0.5 -1 -1.5 O PED ELM BRA PTE HYP. LET HYC HEP PER CHL f ANN SIP NEM HYD TAB CEk • MOLX^^-EPH RHY -2 -1.5 Below Surface 0.8 0.6 0.4 0.2 0 -0.5 CCA 1 MOL 0 0.5 CN < 9. -0.2 TIP BRA LET CHIcHL ANN LIM HYP EPH . HEF -0.4 -0.6 ■0.8 -1 -2 -1.5 SIP PER 44YD- LEP BAE RHY HYC PTE ELM H ►- -0.5 0 CCA 1 0.5 1.5 Fig. 2. Ordination of macroinvertebrate taxa with respect to environmental variables in surface and hyporheic areas. results are more similar to those reported from streams with high proportions of fine sediment in the hyporheos, such as those examined by Poole and Stewart (1976) and Strommer and Smock (1989), who found that total numbers at approximately 30-cm depth were at least 80% less than those near the sur- face. Both studies attributed these differences to high proportions of fines in the hyporheos that altered physical habitat and subsurface water flow. High proportions of fine sediment within the hyporheos of Bear Valley Creek may sig- nificantly decrease available habitat for macroinvertebrates and therefore limit poten- tial secondary production in the stream. Our study suggests that the hyporheos should be included when assessing impacts of sediment additions to stream ecosystems. Since macroinvertebrate assemblages exhibit consis- tent long-term changes to watershed activities that influence substrate characteristics 112 Great Basin Naturalist [Volume 54 (Richards and Minshall 1992), dramatic and potentially persistent effects can be initiated through the introduction of fine sediments into the hyporheos. Acknowledgments Jack Gunderman and Phillip Cernera con- tributed significandy to all phases of this proj- ect. James Wadsworth assisted in the analysis of macroinvertebrate data. Chris Robinson and an anonymous reviewer made helpful suggestions on an earlier draft of the manu- script. Funding for this study was provided in part by Bonneville Power Administration through project No. 83-359. Literature Cited Alexander, G. R., and E. A. Hansen. 1986. Sand bed- load in a brook trout stream. North American Jour- nal of Fisheries Management 6; 9-23. Allan, J. D. 1975. The distributional ecology and diversi- ty of benthic insects in Cement Creek, Colorado. Ecology 56; 1040-1053. BjORNN, T. C, M. a. Brusven, M. p. Molnau, J. H. Mil- LiGAN, R. A. Klamt, E. Chacho, and C. Schaye. 1977. Transport of granitic sediment in streams and its effects on insects and fish. Forest, Wildlife and Range Station, University of Idaho. Technical Report Project B-036-IDA. Bruns, D. a., and G. W. Minshall. 1985. River continu- um relationships in an 8th-order river reach: analy- ses of polar ordination, functional groups, and organic matter parameters. Hydrobiologia 127; 277-285. Bruns, D. A., A. B. Hale, and G. W. Minshall. 1987. Ecological correlates of species richness in three guilds of lotic macroinvertebrates. Journal of Fresh- water Ecology 4: 163-177. Bruns, D. A., G. W. Minshall, J. T. Brock, C. E. Gush- ing, K. W. Cummins, and R. L. Vannote. 1982. Ordination of functional groups and organic matter parameters from the Middle Fork of the Salmon River, Idaho. Freshwater Invertebrate Biology 1; 2-12. Cederholm, C. J., AND L. C. Lestelle. 1974. Observa- tions on the effects of landslide siltation on salmon and trout resources of the Clearwater River, Jeffer- son County, Washington, 1972-73. University of Washington, Fisheries Research Report FRI-UW- 7404. Coleman, M. J., and H. B. N. Hynes. 1970. The vertical distribution of the invertebrate fauna in the bed of a stream. Limnology and Oceanography 15: 31^0. Graham, A. A. 1990. Siltation of stone-surface periphyton in rivers by clay-sized particles from low concentra- tions in suspension. Hydrobiologia 199: 107-115. KonopACKY, R. C. 1984. Sedimentation and productivity in salmonid streams. Unpubished doctoral disserta- tion. University of Idaho, Moscow. KoNOPACKY, R. C, P. J. Cernera, and E. C. Bowles. 1986. Salmon River habitat enhancement. Bon- neville Power Administration, DOE/BP-14383-2, Portland, Oregon. Mackay, R. J. 1992. Colonization by lotic macroinverte- brates: a review of processes and patterns. Journal of the North American Benthological Society 49: 617-628. Minshall, G. W. 1984. Aquatic insect-substratum rela- tionships. Pages 358-400 in V. H. Resh and D. M. Rosenberg, eds.. The ecology of aquatic insects. Praeger Scientific, New York. NuTTALL, P. M. 1972. The effects of sand deposition upon the macroinvertebrate fauna of the River Camel, Comwall. Freshwater Biology 2; 181-186. Palmer, M. A., A. E. Bely, and K. E. Berg. 1992. Response of invertebrates to lotic disturbance: a test of the hyporheic refuge hypothesis. Oecologia 89; 182-194. Pennak, R. W., and E. D. Van Gerpen. 1947. Bottom fauna production and physical nature of the sub- strate in a northern Colorado trout stream. Ecology 28: 42-48. PooLE, W. C. AND K. W. Stewart. 1976. The vertical distribution of macrobenthos within the substratum of the Brazos River, Texas. Hydrobiologia 50; 151-160. Richards, C, and G. W. Minshall. 1992. Spatial and temporal trends in stream macroinvertebrate species assemblages: the influence of watershed dis- turbance. Hydrobiologia 241; 173-184. Richards, C, G. E. Host, and J. W. Arthur. 1993. Identification of predominant environmental factors structuring stream macroinvertebrate communities within a large agricultural catchment. Freshwater Biology 29; 285-294. Robinson, C. T., G. W. Minsh,\ll, and S. R. Rushforth. 1990. Seasonal colonization dynamics of macroin- vertebrates in an Idaho stream. Journal of the North American Benthological Society 9; 240-248. Strommer, J. L., AND L. A. Smock. 1989. Vertical distri- bution and abundance of invertebrates within the sandy substrate of a low-gradient headwater stream. Freshwater Biology 22: 263-274. Tagart, J. V. 1984. Coho salmon survival from egg depo- sition to emergence. In: J. M. Walton and D. B. Houston, eds., Proceedings of the Olympic wild fish conference. Port Angeles, Washington. TER Braak, C. J. F. 1986. Canonical Correspondence Analysis: a new eigenvector technique for multivari- ate direct gradient analysis. Ecolog>' 67: 1167-1179. . 1987. The analysis of vegetation-environment relationship by Canonical Correspondence Analysis. Vegetatio 69; 79-87. TER Braak, C. J., and I. C. Prentice. 1988. A theory of gradient analysis. Advances in Ecological Research 8:271-317. Ward, J. V. 1975. Bottom Huma-substrate relationships in a northern Colorado trout stream: 1945 and 1974. Ecology 56; 1429-1434. . 1989. The four-dimensional nature of lotic ecosys- tems. Journal of the Benthological Society of Ameri- ca 8; 2-8. Wesche, T. a., D. W. Reiser, V. R. Hasfuther, W. A. Hubert, and Q. D. Skinner. 1989. New technique 1994] Macroinvertebrate Colonization of Substrates 113 for measuring fine sediment in streams. North American Journal of Fisheries Management 9; 234-238. Williams, D. D. 1984. The hyporheic zone as a habitat for aquatic insects and associated arthropods. Pages 430-455 in V. H. Resh and D. M. Rosenberg, eds., The ecology of aquatic insects. Praeger Scientific, New York. Williams, D. D., and H. B. N. Hynes. 1974. The occur- rence of benthos deep in the substratum of a stream. Freshwater Biology 4: 233-256. Received 27 January 1992 Accepted 21 September 1993 Great Basin Naturalist 54(2), ©1994. pp. 114-121 RESOURCE OVERLAP BETWEEN MOUNTAIN GOATS AND BIGHORN SHEEP John W. Laundre^ Abstract. — Mountain goat {Oreamnos americanus) and bighorn sheep [Ovis canadensis) ranges overlap substantial- ly in northwestern United States and southwestern Canada. Resource overlap in food and habitat parameters is assumed, but the degree of overlap has not been estimated. Data from published separate and comparative studies on food and habitat use were used to calculate indices of resource overlap for goats and sheep. Indices of overlap for gen- eral forage classes (grasses, forbs, browse) were >0.90 in summer and winter for data based on pooled data from sepa- rate studies and in summer for data from comparative studies. In winter for comparative studies this overlap was 0.64. For studies where forage species were identified, estimates of resource overlap fi-om separate studies were —0.8 but were <0.5 for comparative studies. Indices of overlap for habitat variables were also low (<0.7) for comparative stud- ies. It was concluded that possible overlap in food and habitat use by goats and sheep could be extensive; but in sym- patric populations, resource overlap may be reduced substantially. tionins. Key words: bighorn sheep, mountain goats, Oreamnos americanus, Ovis canadensis, resource overlap, resource parti- In the northwestern United States, Rocky Mountain goats {Oreamnos americanus) his- torically ranged south along the Bitterroot Divide to near the Continental Divide between Idaho and Montana and in the Cas- cade Mountains south to central Washington (Hall 1981). With the gradual extension of European settlement, goats were extirpated from numei^ous areas. Beginning in the early 1900s, goats were transplanted into their his- toric range as well as other suitable habitats, e.g., Colorado, northwest Washington. Many of these areas where goats were not recently found were historic ranges of bighorn sheep {Ovis canadensis). The introduction of moun- tain goats into the range of bighorn sheep raised concerns regarding potential impacts of goats on bighorns. Because goats and sheep are generalist herbivores that use subalpine to alpine environments, it is commonly assumed their food and habitat requirements overlap extensively in these areas. Based on this assumption, some researchers have expressed concern that goats might compete with sheep when introduced into existing sheep range. Although the outcome of this competition is uncertain, goats are thought to be the superior competitors (Whitfield 1983). Concern about potential competition has caused a reevalua- tion of introducing goats into areas beyond their historic range, especially in areas con- taining bighorn sheep. However, resource overlap, which must not be confused with competition, does not justify sweeping gener- alizations about competitive interactions. Additionally, the assumption of extensive resource overlap is based primarily on food and habitat use by goats and sheep from stud- ies separated by space and time. It is unclear whether data from such diverse studies can be used to infer resource overlap of these two species in sympatry. Consequently, the impli- cation of competition between goats and sheep is tenuous and should not be used to influence reintroduction decisions without a review and reassessment of resource overlap between these two species. There are many studies on food habits and habitat use of goats and sheep. However, most are unpublished theses. There has yet to be a comprehensive review of existing literature, nor have any estimates of resource overlap been reported. My objectives were to (1) review and summarize available literature on food and habitat utilization of Rocky Moun- tain goats and bighorn sheep and (2) reevalu- ate and quantify, if possible, the amount of resource overlap that exists between these 'Department of Biological Sciences, Idaho State University, Pocatello, Idaho 83209. 114 1994] Overlap in Goats and Sheep 115 two species. Only with such a review and reevaluation can we proceed to set up rigid experimental designs to address questions concerning potential competition between goats and sheep. Methods I compiled data from 34 separate studies on food habits and habitat use of sheep and goats and from 3 comparative studies of sym- patric populations. Some separate studies were of known allopatric populations because the study areas were outside the range of the corresponding sheep or goat species. Most studies were in areas where both species occurred, but the reports did not indicate whether the corresponding species was in the study area. Because I could not assume allopatry in all studies, this term is not used in reference to these separate studies. Methods of data collection varied among studies. Most authors expressed food habits as percentage of obsei'ved use or occurrence in stomach or feces but did not adjust their esti- mates for forage availability (Rominger and Bailey 1982). Some researchers classified for- age only by classes (grass, forbs, or shrubs); others presented lists of forage species. Most researchers quantified diets separately for either summer or winter or both. A few researchers presented diets for spring and fall, but there were insufficient data to include these seasons in this analysis. I expressed food habits data as percentage of use by category. 1 used these data to calcu- late resource overlap indices (O) based on equation 1 (Lawlor 1980): where: Pij and p]^j are the proportions of resource type j used by species i and k. The index O ranges from 0 (no overlap) to 1.0 (total overlap). Indices of overlap were calcu- lated for all combinations of food categories (forage classes, species), season (winter, sum- mer), and study status (separate, comparative) except specific winter diets from comparative studies for which no data were available. Pooled data from separate studies were used to calculate one set of resource overlap indices; data from comparative studies were used to determine additional indices for each study. Calculations of resource overlap for for- age species from separate studies were limit- ed to studies from subalpine and alpine areas in the northwestern U.S. where sheep and goats co-occur. Food habits reflect availability of food species (Rominger and Bailey 1982), which in more northern areas for goats and more southern areas or lower elevations for sheep can differ greatly, biasing any compar- isons that might be made. Restricting this review to data from subalpine and alpine areas in the northwest region should limit dif- ferences in resource availability to an accept- able level. Several investigators reported percentage of habitat utilized by goats and sheep. Com- paring data from these studies was difficult because habitat classifications were not stan- dardized. In these studies general patterns of habitat use were summarized. For the few comparative studies of sympatric populations, habitat overlap indices were calculated with equation 1. In separate studies some authors measured physical characteristics of the envi- ronment selected by animals, specifically, dis- tance to escape cover, elevation, and slope. Data were not expressed in percent use of dif- ferent categories but were means of observa- tions. These data were compared with t test or analysis of variance designs as appropriate. The null hypothesis was no difference in means for goats and sheep for the tested char- acteristic. Acceptance of the null hypothesis would indicate total resource overlap. Rejec- tion (P < .05) of the null indicates significant statistical separation along the tested resource Results Food Habits Several investigators presented only quali- tative assessments of goat and sheep diets (Davis 1938, Honess and Frost 1942, Spencer 1943, Casebeer 1948, Couey 1950, Smith 1954, McCann 1956, Berwick 1968, Cooper- rider 1969). Diets were similar in summer and winter, with both species relying on grasses and forbs. Where authors estimated diet com- positions, data indicated that goats relied on grasses (52%) and forbs (30%) in summer but shifted to grasses (60%) and shrubs (32%) in winter (Table 1). Sheep (Table 1) used mainly 116 Great Basin Naturalist [Volume 54 Table 1. Summary of general forage classes used by goats and sheep from various studies in alpine and subalpine habitats. Estimates are expressed as percent of total use and are based on either fecal or rumen analyses. Locations of studies are indicated bv standard U.S. Postal Ser\'ice codes. Summer Winter Species Grass Forb Browse Grass Forb Browse Reference Mountain goats Saunders 1955 MT 76 14 2 59 10 30 Hibbs 1967 CO 97 3 0 88 0 12 Peck 1972 MT 22 78 0 90 6 1 Pallister 1974 MT 40 60 0 Johnson et al. 1978 CO 60 29 7 Thompson 1981 CO 84 15 1 Thompson 1981 MT 11 9 79 47 2 51 Stewart 1975 MT 47 53 0 Johnson 1983 WA 44 20 35 31 3 65 Campbell & Johnson 198.3 WA 43 20 36 Adams & Bailey 1983 CO 45 24 30 X 52 30 16 60 8 32 Bighorn sheep Mills 1937 \VY 60 35 5 98 0 0 Moser 1962 CO 75 6 19 Pallister 1974 MT 12 55 32 98 2 0 Frisina 1974 MT 95 4 1 92 6 1 Stewart 1975 MT 44 47 8 40 40 20 Todd 1975 CO 65 6 29 23 11 67 Johnson & Smith 1980 NM 46 50 4 83 10 7 Whitfield 1983 WY 25 12 63 30 32 39 Martin 1985 MT 74 16 10 39 50 10 Estes 1979 WA 30 8 62 62 3 35 Honess & Frost 1942 WY 51 30 19 Harrington 1978 CO 88 12 0 Kasworm et al. 1984 MT 65 12 23 Blood 1967 BC 54 5 40 Constan 1972 MT 72 17 8 Keating et al. 1985 WY 56 7 38 Schallenberger 1966 MT 87 9 2 Oldemeyer et al. 1971 \NY 61 17 22 X 56 23 21 64 15 21 grasses (56%) in summer but used forbs (23%) and shrubs (21%) more equally. Sheep also exhibited a seasonal change to grasses (65%) in winter. Nine studies of goats and 11 studies of sheep contained analyses of forage species used. Although many plant species were used, most were consumed at very low levels (<1% of diet; Laundre 1990). Data were summa- rized for only those 12 genera that occurred > 1% within the diet of at least sheep or goats (Table 2). The main genera used by goats and sheep in the summer were sedges {Carex sp.), wheatgrass {Agropijron sp.), bluegrass {Poa sp.), fescue {Festuca sp.), and bluebells {Mertensia sp.). Winter diets consisted mainly of sedges, wheatgrass, sagebrush {Artemisia sp.), and fescue. Three studies (Pallister 1974, Stewart 1975, Dailey et al. 1984) presented data of food habits from sympatric populations of nonna- tive goats and native sheep. Pallister (1974) and Stewart (1975) primarily studied sheep but also recorded food habits of naturalized mountain goats in their study areas. The goats were descendants of releases made in the 1940s. Pallister (1974) found that summer diets of mountain goats consisted of 40% grasses and 60% forbs. During the same time sheep consumed 12% grasses, 55% forbs, and 32% shrubs. Although both species relied on forbs to a similar level, comparisons of forb species eaten indicated little overlap except clover {Trifolium parriji) (Pallister 1974: 48). Stewart (1975) found a similar reliance on grasses by sheep (44%) and goats (47%), but 1994] Overlap in Goats and Sheep 117 Table 2. Comparison of percent use of preferred plant genera for goats and sheep. The percents are averages of the values reported in the literature/' Sample size (u) is the number of reported values used to calculate the means. Species Agropijron sp. Carex sp. Deschampsia sp. Festuca sp. Foa sp. Koeleria sp. Stipa sp. Artemisia sp. Mertensia sp. Potentilla sp. Salix sp. Trifolium sp. Summer Winter Sheep Goats Sheep Goats {n = 7) in = 7) in = 10) (n = 5) 6 9 15 4 15 10 15 8 2 <1 <1 <1 8 5 12 18 15 14 1 4 <1 5 3 1 <1 <1 2 <1 <1 <1 10 3 2 6 0 <1 4 1 <1 <1 5 4 1 1 5 2 0 <1 a.Mills 1937, Hilibs 1967, Oldemeyer et al. 1971, Constan 1972, Peck 1972, Pallister 1974. Frisina 1974, Stewart 1975, Johnson et al. 1978, Johnson 1983, John- son and Smith 1980, M.J. Thompson 1981, Adams and Bailey 1983, Campbell and Johnson 1983, Whitfield 1983, Kaswomi et al. 1984, Keating et al. 198.5, Mai an 1985 goats relied most on Poa sp. while sheep were more evenly divided among three species: Agropijron, Carex, and Poa (Stewart 1975: 68, 98). Overall forb use by sheep and goats was also similar, 47% for sheep, 53% for goats, but specific use of forbs differed. Sheep relied on a variety of forb species while goat diets con- sisted mostly o( Arnica latifolia and Erigeron sp. Dailey et al. (1984) conducted parallel feeding trials with captive goats and sheep on unoccupied range in Colorado. Their work indicated goats ate more forbs in summer (goats 88%, sheep 70%) and winter (goats 59%, sheep 22%), while sheep consumed more grasses (summer, 30% vs. 11%; winter, 75% vs. 27%). For summer diets expressed in forage classes, overlap indices were high for separate (0.98) and comparative studies (0.93) (Fig. la). Resource overlap in winter, based on data from separate studies, was also high (0.99). For the comparative study from Colorado (Dailey et al. 1984), the winter overlap index was 0.64 (Fig la). For the pooled separate studies where forage species were identified, summer (0.86) and winter (0.80) indices were slightly lower than those for general forage classes (Fig. la). Summer overlap indices (Fig. la) for the two comparative studies in Mon- tana (Fallister 1974, Stewart 1975), however, were substantially lower (0.32 and 0.55) than those for general forage classes (Fig. la). There were insufficient data to determine whether indices from the comparative studies differed statistically from the general diet index. General Habitat Use Oldemeyer et al. (1971) divided habitat used by sheep in Yellowstone National Park into three general types: forest, grass, and shrub. In winter they found that sheep used forest 13%, grass 78%, and shrub 9% of the time. When they divided the area based on terrain, they found sheep used "steep" areas 39%, rocky outcrops 14%, ridgetops 36%, hilly areas 8%, and level areas 4% of the time. Of the numerous structural/vegetational forma- tions defined by Martin (1985) in Montana, sheep spent most of their summertime in the "alpine turf" formation (approximately 50%) and the "sparsely vegetated dirt scree" forma- tion (approximately 28%). In spring, Frisina (1974) found sheep 36% of the time in "rocky reef" and 59% in "bunchgrass" types. Sheep use of the rocky reef type in fall increased to 64% and decreased to 34% in the bunchgrass type. Tilton and Willard (1982) divided their Montana study area into rockland, shrub/ grass, open forest, and closed forest habitat types. They found sheep spending 14% of their time in the rockland type, 46% in the shrub/grass, 40% in the open forest, and 1% in the closed forest types. Peck (1972) divided goat habitat in Mon- tana into four types: timber, sliderock, ledge, and ridge. He found goats spending 4% of their summertime in timber, 36% in sliderock, 54% in ledge, and 6% in ridge areas. In win- ter, goats were seen 16% of the time in timber, 70% in ledge areas, and 14% of the time on ridges. M. J. Thompson (1981) found goats in 118 Great Basin Naturalist [Volume 54 1.5 1.0 O 0.5 0.0 Summer Diets ^^Specific Diet R^ General Diet Winter Diets Separate Separate Sympatric Sympatric 1.0 0.5 0.0 Y77A Pallister (1974) ^ Stewart (1975) Winter Fig. 1. Niclie overlap indices for food habits (a) and habitat selection (b). Overlap indices for food are for pooled data from separate studies and data from compara- tive studies and are based on either general (grass, forbs, shrubs) or specific (to genera) food classifications. The index for sympatric specific summer diets is the mean of indices calculated from Pallister (1974) and Stewart (1974). The inde.x for sympatric general summer diets is the mean from Dailey et al. (1984), Pallister (1974), and Stewart (1974). Indices for habitat selection are all fiom two comparative studies of sympatric sheep and goat pop- ulations. Montana spending 90% of their summertime and 68% of their early wintertime on glacial cirques. In winter in the Bitterroot Moun- tains, Smith (1976) found goats 62% of the time in the "bunchgrass" association. Goats in Colorado spent 85% of their time in the sub- strate type described as "intermittent boulder" by R. W. Thompson (1981). Adams and Bailey (1980) classified the habitat into alpine and subalpine areas. Within the alpine community they identified tundra and rock subcompo- nents. Goats spent 58% of their time during winter in the tundra and 42% in rock areas. The subalpine community was subdivided into rock, shrubs, and trees. Goats were seen 35% of the time in rock areas, 10% in shrub, and 55% in tree areas. Von Elsner-Schack (1986) studied goats in Alberta and divided the study area into rock, gravel, and grass sub- strate types. In spring-summer, goats used the rock substrate 24%, the gravel substrate 26%, and the grass areas 50% of the time. Few studies examined habitat use by sheep and goats simultaneously. Chadwick (1974) found some habitat segregation but did not quantify the differences. Geist (1971) found that goats in winter spent approximately 52% of their time in sheer cliff areas while sheep spent only 28% of their time in these areas. In the two Montana studies of bighorn sheep, Stewart (1975: 68, 96) and Pallister (1974: 28, 56) also recorded habitat use by goats in their study areas. Habitat use overlap indices based on Pallister s (1974) data were low for summer (0.31) and winter (0.50). The summer overlap index (0.33) from Stewart's (1975) data was similar to Pallister's value, but the winter index (0.68) was slightly higher (Fig. lb). Another area of potential overlap between goats and sheep is the physical characteristics of the environment. Several investigators sep- arately quantified habitat use by goats and sheep relative to distance from escape terrain, slope, and elevation (Table 3). The average distance to escape cover in summer was sig- nificantly greater for goats (f = 6.04, n = 9, F < .01). Average slope used did not differ with- in species between winter and summer but did differ significantly between species in both seasons (F = 15.2, n^ = 7, no = 6, F < .01), with goats using steeper areas (Table 3). No difference in use was found between species or seasons for average elevation. Thus, goats preferred steeper slopes and were found fur- ther from escape terrain than were sheep. Discussion Wildlife biologists have been implicitly using data compiled separately on resource use of goats and sheep to formulate views 1994] Overlap in Goats and Sheep 119 Table 3. Means (± SE, n) of physical habitat characteristics by sheep and goats. Distance to escape habitat (DEH) values are the miLxiniinii distances at which > 80% of the animals were found. Values for slope and elevation are the means of data reported in the literature. An asterisk next to a measurement indicates significant {P < .05) differences between sheep and goats. Footnotes list references of original data. Summe ■r Winter Sheep Goats Sheep Goats DEH' Slopeb Elevation'' 120 ± 11.6 m, 4 22 ±2.6°, 4 2655 ± 325.3 m, 4 305 ± 25.6 m, 5* 41 ±5.2°, 4* 2799 ± 320.8 m, 4 278 ± 103 m, 4 24 ±6.5°, 2 2431 ± 306.5 m, 4 47 ±8.4°, 3* 2354 ± 376.7 m, 3 ^Hjeljord 1971. Oldemeyer et al. 1971, Frisina 1974, Pallister 1974, McFetridge 1977, R. W. Thompson 1981, Tilton and Willard 1982, Fo.x 198.3, Whitfield 198.3, Martin 198.5, Smith 1986 bRuck 1973, Frisina 1974, Pallister 1974, Chadwick 1977, Smith 1976, R. W. Thompson 1981, Whitfield 1983, Martin 198.5, Hayden 1989 ^Frisina 1974, Pallister 1974, Smith 1976, Adams and Bailey 1980, M. J. Thompson 1981, R. W. Thompson 1981, Whitfield 1983, Martin 1985, Hayden 1989 concerning competition between the two species. Differences found in this review between results from separate and compara- tive studies indicate a danger in using data from separate studies. Food habits data from separate studies, based on general forage classes and forage species, indicated extensive overlap in goat and sheep diets. Data on habi- tat use from such studies also indicated goats and sheep mutually used "grass" and "tree" habitat types and similar elevations in the subalpine/alpine zones. These data strengthen the commonly held consensus of extensive resource overlap and support concerns that goats and sheep might not coexist if resources are limiting. In contrast, data from compara- tive studies, where specific diet composition and habitat use are considered, indicate sub- stantial reductions in overlap when goats and sheep CO -occur in an area. Consequently, comparisons of data from separate studies might be useful in determin- ing the amount of resource overlap that is possible between similar species but cannot be used to estimate what that overlap would be in sympatr)'. Only results from comparative studies of sympatric populations can be used to predict how two species will interact. Even in such comparative studies, my analysis indi- cates that researchers should avoid the use of general resource categories. Currently, we have only two comparative studies of detailed resource use by goats and sheep. This is hardly a sufficient data base from which to draw valid conclusions con- cerning resource overlap or the potential for competition between goats and sheep. If sci- entifically sound conclusions about interac- tions between goats and sheep are to be for- mulated, additional comparative studies are needed. Only after such studies can we address questions concerning competition and competitive interactions between sheep and goats. If the pattern of reduced resource overlap in sympatiy withstands further study, it may be the result of resource partitioning. Whether this is the case and whether this resource partitioning is in turn a result of competitive interactions cannot be addressed with this data base. If resource partitioning is found to be a major factor in the coexistence of sympatric native goat and sheep popula- tions, the low resource overlap found in the two comparative studies involving nonnative goats indicates goats and sheep may also exhibit such partitioning when one or the other species is an exotic introduction. How- ever, Adams et al. (1982) cautioned that cer- tain conditions (land development, agricultur- al activity, etc.) might limit selection options for one or the other species. In such cases resource partitioning may not be possible, resulting in extensive overlap of resource use between goats and sheep, possibly to the detriment of one of the species if resources are limiting. Acknowledgments My work was supported by the University of Wyoming-National Park Service Research Center (NFS #FX 1200-8-0828). I thank per- sonnel at the center and Yellowstone and Grand Teton national parks for their help and support with the project, including J. Varley, and S. Consolo-Murphy. I also thank F. Dubowy, D. Reed, and T. Reynolds for reviewing drafts of this manuscript. 120 Great Basin Naturalist [Volume 54 Literature Cited Adams, L. C, and J. A. Baili:y. 1980. Winter habitat selection and group size of mountain goats, Sheep Mountain-Gladstone Ridge, Colorado. Biennial Symposium of the Northern Wild Sheep and Goat Council 2: 465-479. . 19' of the Rock Creek bighorn sheep herd, Beartooth Mountains, Montana. Unpublished thesis, Montana State University, Bozeman. 152 pp. McCann, L. J. 1956. Ecology of the mountain sheep. American Midland Naturalist 56: 297-324. McFetridge, R. J. 1977. Strategy of resource use by mountain goat nursery groups. Pages 169-173 in W. Samuel and W. G. Macgregor, eds.. Proceedings of the first international mountain goat symposium, Kalispell, Montana. British Columbia Ministry of Recreation and Gonsenation. Mills, H. B. 1937. A preliminaiy study of the bighorn of Yellowstone National Park. Journal of Mammalogy 18: 205-212. 1994] Overlap in Goats and Sheep 121 MosER, C. a. 1962. The bighorn sheep of Colorado. Col- orado Game and Fish Department Technical Publi- cation 10. Oldemeyer, J. L., W. J. Barmore, and D. L. Gilbert. 1971. Winter ecology of bighorn sheep in Yellow- stone National Park. Journal of Wildlife Manage- ment 35; 257-269. Pallister, G. L. 1974. The seasonal distribution and range use of bighorn sheep in the Beartooth Moun- tains, with special reference to the West Rosebud and Stillwater herds. Unpublished thesis, Montana State University, Bozeman. 65 pp. Peck, S. V. 1972. The ecology of the Rocky Mountain goat in the Spanish Peaks area of southwestern Montana. Unpublished thesis, Montana State Uni- versity, Bozeman. 54 pp. ROMINGER, E. M., AND J. A. Bailey. 1982. Forage prefer- ence indices: adjusting forage availability data for habitat selection. Biennial Symposium of the North- em Wild Sheep and Goat Council 3; 278-286. Saunders, J. K., Jr. 1955. Food habits and range use of the Rocky Mountain goat in the Crazy Mountains, Montana. Journal of Wildlife Management 19: 429-437. Schallenberger, a. D. 1966. Food habits, range use and interspecific relationships of bighorn sheep in the Sun River area, west-central Montana. Unpub- lished thesis, Montana State University, Bozeman. 44 pp. S.viiTH, B. L. 1976. Ecolog\' of RockT Mountain goats in the Bitterroot Mountains, Montana. Unpublished thesis. University of Montana, Missoula. 203 pp. Smith, C. A. 1986. Habitat use by mountain goats in southeastern Alaska. Alaska Department of Fish and Game. Federal Aid in Wildlife Restoration Final Report, Project W-22-1, W-22-2, W-22-3, Job 12.4R. Smith, D. R. 1954. The bighorn sheep in Idaho, its sta- tus, life history and management. Idaho Fish and Game Department Wildlife Bulletin 1. Spencer, C. C. 1943. Notes of the life history of Rocky Mountain bighorn sheep in the Tarryall Mountains of Colorado. Journal of Mammalogy 24: 1-11. Stewart, S. T. 1975. Ecology of the West Rosebud and Stillwater bighorn sheep herds, Beartooth Moun- tains, Montana. Unpublished thesis, Montana State University, Bozeman. 129 pp. Thompson, M.J. 1981. Mountain goat distribution, popu- lation characteristics and habitat use in the Saw- tooth Range, Montana. Unpublished thesis, Mon- tana State University', Bozeman. 80 pp. Thompson, R. W. 1981. Ecology of Rocky Mountain goats introduced to the Eagles Nest Wilderness, Colorado, and some factors of geographic variation in the lambing season of bighorn sheep. Unpub- lished thesis. University of Wyoming, Laramie. 359 pp. Tilton, M. E., and E. E. Willard. 1982. Winter habitat selection by mountain sheep. Journal of Wildlife Management 46: 359-366. Todd, J. W. 1975. Foods of rocky mountain bighorn sheep in southern Colorado. Journal of Wildlife Management 39: 108-111. VON Elsner-Sch,\ck, I. 1986. Habitat use by mountain goats, Oreainnos americanus, on the eastern slopes region of the Rocky Mountains at Mount Hamell, Alberta. Canadian Field-Naturalist 100: 319-324. Whitfield, M. B. 1983. Bighorn sheep histor\', distribu- tions, and habitat relationships in the Teton Moun- tain Range, Wyoming. Unpublished thesis, Idaho State University, Pocatello. 244 pp. Received 12 October 1992 Accepted 2 September 1993 Great Basin Naturalist 54(2), ©1994, pp. 122-129 PREDATION OF ARTIFICIAL SAGE GROUSE NESTS IN TREATED AND UNTREATED SAGEBRUSH Mark E. Ritchiel, Michael L. Wolfel, and Rick Danvir2 Abstiuct. — We measured predation on ]2() artificial Sage Grouse {Centrarciis iir()i)hasiaints} nests in montane sagebrush grassland in northern Utah. We e.xamined nests in areas that had been chained and seeded 25 years previ- ously (treated areas) and in areas that were untreated. Predation rates of artificial nests were higher in areas of untreat- ed sagebrush, even though these areas had greater sagebrush cover, taller shrubs, and greater horizontal plant cover. These results differ from those previously hyi^othesized for treated sagebrush habitat and may reflect a greater abun- dance of other potential prey species, especially lagomorphs, in untreated areas that attracted greater densities of predators. In addition, over 80% of nests were depredated by mammals, which hunt using olfaction and are less likely than avian predators to be affected by nest cover. We conclude that, after treated sagebrush has recovered to some degree, predation rates of Sage Grouse nests may be lower in treated sagebrush. Consecjuently, factors other than nest predation (e.g., winter food, thermal cover, insects, perennial forb abundance) may be more important reasons for pre- serving mature sagebrush stands for Sage Grouse. Keij words: Sage Grouse, Gentrarcus urophasianus, sagebrush, nest, predation, habitat. A key problem in the conservation of wildlife species is fragmentation of large con- tiguous areas of preferred habitat (Lovejoy et al. 1984, Wilcove 1985, Yahner and Scott 1988), a problem that has plagued the man- agement of upland game bird populations in western North America (Vale 1974, Braun et al. 1977). In particular, Sage Grouse popula- tions have declined in some areas, apparently in response to widespread treatment (chain- ing, spraying, burning, etc.) of sagebrush- dominated rangeland to benefit livestock pro- duction (Schneegas 1967, Klebenow 1970, Braun et al. 1977). However, few studies have examined whether such treated areas can re- cover to become suitable Sage Grouse habitat. Sagebrush treatment may reduce Sage Grouse populations by eliminating mature shrubs, which may be important in protecting nests from visual predators (Dalke et al. 1963, Braun et al. 1977, Autenrieth 1981, Gonnelly et al. 1991). In addition, treated areas planted to grass cover (e.g., crested wheatgrass, Agropi/ron desertorum) often recover shrubs slowly (Vale 1974, MacMahon 1987). Sage- brush treatment may therefore permanently reduce nesting cover. For ground-nesting birds in general, dense shrub cover may not always be beneficial; it may increase nest predation by supporting greater populations of alternate prey and attracting greater densities or attention of pred- ators (Groze 1970, Duebbert and Kantrud 1974, Taylor 1977, 1984). Alternate prey, how- ever, may sometimes decrease nest predation by diverting predator effort during nest incu- bation (Byers 1974, Weller 1979, Crabtree and Wolfe 1988). For areas recovering from sage- brush treatment that have relatively low shrub cover, it is not clear whether Sage Grouse nest predation is greater than in untreated areas with greater cover. In this study we tested the hypothesis that artificial Sage Grouse suffer higher predation rates in treated than in untreated sagebrush. We also measured vegetation characteristics associated with nest sites to determine which habitat components might contribute to nest predation. Finally, we measured indices of lagomorph, small mammal, and predator abundance within treated and untreated areas to establish whether higher nest predation rates were associated with a higher density of alternate prey and/or predators. Study Area The study was conducted on the property of Deseret Land and Livestock, an 80,000-ha ranch located in northwestern Utah along the •Department of Fisheries and Wildlife. Utah State University, Lof^an, Utah 84322-5210. ^Deseret Land and Livestock, Box 250, Woodruff', Utah 84086. 122 1994] Sage Grouse Nest Predation 123 Wyoming border (Rich County). We conduct- ed the study on mid-elevation (2000 m) benches dominated by Wyoming big sage- brush {Artemisia tridentata wyomingensis), rabbitbrush [Chrysothamnas viscidiflorus), and several herbaceous species, mainly west- ern wheatgrass {Pascopyrum smithii), needle- and-thread {Stipa comata), Indian ricegrass {Oryzopsis hymenoides), bluegrass {Poa sand- bergi), and Phlox spp. Many separate 1000-5000-ha pastures, totalling nearly 40% of the 32,000 ha of mid- elevation sagebrush grassland on the ranch, were treated by discing or spraying between 1960 and 1965, resulting in a partial loss of sagebrush. These treated areas were seeded with crested wheatgrass {Agropyron deserto- rum) to improve forage for livestock. Thus, two distinct habitats exist on the study area: untreated areas with 5-20% herbaceous cover and 10-40% shrub cover (mostly sagebrush), and treated areas with 5-40% herbaceous cover (mostly crested wheatgrass) and 0-20% shrub cover (mostly sagebrush and rabbit- brush). Treated areas typically had recovered some shrub cover, but shrubs were shorter and less dense than in untreated areas. Alternate prey for potential Sage Grouse nest predators included lagomorphs (white- tailed jackrabbits [Lepus townsendi], moun- tain cottontails [Sylvilagus nuttalli], and pygmy rabbits [Brachylagus idahoensis]) and small mammals {Peromijscus maniculatus and Perognathus parvus). The primary mammalian nest predators were coyotes {Canis latrans), badgers {Taxidea taxiis), and chipmunks {Eutainias minimus). Principal avian nest predators were Common Ravens {Corvus corax). Black-billed Magpies [Pica pica), and California Gulls {Larus californicus) (all Sage Grouse leks [breeding grounds] in the study area were within 10 km of a large gull colony on Neponset Reservoir). Methods Artificial Nests Predation rates of artificial nests were mea- sured in each habitat type during the Sage Grouse nesting season of 1991. We set up arti- ficial nests at 160-m intervals along three 1.6- km transects radiating at random compass bearings and commencing 0.8 km from each of four Sage Grouse leks (two in each habitat type). These locations represented the area most likely used for nesting by females attending each lek (Wallestad and Pyrah 1974, Beck 1977). Thus, we used a total of 120 nests, with 10 nests per transect, 3 transects per lek, and 2 leks per habitat type. To achieve some level of replication, we selected 2 leks in each habitat type so that sampling areas delineated by a 2.2-km radius surrounding each lek did not overlap and included different groups of pastures. Leks in treated areas were located in pasture complexes treated in different years and separated by untreated sagebrush. We drove along each transect in a vehicle to avoid leaving a scent trail. At each 160-m interval, we placed artificial nests under the closest shrub (>10 cm height) to a point at a random distance (10-30 m) along a line per- pendicular (randomly left or right) to the main transect. These precautions were taken to reduce the chance that avian predators could "cue" on artifical nests by following tire marks along the main transect (Galbraith 1987, Maclvor et al. 1990) and the chance that mammalian predators could detect nests by following human scent. However, either type of predator could have followed tracks left by the vehicle. Each "nest" consisted of three unmarked brown chicken eggs. Nests were placed in the field between 30 April and 3 May 1991 during the Sage Grouse nesting period at Deseret Ranch. Nests were checked 15 days later and were considered depi'edated if all eggs were destroyed or missing, or partially depredated if one or two eggs remained. We attempted to identify the nest predator as either mam- malian or avian, based on characteristics of egg remains (Rearden 1951, Patterson 1952). We could identify likely predators at 43 of 57 depredated nests. Habitat Characteristics We measured vegetation characteristics, alternate prey abundance, and badger abun- dance at or near artificial nest transects to evaluate potential differences among habitat types. We measured vegetation characteristics when artificial nests were checked for preda- tion. Specifically, we estimated percent cover of shrubs and herbaceous plants as well as height of the tallest shrub in four Daubenmire (1968) plots at each nest site. These plots were spaced 5 m apart along a 20-m transect 124 Great Basin Naturalist [Volume 54 extending from the nest site and parallel to the main artificial nest transect. We measured horizontal cover by counting the number of 5 X 5-cm scjuares on a 45 X 45-cm board that were obscured by the nest bush to a viewer at 10 m distance and 40 cm height (Jones 1968, Klott and Lindzey 1990). Abundance of alternate prey for potential predators was estimated in July 1991 in both sagebrush habitat types within 1 km of the artificial nest transects. We estimated lago- moiph abundance by counting the number of lagomoq^h fecal pellets in ten 2 X 2-m plots located every 15 m along 150-m transects. We counted fecal pellets along four randomly located transects in each habitat type. We estimated abundance of small mammals by establishing two replicate 200 X 200-m grids of 25 Sherman® live traps placed 50 m apart in each habitat type. Traps were baited with rolled oats and peanut butter and checked for 3 nights (11-13 July). We estimated abundance of badgers, a principal mammalian nest predator, by count- ing the number of active badger holes seen along 2.5-km transects in mid-July 1991. Nine transects were randomly located within 1 km of artificial nest sites in each habitat type. Active badger holes were identified by fresh digging, a large oval hole, and presence of scat and/or tracks. Statistical Tests Proportions of nests depredated were com- pared with chi-square tests for treated vs. untreated areas and with Fisher's Exact Test for mammalian vs. avian predators. We com- pared the mean proportion of depredated nests and vegetation characteristics in treated vs. untreated areas with a nested ANOVA (Dowdy and Wearden 1991) with leks as experimental units and transects as subsam- ples. All proportions were arcsine-square root transformed for statistical tests to equalize variance of proportions (Neter and Wasser- man 1974). Because there were only two replicate leks in each habitat type, the design had a low power to detect differences (Neter and Wasserman 1974). Consequently, we selected an alpha of .10 for significance tests in the nested ANOVA. We compared the abundance of lago- morphs, small mammals, and badgers between habitat types using t tests. Each walked transect was considered a subsample of badger abundance within each habitat type. The relationship between different vegetation characteristics and nest success was analyzed for two sampling units, transect and nest, that measured habitat characteristics at different scales. With transects as sampling units, the relationship between mean vegetation charac- teristics and proportion of nests depredated on each transect was tested using multiple lin- ear regression and partial correlation. With nests as sampling units, the relationship between vegetation characteristics and nest success at individual nest sites was tested with multiple logistic regression. All statistical tests were performed using NCSS (Number Cruncher® Statistical System). Results Female grouse attended leks (8-20 males/ lek) and nested in both treated and untreated sagebrush. Of 22 hens radio-collared on win- tering areas between 1985 and 1989, 9 nested in treated areas the following spring (R. Dan- vir unpublished data). This frequency (40.9%) was not significantly different from the pro- portion of sagebrush grassland on the ranch that had been treated (40%) [X^ = .009, df = 1, P > .90). Overall, artificial nests were depredated significantly less frequently in treated sage- brush (10 of 60) than in untreated sagebrush (33 of 60) (X2 = 19.5, df = 1, P < .001). Mean proportion of nests depredated was greater in untreated than in treated sagebrush and the difference approached significance (F = 6.3, df = 1,2, P = .12; Table 1). Mean proportion of nests depredated differed significantly among leks (F = 4.6, df = 2,8, P = .04; Table 2). The majority of nests were depredated by mammals (37 of 43), with birds accounting for the remaining 6. The proportion of nests depredated by mammals did not differ signifi- cantly among habitat types (treated: 9 of 11; untreated: 28 of 32, Fisher's Exact, P = .63). Differences in nest predation among habitat tyi3es and leks were attributed to differences in vegetation characteristics (Table 1, 2). Horizon- tal cover (% of cover board obscured) and max- imum shrub height were significantly greater in untreated areas, but shrub and herbaceous cover were not (Table 1). Leks varied signifi- cantly in horizontal cover, herbaceous cover. 1994] Sage Grouse Nest Predation 125 Table 1. Artificial nests depredated (%) and habitat characteristics for treated and untreated areas of sagebrush grass- land at Deseret Ranch. Unl treated habitat Tre ated habitat Variable A^ SE N X SE N pa Nests depredated (%) 55.0 16.2 2 16.7 14.3 2 .12 Vegetation characteristics Horizontal cover (%) 89.3 2.5 2 69.4 2.4 2 .08 Shrub cover (%) 27.0 1.8 2 17.0 1.7 2 .15 Herbaceous cover (%) 21.1 1.6 2 18.2 1.4 2 .51 Maximum shrub height (cm) 65.0 3.1 2 36.1 2.8 2 .06 Alternate prey abundance Lagomorph pellets (#/m2)l' 17.0 3.5 4 2.5 3.6 4 .006 Small mammals (#/100 trap-nights) 8.6 0.8 2 12.6 0.8 2 .02 Predator abundance Badger holes (#/km) 2.0 0.52 0.67 0.19 9 .02 ^For nests depredated and vegetation characteristics, nested ANOVA. F < ID sii;riificant, see text; for alternate prey alnindance, ( tests, P < .0.5 significant, '^Data were log-transformed to account for nonnomial distributions. Table 2. Mean proportion of nests depredated and vegetation characteristics (±SE, N = 3) associated with Sage Grouse leks in treated and untreated sagebrush. Lek Untreated habitat Treated habitat Variable Dip Kate Hollow Neponset Alkali Hollow pa Nests depredated (%) Horizontal cover (%) Shrub cover (%) Herbaceous cover {%) Maximum shrub height (cm) 66.7 ± 14.5 89.3 ± 2.2 24.3 ±2.9 26.0 ±3.2 70.7 ±3.3 43.3 ±5.8 87.6 ± 5.4 29.0 ±0.7 17.0 ±0.7 59.9 ±4.9 6.7 ± 6.7 65.3 ±5.3 19.3 ±3.6 17.6 ±0.4 33.4 ±5.6 26.6 ± 12.0 75.6 ±2.8 16.0 ±4.9 15.7 ± 4.0 37.5 ± 5.5 .04 .09 .22 .07 .005 L nested ANOVA, P < .10 significant, see text. and maximum shrub height (Table 2). With transects as samphng units, we regressed pro- portion of nests depredated along the transect against the transect mean of each of the habi- tat variables (Fig. 1). Nest predation increased significantly with each variable except shrub cover Each variable was correlated with each of the others, so we examined the indepen- dent effects of each variable by calculating its partial correlation coefficient (Table 3; Neter and Wasserman 1974). Horizontal cover had a significant positive partial correlation with nest depredation rate, while shrub cover had a significant negative partial correlation. Herbaceous cover showed a positive partial correlation with nest predation that was nearly significant (F = .12), but maximum shrub height showed nearly zero partial correlation with nest predation. Overall, vegetation char- acteristics measured with transects as sam- pling units explained 86% of the variation in nest predation. With nest sites as sampling units, we used logistic regression to analyze the relationship 126 Great Basin Naturalist [Volume 54 (A) Horizontal Cover (B) Shrub Cover 0) *-» (0 -o 0) L. o. o O o uu • • • • 80 RA2 = 0.61 P = 0.003 bO • Untreated ■ Treated ■ 40 20 n ■ --mn — ■ — , — - • SO 60 70 80 90 100 Horizontal Cover (%) 0 10 20 30 40 Shrub Cover (%) (C) Herbaceous Cover 2 0) (0 0) L. a. Q (0 (/) 0) 100 80 60 40 20i • ■ • • ■ ■ RA2 = 0.37 P = 0.04 ■ 10 20 30 Herbaceous Cover (%) 40 ^ (D) Maximum Shrub Height ■D O 4-1 (0 ■D 0) k_ Q. Q (n 2 "20 30 40 50 60 70 80 Shrub Height (cm) uu • • 80 RA2 = 0.47 P = 0.01 60 40 ■ • 20 n- ■ ■ ■ ■-, ■ r»- < — 1 — • — < — Fig. 1. Relationships of proportion of nests depredated (%) and four vegetation characteristics (see text for precise definitions): (A) horizontal cover, (B) mtiximum shmb height, (C) shrub cover, and (D) herbaceous cover. R- values are for linear regressions. Data points represent transect means, presented separately for treated areas (•) and untreated areas (■). between the success (no predation) of individ- ual nests and vegetation characteristics associ- ated with each nest site (Table 4). As was found with transect sampling units, nest pre- dation increased with increasing horizontal cover, herbaceous cover, and maximum shrub height, but not shrub cover. With all four vari- ables considered simultaneously (multiple logistic regression), however, only horizontal cover and maximum shrub height were signif- icant. Overall, vegetation characteristics at nest sites explained only 12% of the variation in nest predation. Abundance of lagomorphs was significantly greater in untreated areas, but abundance of small mammals (primarily Peromyscus manic- ulatus) was greater on treated areas (Table 1). Fresh badger holes, however, were signifi- cantly more common in untreated areas. Con- sequently, greater predation of artificial Sage Grouse nests appeared to be associated with higher abundances of lagomorphs and bad- gers, but lower numbers of small mammals. Discussion Predation rates of artificial nests were higher in untreated sagebrush in spite of greater nest cover. These results contrast with those from studies that found greater preda- tion rates on nests in sparser cover (Wallestad and Pyrah 1974, Connelly et al. 1991). How- ever, Autenrieth (1981) found similar results to ours, namely lower predation rates on nests 1994] Sage Grouse Nest Predation 127 Table 3. Simple and partial correlation coefficients {N = 12) of proportion of nests depredated with four vegeta- tion characteristics when all fom^ variables are included in a multiple regression analysis. Variable Simple r' Partial r pi, Horizontal cover .78 .89 .0004 Shrub cover .20 -.7,5 .01 Herbaceous cover .61 .52 .12 Ma.xinium shrub height .69 -.04 .90 ■'For P vidues, see Figure 1. 'T value for the partial correlation coefficient. in a crested wheatgrass planting with sparse (<5%) sagebrush cover than in untreated sagebrush. In addition, Patterson (1952) found higher nest predation rates under taller, denser sagebrush. At least two patterns emerge that may explain the conflicting results of these studies. First, nest predation was higher when preda- tor densities were higher, regardless of nest cover (Autenrieth 1981, Angelstam 1986, this study). Badgers were the most frequent large mammalian nest predator (Patterson 1952, this study) and were more abundant in untreated areas. Second, nest cover seems to be more important in protecting nests from visually hunting predators, such as ravens, magpies, or gulls (Jones and Hungerford 1972, Picozzi 1975, Autenrieth 1981, Yahner et al. 1989, Sullivan and Dinsmore 1990), than those hunting by olfaction, such as badgers, coyotes, or chipmunks. These two patterns suggest that the type and density of predators may affect the degree to which nest cover reduces nest predation (Bowman and Harris 1980, Angelstam 1986). An additional factor that may explain dis- crepancies among studies is the fact that our treated areas were >25 years old and had recovered some sagebrush. Wallestad and Pyrah (1974) and Connelly et al. (1991) stud- ied areas that had been recently treated and perhaps still contained pretreatment densities of predators. In this study predator densities would have had plenty of time to decline fol- lowing sagebrush treatment. Thus, the effect of habitat on nest predation may be due pri- marily to the densities of predators supported by the habitat. Our conclusions are based on artificial nests; several studies have shown that the fate of artificial nests may not reflect that of natural nests (Angelstam 1986, Storaas 1988, Yahner and Voytko 1989). However, fates of artificial nests are likely to reflect differences in preda- tion rates among habitats and are legitimate tools for testing the hypotheses in this paper Vegetation characteristics associated with increased nest predation depended on the sampling unit used in the analysis. When transects were used as sampling units, increased nest predation was associated wdth increased horizontal and herbaceous cover (Fig. 1, Table 3). When individual nest sites were used, horizontal cover and maximum shrub height were the only significant factors (Table 4). Moreover, vegetation characteristics averaged for a transect explained considerably more variance in nest predation (86%) than did characteristics associated with individual nest sites (12%). Thus, the effect of vegetation characteristics on nest predation may depend on the scale at which they are measured (Bowman and Harris 1980, Allen and Starr 1982). In this case, vegetation characteristics of the overall habitat in which nest sites are located (transect scale) may be more impor- tant than characteristics directly at nest sites. The correlation between horizontal and herbaceous cover and increased nest preda- tion at the transect scale may actually reflect an indirect effect of habitat on nest predation rate: greater horizontal and herbaceous plant cover may be preferred by lagomorphs and other small mammals, which may then attract a greater density of predators. A greater density Table 4. Results of multiple logistic regression of nest predation (0 tion characteristics at individual nest sites. nest destroyed, 1 = no predation) vs. vegeta- Variablc Coefficient SE Variance explained (%) pa Horizontal cover (%) -0.024 0.012 6.1 Shnib cover (%) 0.010 0.020 0.2 Herbaceous cover (%) -0.009 0.021 0.2 Ma.\imum shrub height (cm) -0.021 0.009 6.4 .04 .62 .65 .03 ^Froni X- test within multiple logistic regression, P < .05 significant. 128 Great Basin Naturalist [Volume 54 of predators may then inflict a greater preda- tion rate on Sage Grouse nests. Our data are somewhat consistent with this hypothesis, in that lagomorphs and badgers were more abundant in untreated areas, and badger holes were often associated with burrow systems used by cottontails and pygmy rabbits. How- ever, small mannnals, which provide a lower prey biomass than lagomorphs, were more abundant in treated areas. Nevertheless, the association between vegetation, alternate prey abundance, predator density, and nest preda- tion rates appears to be the most likely hypothesis explaining our results. At the scale of transects, nest predation rate significantly decreased with increasing shrub cover, given horizontal and herbaceous cover (Table 3). However, this pattern was not observed at the scale of individual nest sites (Table 4). Nevertheless, shrub height was important for explaining nest predation at individual nest sites. Shrub cover and height are thought to be most important in prevent- ing predation by visually hunting predators such as birds (Jones and Hungerford 1972, Picozzi 1975, Autenrieth 1981) rather than mammals that hunt by olfaction (Angelstam 1986, Storaas 1988). However, our data sug- gest that increasing shrub cover and height may also help reduce mammalian nest preda- tion. Thus, for a given predator density, increased shrub cover and height may reduce Sage Grouse nest predation (Wallestad and Pyrah 1974, Autenrieth 1981, Connelly et al. 1991). Conclusions Our results suggest that lower nest preda- tion rates for Sage Grouse may occur in recovering treated sagebrush because the sagebrush treatment reduces the long-term density of predators. This result conflicts with the commonly accepted idea (Lovejoy et al. 1984, Wilcove 1985) that habitat fragmenta- tion always increases predation of bird nests. There is little doubt that sagebrush treatment significantly reduces Sage Grouse populations in both the short and long term (Dalke et al. 1963, Braun et al. 1977, Autenrieth 1981). However, the claim that sagebrush treatment increases nest predation rates (Braun et al. 1977, Connelly et al. 1991) is probably not the best reason for preserving contiguous stands of mature big sagebrush. Treating sagebrush may reduce Sage Grouse populations in the long term for reasons other than nest preda- tion (Braun et al. 1977), including elimination of winter habitat (Homer 1990), removal of year-round thermal cover (Moen 1973, Auten- rieth 1981), and reduction of perennial forbs, an important food for hens and chicks (Auten- rieth 1981). Consequently, recommendations to preserve mature sagebrush habitats should probably be made on the basis of these factors rather than nest predation. Acknowledgments We thank Deseret Land and Livestock for permission to do the study on its property and for logistical support. We also thank R. Wharff, K. Clegg, and numerous student vol- unteers from Utah State University for help in doing the fieldwork. Literature Cited Allen, T. F. H., and T. B. Starr. 1982. Hierarchy: per- spectives for ecological complexity. University of Chicago Press, Chicago. Angelstam, P. 1986. Predation on ground-nesting birds' nests in relation to predator densities and habitat edge. Oikos 47: .365-373. Autenrieth, R. E. 1981. Sage Grouse management in Idaho. Wildlife Bulletin 9. Idaho Department of Fish and Game, Boise. 238 pp. Beck, T. D. 1977. Sage Grouse flock characteristics and habitat selection in winter. Journal of Wildlife Man- agement 41: 18-26. Bowman, G. B., and L. D. Harris. 1980. Effect of spatial heterogeneity on ground-nest depredation. Journal of Wildlife Management 44: 806-813. Braun, C. E., T. Britt, and R. O. Walle,stad. 1977. Guidelines for maintenance of Sage Grouse habitats. Wildlife Society Bulletin ,5: 99-106. Byers, S. M. 1974. Predator-prey relationships on an Iowa waterfowl nesting area. Transactions of the North American Wildlife and Natural Resources Conferences 39: 223-229. Connelly, J. W., W. L. Wakkinen, A. D. Apa, and K. P. Reese. 1991. Sage Grouse use of nest sites in south- eastern Idaho. Journal of Wildlife Management 55: 521-524. Crabtree, R. L., and M. L. Wolfe. 1988. Effects of alternate prey on skunk predation of waterfowl nests. Wildlife Society Bulletin 16: 16.3-169. Croze, H. 1970. Searching image in Carrion Crows. Zeitschrift fiir Tierpsychologie Supplement 5: 1-86. Dalke, P. D., D. B. Pyrah, D. C. Stanton, J. E. Craw- ford, and E. F. Sclatterer. 1963. Ecology, pro- ductivity, and management of Sage Grouse in Idaho. Journal of Wildlife Management 27: 811-841. 1994] Sage Grouse Nest Predation 129 Daubenmire, R. F. 1968. Plant communities: a textbook of plant synecology. Harper and Row, New York. 300 pp. Dowdy, S., and S. Wearden. 1991. Statistics for research. Wiley Press, New York. DUEBBERT, H. F., A.ND H. A. Kantrud. 1974. Upland duck nesting related to land use and predator reduction. Journal of Wildlife Management 38: 257-265. Galbraith, H. 1987. Marking and visiting lapwing {Vanellus vanellus) nests does not affect clutch sur- vival. Bird Study 34: 137-138. Homer, C. G. 1990. Use of Landsat TM data in a GIS to model Sage Grouse winter habitat. Unpublished master's thesis, Utah State University, Logan. Jones, R. E. 1968. A board to measure cover used by Prairie Grouse. Journal of Wildlife Management 32: 28-31. Jones, R. E., and K. E. Hungerford. 1972. Evaluation of nesting cover as protection from magpie preda- tion. Journal of Wildlife Management 36; 727-732. Klebenow, D. a. 1970. Sage Grouse versus sagebrush control in Idaho. Journal of Range Management 23: 396-400. Klott, J. H., and F. G. Lindzey. 1990. Brood habitats of sympatric Sage Grouse and Columbian Sharp-tailed Grouse in Wyoming. Journal of Wildlife Manage- ment 54: 84-88. Lovejoy, T. E., J. M. Rankin, R. O. Bieregaard, K. S. Brown, L. H. Emmons, and M. E. Van Der VooRT. 1984. Ecosystem decay of Amazon forest fragments. Pages 296-325 in M. H. Nitecki, ed., Extinction. University of Chicago Press, Chicago. MacIvor, L. H., S. M. Melvin, and C. R. Griffin. 1990. Effects of research activity on Piping Plover nest predation. Journal of Wildlife Management 54: 443-447. MacMahon, J. A. 1987. Disturbed lands and ecological theory: an essay about a mutualistic association. Pages 221-238 in W. R. Jordan, M. E. Gilpin, and J. D. Aber, eds.. Restoration ecology: a synthetic approach to ecological research. Cambridge Univer- sity Press, New York. 342 pp. Moen, a. N. 1973. Wildlife ecology: an analytical approach. W. H. Freeman, San Francisco. Neter, J., and W. Wasserman. 1974. Applied linear sta- tistical models. Irwin Press, New York. 841 pp. Patterson, R. L. 1952. The Sage Grouse of Wyoming. Sage Books, Inc., Denver, Colorado. 341 pp. Picozzi, N. 1975. Crow predation on marked nests. Jour- nal of Wildlife Management 39: 151-155. Rearden, J. D. 1951. Identification of waterfowl nest predators. Journal of Wildlife Management 15: 386-395. ScHNEEGAS, E. R. 1967. Sage Grouse and sagebrush con- trol. Transactions of the North American Wildlife and Natural Resources Conference 32: 270-274. Storaas, T. 1988. A comparison of losses in artificial and naturally occurring capercaillie nests. Journal of Wildlife Management 52; 12.3-126. Sullivan, B. D., and J. J. Dinsmore. 1990. Factors affecting egg predation by American Crows. Journal of Wildlife Management .54: 433-437. Taylor, R. J. 1977. The value of clumping to prey: exper- iments with a mammalian predator. Oecologia 30: 285-294. . 1984. Predation. Chapman and Hall, New York. 140 pp. Vale, T. R. 1974. Sagebnish conversion projects; an ele- ment of contemporarN' environmental change in the western United States. Biological Conservation 6: 274-284. Wallestad, R. O., and D. B. Pyrah. 1974. Movement and nesting of Sage Grouse hens in central Mon- tana. Journal of Wildlife Management 38; 630-633. Weller, M. W. 1979. Density and habitat relationships of Blue-winged Teal nesting in northwesteiTi Iowa. Journal of Wildlife Management 43: 367-364. Wilcove, D. S. 1985. Nest predation in forest tracts and the decline of migratorv songbirds. Ecologv 66: 1211-1214. Yahner, R. H., and D. p. Scott. 1988. Effects of forest fragmentation on depredation of artificial nests. Journal of Wildlife Management 52; 158-161. Yahner, R. H., and R. A. Voytko. 1989. Effects of nest site selection on depredation of artificial nests. Jour- nal of Wildlife Management 53: 21-25. Yahner, R. H., T. E. Morrell, and J. S. Rachael. 1989. The effects of edge contrast on depredation of artifi- cial nests. Journal of Wildlife Management 53; 1135-1138. Received 25 February 1993 Accepted 2 September 1993 Great Basin Naturalist 54(2), ©1994, pp. 130-141 TIMING, DISTRIBUTION, AND ABUNDANCE OF KOKANEES SPAWNING IN A LAKE TAHOE TRIBUTARY David A. Beauchamp', Phaedra E. Budy^, Brant C. Allen'^, and Jeffrey M. Godfrey^ Abstract. — We counted kokanee spawners and carcasses every 1-7 days from mid-September through mid- November in 1991 and 1992 in Taylor Creek, a tributary to Lake Tahoe, California-Nevada. Less than 1% of the spawn- ing run entered Taylor Creek before flow from Fallen Leaf Lake was increased on 2 October 1991; in 1992 the peak occurred on 30 September or 1 October after flows increased on 29 September. In both years spawners concentrated in the middle three of five stream reaches below the impassable Fallen Leaf Lake dam. From tag-and-recovery experi- ments, the average longevity of male spawners in the stream was 3.5 days in 1991 and 2.8 days in 1992, whereas the average female longevit\' was 2.0 days in 1991 and 2.3 days in 1992. Observed carcasses accounted for less than 10% of spawners counted, suggesting removal by scavengers or high predation on prespawners. An estimated 1928 males and 1309 females spawned in 1991, and 8021 males and 8712 females spawned in 1992. Our estimate of 3237 spawners in 1991 compared favorably to our estimate of 3520 ± 1474 prespawners staging in Lake Tahoe in mid-September. An index of kokanee abundance in Lake Tahoe has historically been based on 1-day sui-veys every 1 November since 1960; however, estimated total spawner abundance was 19 times higher than the annual index of 158 spawners in 1991, and 141 times higher than the index count of 100 spawners in 1992. The index count and mean fork lengths of spawners (278 ± 10 mm [2 SE] for males, and 248 ± 3 mm for females) in 1991 and 1992 were the lowest on record. Key words: kokanee, spawner. abundance, life span, turnover rate. Kokanee salmon {Oncorhynchus nerka) rep- resent an important food source for lake trout {Salveliniis nainaycush) (Frantz and Cordone 1970) and ovei-wintering Bald Eagles {Haliaee- tus leucocephalus) (U.S. Forest Service [USFS] 1979), are important zooplanktivores, and pro- vide a valuable sport fishery in Lake Tahoe (Cordone et al. 1971). Despite these promi- nent recreational and ecological roles, little is known of the survival or abundance of this population. The long-term record consists of 1-day spawner surveys of Taylor Creek on or near 1 November every year, and this has served as an index of abundance. Population trends have been inferred from these data, but the relationship of this index to actual abun- dance has not been evaluated. Interannual dif- ferences in run-timing could violate the criti- cal assumption that these annual spawner counts represent some constant, but unknown, fraction of the total run size. Therefore, the purpose of this study was to estimate absolute abundance of spawners, compare annual 1 November index counts to absolute abun- dance estimates, and determine whether index counts represented a constant fraction of the total spawning population among years. If the index count proves inadequate, then some improved, but streamlined, monitoring program should replace it. Therefore we also examined whether, over the spawning season in Taylor Creek, changes occurred in the sex ratio or distribution of spawners among reaches. Tem- poral changes in either of these could bias population estimates and must be accounted for in the design of future kokanee spawning sui^veys. Study Site Kokanees were originally introduced into the lake in 1944, but the population remained small until 1960 when a popular summer sport fishery emerged (Cordone et al. 1971). Most natural reproduction occurs in Taylor Creek, but the population has also been supplement- ed irregularly by stocking fry into the creek or fingerlings into various regions of the lake 'Utah Cooperative Fish and WildUfe Research Unit, Department of Fisheries and Wildhfe/Ecology Center, Utah State University, Logan, Utah 84322- 5210. The unit is jointly sponsored by Utah State University. U.S. Fish and Wildhfe Service. Utah Division of Wildhfe Resources, and the Wildlife Management Institute. ^Department of Fisheries and Wildlife, Utah State University. Logan. Utah 84322-5210. •^Tidioe Research Group, Box 633. Tahoe City. California 96145. 130 1994] Spawning of Lake Tahoe Kokanees 131 (T. Frantz personal communication). Kokanees spawn and die in Taylor Creek in autumn. The eggs and alevins incubate in gravel redds until spring, emerge and migrate to Lake Tahoe, then reside in the lake 3-4 years before spawning in the natal stream. Although annual discharge fiom Taylor Creek ranks only fourth among the lake tributaries (Byron and Gold- man 1989, USFS unpublished data), kokanees spawn there because elevated autumn flows have been maintained since 1959. USFS gen- erally increases flow from the dam on Fallen Leaf Lake (Fig. 1) in early October "to protect and enhance the habitat for fish ... to insure an adequate and increasing food supply for wintering eagles " (USFS 1979). In general, flows remain low during summer but are increased during the first week in October to provide adequate flows during spawning and incubation periods. Additional spawning occurs along the western and eastern shore- lines and in several other tributaries, but these represent a small fraction of the total kokanee population (Cordone et al. 1971, B. Allen unpublished data). Taylor Creek flows 2.6 km from a con- trolled dam on Fallen Leaf Lake through a forested valley into the southwest corner of Lake Tahoe (Fig. 1). We divided the stream into five survey sections, based on gradient differences, between the mouth and base of the dam to determine spatial distribution of kokanees during the spawning season (Table 1, Fig. 1). From the mouth of Taylor Creek, the lake bottom slopes gradually to 6 m deep about 0.4 km offshore, then drops steeply to 30 m deep approximately 0.6 km offshore. The lake bottom is much steeper near the mouth of Cascade Creek (deeper than 60 m within 100 m of shore) approximately 2.2 km north- west of Taylor Creek (Fig. 1). During the 1991 and 1992 spawning seasons, the lake level was 2 m lower than normal and 0.3 m below its natural rim due to six years of drought; this did not block spawners' access into the stream. Methods We monitored the relative abundance of kokanees staging offshore in Lake Tahoe prior to and during the spawning season using scuba surveys, echosounding, and undenvater video with a remotely operated vehicle (ROV; see Beauchamp et al. 1992). We searched the Fig. 1. Map of southwestern Lake Tahoe (the dashed box inside the inset), showing lake bathymetn' near Tay- lor and Cascade creeks, California. Locations of stream sections in Taylor Creek between Fallen Leaf Lake and Lake Tahoe during the 1991 spawning season are num- bered. The viewing chamber at the USFS Interpretive Center, Cascade Lake, and Emerald Bav are indicated. lake from shore to 30 m deep for the 2.2 km between Taylor and Cascade creeks, and the area within a 200-m radius of the creek mouths (Fig. 1). Although we infrequently searched other streams and shoreline areas reported to have spawning activity in past years, we found no spawners or carcasses in either year. A school of adult kokanee was videotaped offshore Cascade Creek on 12 September 1991. The tape was later rerun as a series of freeze frames. Each frame was divid- ed into 16 vertical strips (25 mm wide by 295 mm long), marked on the monitor, and koka- nees were counted within each strip. Each frame sampled different segments of the school. The school was further stratified into 132 Ghkat Basin Naturalist [Volume 54 low- and liigh-densih' repons, and mean den- sit)' and standard deviation of kokanees were estimated from the strips in each region; mean densities were then multiplied by the total number of low- and high-density strips, and products were sunuiied to estimate the total number in the school. In 1991 we surveyed all reaches of Taylor Creek weekly from 17 September to 19 November, recording the number of salmon in each section that were alive or dead, male or female. A weir, installed 40 m upstream of the lake on 16 October 1991 to capture spawners, was opened 2 days later because of low catch- es. It was closed again on 28 October, and in one night it trapped sufficient spawners for tagging in a tag-and-recover estimate of longevity in the stream. In 1992 USFS personnel first reported spawners in Taylor Creek on 24 September. Tahoe Research Group or USFS personnel counted spawners in sections 1 and 2 twice during the first week of the spawning run. Since counts in section 2 represented at least half the spawners during the early part of the run, we doubled these counts to approximate total number of spawners in the stream on days of the counts. We surveyed all reaches of Taylor Creek either daily or every other day from 1 October to 4 November. Numbers of spawners during intei-vening days were esti- mated by linear interpolation. To estimate the total spawning escapement of males and females into Taylor Creek, we used the relationship: T i = 1 N: = where N: = total abundance of sex J, N,, = number of sex j spawners counted (or inteipo- lated) on day i, T is the total number of days in the spawning season, and D is the average gender- specific longevity (in days) after enter- ing the stream. We estimated average gender-specific longevity, D-, under controlled and natural conditions. First, five males and five females were trapped at the mouth of Taylor Creek overnight on 28-29 October 1991, tagged with orange dart tags, and released into the stream profile viewing chamber at the USFS Inter- Table 1. Lengths and gradients of the stream sections surveyed for spawning kokanee sahnon in Taylor Creek, the major spawning trihutary to Lake Tahoe, California- Nevada (estimated from USGS 7.5 minute topographic map). Section Length (i (Gradient 290 550 660 660 620 0.3% 0.9% 3.0% L9% 1.0% pretive Center (Fig. 1). USFS personnel also moved additional untagged kokanee salmon into the viewing chamber periodically. By observing behavior, relative health, and mor- tality of tagged and untagged spawners in the viewing chamber, we concluded that tags had no apparent detrimental effect, and no tags were lost. We surveyed the stream and cham- ber every 1-3 days until all tagged fish had died. We obtained a second estimate of longevity using the same procedure on untagged spawners in the chamber during two periods when distinct groups of spawners could be tracked over time. Finally, to esti- mate average longevity in the stream, we tagged and released 38 males and 21 females upstream of the weir at dusk on 29 October 1991. High winds and dense floating leaf litter prevented an adequate survey in the stream on the day after the tag release; few tagged fish remained in the stream when conditions finally permitted a survey 2 days later. We repeated the tagging experiment in Taylor Creek in 1992. Thirty-nine females and 5 males were captured overnight in a fyke net set at the mouth of the stream on 15 October Each fish was anesthetized with MS -222, sexed, weighed, measured, and given two tags on the right side below the posterior insertion of the dorsal fin. Fish were allowed 8 h to recover, then were released upstream. One male died during the holding period, and 2 females were eaten by gulls (species un- known) before they could swim to cover upstream. Thus, 4 males and 37 females remained. The remaining fish were counted in the stream on 16-19, 22-23, 25, 27, 29, and 30 October. A male and female each lost one of their two tags while still alive in the stream. Average longevity, D-, for each study group was computed using time-density (both years) and regression (1991 only) approaches. In the 1994] Spawning of Lake Tahoe Kokanees 133 time-density estimate, the number of live spawners from each group was counted on eveiy survey day until all were dead. We esti- mated counts between survey dates by linear interpolation. For each sex, daily counts were summed (total spawner days) and divided by the initial number released to estimate the average number of days D- individuals sur- vived in the stream. Regressions were com- puted for the percentage of tagged males and females (and the two untagged groups in the viewing chamber) surviving as a function of time. The resulting equations were solved for number of days until survival equaled 50% to estimate D Results In 1991 kokanees staged in one large pre- spawning aggregation off the mouth of Cascade Creek. The creek is inaccessible to spawners due to an extremely high gradient and low autumn flows. The aggregation concentrated in a dense, 2-3-m-high band 3-10 m offshore from where the metalimnion (28-33 m deep in 10-1 1°C) intersects the steeply sloping lake bottom. From video samples we estimated that the aggregation contained 3520 ± 1474 (mean ± SD) mature (red-colored) kokanees. No other aggregations were found closer to Taylor Creek at any time tliroughout the season. In 1991 only 8-29 spawners were counted per day in Taylor Creek during the last two weeks of September; the major portion of the run began after stream flows were increased (from 4 cfs to 8-9 cfs) on 2 October, peaked at 600 spawners on 14 October, and declined until only 14 live fish remained on 12 Novem- ber (Fig. 2). All spawners and carcasses disap- peared between 14 October and 18 October, 2 days after the weir was installed, but daily counts exceeded 400 spawners in the descending limb of the run after upstream access was reinstated on 18 October (Fig. 3). The run was larger and peaked earlier in 1992 (Fig. 2). The largest number of spawners was counted on 1 October. We might have missed the actual peak, but spawner counts in section 2 on 27 and 29 September were 50-75% smaller than on 1 October. Maximum densities probably occurred on 30 September or 1 October, and sensitivity analysis of likely alternatives indicated that errors would at most increase our abundance estimate by only 1000 fish (a 6% underestimate). Stream sections 2, 3, and 4 were the most heavily used reaches both years, whereas sec- tion 1 was lightly occupied (Fig. 3). No fish were ever seen in section 5 in 1991, but nearly 100 spawners were counted there throughout the 1992 season (Fig. 3). Proportional alloca- tion of spawners among sections was relatively constant during both years. Temporal distribu- tion of live and dead fish indicated little or no spawning in section 1. Only live males and carcasses of both sexes were found there, mostly during the first half of the run (Fig. 3). Since kokanees die after spawning, live fish should be replaced by an equal number of car- casses, but carcass counts could account for only 5-10% of the number expected from declines in daily spawner counts. In three sur- veys divers found only 3-17 carcasses on the lake bottom within 200 m of the creek mouth. Few carcasses were washed down to the weir in 1991. Thus, emigration of spawned-out fish and flushing of carcasses could not account for the discrepancy between declining spawner counts and carcasses. Male kokanees lived longer in the stream than females (Table 2). Consequently, males would appear more abundant because of accu- mulation over a longer period. In 1991 one tagged male lived in the stream for 14 days, whereas none of 21 tagged females remained after the first 3 days. Since we did not know when females disappeared over that 3-day period, we assumed a conservative linear decline to zero on the third day. In 1991, counts and linear interpolations summed to 132.0 total spawner days for the 38 tagged males; thus, males lived an average 3.5 days in the stream (Table 2). Similar analyses yielded longevity estimates of 2.0 days per female in 1991, 2.3 days per female in 1992, and 2.8 days per male in 1992 (Table 2). Longevity estimates differed between sexes and among groups in the viewing chamber and stream in 1991. Exponential regressions of sui-vival as a function of time in the cham- ber were significant for tagged males (r = .917, P < .05), untagged males (r = .918, P < .05), tagged females (r = .867, P < .05), and untagged females (r = .972, P < .05). Result- ing longevity estimates were 8.3 days for tagged males, 7.5 days for tagged females, 4.4 days for untagged males, and 3.3 days for 134 Great Basin Naturalist [Volume 54 1000 17-Sep 24-Sep 01-Oct 08-Oct 15-Oct 22-Oct 29-Oct 05-Nov 12-Nov DATE 2500 2000 1500 1000 17-Sep 24-Sep 01-Oct 08-Oct 15-Oct 22-Oct 29-Oct 05-Nov 12-Nov DATE Fig. 2. Daily abundance of male and female spawners in Taylor Creek, California, in 1991 and 1992. A weir installed on 16 October 1991 prevented upstream passage from Lake Tahoe for 2 days (16-18 October). untagged females. However, these longevities were not consistent with the rapid disappear- ance of spawners between 14 and 18 October, following closure of the weir on 16 October (Fig. 2), suggesting that longevity was shorter than indicated by the chamber survival exper- iment. In contrast to spawners in the stream, tagged and untagged fish in the chamber were passive, rarely displaying spawning or aggres- sive behavior Using the exponential model for tagged males released upstream in 1991 (r = .831, P < .10) resulted in an average longevity of 3.0 days, compared to the time-density esti- mate of 3.5 days in Table 2. Although the exponential model accounted for 69% of the variabilitv in survival over time, it was not 1994] Spawning of Lake Tahoe Kokanees 135 0\ N^mm8M^ On ^^lE ^lE o o o CO o o o SIMQOD ^HNMVdS SXMnOD SSV0HV3 z z o o 0 b t/3 tfJ o o o o O 00 o o o o o o SINQOD HHMMVdS SlNflOD SSV3HVD b 136 Great Basin Naturalist [Volume 54 Table 2. Time-densit\' estimates of longevity from 1991 and 1992 in-stream tagging experiments (mean days spawn- ers survived) with male and female kokanees. Tagged males Date Actual tag counts Actual and interpolated spawner davs fagged females Actual Actual and tag interpolated counts spawner days 21 21.0 14.0 7.0 0 0.0 0.0 0.0 0.0 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0.0 0 0.0 1991 29 Oct 30 Oct 31 Oct 01 Nov 02 Nov 03 Nov 04 Nov 05 Nov 06 Nov 07 Nov 08 Nov 09 Nov 10 Nov 11 Nov 12 Nov 13 Nov 38 38.0 27.7 17.3 7.0 6.5 6.0 5.5 5.0 4.4 3.9 3.3 2.7 2.1 1.6 1.0 0.0 Total tagged spawner days Number of tags released Average longevity (days) 132.0 38 3.5 42.0 21 2.0 1992 15 Oct 16 Oct 17 Oct 18 Oct 19 Oct 20 Oct 21 Oct 22 Oct 23 Oct 24 Oct 25 Oct 26 Oct 27 Oct 28 Oct 29 Oct 30 Oct 4.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 37 6 6 6 6 37.0 6.0 6.0 6.0 6.0 5.3 4.7 4.0 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0 Total tagged spawner days Number of tags released Average longe\ity (days) 11.0 4.0 2.8 85.5 37.0 statistically significant (P > .05). All longevity estimates from the chamber and stream exper- iments were comparable to other spawning populations of kokanee and sockeye salmon in North America (Table 3), but estimates from tagged fish in the stream were consistent with the disappearance of spawners between 14 and 18 October 1991 and best represented the natural condition in Taylor Creek. For males in 1991 the regression method yielded a 20% higher population estimate than did the time- density method. Time-density estimates of gender-specific longevities were combined with stream sui-vey data, using equation 1, to estimate total spawning populations in Taylor Creek in 1991 and 1992: 1788 male and 1200 female spawn- ers in 1991, and 6927 males and 7167 females in 1992. The total population estimate of 2988 spawners in 1991 was 19 times higher than the annual index survey count of 158, and the 1992 estimate of 14,094 spawners was 141 times higher than the index count of 100. The 1991 estimate of 2988 compared favorably to 1994] Spawning of Lake Tahoe Kokanees 137 Table 3. Mean spawning ground longevities from different North American populations of kokanees and sockeye salmon. Population/Locution Karluk Lake Brooks Lake Wood River Lakes Pick Creek Kvichak Lake Taylor Creek males Taylor Creek females Viewing chamber males Viewing chamber females Viewing chamber males Viewing chamber females Longevit>' (days) Reference 7.0 Gangmark and Fulton 1952 3.0 Eicher 1951 LO Mathisen 1955 3-5 Mathisen 1955 7.6 Hartman 1959 2.8-3.5 this studv, from tagged fish in the stream 2.0-2.3 this studv, from tagged fish in the stream 8.3 this studv, tagged fish 7.5 this studv, tagged fish 4.4 this study, untagged fish 3.3 this study, untagged fish our video estimate of 3520 ± 1474 prespawii- ers aggregated near Cascade Creek in mid- September Spawning runs in 1991 and 1992 were the smallest on record, as were the mean body lengths of spawners (Fig. 4). Males were larger than females in both years, but the size range of males was much narrower in 1992 than in 1991 and previous years (Fig. 5). Discussion Although 1991 and 1992 spawning popula- tions were the smallest on record, estimates of their timing, abundance, and distribution will be veiy useful. An important finding was the variability and degree to which the 1 Novem- ber spawning index surveys underestimated actual spawner abundance. Short spawner longevity caused much higher turnover than was previously believed and resulted in popu- lation estimates that were orders of magnitude higher than indicated by a single snapshot of the population. Since daily abundance changed over time, timing of the index survey relative to its position on the abundance curve (and the shape of the curve) is important to the consistency and accuracy of the index. The index represented 5% of the total run in 1991, but only 0.7% in 1992. In both years the 1 November index counts were made near the end of the runs (Fig. 2), and peaks for the 2 years differed by 2 weeks. An index survey should be conducted at or near the peak of the run for several reasons. First, during peak spawning the sex ratio observed from one survey is closer to the actual sex ratio of the whole run. Second, counts near the peak capture a higher per- centage of the total run. Counts during the peaks in 1991 and 1992 represented 14-20% of the total run; thus, an expansion factor of 5-7 times an index count during peak spawn- ing could estimate total run size. In contrast, the 1 November index counts would have required expansion factors of 19-141; the magnitude and volatility of these latter factors severely compromise accuracy, precision, and thus the utility of the 1 November index. Third, variation among daily counts was lower near the spawning peak: the slope of the cui-ve was relatively flat near the peak but became steep 1 week later Consistency of the index, as some range of fractions of the total popula- tion, would have improved had index counts occurred near the peaks in 1991 and 1992. If three index sui-veys were made on 1, 8, and 15 October, the largest of these spawner counts should represent 14-20% of the total spawner abundance. If only one survey is performed, we suggest that it occur on 8 October every year. The total spawner estimate depended on counts every 1-8 days and on estimates of longevity. The latter was more difficult to esti- mate with certainty. Since D,- appears in the denominator of the population estimator, the shorter the longevity', the greater the popula- tion estimate. Survival rates in the viewing chamber were higher for tagged than untagged fish, and those in the viewing cham- ber generally lived longer than those in the stream. Some untagged fish might have been in the stream for one or more days before transfer to the chamber, potentially causing an underestimate of longevity. Normal spawning activity and aggression were infi-equent in the chamber, and these fish might have lived longer due to lower energy expenditures and stress than was experienced in the stream. 138 Great Basin Naturalist [Volume 54 I960 1965 1970 1975 1980 SPAWNING YEAR I COUNTS MALES 1985 1990 FEMALES Fig. 4. Historical record of annual index sui"vey counts of spawning kokanees in Taylor Creek, California, on or near 1 November (bars), and corresponding mean fork lengths (right vertical tixis) of males (squares) and females (diamonds). Mathisen (1955) reported that spawning females performed digging activities at least once per minute, and that males actively defended females. These activities would pre- sumably burn the spawners remaining energy reserves much faster than the activity observed in the viewing chamber. Disappearance of all spawners in the stream between the peak of the 1991 run on 14 October and 18 October, 2 days after a weir blocked passage, suggested a faster turnover rate for spawners in the stream than was esti- mated from the viewing chamber experiment. An overestimate of longevity causes an under- estimate of total abundance, and we consider the estimates derived from viewing chamber experiments to represent maxima of longevity and result in minimum estimates of spawner abundance. The spawning population of kokanee salmon in Taylor Creek was extremely low in 1991. The annual index count on 1 November was 158 compared to counts of 3000-49,000 on similar dates during 1985-90. Index counts for the parents of the 1991 run (1987 for 4-yr- olds or 1988 for 3-yr-olds) were the largest (44,000-49,000) on record for T\iylor Creek. The 1987-88 brood populations were supple- mented with 590,000-850,000 frv stocked in 1984 and 1985 (Russ Wickwire, California Department of Fish and Game, Tahoe City, California, unpublished data). Reasons for the failure of these large brood years to produce large numbers of recruits are unknown and will require further study. The magnitude of the historical index's underestimate of abundance has several important implications. Morgan et al. (1978) published a preliminary spawner-recruit rela- tionship, based on annual index counts, and suggested that sustainable population levels dropped from 15,000 to 7000 spawners due to declines in their cladoceran prey. Although the direction and relative change in the sus- tainable population may be correct, actual abundance of spawners (and recruits) was severely underestimated because uncorrected index counts were used instead of estimates of absolute abundance. If a consistent relation- ship could be found between index counts and absolute abundance, then historical data could be adjusted and compared with contemporary' data. Taylor Creek supports most of the natural reproduction (>90%) in Lake Tahoe, but the relative contribution of other spawning areas to the lake population needs further examina- tion. Because of severe drought conditions 1994] Spawning of Lake Tahoe Kokanees 139 1991 KOKANEE SPAWNER SIZE 50% 325 350 375 400 FORK LENGTH (mm) 1992 KOKANEE SPAWNER SIZE 225 250 275 300 325 350 375 400 425 450 475 500 FORK LENGTH (mm) Fig. 5. Length frequency histograms for male and female spawners trapped at the mouth of Taylor Creek, California, in October 1991 (top) and 1992 (bottom). during this study, most other streams were inaccessible to spawners due to low flow. Moreover, the 2-m drop in lake level reduced the submerged gravel habitat for shore spawn- ers by 80% (Beauchamp et al. 1994). Absence of shore spawners could be attributed either to low overall abundance (and our inability to detect low numbers of spawners) or to low lake levels. Spawners concentrated in the middle three sections of Taylor Creek. Given the low abun- dance of spawners, this distribution probably indicated preferred spawning reaches since competition for redd sites was low. Only low numbers of live males and carcasses of both sexes were found in the first reach, particular- ly during the first half of the run. Slower cur- rent in this section offered a "first stop" for 140 Great Basin Naturalist [Volume 54 new spa\\niers and a place for carcasses to set- tle, but slow current and fine substrate pro- vided inferior spawning habitat (Parsons and Hul^ert 1988). High turnover rate of spawners and low (5-10%) percentage of carcasses obsei^ved, rel- ative to the number expected, suggested a high demand for carcasses and perhaps for liv- ing prespawners by scavengers and predators. Cederholm et al. (1989) found that most coho salmon {Oncorhynchm kisutch) carcasses were retained by low-gradient streams and the adjoining forest, that few were flushed very far (<600 m) downstream, and that most (88%) carcasses removed from the stream remained within 15 m of the bank. They noted further that carcass retention was correlated positively with organic debris load and nega- tively with carnivore scavenging; they pre- sented a list of 22 birds and mammals that consume carcasses. Since stream flows are regulated at relatively low discharge rates throughout the spawning season, "missing" carcasses were probably not washed out of the stream. This assumption was supported by the low number of carcasses washed against the weir or along the lake bottom in the vicinity of Taylor Creek. The concern here is that preda- tors might remove large quantities of koka- nees before they spawn and die naturally. This could be exacerbated when run sizes are low, because scavengers, prompted by the lack of carcasses, might pursue live salmon. Small spawning runs would suffer an additional bur- den of depensatoiT mortality that would accel- erate the decline of the population. This dis- appearance of carcasses has important impli- cations for temporal and spatial nutrient cycling of carcasses (Richey et al. 1975) and for reproductive potential of kokanees in Lake Tahoe. The combination of record low body size and index counts for Taylor Creek spawners in 1991 and 1992 is puzzling. Abundance could be affected by parental abundance and mortal- ity in either Taylor Creek or Lake Tahoe, but since all feeding occurs in the lake, growth can only be affected by in-lake processes. When kokanee abundance is low, density- dependent growth should produce larger spawners (Rieman and Myers 1992), but we observed the opposite. These counterintuitive results suggest that complex trophic interac- tions are occurring in the lake and will require careful study to identify processes that are important to kokanee productivity. Since fecundity increases with female body size (Foerster 1968), depressed growth and low population densities could severely hamper the population's ability to rebound naturally, because fewer females are each producing fewer eggs. Acknowledgments Funding was provided by the Sport Fish Restoration Act through Nevada Division of Wildlife. The Tahoe Research Group, under the direction of Dr. Charles Goldman, Univer- sity of California-Davis, and California Department of Fish and Game provided labora- toiy facilities. This article is dedicated to Ted Frantz, deceased, for his lifetime contribution to our knowledge of Lake Tahoe fish popula- tions. Literature Cited Beauchamp, D. a., B. C. Allen, R. C. Richards, W. A. WURTSBAUGH, AND C. R. GOLDMAN. 1992. Lake trout spawning in Lake Tahoe: egg incubation in deep water macrophytes. North American Journal of Fisheries Management 12: 442-449. Beauchamp, D. A., E. R. Byron, and W. A. Wurts- BAUGH. 1994. Summer habitat use of httoral fishes in Lake Tahoe and the effects of shorehne structures. North American Journal of Fisheries Management 14: In press. Byron, E. R., and C. R. Goldman. 1989. Land-use and water quality in tributary streams of Lake Tahoe, California- Nevada. Journal of Environmental Quali- ty 18: 84-88. Cederholm, C. J., D. B. Houston, D. L. Cole, and W. J. Scarlett. 1989. Fate of coho salmon [Oncorhynchus kisutch) carcasses in spawning streams. Canadian Journal of Fisheries and Aquatic Sciences 46; 1.347-1355. Cordone a. J., S. J. Nicola, P. H. Baker, and T. C. Frantz. 1971. The kokanee salmon in Lake Tahoe. California Fish and Game 57: 28-43. Eicher, G. J., Jr. 1951. Effect of tagging on the subse- quent behavior and condition of red salmon. U.S. Fish and Wildlife Service, Special Scientific Report Number 64. 52 pp. Frantz, T. C, and A. J. Cordone. 1970. Food of lake trout in Lake Tahoe. CalifoiTiia Fish and Game 56: 21-35. Gangmark, H. a., and L. A. Fulton. 1952. Status of the Columbia River blueback salmon runs, 1951. U.S. Fish and Wildlife Service, Special Scientific Report Number 74. 29 pp. Hartman, W. L. 1959. Red salmon spawning behavior. Science Alaska: proceedings of the northern Alaska science conference. College, Alaska 1958: 48-49. Mathisen, O. a. 1955. Studies on the spawning biology of the red salmon, Oncorhynchus nerka (Walbaum), 1994] Spawning of Lake Tahoe Kokanees 141 with special reference to the effect of altered sex ratios. Unpublished dissertation, University of Washington, Seattle. 285 pp. Morgan, M. D., S. T. Threlkeld, and C. R. Goldman. 1978. Impact of the introduction of kokanee (Qncorhijnchus nerka) and opossum shrimp (Mysis relicta) on a subalpine lake. Journal of the Fisheries Research Board of Canada .35: 1572-1579. Parsons, B. G., and W. A. Hubert. 1988. Influence of habitat availability on spawning site selection by kokanees in streams. North American Journal of Fisheries Management 8; 426-431. RiCHEY, J. E., M. A. Perkins, and C. R. Goldman. 1975. Effects of kokanee salmon {Oncorhynchtis nerka) decomposition on the ecology of a subalpine stream. Journal of the Fisheries Research Board of Canada 32: 817-820. Rieman, B. E., and D. L. Myers. 1992. Influence of fish density and relative productivity on growth of koka- nee in ten oligotrophic lakes and reservoirs in Idaho. Transactions of the American Fisheries Society 121: 178-191. United States Forest Service (USES). 1979. Bald Eagle winter management plan for the Lake Tahoe Basin Management Unit. Received 18 August 1993 Accepted 2 December 1993 Great Basin Naturalist 54(2), ©1994, pp. 142-149 FULL-GLACIAL SHORELINE VEGETATION DURING THE MAXIMUM HIGHSTAND AT OWENS LAKE, CALIFORNIA Peter A. Koehler' and R. Scott Anderson^^ Abstract. — Owens Valley, California, was markedly different during the Wisconsin glacial stage from what it is today. Alpine glaciers hounded the Sierra Nevada, and pluvial Owens Lake reached highstands and overflowed its nat- ural basin. We analyzed three layers from two packrat middens, dated to ca 23,000-14,500 yr BP, obtained from Haystack Mountain (1155 m) only 10 m above and <100 m from the highstand strandline of pluvial Owens Lake. Dur- ing this period vegetation near Owens Lake reflects the influence of the Tioga glacial advance and retreat on lake levels, and microclimatic effects on shoreline vegetation. Between ca 23,000 and 17,500 yr BP a Utah juniper (Juniperus osteosperma) and single-needle pinyon pine {Piniis monophijlla) woodland existed at the site. In the layers dated to ca 17,500 and 16,000 yr BP, macrofossils document the presence of Rocky Mountain juniper (Juniperus scopulorurn), a species that no longer occurs in California. It is suggested that meltwater from the retreating glacial ice inundated the Owens River Lake chain causing pluvial Owens Lake to reach its highstand. This caused an increase in effective mois- ture, due to high groundwater, allowing the mesophytic Rocky Mountain juniper to exist at the site. Key words: paleoecology, packrat middens. Rocky Mountain juniper, Juniperus scopulorum, pluvial Owens Lake, Tioga glacial stage, California. Few places in western North America record such a full range of Quaternary events as found in the Owens Valley of eastern Cali- fornia. Within the confines of the narrow Owens River corridor, never more than 33 km wide, is found evidence of late-Quaternary glacier expansions (Birkeland and Burke 1988, Bursik and Gillespie 1993), volcanic eruptions (Pakiser et al. 1964), and expansion and con- traction of large "pluvial" lakes (Lajoie 1968, Smith and Street-Perrott 1983, Benson et al. 1990). Such deposits are the manifestations of great climatic and environmental changes that have occurred during the late Quaternary. Less studied but equally striking is the record of biological changes contemporaneous and associated with pervasive changes in the physical system. Pollen from pluvial lake sedi- ments (Leopold 1967, Batchelder 1970, Davis unpublished) has been used to reconstruct the broadscale, regional changes in vegetation. Other studies (Koehler and Anderson 1990, Jennings and Elliott-Fisk 1993, Koehler unpublished) have relied on packrat [Neotoma) middens, which record local vegetation changes. A combination of all proxy indicators will ultimately allow a comprehensive picture of environmental change to be revealed. The goal of this study was to investigate the pleni- to late-glacial vegetation communities near pluvial Owens Lake (Fig. 1), which fluc- tuated considerably during this period. Increased effective moisture and glacial runoff during the late Wisconsin initiated a series of overflow events in the lakes that define the Owens River system (Fig. 1). The chain began in the Mono Lake Basin where Pleistocene Lake Russell (Putnam 1950) overflowed when it filled to 2175 m elevation (Lajoie 1968, Ben- son et al. 1990). Owens River then flowed through the Adobe Valley and mixed with waters from Long Valley en route to Owens Lake. Owens Lake periodically filled and overflowed at 1145 m. Eventually, runoff flowed to China and Searles lakes, then into Lake Manly in Death Valley (Smith and Street-Perrott 1983). Lake levels fluctuated considerably between ca 24,000 and 21,000 yr BP followed by high and relatively stable lake levels between 21,000 and 14,000 yr BP (Smith and Street-Perrott 1983). Six layers from two packrat middens found at 1155 m elevation in Owens Valley, Inyo County, California (Fig. 1), document full- glacial vegetation changes during the period between ca 23,000 and 14,500 yr BP The 'Quaternary Studies Program, Northern Arizona Univer.sity, Flagstaff, Arizona 86011. ^Environmental Sciences Program, Northern Arizona University, Flagstaff, Arizona 86011. 142 1994] Shoreline Vegetation During Highstand 143 121' 119° 41° — 39° 37° — 35° — Legend .--./ Pleistocene rwer pathways EM Pleistooene lakes A Packral midden locations O Pollen locations State boundary + Other locations • Towns \Vf^ Mountain crest 36°5a- Lake Russell Ax Adobe Valley Long Valley \Volcanic Tablelands 118° _J_ 36°5a + Mt.Whitney '^ 4418m 118° Cottonwood Basin +, Eureka View A N Devifs Eleana Range "^.Hole ^ rrySfc A Specter Range Owen's (vi^^A Skeleton Hills ^LakeSite^^: 1 \ kDeath Valley '-^ A \\ ^. Lake "^ I Manly ^ A Sheep Range Tule Springs O \ Charleston Mountain \ N A Clark Mountain \ \ ^/ 200 kilometers 121° 119° T 117° 115° Fig. 1. Map showing the Owens Lake chain and sites discussed in the text (map is after Smith and Street-Perrott 1983). midden assemblages described here are impor- tant in deducing paleoenvironments of the region for several reasons. First, the middens occur within 10 m elevation of and <100 m away from the pluvial Owens Lake maximum highstand strandline. Second, the occurrence within several middens of plant macrofossils of Rocky Mountain juniper (Jimiperus scopulo- nim), a tree not found today in California, sug- gests that pluvial highstands during glacial retreats produced a unique microclimate at this nearshore location. The Site Owens Valley lies between the massive Sierra Nevada to the west and the Inyo-White Mountain ranges to the east. The Sierra Nevada 144 Great Basin Naturalist [Volume 54 rise from California's Great Valley, with a gen- tle westward gradient of 6% toward a lofty crest that contains some 500 summits over 3660 m, with 11 peaks over 4260 m. The east- ern escarpment of the Sierra plummets, with a ca 14% gradient, as much as 3050 m into the Owens Valley. The graben forming Owens Val- ley ranges ca 13-33 km wide, with an average elevation of only 1160 m. The eastern flank of the Owens Valley is bounded by the Inyo- White Mountain chain, with a crest elevation that averages ca 2900 m. Two indurated packrat middens, containing six stratigraphic units, were found at a single location on Haystack Mountain, ca 3 km east of the Inyo Mountains (36°36'N, 118°05'W; Fig. 1). The outcrop where the middens were found is of a spheroidally weathered Greta- ceous granite (Ross 1967) and faces southeast. At 1155 m the site is located ca 100 m north and 10 m above the Owens Lake highstand strandline (1145 m). Gurrently the local vegetation is dominated by a saltbush {Atriplex spp.)/hopsage {Graijia 67;jnos«)/sagebrush {Artemisia spp.) communi- ty on the valley floor, while wolfberry {Lyciiim andersonii), mallow {Sphaeralcea spp.), and various species of the grass family occupy the immediate rock outcrop. Greosote bush {Lar- rea tridentata) and bursage {Ambrosia dumosa) are found locally on well-drained sites, and greasewood {Sarcobatus vermiculatus) occurs on sites with alkaline soils and high water tables. Vegetation that occurs from 1150 m on the alluvial fans to 1950 m is represented by a Mojave Desert community dominated by cre- osote bush and bursage. Joshua tree {Yucca brevifolia), single-needle pinyon pine {Pinus monophijlla), and spiny menodora {Menodora spinescens) occur in a transition zone (1950- 2100 m) that trends into a pinyon-Utah juniper {Jiiniperus osteosperma) woodland at ca 2100-2900 m. In the southern Inyo Moun- tains subalpine trees are found only on peaks above ca 2900 m. Limber pine {Pimis flexilis) is common in this region, with lesser amounts of bristlecone pine {Pinus longaeva), fernbush {Chamaebatiaria millefoUium), and sagebrush. Several individuals of Sierra juniper {Junipe- 7'us occidentalis var australis) also grow in the Inyo Mountains (Vasek 1966). The nearest weather stations. Lone Pine (1120 m, 8 km west of the site) and Keeler (1100 m, 16 km southeast of the site), record an annual precipitation of 127 mm/yr and 80 mm/yr (Lee 1912, Elford 1970), respectively. Precipitation occurs primarily in the winter months with some rainfall in the summer months as isolated thunderstorms. Precipita- tion of 203-304 mm/yr has been estimated for the pinyon woodlands of the White Mountains at elevations of 1525-2135 m (St. Andre 1965). Methods Three layers (A-G) were found in each of the two middens (HMl and HM2) and were separated along stratigraphic planes. Once separated, the samples were disaggregated in distilled water in a covered bucket. This was done to prevent contamination by modern pollen. The disaggregated middens were sieved through a number 20-mesh (0.85-mm) screen with the decant saved for pollen analysis. The midden debris was air-dried and hand- sorted under a dissecting microscope (7-40X magnification). Plant macrofossils were identi- fied from reference materials. Interpretation of the macroremains is based on a relative abundance scale with 5 = >200 macrore- mains, 4 = 100-200, 3 = 30-99, 2 = 2-29, and a single specimen = 1 (Van Devender et al. 1987). Radiocarbon data were obtained pri- marily on fecal pellets (Table 1). Processing for fossil pollen followed Faegri and Iverson (1989) and included the addition of Lycopodium tracer tables (Stockmarr 1971), acetolysis, staining, and suspension in silicon oil. A 300-grain count (range = 262-361; Table 3) of terrestrial pollen types was made at 400X magnification. The count excluded tracer, deteriorated, and aquatic pollen types. Pollen percentages were calculated based on the total terrestrial pollen counted in each sample. Many of the pollen types were identified to Table 1. Radiocarbon analysis of the Owens Lake site (1155 111), Inyo Coimty, California. Sample Radiocarbon Dated Lab years B.P. material number HM2A 14,870 ± 130 Dung Beta-.39274 HM2C 16,010 ± 330 Dung Beta-36732 HM2B 17,680 ± 1.50 Dung Beta-35503 HMIA 20,590 ±210 Debris Beta-40000 HMIB 20,960 ± 240 Dung Beta-34833 HMIC 22,900 ± 270 Dung Beta-39273 1994] Shoreline Vegetation During Highstand 145 family; however, some types were broken into morphological categories. Piniis pollen was separated into the haploxylon (white pine) and diploxylon (yellow pine) groups. Ephedra pollen was divided into E. viridis and E. cali- fornica pollen types. Piirshia-Cercocarpiis pollen types were discriminated from Rosaceae, and Sarcobatus was separated from other members of the Chenopodiaceae-Afno- ranthus (Cheno-am) group. Results Macrofossils recovered from the middens document plants typically found in the pin- yon-juniper zone of the Inyo Mountains. The exception to this is the occurrence of Rocky Mountain juniper {Juniperus scopulorum), which does not occur today in California. Fos- sil pollen recovered from the middens repre- sents plants found within the midden as well as local species that either are avoided by packrats or occur beyond their foraging range (Anderson and Van Devender 1991). Midden macrofossils are represented by the presence of Utah juniper in all samples (Table 2). Green ephedra {Ephedra viridis), wild rose {Rosa woodsii), Menodora, and pin- yon pine occur in most of the other samples. Nevada greasebush {ForseUesia nevadensis) and Joshua tree also occur in several of the older middens (ca 22,900-20,590 yr BP). Rocky Mountain juniper is present in two middens dated to ca 17,680 and 16,010 yr BE Pollen identified from the middens gener- ally supports macrofossil evidence (Table 3). Exceptions to this are the high amounts of Artemisia (ca 6-50%) and moderate amounts of Cheno-ams (ca 5.5-18%). High variability within pollen percentages may be due to the uncertain association with deposition time (months to centuries) and the year-to-year variability in pollen production. Discussion During the Pleistocene several alpine glac- ier advances sculpted the Sierra Nevada, with at least three stages recorded during the late Wisconsin (Bursik and Gillespie 1993). The most recent episode, the Tioga advance, occurred during the full-glacial period, ca 21,000-18,000 yr BP Significant advances in glacial chronology have been made in the last decade. Experimental analysis of the accumu- lation of cosmogenic Cl-36 suggests that maxi- mum Tioga glaciation occurred prior to ca 21,000 yr BP (Phillips et al. 1990). Radiocar- bon dates of 21,000 ± 130 yr BP (Lebetkin 1980) from tufa underlying an alluvial fan of inferred Tioga-age at Owens Lake (Fullerton 1986) and of 19,050 ±210 yr BP on basal rock varnish from an outermost Tioga moraine in Pine Creek (Dorn et al. 1987) also support a maximum advance before this time. Timing on Sierra Nevada deglaciation is recorded by dates of glaciolacustrine sediments from mid- elevation west-side lakes (Swamp Lake ca 15,565 yr BP [1957 m; Batchelder 1980], Lake Moran ca 14,750 yr BP [2018 m; Edlund and Byrne 1991], Swamp Lake Yosemite ca 13,350 yr BP [1554 m; Smith and Anderson 1992]) and rock varnish dates on recessional Tioga moraines of ca 13,910 yr BP from Pine Creek (1830 m; Dorn et al. 1987). Dates from near the Sierran crest at Barrett Lake (2816 m) of ca 12,500 yr BP (Anderson 1990) and ca 10,300 yr BP in the Cottonwood Basin (ca 3000 m elevation; Mezger 1986) document high-elevation deglaciation on the east side. The presence of ice in the Sierra Nevada had a significant impact on paleoenvironments within Owens Valley. The Owens River water- shed covers ca 8500 km^, with nearly all of its runoff originating in the 16% of this area lying in the eastern Sierra Nevada (Lee 1912, Smith and Street-Perrott 1983). Thus, as melting glaciers retreated, lakes within the valley would periodically fill, overflowing to a down- stream lake in the chain. Based on glacial fea- tures, the glacial ice west of the Sierra Nevada crest increased the average elevation by ca 50 m in the south (Gillespie 1982, Mezger 1986) to as much as 300 m in the Yosemite National Park area (Alpha et al. 1987). Elevational in- creases east of the crest were insignificant because their glaciers were largely restricted to steep valleys. During the period of maxi- mum ice extent within the Sieira Nevada, the increased average elevation of the range, caused by the combination of upwards of ca 600 m of ice plus the ca 100 m lowering of sea levels, may have had two effects on the Owens Valley and Inyo-White Mountains to the east. First, the higher average elevation of the Sier- ra Nevada intensified the rainshadow effect, as witnessed by the limited glaciation within the Inyo-White Mountains (Elliott-Fisk 1985, Swanson et al. 1993). Second, accumulation of 146 Great Basin Naturalist [Volume 54 Table 2. Plant macrofossils identified fi-om the Owens Lake site (1155 m), Inyo County, California. Relati\e abun- dance is based on >200 specimens = 5, l()()-2()() = 4, 30-99 = 3, 2-29 = 2, and a single specimen = 1. Sample unit HM2A HM2C HM2B HMIA HMIB HMIC Sample age yr B.P. 14,870 16,010 17.680 20,590 20,960 22,900 Trees/shrubs JtiniiH-nis osfcosperrna 4 4 4 3 5 5 Jiinipcrii.s scopulorum — 3 2 — — — I'iiiits tnonoplujUa — — 2 2 3 2 E})lu'dra lirklis 2 — 2 5 5 5 Mcnodora spinescens 2 2 — — 2 2 Mirahilis bigelovii — — — 2 2 3 Eriogonmn d.fasiculatum — — 2 — — — Forsellesia nemdensis — — — 3 2 5 Artemisia tridcntata — — — — — 1 Chnjsothammis teretifolius 2 2 — 2 2 2 Ericameria cuneata 2 2 2 2 2 2 Tetradymia sp. — 2 — — — — Atriplex pohjcarpa — — — — 1 — Atriplex confertifolia 1 — — — — — Graijia spinosa — 1 — — — — Rosa woodii 2 2 2 — — 2 Coleogyne ramosissima — — — — 1 — Yucca cf. brevifolia — — — — 2 — Herbs Sphaeralcea ambigua 2 — — — 2 2 Cirsium sp. — — 2 1 2 — Boraginaceae — — — 2 — — Amsinkia sp. — — — — 2 — Cnjptantha sp. — — — — 2 — Plagijhothnjs spp. — — — — — 2 Salvia sp. — — — — 1 — Orthocarpiis sp. — 2 — — — — Succulent Opuntia hasilaris 1 — 2 2 4 4 Grass Oryzopsis hymenoides — — — 2 2 2 ice in the central Sierra Nevada probably deflected storm tracks further south than today and at a more frequent rate, as wit- nessed by wetter conditions in the modern Mojave Desert at that time (Spaulding and Graumlich 1986). While the lake-level fluctuations at Owens Lake are poorly known, the periods of high- stands and overflow can be estimated from the detailed records of pluvial Lake Russell and Searles Lake (Smith and Street-Perrott 1983, Benson et al. 1990). Owens Lake either received overflow from (Lake Russell) or con- tributed to (Searles Lake) pluvial lakes. Lake levels at Searles were generally high to over- flowing between ca 25,000 and 10,000 yr BP Between ca 21,000 and 15,000 yr BP a contin- uous highstand is inferred. Lake levels then returned to moderately low levels after ca 15,000 yr BP (Smith and Street-Perrott 1983) or ca 14,000 yr BP (Benson et al. 1990). For Lake Russell, lake-level chronologies suggest intermediate levels from at least 35,000 yr BP until a highstand after 15,000 (ca 14,000 or 13,000 yr BP; Benson et al. 1990). During the full-glacial, the Owens Lake midden site was located in a transitional posi- tion between the full-glacial single-needle pinyon-juniper woodlands of the Mojave Desert and the Utah juniper-limber pine woodland of the southern Great Basin. In the rain shadow of the Sierra Nevada, the Eleana Range (1810 m) records limber pine and steppe shrubs (Spaulding 1990). North of the Owens Lake site at slightly higher elevations, colder conditions are recorded by the occur- rence of Utah juniper and sparse limber pine at Eureka View (1430 m) at ca 14,700 yr BP (Spaulding 1990) and Utah juniper and Great Basin desert shrubs at the Volcanic Tablelands (Jennings and Elliott-Fisk 1993). Pinyon pine was not found in Death Valley where Utah juniper existed with a yucca semidesert (260- 1280 m; Wells and Woodcock 1985). South of 1994] Shoreline Vegetation During Highstand 147 Table 3. Percentages of identified pollen t\'pes fi-om the Owens Lakes site (1155 m), Inyo County, California. Sample unit IIM2A HM2C HM2B HMIA HMIB HMIC Sample age yr B.P. 14,870 16,010 17,680 20,590 20,960 22,900 Tracer 27.0 20.0 16.0 42.0 85.0 76.0 Deteriorated 5.0 30.0 10.0 15.0 13.0 3.0 Abies 0.0 0.3 0.0 0.0 0.3 0.0 Pimis haploxylon 9.4 9.7 3.0 9.2 30.4 9.0 Piiuis diploxylon 12.2 1.7 1.5 1.5 0.9 9.0 Cupressaceae 13.6 23.4 32.5 22.5 36.6 34.7 Ephedra viridis-type 0.6 2.5 0.6 0.4 1.2 1.0 Ephedra califonnca-type 0.8 1.4 0.0 0.0 0.0 0.3 Menodora 0.0 0.3 0.0 0.0 0.3 0.0 Syinplioricarpos 0.0 0.0 0.0 0.0 0.3 0.0 Qiierciis 0.0 0.0 0.0 0.0 0.3 0.0 Ambrosia 0.6 1.1 0.0 3.4 1.6 0.3 Artemisia 50.1 39.6 43.8 45.0 5.9 17.0 Cirsittm 0.3 0.0 0.0 0.0 0.0 0.0 Other Compositae 1.1 1.7 1.8 5.3 7.1 6.4 Cheno-ams 5.5 10.6 6.4 7.6 9.6 18.0 Sareobatus 4.7 5.6 8.8 4.6 2.8 4.2 Rosaceae 0.0 0.0 0.0 0.0 0.3 0.0 Purshia 0.0 0.0 0.0 0.0 0.9 0.0 Ceanothus 0.0 0.0 0.0 0.0 0.3 0.0 Eriogonum 0.3 0.6 0.3 0.0 0.0 0.0 Solanaceae 0.0 0.0 0.9 0.0 0.0 0.0 Cruciferae 0.0 1.1 0.0 0.0 0.3 0.0 Leguminosae 0.6 0.0 0.0 0.4 0.6 0.0 Polemonaceae 0.0 0.0 0.3 0.0 0.0 0.0 Gramineae 0.3 0.6 0.0 0.0 0.0 0.0 Terrestrial total 361 359 329 262 322 311 Aquatics Typha 0.0 0.0 0.0 0.0 1.0 0.0 the Owens Lake site, the Mojave Desert full- glacial vegetation records the widespread occurrence of a pinyon-juniper woodland (Spaulding 1990). Records from the Owens Lake site (1155 m) and Skeleton Hills (925 m; Spaulding 1990) are the only documentation of pinyon pine during the full-glacial at this latitude. The lower limit of pinyon pine is recorded in the Skeleton Hills at 925 m. In Owens Valley the upper limit of pinyon is constrained between 1155 m (this report) and 50 km north in the Volcanic Tablelands at 1340 m (Jennings and Elliott-Fisk 1993). Despite the absence of pinyon at Death Valley, these sites define the northern distribution of pinyon in the Mojave Desert during the full-glacial. The most interesting macrofossil found in the midden series dating 17,680 and 16,010 yr BP is Rocky Mountain juniper. This tree is not found in California today but occurs in the Charleston Mountain area of southwestern Nevada, ca 225 km east of the site. The eleva- tional and latitudinal migration of Rocky Mountain juniper is well understood in the southeastern and central Great Basin (Thomp- son 1990). Using terpene variations, Adams (1983) provided evidence for Rocky Mountain juniper colonization in post-glacial environ- ments within the extreme northern and south- ern extensions of its range, suggesting migra- tion routes along pluvial lake corridors. Rocky Mountain juniper is generally restricted to regions that lack pronounced summer droughts (West et al. 1978, Thompson 1988). In the southern part of its range, Rocky Mountain juniper is restricted to riparian set- tings or areas of shallow groundwater and springs (Adams 1983). This information is ger- mane to the history of lake-level fluctuations within the Owens Valley area. The occurrence of Rocky Mountain juniper at the Owens Lake site is thus partially ex- plained by local climatic factors associated with pluvial lake highstands. Its existence around Owens Lake between ca 17,680 and 16,010 yr BP was probably influenced by relatively high water tables or locally humid conditions 148 Great Basin Naturalist [Volume 54 associated with the highstand at that time. Its suhsequent absence by 14,870 \r BP was probably a result of declining lake levels. Conclusions The Owens Lake midden site provides evi- dence for paleoenvironmental change along the shore of Owens Lake spanning the full- glacial Tioga advance. The midden sequence from ca 23,000 to 17,680 \r BP records a juni- per-pinyon woodland with associated xeric upland desert scrub and possible Joshua tree. The presence of Rocky Mountain juniper at 17,680 and 16,010 yr BP suggests a mesophyt- ic association due to the presence of Owens Lake. In apparent contradiction, drier condi- tions are recorded after ca 17,500 yr BP at nearby locations (Death Valley, Wells and Woodcock 1985; Skeleton Hills, Spaulding 1990; Sheep Range, Spaulding 1981), and this is supported at the Owens Lake site as pinyon pine is not recorded after 17,680 yr BP Dated moraines record the timing of the Tioga glaciation (Dorn et al. 1987, Bursik and Gillespie 1993). Prior to ca 19,000 yr BP plu- vial lake highstands are not recorded (Bursik and Gillespie 1993), as some available mois- ture was sequestered to the Owens Lake chain in the alpine ice of the Sierra Nevada. A deglaciation with possible readvances, between ca 19,000 and 13,000 yr BP caused the lakes of the Owens River to achieve high- stands. The close proximity of the Owens Lake highstand allowed sufficient effective moisture for Rock)' Mountain juniper to exist close to the midden site (within 10 m eleva- tion and <100 m distance from the lake). Acknowledgments We wish to thank R. S. Thompson and T R. Van Devender for assistance in identifying several of the plant macrofossils, and Julio Betancourt, Paul Tueller, Susie Smith, and an anonymous reviewer for their helpful editorial suggestions. Robyn O'Rielly drafted Figure 1, and Chris Force provided identification of Neotoma sp. This is Laboratoiy of Paleoecolo- gy contribution 43. Literature Cited Adams, R. P. 1983. Infraspecific terpenoid variation in Jiiniix'nis scoptilorum: evidence for Pleistocene rcfugia and recolonization in western North Ameri- ca. Taxon 32: 30-46. Alpha, T. R., C. Wahrhaftig, and N. K. Hubek. 1987. Oblique map showing maximum extent of 20,000- year-old (Tioga) glacier, Yosemite National Park, central Sierra Nevada, California. USGS Miscella- neous Investigations Series Map 1-188.5. Anderson, R. S. 1990 Holocene forest develoi^ment and paleoclimates within the central Sierra Nevada, Cal- ifornia. Journal of Ecology 78: 470-489. Anderson, R. S., and T. R. Van Devender. 1991. Com- parison of pollen and macrofossils in packrat (Neotoma) middens: a chronological sequence from the Waterman Mountains of southern Arizona, USA. Review of Paleobotan\^ and Pal\ nology 68: 1-28. Batchelder, G. L. 1970. Post-glacial fluctuations of lake levels in Adobe Valley, Mono County, California. Page 7 in Abstracts of the first meeting of the Ameri- can Quatemaiy Association. . 1980. A late Wisconsinan and early Holocene lacustrine stratigraphy and pollen record from the west slope of the Sierra Nevada, California. Page 13 in Abstracts and programs — American Quaternary Association. Sixth biennial meeting, Orono, Maine. Benson, L. V., et al. 1990. Chronolog\' of expansion and contraction of four Great Basin lake systems during the past 3.5,000 years. Palaeogeography, Palaeocli- matology, Palaeoecologv' 78: 241-286. Birkeuand, p. W. and R. M. Burke. 1988. Soil catena chronosequences on eastern Sierra Nevada moraines, California, USA. Arctic and Alpine Research 20: 473-184. Bursik, M. I., and A. R. Gillespie. 1993. Late Pleis- tocene glaciation of Mono Basin, California. Quater- nary' Research 39: 24-35. Dorn, R. I., B. D. Turrin, A. J. T. Jull, T. W. Linick, and D. J. Donahue. 1987. Radiocarbon and cation- ratio ages for rock varnish on Tioga-Tahoe morainal boulders of Pine Creek, eastern Siena Ne\ada, Cali- fornia, and their paleoclimatic implications. Quater- nary Research 28: 38-49. Edlund, E. G., and R. Byrne. 1991. Climate, fire, and late Quaternary vegetation change in the central Sierra Nevada. Pages 390-396 in S. C. Nodvin and T. a. Waldrop, eds.. Fire and the environment: eco- logical and cultural perspectives. Department of the Interior, National Park Sendee, Washington, D.C. Elford, C. R. 1970. The climate of California. Pages .538-594 in NOAA, ed.. Climates of the United States. Vol. 2. Elliott-Fisk, D. L. 1985. Quaternary' dynamics of the White Mountains. Pages 47-.50 in C. A. Hall, Jr., and D. J. Young, eds.. Natural histon- of the White- Inyo Range, eastern California and western Nevada, and high altitude physiology'. University of Califor- nia, White Mountain Research Station Symposium. Vol. 1. Faegri, K., and J. IVERSEN. 1989. Textbook of pollen analysis. Hafner Press, New York. FULLERTON, D. S. 1986. Chronology and correlation of glacial deposits in the Sierra Nevada, California. Quaternary' Science Review 5: 161-169. Gillespie, A. R. 1982. Quatemaiy glaciation and tecton- ism in the southeastern Sierra Nevada, Inyo County, California. Unpublished doctoral dissertation, Cali- fornia Institute of Technology-, Pasadena. 1994] Shoreline Vegetation During Highstand 149 Jennings, S. A., and D. L. Elliott-Fisk. 1993. Packrat midden evidence of late Quaternary vegetation change in the White Mountains, CaHfornia-Nevada. Quaternaiy Research 39: 214-221. KOEHLER, p. A., AND R. S. ANDERSON. 1990. Abstract: Late Wisconsin to recent vegetation changes from the Alabama Hills, Owens Valley, California. San Bernardino County Museum Association Quarterly 36. Lajoie, K. R. 1968. Quatenian,' stratigraphy and geologic liistorA' of Mono Basin, eastern California. Unpub- lished doctoral dissertation, University of California, Berkeley. Lebetkin, L. K. 1980. Late Quatemaiy activity' along the Lone Pine fault, Owens Valley fault zone, California. Unpublished master's thesis, Stanford University, Palo Alto, California. Lee, C. H. 1912. An intensive study of the water resources of a part of Owens Valley, California. USGS Water Supply Paper 124. Leopold, E. B. 1967. Summary of palynological data from Searles Lake. Pages 52-66 in Pleistocene geol- ogy and palynolog>', Searles Lake Valley, California. Friends of the Pleistocene guidebook. Mezger, E. B. 1986. Pleistocene glaciation of Cotton- wood Basin, southeast Sierra Nevada, California. Unpublished masters thesis, Universitx' of Southern Calitomia. Pakiser, L. C, M. F. Kane, and W. H. Jackson. 1964. Structural geology and volcanism of Owens Valley region, California — a geophxsieal study. USGS Pro- fessional Paper 348. Phillips, F. M., M. G. Zreda, S. S. S.mith, D. Elmore, P. W. KUBIK, AND P. Sharma. 1990. Cosmogenic chlorine-36 chronology for glacial deposits at Bloody Canyon, eastern Sierra Nevada. Science 248; 1529-1532. Putnam, W. C. 1950. Moraine and shoreline relationships at Mono Lake, California. Geological Society of America Bulletin 61: 115-122. Ross, D. C. 1967. Generalized geologic map of the Inyo Mountains region, California. USGS Miscellaneous Geologic Investigations Map 1-506. Smith, G. L, and F. A. Street-Perrott. 1983. Pluvial lakes of the western United States. Pages 190-212 /'/i S. C. Porter, ed., Late Quaternary of the United States. University of Minnesota Press. Smith, S. J. and R. S. Anderson. 1992. Late Wisconsin paleoecologic record from Swamp Lake, Yosemite National Park, California. Quaternary Research 38: 91-102. Spaulding, W. G. 1981. The late Quaternary vegetation of a southern Nevada mountain range. Unpublished doctoral dissertation. University of Arizona, Tucson. . 1990. Vegetational and climatic development of the Mojave Desert: the last glacial maximum to the present. Pages 166-199 in J. L. Betancourt, T. R. Van Devender, and P. S. Martin, eds., Packrat mid- dens. The last 40,000 years of biotic change. Univer- sity of Arizona Press, Tucson. Spaulding, W. G., and L. J. Graumlich. 1986. The last pluvial climatic episodes in the deserts of southwest- ern North America. Nature 320: 441-444. St. Andre, G. L. 1965. The pinyon woodland zone in the White Mountains of California. American Midland Naturalist 73: 257-239. Stock.marr, J. 1971. Tablets with spores used in absolute pollen analysis. Pollen et Spores 13: 615-621. Swanson, T. W., D. L. Elliott-Fisk, and R. S. Southard. 1993. Soil development parameters in the absence of a elironosequence in a glaciated basin of the White Mountains, California-Nevada. Quater- nary Research 39: 186-200. Thompson, R. S. 1988. Vegetation dynamics in the west- ern United States: modes of response to climate fluctuations. In: B. Huntley and T. Webb III, eds., Vegetation histor\'. Kluwer Academic Publishers. . 1990. Late Quaternary vegetation and climate in the Great Basin. Pages 200-239 in J. L. Betancourt, T. R. Van Devender, and P. S. Martin, eds., Packrat middens. The last 40,000 years of biotic change. University of Arizona Press, Tucson. Van Devender, T. R., R. S. Thompson, and J. L. Betan- court. 1987. Vegetation history in the southwest: the nature and timing of the late Wisconsin- Holocene transition. Pages 323-352 in W. F. Ruddi- man and H. E. Wright, Jr., eds.. North America and adjacent oceans during the last deglaciation. GSA, Boulder. Vasek, F. C. 1966. The distribution and taxonomy of three western junipers. Brittionia 18: 350-372. Wells, P. V., and D. Woodcock. 1985. Full-glacial veg- etation of Death Valley, California: juniper wood- land opening to yucca semidesert. Madroiio 32: 11-23. West, N. E., R. J. Tausch, K. H. Rea, and P. Tueller. 1978. Phytogeographical variation within juniper- pinyon woodlands of the Great Basin. Great Basin Naturalist Memoirs 2: 119-136. Received 17 June 1993 Accepted 4 J aniianj 1994 Great Basin Naturalist 54(2), ©1994, pp. 150-155 REDBAND TROUT RESPONSE TO HYPOXIA IN A NATURAL ENVIRONMENT Mark Vinson^ and Steve Levesque^ Abstract. — Redband trout iOucorhijnchtis mykiss gairdneri) were observed appro.ximately every 2 weeks in an intermittent southwest Idaho stream between August and December 1991. Instantaneous daytime dissolved o.xygen concentration and water temperature declined from 4.0 to <2 mg/L and 17 to 2°C, respectively, during this period. Redband trout declined from a maximum captured of 48 on 28 August to I on 8 November in one series of pools. As conditions approached hypoxia, trout exhibited little movement and positioned themselves in water just deep enough to cover their dorsal fin. High densities of speckled dace {Rliinichthijs osculus) were also present in each pool until dr\'ing. The response of these fish to such extreme habitat conditions is probably a primary factor accounting for their distribu- tion within arid landscapes. Key words: Oncorhynchus mykiss gairdneri, redband trout. Rliinichthys osculus, oxygen tolerance, aquatic surface respiration, intermittent streams, desert fishes. Headwater streams often become intermit- tent during summer (Williams 1987). Fish unable to migrate to perennial reaches are trapped in isolated pools where they may be subjected to lethal conditions. Observations of behavioral responses and survival of stream fishes in these conditions are few (Tramer 1977, Matthews et al. 1982, Mundahl 1990). Conditions associated with isolated pools in intermittent streams include lack of space and cover (Capone and Kushlan 1991), widely fluc- tuating and often lethal pH (Capone and Kushlan 1991), temperature (Huntsman 1942, 1946, Bailey 1955, John 1964, Matthews et al. 1982, Mundahl 1990), and dissolved oxygen (Tramer 1977). The ability of fish to survive harsh habitat conditions {sensu Matthews 1987) has been attributed to physiological (Matthews 1987) and behavioral adaptations (Kramer 1987). Arid-land rainbow trout {Oncorhynchus mykiss) occur in southern Oregon, southwest- ern Idaho, and northern Nevada. Behnke (1992) believes these fish are a form of Colum- bia River redband trout (O. m. gairdneri) that have adapted to arid-land streams character- ized by extremes in stream flow, temperature, and dissolved oxygen; little is known of their life history or ecology (Kunkel 1976, Behnke 1992). This paper describes the demise of most and the survival of a few redband trout under low dissolved oxygen concentrations in an intermittent stream in southwest Idaho. Study Area The study was conducted on Sinker Creek, a second-order tributary of the Snake River in southwestern Idaho, approximately 1 km from the confluence with the East Fork of Sinker Creek (T4S, R2W, Sec. 19, 20, Owhyee Coun- ty, Idaho). Elevation of the study area is 1100 m. The geomorphology of the area is charac- terized by coarse alluvial fill interspersed with bedrock in a basalt canyon. Riparian areas are mostly unvegetated, except immediately adja- cent to pools where willow {Salix sp.) clumps overhang the stream channel. Dewatered streambed areas are unvegetated. The water- shed is subjected to summer livestock grazing. During 1991, streams throughout south- western Idaho were flowing at far below nor- mal levels (U.S. Geological Survey 1992). Sinker Creek upstream of the confluence with the East Fork of Sinker Creek was dry with the exception of a few isolated reaches. Data are presented from the largest wetted reach that consisted of five pools (A-E, sequentially, downstream) separated by shallow riffle areas along a 300-m reach. ^USDI Bureau of Land Management, Acjuatic Ecosystem Laboraton , P'isheries and Wildlife Department, Utiih State Ur 2R.D. 1, Gillead Road, Randolph, Vermont 05060. iit>'. Logan, Utah 84322-5210. 150 1994] Redband Trout Hypoxia 151 Methods Fish populations and habitat conditions were monitored approximately eveiy 2 weeks between 14 August and 5 December 1991. Fish behavior was observed for 30-60 min prior to sampling. High-water clarity and lim- ited pool area and depth facilitated direct observations, especially later in the study. Beginning 28 August, a Smith-Root battery- powered backpack electrofisher (Model llA) was used to collect all redband trout in each pool. Total length and weight were recorded (weights were not taken prior to 24 October), and trout >100 mm were differentially fin clipped to identify fish from individual pools (there was a distinct size class break at 100 mm; trout < 100 mm are herein referred to as small and trout > 100 mm as large). During each visit water temperature and dissolved oxygen (DO, mg O2/L) were mea- sured with a YSI (Yellow Springs Instrument Company, Inc.) Model 57. Measurements were made between 1000 and 1430 h on all dates at several locations throughout each pool at surface, mid-depth, and bottom. Pool surface area was determined as a rec- tangle by averaging several length and width measurements; depth was measured at the same locations as temperature and DO. Mea- surements continued until a pool dried or until 5 December when continuous flow returned to the stream. On 14 August 1991 pieces of plastic flagging were placed in the dry streambed between pools and weighted down with leaves so flow events between sam- pling dates could be detected; none were detected prior to 5 December Results On 14 August 1991 one redband trout was observed in pool B and three in pool D; numerous speckled dace {RJiinichthys osculus) were observed in all pools. Water temperature and DO ranged from 7°C and 1.5 mg/L, where water emerged from the substrate, to 28.5 °C and 11.5 mg/L at the end of the wet- ted reach downstream of pool E. On 28 August 48 redband trout were cap- tured: 1 in pool A, 23 in pool B, 9 in pool C, 15 in pool D, and 0 in pool E. Pool areas ranged from 8 to 27 m^. Water temperature and DO ranged from 16.8 to 18.5 °C, and 1.7 to 4.6 mg/L, respectively, at the up- and downstream ends of the wetted reach. Between 28 August and 11 October all pools and riffles went dry with the exception of pool D. No trout were found in any pool other than pool D after 28 August. Numerous (>100) speckled dace (mean length 25 mm) were present in each pool until drying. Between 28 August and 5 December the volume and surface area of pool D decreased from 10.5 to 0.5 m-^ and 27 to 4 m^, respec- tively. Maximum pool depth decreased from 81 to 18 cm, and mean pool depth went from 56 to 13 cm. Dissolved oxygen decreased from 3.7 mg/L on 28 August to 1.6 mg/L on 1, 6, and 11 October Between 24 October and 22 November, DO increased slightly to 2.0-3.0 mg/L. Water temperature dropped throughout the study period (Fig. 1) and was less than 10 °C after 6 October. On all dates DO was slightly higher at the surface than at the bot- tom (mean difference = 0.2 mg/L). On 22 November a thin layer of ice covered half of pool D. On 5 December, when continuous flow returned to Sinker Creek, water temper- ature was 5°C and DO was 9.5 mg/L (Fig. 1). Redband trout in pool D declined from a maximum of 33 captured on 20 September to 1 on 16 November (Fig. 1). The increase in trout captured between 28 August and 20 September was probably due to increased capture efficiency caused by a reduction in pool area and not from relocation of trout from other pools. No marked trout were captured in a pool different from the one in which they were originally captured. Recapture success was ca 50% in September and October and 100% in November. In November there was litde space left in pool D for redband trout to hide, and the pool was more accessible to electrofishing. On several occasions dead redband trout were found along the edge of pool D: one on 1 October, two on 11 October, and five on 8 November. No dead redband trout were recovered from any other pools. Dead red- band trout had not been scavenged, and no sign of scavengers was apparent during any site visit. Of the missing redband trout, most were not accounted for, and those that were found were all > 140 mm. Smaller trout may have been buried in the extensive decaying leaf material present in each pool. On 1 Octo- ber two trout died and on 11 October one trout died as a result of electrofishing. 152 Great Basin Naturalist [Volume 54 • POOL AREA A POOL VOLUME Fig. 1. Sinker Creek pool D volume, area, mean depth, air and water temperature, dissolved oxygen concentration and percent saturation, and trout numbers and densities (#/ni-) between 14 August and 5 December 1991. Trout num- bers and densities on 8, 16, and 22 November include trout transferred from the East Fork of Sinker Creek, n are trout >100 mm (large). H are trout <100 nmi (small). There was a distinct size class break at 100 mm. Numbers above bars are recaptures of previously marked trout (>100 mm). Pool dimensions were not measured on 20 September, 6 Octo- ber, and 5 December. Trout were not collected on 14 August, 16 September, 6 October, 12 November, and 5 Decem- ber. 1994] Redband Trout Hypoxia 153 Prior to 24 October redband trout were observed swimming and surfacing throughout the pool. Approximately 60% of the pool could be effectively observed up to this date. Between 11 October and 24 October pool D shrank by 83% (Fig. 1), and the entire pool could now be observed. From this date on, redband trout were generally biding beneath small boulders in the middle of the pool or among the leaf debris along the shallow pool margins. Trout were rarely seen swimming around the pool. After being captured and returned to the pool, trout typically moved to shallow water along the shoreline and faced outward. While positioned along the pool edge, they occupied water just deep enough to cover their backs and "gulped" at the air- water interface. They remained in this posi- tion for at least 1 h. This post-capture behav- ior was probably a response to the stress of electrofishing. Speckled dace, whose densities were high throughout the study, showed no obvious change in behavior during the four months of obsei"vation. On 8 November six large redband trout (145-222 mm) were captured by electrofish- ing and transferred from the perennial East Fork of Sinker Creek to pool D. Water tem- perature in the East Fork was 8.5 °C and DO was 9.8 mg/L. In pool D water temperature was 6.5 °C and DO was 3.0 mg/L. This trans- fer took ca 15 min. Immediately upon being placed in pool D, the redband trout moved to shallow margins of the pool and began "gulp- ing' at the air- water interface. Over the next hour they remained along the edge of the pool facing outward with their dorsal fin just break- ing the water surface. Opercular movements remained rapid during this time. The one original pool D large redband trout remained under a boulder in the middle of the pool and did not interact with the transplanted redband trout; the two remaining small trout were not observed. On 12 November water temperature was 5.5 °C and DO was 2.0 in pool D. Because the electrofishing unit was not functional, no red- band trout were captured, but several were observed along the pool margins facing out- ward. On 16 November seven large redband trout were collected (one original pool D trout and the six East Fork redband trout). The two remaining small redband trout could not be found. The six transplanted East Fork redband trout had all lost weight (1-8 g) since being transferred 8 days earlier; the accuracy of the scale was ca ±1 g. On 22 November three of the transplanted redband trout maintained the same weight as on 16 November, one lost 1 g, and two could not be found. The weight of the original pool D redband trout declined from 66 g on 11 and 24 October to 63 g on 8 November, 62 g on 16 November, and 60 g on 22 November Discussion One of 48 redband trout sun'ived at least 114 days in DO <4 mg/L, and 4 redband trout transplanted from the perennial, high DO (9.8 mg/L) East Fork of Sinker Creek sur- vived 43 days at DO <2.5 mg/L. Water tem- perature declined from 17 to 2°C during this period. Additional survival would probably have occurred if not for repeated electrofish- ing. The ability of arid-land redband trout to withstand harsh habitat conditions has been suggested by Wishard et al. (1984) and Behnke (1992). Behnke (1992) reported fish- ing for and catching arid-land redband trout in intermittent stagnant pools in Oregon. There are a few studies on water temperature tolerance of native western trout species (e.g., Lee and Rinne 1980), but no field observa- tions have been previously made of native trout responses to low DO. Effects of low DO on rainbow trout were summarized by Davis (1975). Negative effects of low DO first become apparent at 5-6 mg/L (Davis 1975). Responses of adult rainbow trout to DO <5 mg/L or <50% saturation include elevated breathing amplitude and buccal pressure (Hughes and Saunders 1970), reduced heart rate (Randall and Smith 1967), reduction in swimming speeds (Kutty 1968), and reduced capacity for anaerobic metabo- lism (Kutty 1968). A possible explanation for the survival of redband trout in this study might be that a seep with higher DO was entering this pool. We looked for such a source by measuring temperature and DO throughout the pool on each sampling date and by digging shallow groundwater wells into the streambed on 8 November There was little variation in tem- perature and DO within pools for any date. Groundwater temperature was 2.5 °C and DO 154 Great Basin Naturalist [Volume 54 was 1 mg/L. Groundwater t\picall\' contains little or no dissolved oxygen (Hem 1985). Probable factors contributing to fish sur- vival included a long acclimation period (Shepard 1955, Davis 1975), especially when compared to laboratory studies, sedentary behavior (Davis 1975), declining air and water temperatures as the study progressed (Fig. 1), and lack of water velocity in either pool so that energy expenditures would have been minimal (Davis 1975). Sinker Creek probably became intermittent in iMay or June, after snowmelt runoff. Fish unable to escape down- stream to the perennial East Fork of Sinker Creek had many weeks to become acclimated to the gradually deteriorating conditions. Increased acclimation improves trout sur- vival in low DO. Shepard (1955) showed that with sufficient acclimation time lethal DO lev- els for brook trout {Salveliniis fontinalis) lai-vae could be reduced to 1.05 mg/L at 8°C. Com- plete survival of larvae was achieved at the approximate acclimation rate of 70 h per each 1.0 mg/L decrease in O2 at 9°C; fish in Sinker Creek probably had a more gradual acclima- tion. In the laboratory most fish species tend to move away from low DO areas and occupy higher locations in the water column (Kramer 1987), where DO is generally higher due to diffusion at the air-water interface. Early in the study redband trout were frequently seen swimming near the surface. As pool D shrank, most trout were observed in shallow water along the edge of the pool. Tramer (1977) observed Johnny darters {Etheostoma nigrum) lined up around pool margins with their heads facing outward in an oxygen-depleted pool. This position would provide easier access to the surface film where diffusion maintains higher levels of dissolved oxygen (Lewis 1970). The use of the surface film by fish is a response to low dissolved oxygen levels and is known as aquatic surface respiration (ASR, Lewis 1970). The ability to perform ASR is common among fishes living in hypoxia-prone waters (Kramer 1983), but it has not been reported for salmonids. Shepard's (1955) observations of trout immediately surfacing upon the introduction of oxygen-deficient water to their tanks is the only evidence we could locate of salmonids performing ASR-type behavior. Gee et al. (1978) subjected rainbow trout, arctic char (Salvelimis (il])ituis), and lake whitefish (Core- gonus clupcafuiiiiis) to progressive hypoxia, and none exhibited ASR behavior. They felt salmonids might not have evolved this behav- ior because they typically occupy well-oxy- genated waters. Arid-land redband trout may be an exception (Behnke 1992). A major factor enhancing redband trout survival was that as the pool shrank, the pro- portion of groundwater flow to surface water flow increased and air temperatures declined. Increased influence of groundwater and lower air temperatures reduced water temperatures, diel fluctuations in dissolved oxygen, and fish metabolic rates. Tramer (1977) found that fish mortality in isolated pools was highest during periods of maximum water temperature and diel DO fluctuation. Whitmore et al. (1960) observed juvenile chinook salmon (O. tshawytscha) strongly avoiding low DO (1.5-4.5 mg/L) areas in the summer when water temperatures were high but not in the autumn when water temperatures were cool- er. The ability of transplanted East Fork red- band trout to suiA/ive in pool D was especially surprising. We initially expected that with their lack of acclimation to low DO levels, pool D would be lethal (Davis 1975). Although they appeared stressed initially, four of the six trout survived 43 days until perennial flow returned. Immediately upon being placed in pool D, these fish appeared to perform ASR. On subsequent visits they were usually seen lying along the pool margin partially covered by fallen leaves. Their breathing rate appeared relaxed and was much slower than when they were introduced into pool D. Lowe et al. (1967) found greater sundval of smaller-bodied native Arizona fishes in low oxygen concentrations than larger-bodied fish- es. Shepard (1955), however, found that larger trout (12 g) tended to live longer at lethal DO levels than larval fish (1 g). In this study a large (146 mm, 60 g) redband trout oudived all small (<100 mm) redband trout in pool D, while just downstream in another pool of simi- lar size eight small and no large redband trout survived (Vinson unpublished data). Speckled dace survived in high numbers in all pools until dning. Although most of the redband trout died during our study, their ability to survive even a short time in these extreme habitat conditions. 1994] Redband Trout Hypoxia 155 including our repeated electrofishing, which we feel may have been a source of delayed mortality, is noteworthy. The use of surface film water (ASR-type behavior) may be a strat- egy these fish use to survive periods of hypox- ia in their harsh desert stream environment. Future study is needed to better describe the physiological tolerances of this desert fish species. Acknowledgments We thank Ted Angradi, Robert Behnke, Jeff Kershner, Peter Moyle, and Jack Williams for their constructive comments on earlier drafts of this manuscript. Literature Cited Bailey, R. M. 1955. Differential mortality from high tem- perature in a mixed population of fishes in southern Michigan. Ecology 36: 526-528. Behnke, R. J. 1992. Native trout of western North Ameri- ca. American Fisheries Society Monograph 6. Bethesda, Maryland. Capone, T. a., and J. A. KusHLAN. 1991. Fish community structure in dry-season stream pools. Ecology 72: 983-992. Davis, J. C. 1975. Review of the o.xygen requirements of aquatic life. Journal of the Fisheries Research Board ofCanada 32: 2295-2332. Gee, J. H., R. F. Tallm,\n, and H. J. Smart. 1978. Reac- tions of some Great Plains fishes to progressive hypoxia. Ganadian Journal of Zoology 56: 1962-1966. Hem, J. D. 1985. Study and interpretation of the chemical characteristics of natural water. United States Geo- logical Survey Water Supply Paper 2254. Hughes, G. M., and R. L. Saunders. 1970. Responses of the respiratory pump to hypoxia in the rainbow trout [Sahno gairdneri). Journal of Experimental Biology 53: 529-545. Huntsman, A. G. 1942. Death of salmon and trout with high temperature. Journal of the Fisheries Research Board ofCanada 5: 485-501. . 1946. Heat stroke in Canadian Maritime stream fishes. Journal of the Fisheries Research Board of Canada 6: 476-482. John, K. R. 1964. Survival of fish in intermittent streams of the Chiricahua Mountains, Arizona. Ecology 45: 112-119. Kramer, D. L. 1983. Aquatic surface respiration in the fishes of Panama: distribution in relation to risk of hypoxia. Environmental Biology of Fishes 8: 49-54. . 1987. Dissolved oxygen and fish behavior. Envi- ronmental Biology of Fishes 18; 81-92. KUNKEL, C. M. 1976. Biology and production of the red- band trout [Sahno sp.) in four southwesteiTi Oregon streams. Unpublished master's thesis, Oregon State University, Con'allis. KUTTY, M. N. 1968. Respiratory quotients in goldfish and rainbow trout. Canadian Journal of Zoology 46: 647-653. Lee, R. M., and J. R. Rinne. 1980. Critical thermal maxi- ma of five trout species in the southwestern United States. Transactions of the American Fisheries Soci- ety 109: 632-635. Lewis, W. M., Jr. 1970. Morphological adaptations of cyprinodontoids for inhabiting oxygen deficient waters. Copeia 1970: 319-326. Lowe, C, F. Hinds, and E. Halpern. 1967. Experimen- tal catastrophic selection and tolerances to low oxy- gen concentration in native Arizona freshwater fish- es. Ecology 48: 1013-1017. Matthews, W. G. 1987. Physicochemical tolerance and selectivity of stream fish as related to their geo- graphic ranges and local distributions. Pages 111-120 in W. J. Matthews and D. C. Heins, eds.. Community and evolutionary ecology of North American stream fish. University of Oklahoma Press, Norman. Matthews, W. G., E. Surat, and L. G. Hill. 1982. Heat death of the orangethroat darter Etheostoma spectahile (Percidae) in a natural environment. Southwest Naturalist 27; 21&-217. Mundahl, N. D. 1990. Heat death of fish in shrinking pools. American Midland Naturalist 123: 40—46. Randall, D. J., and J. C. Smith. 1967. The regulation of cardiac activity in fish in a hypoxic environment. Physiology Zoology 40: 104-113. Shepard, M. p. 1955. Resistance and tolerance of young speckled trout (Salvelinus fontinalis) to oxygen lack, with special reference to low oxygen acclimation. Journal of the Fisheries Research Board of Canada 12: 387-446. TR.A.MER, E. J. 1977. Catastrophic mortality of stream fish- es trapped in shrinking pools. American Midland Naturalist 97; 469-478. U. S. Geological Survey. 1992. USGS water resources data — Idaho. U.S. Geological Survey, Water Resources Division (USGS/WRD/HD-92/278), Boise, Idaho. Whitmore, C. M., C. E. Warren, and P. Doudoroff. 1960. Avoidance reactions of salmonids and centrar- chid fishes to low o.xygen concentrations. Transac- tions of the American Fisheries Society 89; 17-26. Williams, D. D. 1987. The ecology of temporar>' waters. Timber Press, Portland, Oregon. WiSHARD, L. N., J. E. Seeb, and F. M. Utter. 1984. A genetic investigation of suspected redband trout populations. Copeia 1984; 120-132. Received 16 February 1993 Accepted 3 December 1993 Great Basin Naturalist 54(2), ©1994. pp. 156-161 FIELD STUDY OF PLANT SURVIVAL AS AFFECTED BY AMENDMENTS TO BENTONITE SPOIL Daniel W. Uresk^ and Teruo Yanuinioto^ AliSTIUCT. — Efforts to rt-claim aiiicMided and raw bentonite spoils with six plant species (two forbs, three shrubs, and one tree) were evaluated over a 4-year period. Plant species included fourwing saltbush {Atriplex canescens [Pursh] Nutt.), big sagebrush (Artemisia triclcntata tridentuta Nutt.), Rocky Mountain juniper ijimipenis scopiilorum Sarg.), Russian olive (Elaeagnus an'^mtifolia L.), common yarrow [Achillea millifnlium L.), and scarlet globemallow {Sphaeralcea coccinea [Pursh] Rydb.). Spoil treatments included addition of gypsum, sawdust, perlite. straw, and venniculite; the con- trol treatment was unamended. Founving saltbush had 52% survival across all spoil treatments, with greatest survival occurring on perlite-treated spoil (80%), followed by gypsum (70%) and venniculite amendments (70%). Survival of other plant species ranged from 0 to 3% averaged across all treatments after 4 years. No differences in plant survival occurred among amendments when all species were considered. Key words: shrubs, forbs, trees, Wyoininp., mining, reclamation. Bentonite, a montmorillonite clay, is a term referring to an altered deposit of volcanic ash (Barchardt 1977). In the northern Great Plains, bentonite is strip-mined from the Cre- taceous marine Pierre and Mowry shale and Belle Fourche formations. Because of the spoils' saline sodic quality, high shrink-swell characteristics, limited internal drainage, arid and semiarid climate of the region, and absence of irrigation water, attempts at reveg- etation have faced restrictive problems (Hem- mer et al. 1977, Bjugstad 1979). In the absence of drainage and leveling possibilities, Russell (1973) suggested the planting of salt-tolerant Atriplex species. Voorhees et al. (1987) reported successful growth of rillscale (A. stickleyi [Torrey] Rydb.), a native annual, on bentonite mine spoil with amendments and supported use of this species for early revegetation on such spoils. Nutri- tional qualities of rillscale are generally ade- quate for livestock and wildlife (Voorhees 1990). Sieg et al. (1983) also reported that rillscale was the most successful invader of bentonite-mined land spoil in southeastern Montana. Other investigators using various species of plants (Hemmer et al. 1977, Bjugstad 1979, Dollhopf et al. 1980, Dollhopf and Bauman 1981, Smith et al. 1985, 1986, Voorhees et al. 1987) have reported varying degrees of success on bentonite mine spoils using topsoil and spoil amendments to pro- mote growth and establishment of plants. However, with the limited quality and quanti- ty of topsoil on shale overburden of bentonite mine lands, a realistic and practical approach is to use salt- and drought-tolerant plants (Shannon 1979), fertilizers, and both organic and chemical amendments on raw spoils. Organic and chemical amendments are pri- marily intended to promote development and aggregation of the spoil material and to ame- liorate the dispersion effect of excessive sodi- um. In an earlier greenhouse study, Uresk and Yamamoto (1986) showed that salt- or drought- tolerant plants can survive in amended or untreated bentonite mine spoil. The objective of this study was to obtain field verification of the greenhouse study and to test effects of amendments on raw bentonite mine spoil for survival of salt- and drought-tolerant plants. Materials and Methods Study plots are located near Upton, Wyoming, on property owned by the Ameri- can Colloid Company. Average precipitation is 380 mm per year, with highest rainfall occur- ring during the period from May through July. Average annual air temperature is 6°C, with an average low of-17°C in January and an 'USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Rapid City, South Dakota 57701. ^Retired from USDA Forest Service, Rocky Mountain Forest and Range E.xperiment Station, Rapid City, South Dakota 57701. 156 1994] Plant Survival on Bentonite Spoil 157 average high of 23°C in July. The cHniate can be characterized as dr)', hot, and windy in the summer and windy and cokl in the winter. Stands of ponderosa pine {Pinus ponderosa Laws.) and sagebrush (A/-t(?/ni.s/V/)/grass vegeta- tion characterize areas around the bentonite mines. Bentonite mine spoils are essentially derived from the Mowry shale formation. Study plot surfaces were barren with 15 cm of loose shale spoil or exposed Mowry shale beds. These shale spoils are extremely hard and crusty when diy Mowry shale is classed as siliceous shale (Pettijohn 1957), with an abnormally high silica content of 85% vs. 58% for normal shales. Its siliceous character is attributed to its volcanic ash origin (Rubey 1929). Uresk and Yamamoto (1986) previously reported on the spoil properties (0-20 cm sampling depth) of the study site. Spoil is ade- quate in NPK status with nitrate nitrogen at 19 kg/ha, ammonia-nitrogen at 55 kg/ha, phos- phoiois and potassium at 39 /xg/g and 170 ^tg/g, respectively. Soluble salt concentrations vary markedly, but the spoils are saline (EC of 9.2 mmhos/cm) and sodic (SAR 33.1). Spoil pH averaged 6.9 but ranged from 4.1 to 8.0. X-ray diffraction of bentonite spoil revealed a mix- ture of silicate clay types wath a dominance of illite based on the CEC of 30 meq/lOOg, rather than montmorillonite as sometimes believed (Uresk and Yamamoto 1986). Other clays included montmorillonite and kaolinite. Additional information on chemical character- istics of raw bentonite spoil from the study area is reported by Voorhees et al. (1987). Five different amendments to raw spoil materials were applied in 1979 (Table 1). Four spoil treatments included ponderosa pine sawdust, wheat straw, perlite, and vermiculite. Fertilizer (NPK, 11-5-6) was added at 84 kg/lia to all treatments except the control. Additional nitrogen (dry pellet) was added to sawdust and straw spoil treatments at 12 kg/mt and 8 kg/mt, respectively, at time of planting. Each amendment was mixed with bentonite spoil at a 50:50 volume ratio with added gypsum at 10 mt/ha-30 cm depth. The fifth treatment, with- out any physical amendment, was gypsum at 20 mt/ha-30 cm depth (USSLS 1954). The sixth treatment was a control (untreated). All treatments except the control were surface mulched with ponderosa pine woodchips to a depth of 2 cm. Six plant species were selected on the basis of drought and saline-alkali tolerance (Gill 1949, Wright and Bretz 1949, McKell 1978): foui-wing saltbush {Atriplex canescens [Pursh] Nutt.), big sagebrush {AHemisia tridentafa tri- dentata Nutt.), Rocky Mountain juniper (Junipertis scopidonim Sarg.), Russian olive {Elaeagniis angtistifolia L.), common yarrow {Achdlea mdUfoliwn L.), and scarlet globemal- low {Sphaeralcea coccinea [Pursh] Rydb.). Russian olive and Rocky Mountain juniper were 3- and 1 1/2-year-old bare root seedlings, respectively; the remaining species were 1- year-old container- grown seedlings. All plants were planted in mid-May 1979. Survival of plants was evaluated twice per year (spring and fall) from 1979 to 1982. The experimental design was a 6 x 6 (6 species X 6 treatments) factorial arrangement with two replications accomplished through randomized blocks. Each block consisted of 6 treatments (columns) per species with 10 plants per treatment. Twenty plants for each species were evaluated per treatment in this design. Plants were spaced 1 m apart, within and between columns. Each plant was planted Table L Experimental design with six treatments and added supplements. Treatments Supplements Control' G\psum Sawdust Perlite Stra\\' \'emiiculite NPK(ll-5-6) 84 kg/ha X X X X X N 12 kg/mt 8 kg/mt Gvpsum X X X X X (10 mt/ha-30 cm depth) Mulch (2 cm depth) x X X X X 'Raw bentonite spuil onK 158 Great Basin Naturalist [Volume 54 to the depth of the root erown in a hole 30 em in diameter and 35 em deep with the bottom 5 cm filled with sawdust. The hole was then backfilled with spoil, chemical and physical amendments. Each plant was gently watered (deionized tap water) to saturation immediate- ly after planting. The study site was visited once a week for watering (1 L per plant) dur- ing the first month after transplanting and biweekly thereafter from May to September for two growing seasons (1979 and 1980). Differences in survival among species and treatments were analyzed by chi-square pro- cedures for comparison of proportions from many samples (Fleiss 1973). All differences were evaluated at a = .10. Results and Discussion Mortality of all plant species was very high after the first growing season. All Russian olives died after 2 weeks, and further docu- mentation of this species was discontinued. (1) Survival and growth of these species on field plots of bentonite spoil was different from that recorded in the greenhouse (Uresk and Yamamoto 1986). (2) Fourwing saltbush per- formed well, with an overall survival rate of 52% after four growing seasons (Table 2). Sur- vival rate for this species was greatest on the perlite spoil amendment (80%) followed by gypsum (70%) and vermiculite (70%) amend- ments. Survival of yarrow, scarlet globemal- low, and big sagebrush was very low but did indicate some adaptability to amended spoils. The halophytic capabilities of many mem- bers of the genus Atriplex are well document- ed in literature (Waisel 1972, Osmond et al. 1980, Richardson and McKell 1980, Tiede- mann et al. 1984). A 40% survival of fourwing saltbush on the control treatment (Table 2) demonstrates the natural adaptability of this species to saline-alkali conditions. Survival rates increased 40 percentage units when the perlite amendment was utilized. Gypsum and vermiculite showed survival rates increased 30 percentage units but were not significantly different from the control. This indicated that some gains in survival could be attained by amendment of raw bentonite spoil, although not significant at F = .10. In this study overall plant sui-vival was greatest with perlite. Adding sawdust has been the most accept- ed amendment for reclaiming raw bentonite spoils; however, there are differences in opin- ion on the desirability of using sawdust. It has shown some beneficial results, particularly when applied as mulch and eventually worked into the soil (Lunt 1955). Others have tried applying sawdust or shavings to their soils with disappointing and sometimes disastrous results (Lunt 1955). Most researchers have observed no toxicity when adequate nitrogen was used with sawdust or other wood prod- ucts (Allison and Anderson 1951). However, tannins, resins, and secondary compounds Table 2. Survival of transplanted plant species in bentonite mine spoil treatments after four growing seasons at Upton, Wyoming. Spoil treatment Species Control (%) Gypsum {%) Sawdust (%) Perlite (%) Straw (%) Vermiculite (%) Average survival! (%) Yarrow 0 0 L5 5 0 0 3" Rabbitbrush 0 0 0 0 0 0 0" Scarlet globemallow 0 0 0 0 0 5 P Fourwing saltbush 40ab 70ab 30"'' 801' 20'' 70ab 52b Big sagebrush 0 0 0 10 10 0 3" Rocky Mountain juniper 0 0 0 0 0 0 0-' Average survival 7^ 12" 8" 16» 5" L3" 10 ^ Means in row oicolnnin folic iwed liy the same letter are not different at a = .10. Absence of letters indicates no statistical significance. 1994] Plant Survival on Bentonite Spoil 159 PERCENT SURVIVAL F S F S F S F 100 90 - 80 - 70 - 60 - 50 I- 40 30 20 10 0 SAWDUST Ua a Mi a a a a F s F s F s F 79 80 81 82 YEARS YEARS Fig. 1. Sun'ival of fourvving saltbush during fall (F) and spring (S) measurements, 1979-82, at a=.10. 160 Great Basin Naturalist [Volume 54 from fresh sawdust are possibK' toxic to plants. Also, increased soil acidity fiom sawdust may decrease survival (Allison and Anderson 1951). Sawdust added to raw bentonite spoils increased water infiltration (Voorhees 1986). Infiltration rates have not been evaluated against perlite or vermiculite as amendments. In addition, Voorhees et al. (1987, 1991) and Voorhees and Uresk (1990) found that ben- tonite spoil amended with sawdust, alone or in combination with other amendments, increased growth of rillscale through two growing seasons. Schuman and Sedbrook (1984), Smith et al. (1985), and Belden et al. (1990) showed that sawdust added to spoil, with wood chips added to the surface, promot- ed vegetation establishment over a 4-year period. Measurements of spoil pH from the four- wing saltbush plots after four growing seasons langed from 6.1 to 6.8 for the various treat- ments. These values were generally 1-2 units lower than pH values of bentonite spoils mea- sured at the termination of the greenhouse study (Uresk and Ycimamoto 1986). Lowest pH values were found in samples from plots that had been amended with sawdust, straw, or vermiculite. Since vermiculite amendment was associated with higher survival trends of fourwing saltbush, and sawdust and straw amendments with the lowest survival, pH may not have been a factor in survival of foui^wing saltbush. Examination of survival for fouiAving salt- bush (Fig. 1) showed that greatest mortality for all treatments except the control occurred during the first growing season prior to fall evaluation. Thereafter, no significant mortality occurred, except on the control treatment. Greatest survival rates of fourwing saltbush were with perlite, vermiculite, and gypsum. These materials and their mixes are well known in plant nurseries and in the horticul- tural industiy as pot mixes. Apparently, they mix well with bentonite spoils, improve plant survival rates, and may indicate that water infiltration, aeration, and bulk densit)' charac- teristics are improved with their addition to spoils. With the exclusion of fourwing saltbush, overall average sunival rates of selected plant species across amendment-treated spoils after four growing seasons were 0—3%. Fourwing saltbush demonstrated a natural adaptability to establishment on saline-alkali bentonite spoil, with an overall survival rate of 52%. Per- lite, vermiculite, and gypsum amendments enhanced survival rates for founving saltbush. Sawdust and straw amendments did not increase sui-vival as much as perlite or vermi- culite amendments, but plant survival on these amendments was very stable. Sawdust and straw are materials that are readily avail- able in the Black Hills and in bentonite min- ing areas. Further experimentation with differ- ing rates, type of applications, bed prepara- tions, and times of planting and seeding are needed, especially for shrubs and forbs. Literature Cited Allison, F. E., and M. S. Anderson. 19.51. The use of sawdust for mulches and soil improvement. Circular 891. U.S. Department of Agriculture, Washington, D.C. 19 pp. Barchardt, G. a. 1977. Montmorillonite and other smec- tite minerals. Pages 293-.325 in Soil environment. Soil Science Society of America, Madison, Wisconsin. 948 pp. Belden, S. E., G. E. Schuman, and E. J. Depuit. 1990. Salinity and moisture responses in wood residue amended bentonite mine spoil. Soil Science 150: 874-882. BjUGSTAD, A. J. 1979. Bentonite mine spoil and pit recla- mation: a major research problem. Pages 1-19 in Proceedings of the Mineral Waste Stabilization Liai- son Committee. Erie Mining Co., Eleventh, Min- nesota. DoLLHOPF, D. J.. E. J. Depuit, and M. G. Kl.'VGES. 1980. Chemical amendment and irrigation effects on sodi- um migration and vegetation characteristics in sodic minesoils in Montana. Bulletin 736. Reclamation Research Technology, Montana Agricultural E.xperi- ment Station, Montana State Uni\ersit\\ Bozeman. DoLLHOPF, D. J., and B. J. Bauman. 1981. Bentonite mine land reclamation in the northern Great Plains. Research Report 179. Montana Agricultural E.xperi- ment Station, Montana State Universit>\ Bozeman. 42 pp. Fleiss, J. L. 1973. Statistical methods for rates and pro- portions. John Wile\' and Sons, New York. Gill, L. S. 1949. Shade trees for the Rockies. Pages 72-76 in Trees: yearbook of agriculture 1949. US DA, U.S. Government Printing Office, Washing- ton, D.C. Hemmer, D., S. Johnson, and R. Beck. 1977. Bentonite mining related reclamation problems in the north- western states. Old West Regional Commission Report. Montana Department of State Lands, Hele- na. LUNT, H. A. 1955. The use of wood chips and other wood fragments on soil amendments. Connecticut Agri- cultural Station Bulletin 593. New Haven, Connecti- cut. 46 pp. McKell, C. M. 1978. Establishment of native plants for the rehabilitation of paraho processed oil shale in an arid environment. Pages 13-32 in Robert A. Wright, 1994] Plant Survival on Bentonite Spoil 161 ed.. The reclamation of disturbed arid lands. Uni- versity of New Mexico Press, Albuquerque. Osmond, C. B., Bjorkman, O., and D. J. Anderson. 1980. Physiological processes in plant ecology. Toward a synthesis with Atriplex. Springer- Verlag, Berlin-Heidelburg-New York. 461 pp. Pettijohn, F. J. 1957. Sedimentary rocks. Harper and Brothers, New York. 718 pp. Richardson, S. G., and C. M. McKell. 1980. Water relations of Atriplex canescens as affected by the salinity and moisture percentage of processed oil shale. Agronomy Journal 72: 946-950. RUBEY, W. W. 1929. Origin of the siliceous Mowr>' shale of the Black Hills region. U.S. Geological Survey Paper 154D. Russell, E. W. 197.3. Soil conditions and plant growth. 10th ed. Longman Group Limited. 849 pp. Schuman, G. E., and T. a. Sedbrook 1984. Sawmill wood residue for reclaiming bentonite spoils. Forest Products Journal .34; 65-68. Shannon, M. C. 1979. In quest of rapid screening tech- niques for plant salt tolerance. Horticultural Science 14: 587-589. SiEG, C. H., D. W. Uresk, and R. M. Hansen. 1983. Plant-soil relationships on bentonite mine spoils and sagebrush-grassland in the northern High Plains. Journal of Range Management 36: 289-294. Smith, J. A., E. O. Depuit, and G. E. Schuman. 1986. Wood residue and fertilizer amendment on ben- tonite mine spoils, II: plant species responses. Jour- nal of Environmental Quality 15: 427-435. Smith, J. A., G. E. Soberman, E. J. Depuit, and T. A. Sedbrook. 1985. Wood residue and fertilizer amendment of bentonite mine spoils, I: spoil and general vegetation responses. Journal of Environ- mental Quality 14: 575-580. Tiedemann, a. R., E. D. McArthur, H. C. Stutz, R. Stevens, ,\nd K. L. Johnson, compilers. 1984. Pro- ceedings— symposium on the biology of Atriplex and related chenopods, 2-6 May 1983, Provo, Utali. Gen- eral Technical Report INT-172. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range E.xperiment Station. Ogden, Utah. .309 pp. United States Salinity Laboratory Staff (USSLS). 1954. Diagnosis and improvement of saline and alkali soils. U.S. Department of Agriculture Hand- book 60. Washington, D.C. Uresk, D. W., and Yamamoto, T. 1986. Growth of forbs, shrubs, and trees on bentonite spoil under green- house conditions. Journal of Range Management .39: 113-117. Voorhees, M. E. 1986. Infiltration rate of bentonite mine spoil as affected b\' amendments of gypsum, sawdust and inorganic fertilizer. Reclamation and Revegeta- tion Research 5: 483^90. . 1990. Forage quality of rillscale {Atriplex suckleyi) grown on amended bentonite mine spoil. Great Basin Naturalist 50: 57-62. Voorhees, M. E., and D. W. Uresk. 1990. Effect of amendments on chemical properties of bentonite mine spoil. Soil Science 150: 66.3—670. Voorhees, M. E., M. J. Trlica, and D. W. Uresk. 1987. Growth of rillscale on bentonite mine spoil as influ- enced by amendments. Journal of Environmental Quality 16:411-416. Voorhees, M. E., D. W. Uresk, and M. J. Trlica. 1991. Substrate relations for rillscale {Atriplex suckleyi) on bentonite mine spoil. Journal of Range Management 44: 34-38. Waisel, Y. 1972. Biology of halophytes. Academic Press, New York. 395 pp. Wright, E., and T. W. Bretz. 1949. Shade trees for the plains. Pages 65-72 in Trees: yearbook of agriculture 1949. USDA, U.S. Government Printing Office, Washington, D.C. Received 24 August 1993 Accepted 25 October 1993 Great Basin Naturalist 54(2), © 1994, pp. 1 62- 1 69 POPULATION STRUCTURE AND ECOLOGICAL EFFECTS OF THE CRAYFISH PACIFASTACUS LENIUSCULUS IN CASTLE LAKE, CALIFORNIA James J. Elser^, Christopher Junge^, and Charles R. Goldman^ Abstract. — The recent appearance of the "California crayfish," Pacifastacus li'niii.scuhis, in Castle Lake, California, and interest in its potential impacts on the lake ecosystem provided motivation for a study of the population structure and habitat use of this species and its effects on aquatic macrophytes. Mark- recapture studies indicated that the total number of adult (3+ yr or older) crayfish in the lake was ca 10,100 individuals, yielding an estimate of lakewide crayfish densit\' in preferred crayfish habitats of 0.13 adults m~2. Using mean body mass of individuals, we estimated that ambi- ent biomass density was 5.9 g m~^. Length-weight relationships determined for captured individuals were sex depen- dent, with males having greater body mass for a given carapace length. Length-fiequency and weight-fiequency dia- grams indicated that P. leniusciihis reaches larger sizes in Castle Lake than do populations of P. leniusculus in ultraolig- otrophic Lake Tahoe. Population-wide, males were significantly larger in both carapace length and body mass than females. We also examined sex dependence of interhabitat differences in crayfish body size by comparing animals trapped in rocky areas with those from areas with macrophytes and soft sediments. No significant differences in overall body size were found between habitats, but a significant habitat-sex interaction term occurred because the sex-depen- dent size differences were more pronounced in sediment than in rock-y areas. Exclosure and enclosure experiments indicated that crayfish had large but differential impacts on Castle Lake macrophyte species, as the abundance of two of the dominant species {Chara sp., Potamogeton richardsonii) declined in the presence of crayfish and, in one case, increased in exclosures. These effects occurred via both consumptive and nonconsumptive mechanisms. These studies indicate that an expanding population of P. leiuuscitlus in Castle Lake may be producing sizable impacts on the littoral zone habitat. Key words: crayfish, herbivory, macrophytes, Pacifastacus leniusculus. Littoral zones are important to the dynam- macrophyte species (Lodge 1991). Crayfish ics of lake ecosystems (Wetzel 1983, Caipen- may be particularly important in influencing ter and Lodge 1986). Vascular plant communi- the dynamics of littoral zone plant communi- ties (macrophytes) are particularly important ties because of their diverse feeding modes; in the littoral zone, providing food resources crayfish may act as predators (consuming for herbivores, attachment substrata for peri- other littoral zone invertebrates), as herbi- phyton, and cover for both predators and prey, vores, or as detritivores. Past studies of cray- Macrophytes are also particularly important in fish impacts on macrophytes indicate that fueling detritus food chains (Wetzel 1983). their effects occur directly (via both consump- Traditionally, studies have generally empha- tive and nonconsumptive mortality; Lodge sized the influence of physical (e.g., light and Lorman 1987) as well as indirectly via availability or wave action; Spence 1982) and predation on other potential herbivores (Han- chemical (inorganic carbon; Sand- Jensen son et al. 1990). However, only a few of the 11 1978) factors in regulating littoral zone macro- genera and 300 named species and subspecies phyte communities; as a result, biotic interac- of crayfish in North America (Bouchard 1978) tions, particularly herbivory, have been con- have been studied with respect to their poten- sidered less important (Gregory 1983, Wetzel tial impacts on littoral zone vegetation. 1983). The "California crayfish," Pacifastacus However, recent experimental studies indi- leniusciihis, is a member of family Astacidae cate that invertebrate and vertebrate herbi- and includes three subspecies with a range vores can have large impacts on macrophytes that encompasses northern California and and that herbivory impacts vary for different much of the Pacific Northwest (Miller 1960). ^Department of Zoology, Arizona State University, Tempe, Arizona 85287. ^Division of Environmental Studies, University of California, Davis, California 95616. 162 1994] Crayfish Herbivory 163 Abrahamsson and Goldman (1970) suggested that P. leniusculus, introduced to Lake Tahoe at the turn of the century, were important in determining the distribution of macrophyte communities at depths less than 50 m. Flint and Goldman (1975) subsequently supported this suggestion experimentally, showing that P. leniusculus controlled Mijriophijllutn at shal- low depths. Flint and Goldman also demon- strated that low levels of crayfish grazing enhanced primary productivity by attached algae. In 1988 researchers obsei-ved individuals of P. leniusculus in the littoral zone of Castle Lake, where it had been unrecorded previous- ly during nearly 30 years of ecological research (Beatty 1968, Swift 1968, Neame 1975, Carlton 1982, Paulsen 1987, Hagley 1988). Therefore, in the summer of 1990 we began studies designed to evaluate population size and structure, habitat utilization, and potential effects on macrophyte communities oiP. leniusculus in this subalpine lake. Materials and Methods Study Site Castle Lake is a relatively small (20 ha) but deep (35 m maximum depth) subalpine lake located in the Siskiyou Mountains of northern California (Siskiyou County), USA (41°13'N, 122°22'W). The main basin of the lake is rela- tively steep sided, but the lake contains an extensive shallow (<4 m) shelf on its northeast side. Bottom substrates are diverse and include steep-sided rock faces and boulders in the vicinity of the lake's cirque face, coarse particulate-dominated sediments in the vicin- ity of forested slopes, and flocculent, low-den- sity organic sediments that cover most of the bottom of the main basin below depths of 10 m as well as much of the shallow shelf. According to recent work of Hagley (1988), the most abundant macrophyte species of Cas- tle Lake are Isoetes occidentalis (Henders.), Chara sp., and Potamogeton sp. {P. richardsonii [(Benn.) Rydb.] is dominant but P. gramineus [L.] is also present.) Population Estimates Estimates of crayfish population abun- dance were made via the multiple-recapture Schnabel method (Schnabel 1938), with cap- turing, marking, releasing, and resampling of animals occurring at biweekly intervals from July through mid-September 1990. Animals were captured using cylindrical nylon-mesh traps with funneled entrances and baited with dead fish. Traps were set overnight in shallow waters (<10 m) in all areas of the lake in late afternoon or early evening and retrieved in early morning. To estimate depth distribution of crayfish in the Castle Lake littoral zone, on five occasions we established transect trap lines across the depth contours of the lake to determine the maximum depths at which crayfish could be found; each transect sample consisted of 15 traps placed at 10-m intervals along a nylon line. These extended to a depth of ca 25-30 m. Sex, carapace length, and wet weight of each animal were recorded. Animals were also classified with respect to areas from which they were obtained, i.e., rocky bottoms vs. those with organic sediments. All large ani- mals (generally >35 mm carapace length [CL], age 3''" yr according to Abrahamsson and Goldman [1970], although similarly sized individuals in Castle Lake may be younger than those growing in ultraoligotrophic Lake Tahoe) were given unique marks via cauteriza- tion of the carapace (Abrahamsson 1965) and returned to the lake. We rarely captured recently moulted crayfish, indicating that the majority of moulting had occurred prior to our sampling period. Thus, our results are not likely to be complicated by potential changes in trapability in response to moulting events. A total of 750 animals were eventually marked during the sampling period. Exclosure/Enclosure Studies of the Effects off! leniusculus on Macrophytes To examine potential ecological effects of crayfish on the macrophyte communities of Castle Lake, we performed an 8-wk exclo- sure/enclosure experiment. Cages consisted of 1 X 1 X 1-m wood-framed cages covered with 0.9-cm mesh nylon netting on sides and top. Replicate (n = 4) enclosures received either one or three adult male crayfish (50-52 mm carapace length, ca 47 g body weight), equiva- lent to densities of 47 g m-2 and 141 g m-2, respectively. Adult crayfish were used because earlier studies suggested that smaller crayfish tend to be more carnivorous than herbivorous, while adults are usually primarily herbivorous 164 Great Basin Naturalist [Volume 54 (Abrahanisson 1966). Logistical constraints on the size of enclosures we could use meant that enclosure densities were likely higher than ambient crayfish densities (see Results). Thus, we also decided to maintain four e.xclosures that received no additions of crayfish and were inspected to ensure that no animals had been enclosed. Results of exclosure treat- ments thus are critical in assessing whether potential herbivore impacts detected in enclo- sures with artificially high animal densities are likely to be operating in the lake itself Sup- port for the hx'pothesis that crayfish are exert- ing an impact on macrophytes in the lake itself would come not only from depressed biomass of macrophytes in enclosure treatments but more importantly from increases in macro- phyte biomass in exclosures where macro- phytes are protected from ambient grazing intensity. When each exclosure or enclosure was positioned, a control section of equal area was also delineated to be sampled at the end of the experiment to enhance the power of sta- tistical analyses. Without paired control areas, high site-to-site variation in local macrophyte abundance might overwhelm treatment effects even if treatments substantially altered local macrophyte abundance. Thus, a total of eight enclosures and four exclosures, each with a paired control area, were monitored. Cages, with control areas, were placed along the 3-5-m contour interval within vege- tated areas of the lake. Cages were checked at weekly intervals via scuba, and crayfish were added to enclosures from which animals had escaped (this happened twice); no crayfish were observed inside exclosures. At the end of the S-wk period, the above-sediment portions of all submersed macrophytes in each cage and control area were harvested, sorted by species, drained, and weighed to the nearest gram. Results and Discussion The estimated population size of adult P. leniusculiis (i.e., individuals > 3 yr) in Castle Lake obtained using the Schnabel method was 10,100 ± 23 (SD) individuals. Our trap tran- sects across depth contours in the lake indicat- ed that crayfish did not generally inhabit bot- tom areas below 10 m, as at these depths the bottom is dominated by soft, flocculent sedi- ments. Likewise, animals were rarely caught in traps placed in much of the shallow shelf area of the lake, which is also dominated by soft sediments lacking macrophyte develop- ment. Thus, we then estimated the total amount of crayfish habitat as the total bottom area shallower than 10 m (117,000 m-) minus the estimated area of the shallow shelf domi- nated by soft sediments (ca 42,400 m^), yield- ing a total habitat area of ca 74,600 m^. Thus, average crayfish densities in Castle Lake were approximately 0.13 adults m--. This estimate is somewhat lower than, for example, the esti- mates of densities of P. leniusculiis made by scuba census and trapping efficiency in ultra- oligotrophic Lake Tahoe (0.16-5.85 adults m--) and oligotrophic Donner Lake (0.23-0.44 adults m-2) reported by Abrahamsson and Goldman (1970), Flint and Goldman (1977), and Goldman and Rundquist (1977). Some of this discrepancy may reflect differences in methodology (mark-recapture, which focused only on adults, vs. scuba census or trapping efficiency methods, which included juvenile animals). In addition, our estimate is likely to be on the low side as we conservatively included bottom areas down to 10 m although our tran- sect data indicate that a majority of catches were made at depths shallower than 5 m. Therefore densities in the habitat areas pre- dominantly used by P. leniusculiis in Castle Lake are likely higher and likely approach the lower end of the range reported for Lake Tahoe (ca 0.2 adults m-^). By stratifying the population estimates based on relative catches in different parts of the lake, we estimated populations of 8000 individuals in rocky-bottomed areas of the lake and 2100 in macrophyte-dominated areas. Average body mass (male and female) of sampled crayfish was 45.6 g, and therefore areal biomass of crayfish in Castle Lake was 5.9 g m-2. The sex ratio for animals in all catches was 0.90 (female:male). Length and wet-weight measurements were made on approximately 1188 animals during the course of the sampling season. Length-weight relationships differed signifi- cantly for males and females (based on confi- dence limits on the slopes of the length- weight relationships), with males having greater body mass for a given carapace length, especially at larger sizes (Fig. 1). This may reflect the fact that male chelae undergo allo- metric growth during ontogeny, while female 1994] Crayfish Herbivory 165 4.5- A Males J 4.0- • • jB[>. OD Br* 1/5 a E 3.5- ■i f c 3.0- 2.5- /tj* 2.0- 1.5- / / In (W) = 3.04 (In L) - 8.26 / r-sq = 0.85 1 ■■-r 1 , 3.0 3.5 4.0 In (length, mm) 4.5 5.0 4.5- 4.0 ^ 35- 3.0- 2.5- 2.0- 1.5 B Females In (W) = 2.53 (in L) - 6.33 r-sq = 0.79 3.0 3.5 4.0 In (length, mm) 4.5 Fig. 1. Length-wet weight relationships for male (A) and female (B) crayfish. Allometric relationships fit to the data are given. chelae grow isometrically (Mason 1975). In addition, males were on the whole larger in carapace length (and body mass) than females (P < .05 based on a t test; Fig. 2), possibly reflecting the fact that females must necessari- ly invest substantial portions of their energy budgets in reproductive output. The size fre- quency distribution (Fig. 2) was quite broad, with largest individuals reaching 70"*" mm CL. This contrasts with the results of the studies by Abrahamsson and Goldman (1970) and Flint (1975) of P. leniusculus in ultraolig- otrophic Lake Tahoe, where the largest ani- mals observed were around 55 mm CL. With- out more detailed sampling and cohort analy- ses it is not possible to determine whether this difference reflects faster growth rates or longer life span in Castle Lake relative to Lake Tahoe. We also used ANOVA to compare male and female sizes in different habitat categories (rocky vs. sediment/macrophyte areas; Fig. 3). While there was no significant main effect of habitat type on crayfish body size, there was a significant habitat X sex interaction, as male body size in sediment/macrophyte areas was higher than in rocky areas, with the opposite being true for female crayfish. It is important to note that baited traps are biased in favor of large males (Brown and Brewis 1978); thus, interpretation of data on sex-dependent differ- ences in body size between habitats is compli- cated by the possibility that trap bias operates differently at different locations. Bearing this in mind, the patterns illustrated in Figure 3 may reflect interacting influences of burrow availability in different habitats and size- and sex-dependent dominance patterns between individual crayfish. Behavioral studies of P. leniusculus are limited, but the study of Momot and Leering (1986) indicated that, within sexes, large P. leniusculus are dominant over small but that, for animals of the same size, females dominate over males. Given this suggestion and our lack of knowledge about general burrow availability in different bottom types in Castle Lake, we can only speculate about the factors contributing to observed sex- dependent differences in body size between habitat types. It is possible, for example, that general burrow shortages in sediment regions cause these regions to be dominated by the largest (and therefore most dominant) male crayfish able to successfully defend the limit- ed number of available burrows, resulting in a larger mean body size for males in sediment areas. However, the observation that the opposite pattern was true for females (females tended to be larger in rock>' areas than in sedi- ment areas) suggests that simple dominance patterns and general burrow shortages are insufficient to explain these data, as it is possi- ble, for example, that the quality of available burrows may be different for females than for males. Experimental studies of burrow choice 166 Great Basin Naturalist [Volume 54 220 Carapace Length (mm) Weight (g) Fig. 2. Length-frequency (A) and weight-frequency (B) diagrams for male and female crayfish. Males were significant- ly (P < .05) larger and heavier than females. and defense and aggressive displacement in different habitats, as well as the sex-depen- dence of these phenomena, are needed to elaborate on these possibilities. Our enclosure/exclosure experiment indi- cates that P. leniuscuhis exerts substantial grazing pressure on the afjuatic macrophytes of Castle Lake, as both total biomass and bio- mass of certain species responded to the pres- ence or absence of crayfish (Fig. 4). However, densities of crayfish used in the experimental enclosures (one or three adult male crayfish per square-meter enclosure) exceeded our estimates of ambient densities of adult cray- fish in the lake (ca 0.15-0.20 crayfish m-^) by more than fivefold. High densities of crayfish in enclosures were the necessary result of hav- ing to construct enclosures of manageable size that would not disturb, or be disturbed by, other human users of the lake. The observation 1994] Crayfish Herbivory 167 C a o. 03 u 03 b8- A O Rocky Habitat • Sediment/Macrophytes 56- H^ 54- ^-'.''■' N^ 52- ^'\ 50- Habitat Sex: p n.s. < .001 '■■f /IC- HxS:f < .03 1 Male Female o /o- B _ O Rocky Habitat • Sediment/Macrophytes 60- <► 50- ^-.'^ '^N 40- iri - Habitat: n.s. Sex: p< .001 HxS:p< .001 Male Female Fig. 3. Body size of male and female crayfish in differ- ent habitat types: A, carapace length; B, body weight. Analysis of variance indicated no significant main effect of habitat on animal size, but there was a significant habitat- se.x interaction [P < .001), reflecting a pronounced differ- ence between habitats in males but an opposite, and less- er, difference for females. Error bars represent ±1 SE for each mean. that crayfish impacts did not increase in three- vs. one-crayfish treatments suggests that the artificially high crayfish densities in enclo- sures created unrealistic intensities of her- bivory. Thus, the potential exists that reduc- tions in macrophytes relative to control areas reflect these artificially high levels of crayfish and that crayfish grazing at ambient levels is inconsequential for macrophytes. However, this conclusion is not supported by the obser- vation that macrophyte biomass increased greatly in exclosures, which prevent ambient crayfish grazing, relative to control areas exposed to crayfish. Chara biomass increased in the enclosure treatment by a factor even greater than its decline in the exclosure treat- ments, consistent with our prediction that, if crayfish grazing is important under natural lake conditions, macrophyte biomass should increase strongly when ambient grazers are excluded. We also observed that Chara canopies in experimental exclosures were more branched, taller, and more open than Chara stands in both the crayfish enclosures and the lake itself, suggesting that crayfish grazing is important in the natural lake condi- tion, a conclusion supported by our observa- tion during nighttime scuba dives that most of the observable crayfish could be found feed- ing on the Chara beds. The impact of P. leniusculus on macrophytes was clearly species dependent. No effect of either exclosui"es or enclosures was observed for Isoetes, a pteridiophyte with relatively tough leaves arranged in a basal rosette. No statistically significant impacts were detected for Potamogeton either However, these results are misleading, as they are largely an artifact of the general scarcity of Potamogeton in the Castle Lake littoral zone. This scarcity result- ed in a poor representation of Potamogeton among experimental treatments (only one exclosure and a single one-crayfish enclosure had Potamogeton initially and in the adjacent control area). In fact, in all four cases where Potamogeton occurred in enclosures with cray- fish, abundance was reduced to zero as cray- fish snipped off the single-stemmed plants at the base. This example of nonconsumptive mortality is similar to that observed by Lodge and Lorman (1987) for Orconectes rusticus feeding on Megalodonta beckii and Vallisneria americana. The impact of such whole-plant mortality is undoubtedly more extreme in its impact than partial consumption of individual plant parts and may account for the general scarcity of Potamogeton in Castle Lake. The low impact of P. leniuscidus on Isoetes may reflect the low potential food values of this species. Hagley (1988) reports a high C:N ratio for Isoetes in Casde Lake (ca 14-18:1 by weight vs. 10-11:1 for Chara and Potamogeton vegetative shoots); low C:N ratios in plant materials are generally considered indicative of nutritionally superior foods for a wide vari- ety of herbivores (Mattson 1980, Crawley 1983). Biomass of the macroalga Chara changed significantly in both exclosures and enclo- sures. This response reflects consumptive grazing by crayfish, as we never saw severed Chara "stems" inside cravfish enclosures. The 168 Great Basin Naturalist [Volume 54 1000 500- M) c o C/} U c« W o £ -tad o ^ n c ^--^ •"■ V bC e CQ fii -500- ■1000 I I Exclosure 1 Crayfish 3 Crayfish Chara,p< .001 Isoetes n.s. Potamogeton n.s. All Species p< .01 Fig. 4. Results of crayfish exclosure/enclosure experiments. Analysis of variance indicated the presence of crayfish had significant effects on total macrophyte abundance (P < .01) and on Chara {P < .001). Although Potamogeton was nonconsumptively eliminated in all enclosures in which it occurred with crayfish, no statistically significant effect was observed for Potamogeton, largely a result of poor representation of this species among experimental units. Isoetes was unaffected by crayfish. EiTor bars represent ±1 SE for each mean. EiTor bars are missing for the exclosure and one- crayfish treatments for Potamogeton because Potamogeton was present in both treatment and control areas in only one replicate pair for these treatments. substantial and rapid increase in Chara bio- mass in exclosures indicates that crayfish potentially regulate the natural abundance of Chara in Castle Lake; high Chara growth rates may permit it to persist in the face of this consumption. Overall distributions of these macrophytes are consistent with the differen- tial impacts of crayfish just described: Isoetes dominates bottom areas with large crayfish populations while Potamogeton is confined to sediment-dominated littoral zone areas where crayfish abundance is lowest. In sum, our observations of the abilities of Pacifastaciis leniusculus to differentially regulate macro- phyte species in this lake lend further support to the conclusions of Lodge (1991) that macro- phytes are actively engaged in aquatic food webs via direct consumption by herbivores, in addition to their role as contributors to detri- tal -based trophic pathways. Given the potential influences of crayfish- induced mortality on Castle Lake macro- phytes demonstrated by the cage experiments, it would be of interest to know the history of the P. leniuscuhis population in this system. Previous thorough investigations of the Castle Lake littoral zone do not report any crayfish. However, crayfish carapaces were observed in the lake as early as 1986 (E. Marzolf personal communication), and a substantial population was verified during gill net studies of Castle Lake rainbow trout begun in 1988. This places the date of potential introduction of crayfish in the mid-1980s, as a considerable amount of littoral zone research occurred in Castle Lake in the earlv 1980s with no report of crayfish (e.g., Paul'sen 1987, Hagley 1988). In the absence of a more thorough evaluation of present-day species composition, spatial dis- tribution, and biomass development of Castle Lake macrophytes, it is not possible to evaluate whether these assemblages changed during the period between the macrophyte studies of Hagley (1988) prior to crayfish introduction and 1990 when our study was performed. However, given that invasions of new species into unoccupied habitats are often explosive, population densities o{ P. h'niusciihis in Castle Lake may increase even further and approach those densities actually used in our experi- mental enclosures; this possibility' is support- ed by recent (1992-93) crayfish trapping, which indicates that catch-per-unit-efifort may have increased by a factor of 2-3 since our 1990 study (J. J. Elser personal observation). Thus, our experimental studies likely yield 1994] Crayfish Herbivory 169 some insights into the impacts on macro- phytes of further population development of P. leniusculus, providing an additional iUustra- tion of the effects of invading species on aquatic ecosystems. Acknowledgments This research was supported by NSF grants BSR-9017579 to J. J. Elser and BSR- 9006623 to C. R. Goldman. C. Junge was sup- ported by an NSF Research Experiences for Undergraduates supplement to BSR-9006623. We are grateful to E. R. Marzolf and F. S. Lubnow for advice and assistance in the field. D. M. Lodge, W. T. Momot, and C. Richards provided helpful comments on the manu- script. Literature Cited Abrahamsson, S. a. a. 1965. A method of marking cray- fish Astacus astacus in population studies. Oikos 16: 228-231. . 1966. Dynamics of an isolated population of the crayfish Astacus astaciis. Oikos 17: 96-107. Abr.\hamsson, S. a. A., and C. R. Goldman. 1970. Dis- tribution, density, and production of the crayfish Pacifastacus leniuscuhis Dana in Lake Tahoe, Cali- fornia-Nevada. Oikos 21: 83-91. BEATTi', K. W. 1968. An ecological study of the benthos of Castle Lake, CA. Unpublished doctoral dissertation. University of California, Davis. 94 pp. Bouchard, R. W. 1978. Ta.xonomy, distribution, and gen- eral ecology of the genera of North American cray- fish. Fisheries 3: 11-16. Brown, D. J., and J. M. Brewis. 1978. A critical look at trapping as a method of sampling a population of Austropotamobius pallipes (Lereboullet) in a mark and recapture study. Freshwater Crayfish 4: 159-164. Carlton, R. G. 1982. The role of sediments in the nutri- ent dynamics of Castle Lake, California. Unpub- lished master's thesis. University of California, Davis. 145 pp. Carpenter, S. R., and D. M. Lodge. 1986. Effects of submersed aquatic macrophytes on ecosystem processes. Aquatic Botany 26: 341-370. Crawley, M. J. 1983. Herbivory: the dynamics of animal- plant interactions. University of California Press, Berkeley. 437 pp. Flint, R. W. 1975. Growth in a population of the crayfish Pacifastacus leniusculus from a subalpine lacustrine environment. Journal of the Fisheries Research Board of Canada 32: 2433-2440. Flint, R. W., and C. R. Goldman. 1975. The effects of a benthic grazer on the primary productivity of the lit- toral zone of Lake Tahoe. Limnolog>' and Oceanog- raphy 20: 935-944. . 1977. Crayfish growth in Lake Tahoe: effects of habitat variation. Journal of the Fisheries Research Board of Canada 34; 155-159. Goldman, C. R., and J. C. Rundquist. 1977. A compara- tive ecological study of the California crayfish, Paci- fastacus leniusculus (Dana), from two subalpine lakes (Lake Tahoe and Lake Donner). Freshwater Crayfish 3: 51-80. Gregory, S. V. 1983. Plant-herbivore interactions in stream systems. Pages 157-189 in J. R. Barnes and G. W. Minshall, eds.. Stream ecology: application and testing of general ecological theory. Plenum Press, New York. Hagley, C. 1988. The ecology and nutrient cycling of macrophytes in Castle Lake, California. Unpub- lished master's thesis. University of California, Davis. 143 pp. Hanson, J. M., P. A. Chambers, and E. E. Prepas. 1990. Selective foraging by the crayfish Orconectes virilis and its impact on macroinvertebrates. Freshwater Biology 24: 69-80. Lodge, D. M. 1991. Herbivory on freshwater macro- phytes. Aquatic Botany 41: 19.5-224. Lodge, D. M., and J. G. Lorman. 1987. Reductions in submersed macrophyte biomass and species rich- ness by the crayfish Orconectes rusticus. Canadian Journal of Fisheries and Aquatic Sciences 44: 591-597. Mason, J. C. 1975. Crayfish production in a small wood- land stream. Freshwater Crayfish 2: 449-479. M.'^ttson, W. J. 1980. Herbivor)' in relation to plant nitro- gen content. Annual Review of Ecology and System- atics 11: 119-161. Miller, G. C. 1960. The taxonomy and certain biological aspects of the crayfish of Oregon and Washington. Unpublished master's thesis, Oregon State Universi- ty, Corvallis. 216 pp. Momot, W. T., and G. M. Leering. 1986. Aggressive interaction between Pacifastacus leniusculus and Orconectes virilis under laboratory conditions. Freshwater Crayfish 6: 87-93. Neame, p. a. 1975. Benthic o.xygen and phosphorus dynamics in Castle Lake, CA. Unpublished doctoral dissertation. University of California, Davis. 234 pp. Paulsen, S. 1987. Contributions of sediment denitrifica- tion to the nitrogen cycle of Castle Lake, CA. Unpublished doctoral dissertation. University of California, Davis. 14.5 pp. Sand-Jensen, K. 1978. Metabolic adaptations and vertical zonation of Littorella uniflora (L.) Ashers and Isoetes lacustris L. Aquatic Botany 4: 1-10. Schnabel, Z. E. 1938. The estimation of the total fish population of a lake. American Mathematics Mono- graphs 45: 348-3.52. Spence, D. H. N. 1982. The zonation of plants in fresh- water lakes. Advances in Ecological Research 12: 37-125. Swift, M. C. 1968. A quantitative and qualitative study of trout food in Castle Lake, California. Unpublished masters thesis, University of California, Davis. .54 pp. Wetyel, R. G. 1983. Limnologv'. Saunders College Publish- ing, New York. 760 pp. Received 9 September 1993 Accepted 15 November 1993 Great Basin Nahiralist 54(2), ©1994, pp. 170-176 BROOD HABITAT USE BY SAGE GROUSE IN OREGON Martin S. Drut^ John A. Crawford^ and Michael A. Gregg^ Abstract. — Habitat use by Sage Grouse {Centrocercus urophasianus) hens with broods was examined at Jackass Creek and Hart Mountain, Oregon, from 1989 through 1991. Sage Grouse hens initially selected low sagebrush {Artemisia spp.) cover t>'pes during early brood-rearing, big sagebrush cover types later in the brood-rearing period, and ultimateh- concentrated use in and near lakebeds and meadows. Areas used by Sage Grouse broods typically had greater I'orb f're(iuency than did random sites. Hens at Jackass Creek selected sites with forb cover similar to that generally available to broods at Hart Mountain, but home ranges were larger at Jackass Creek because of lower availability' of suit- able brood-rearing habitat. Differences in habitat use by broods on the two areas were reflected in dietary differences; at Hart Mountain, chicks primarily ate forbs and insects, whereas at Jackass Creek most of the diet was sagebrush. Larger home ranges, differences in diets, and differences in availability of forb-rich habitats possibly were related to differences in abundance and productivity between areas. Key words: broods. Centrocercus urophasianus, habitat, Oregon, Sage Grouse. Habitat factors, including resource avail- ability, may limit Sage Grouse [Centrocercus urophasianus) populations through reduced recruitment of young (Klebenow 1969, Blake 1970, Wallestad 1975, Autenrieth 1981). Stand structure and food availability are characteris- tics most frequently associated with habitat selection by hens with broods (Klebenow 1969, Peterson 1970, Wallestad 1971, Autenrieth 1981). Dunn and Braun (1986) found that veg- etative cover and extent of habitat intersper- sion are the most important factors influencing summer habitat use by Sage Grouse. Forbs and insects typically constitute the primary food of chicks (Klebenow and Gray 1968, Peterson 1970, Drut et al. 1994), and forb cover is often greater at sites used by broods than at random locations (Klebenow 1969, Autenrieth 1981, Dunn and Braun 1986). Shrubs, particu- larly sagebrush {Artemisia spp.), provide escape and thermal cover (Klebenow and Gray 1968) but are not a primary component of chick diets except where forbs and insects are limited in availability (Drut et al. 1994). Peter- son (1970) noted decreased use of sagebrush/ grassland cover types as broods mature and ascribed these changes to differential avail- ability of succulent forbs. Martin (1970) ob- served that broods typically use big sagebrush (A. tridentata) stands during early brood-rearing and that broods <6 weeks old use areas with lower densities of sagebrush than do older broods. Despite numerous studies of Sage Grouse summer habitat use, knowledge of habitat use and selection by Sage Grouse hens with broods is incomplete because of small sample sizes, lack of information about use and availability of cover types and habitat components within cover types used by hens with broods, failure to distinguish habitat use by hens with broods from other adults, or no provision of informa- tion regarding population status and habitat use. Information that relates population status and habitat use is critical for Oregon because the western subspecies (C. u. phaios), which inhabits most of the Sage Grouse range in the state, was listed as a candidate for threatened and endangered status by the Department of Interior in 1985. This listing resulted from declines in abundance caused by depressed productivity (Crawford and Lutz 1985). The objective of the study was to determine use of cover types and habitat components by Sage Grouse hens with broods during two brood- rearing periods on two study areas with differ- ent Sage Grouse population characteristics in southeastern Oregon. •Department of Fisheries and Wildlife, Oregon State Universit\-, Cor\allis, Oregon 97.331-3803. 170 1994] Brood Habitat Use by Sage Grouse 171 Study Areas The study was conducted at Jackass Creek, administered by the Bureau of Land Manage- ment, and Hart Mountain National Antelope Refuge, administered by the U.S. Fish and Wildlife Service. Estimates of Sage Grouse abundance since 1980 indicated approximate- ly 2.5 birds/km2 and 1.5 birds/km^ at Hart Mountain and Jackass Creek, respectively (J. Lemos, Oregon Department of Fish and Wildlife, unpublished data; W. H. Pyle, U.S. Fish and Wildlife Service, unpublished data). Summer productivity counts from 1985 through 1992, the only period for which com- parable data were available, averaged 1.6 and 0.9 chicks/hen (p < .05) at Hart Mountain and Jackass Creek, respectively. The Jackass Creek study area, approxi- mately 70 km southwest of Burns, Harney County, Oregon, comprises nearly 39,000 ha. Prominent shrubs are low sagebrush (A. arhuscula) and big sagebrush (A. tridentata). Western junipers (Juniperus occidentalis) are present on the eastern portion of the study area. Common annual and perennial forbs include mountain dandelion {Agoseris spp.), hawksbeard {Crepis spp.), lupine {Lupinus spp.), and phlox {Phlox spp.). Grasses are prin- cipally bluegrass {Poa spp.) and fescue [Festii- ca spp.). Annual temperature averages 10°C, and mean precipitation is 25 cm. The Hart Mountain National Antelope Refuge study area is 100 km southwest of Jackass Creek in Lake County, Oregon, and is 89,000 ha in size. Dominant cover consists of low sagebrush, big sagebrush, and antelope bitterbrush {Purshia tridentata). Areas >2000 m in elevation contain curl-leaf mountain- mahogany {Cercocarpus ledifolius) and trem- bling aspen {Populus tremuloides). Forb and grass composition is similar to Jackass Creek. At refuge headquarters (elevation 1700 m) annual temperature averages 6°C, and mean precipitation is 29 cm. Plant nomenclature fol- lows Hitchcock and Cronquist (1987). Methods Sage Grouse hens were radio-marked in 1989-91 (Gregg et al. 1994). At the conclusion of each field season, marked hens were recap- tured, radio transmitters were removed, and a sample of previously unmarked hens was equipped with radios to maintain indepen- dence of samples among years. Radios were attached with herculite ponchos (Amstrup 1980), and all hens were fitted with numbered leg bands. Locations of radio-marked hens were obtained with portable receivers and two-element, hand-held antennae. Cover types and habitat components used for rearing broods were identified from loca- tions of radio-marked hens with broods. Radio-marked hens with broods were located four times weekly to identify cover types used. Monitoring of broods continued until a hen lost her brood or brood integrity disintegrated (approximately 1 August each year). We classified cover at brood sites into one of seven cover types: big sagebrush, low sage- brush, mixed sagebrush, lakebed/meadow, mountain shrub, grassland, and juniper/aspen. Cover type descriptions were based on Soil Conservation Service information (J. Kinzel, U.S. Department of Agriculture, Soil Conser- vation Service, unpublished data) and previ- ous descriptions at Jackass Creek (Trainer et al. 1983, Gregg 1992). Study area boundaries, based on locations of radio-marked hens with broods, were determined each year with the minimum con- vex polygon method (Mohr 1947, Odum and Kuenzler 1955). Proportions of cover types within the area used for rearing broods were determined with a dot grid system (Avery 1977). Each brood location was marked and served as a site for habitat sampling, which was completed within 2 days after location of a brood. Percent cover of forbs, grasses, and shrubs and frequency of occurrence of ground-dwelling insects were measured at all brood locations. We established two 10-m per- pendicular transects intersecting at each brood location. The position of the first tran- sect was determined from a randomly selected compass bearing. The intercept distance (cm) of all species of shi-ubs along each transect was recorded to determine canopy cover (Canfield 1941). Heights of shrubs intercepted were measured from the ground to the top of the shrub canopy and placed into one of three classes: short (<40 cm), medium (40-80 cm), or tall (>80 cm). Canopy cover of shrubs was recorded separately for each height class. Per- cent cover of forbs was estimated from five uniformly spaced rectangular plots (20 X 50 cm) on each transect (Daubenmire 1959). 172 Great Basin Naturalist [Volume 54 Sampling intensity was determined by con- structing a species area curve with data col- lected from initial sampling (Pieper 1978:12). Occurrence of ground-dwelling arthropods was established from 12 pitfall traps (Morill 1975) arranged systematically along each 23-m transect, 36 at Hart Mountain and 28 at Jack- ass Creek, in cover types used by broods (see Drut et al. 1994). Arthropods were classified into Scarabeidae (June beetles), Tenebrionidae (darkling beetles), Formicidae (ants), and other. Vegetative structure of habitats available to Sage Grouse broods was characterized at ran- domly selected locations within cover types on each study area during the brood-rearing period. Sampling of random locations, which was concurrent with measurements taken at sites used by broods, was conducted during May and June of each year. Number of ran- dom locations sampled in each cover type was based on canopy cover of sagebrush, which represented the least variable habitat compo- nent, and was determined with the "n-test " (Snedecor and Cochran 1980:210). Home ranges for hens with broods were determined with the McPaal home range pro- gram (Stuwe and Blohowiak 1983). Home ranges were compared for two brood-rearing periods (early: hatching to 6 weeks; and late: 7 to 12 weeks after hatching) within and between study areas with chi- square analysis (Snedecor and Cochran 1980:20). Six-week intervals were based on data from Martin (1970), which indicated hens with broods changed habitat use at this time, and from Peterson (1970), which revealed differences in foods consumed by juveniles beginning approximately 6 weeks after hatching. Within study areas, cover types used by Sage Grouse for rearing broods were compared with availability of cover types. Between study areas, cover type availability and use were compared. We arranged data in contingency tables and analyzed them with chi-square analysis; cover types with <5 brood locations were combined and analyzed collectively. If differences were detected, confidence interval testing (Neu et al. 1974, Byers et al. 1984) was used to identify cover types used selectively. Use of cover types by hens with broods of dif- ferent ages was compared with chi-square to assess possible changes in habitat use associat- ed with age of broods. Cover types used for nesting by hens that successfully hatched clutches were compared with cover types used by hens with broods during the first 6 weeks after hatching. Habitat components measured at brood sites were compared by chi-square analysis to random sites within the same cover types for each study area to identify which vegetative components were selected. Analysis of vari- ance was used to test among cover types and between study areas for differences in avail- ability (random locations) and use (brood loca- tions) of vegetative cover (Snedecor and Cochran 1980:258). The least significant dif- ference test was used to separate means (Snedecor and Cochran 1980:272). Results were considered significant at the 95% level. Results Most broods (13) were produced in the big sagebrush cover type, but during early brood- rearing (hatching-6 weeks), hens with broods were most frequently found (54-67% of Table I. Use and availability of cover types in which Sage Grouse broods were produced and those used for early (hatching-6 weeks) and late (7-12 weeks) brood-rearing periods at jackass Creek and Hart Mountain, Oregon, 1989-91. J^ ickass Creek Hart Mountain Used Available Used Avail able (% frequency) (%ofs irea) (% frequency) (%of Earlv area) Hatched Earlv Late Earlv Late Hatched Early Late Late Cover t^^pe (.V = 7) {N = imy (.V = 3/40)^' (,V= 11) (.V= 11/89)^ (.V = 4/40H' Big sagebrush 42 17 45 54 30 91 32 52 30 57 Low sagebrush 29 53 17 32 30 9 67 38 48 16 Mixed sagebnish 29 29 20 9 15 0 0 0 1 1 Lakebed/ meadow 0 0 15 3 23 0 0 8 3 5 Other 0 0 3 2 2 0 1 2 18 21 "Number of broods/iuiiulicr Dflocations. 1994] Brood Habitat Use by Sage Grouse 173 observations) in low sagebrush cover (Table 1). Three cover types were used differentially during early brood-rearing: low sagebrush was used more (/:> < .05) than expected on both areas, mixed sagebrush was used in greater proportion {p < .05) than available at Jackass Creek, and big sagebrush was used to a lesser extent {p < .05) than available at Jackass Creek. None of the other cover types was used during the early brood-rearing period. During late brood-rearing (7-12 weeks) habitat use shifted to predominantly big sage- brush (45-52% of observations). Use of low sagebrush declined on both areas (Table 1). Availability of low sagebrush within areas used by hens with broods declined from 48 to 16% at Hart Mountain as hens with broods moved away from low-sagebrush-dominated areas. Also, during late brood-rearing, use of lakebeds and meadows increased; these habi- tats received the greatest use after brood break-up in August. Forb cover ranged from 10 to 14% at sites used by hens with broods during the early brood-rearing period (Table 2) and was greater ip < .01) at sites used by broods than at ran- dom locations at Jackass Creek. At Hart Mountain, forb cover was used in proportion to availability during early brood-rearing (Tables 2, 3). During late brood-rearing, forbs were used in greater {p < .01) proportion than available at Hart Mountain, where sites used by broods had 19-27% forb cover No use pat- tern in relation to forb availability was evident at Jackass Creek during late brood-rearing. There were no differences in use and avail- ability for any shrub cover category in low (p > .50), big (p > .20), or mixed {p > .20) sage- brush stands. Only in lakebed/meadow habitat at Jackass Creek during the late brood-rearing period were use and availability of shrub cover different {p = .05). In that instance, cover of short and medium shrubs was approximately twice as great at sites used by broods as at random locations (Tables 2, 3). Hart Mountain had more forb cover (p < .05) and less tall shrub cover (p < .05) than Jackass Creek (Table 3). In addition, there was more {p < .05) short shrub cover available during the early brood-rearing period at Jack- ass Creek than at Hart Mountain. The greatest availabilit}' of forb cover on both areas was in lakebed/meadow habitat during late-brood- rearing (14 and 21% at Jackass Creek and Hart Mountain, respectively). Hart Mountain sup- ported greater [p < .01) frequencies of ground-dwelling arthropods than did Jackass Creek, but no differences were found within study areas between time periods or cover types except at Jackass Creek, where mixed sagebrush had a greater (p = .05) frequency of invertebrates during the early period than did low sagebrush (Table 4). At Hart Mountain, big sagebrush and lakebed/meadow habitats supported more (p < .05) forbs than did low sagebrush during late brood-rearing (Table 3). At Jackass Creek low and big sagebrush supported the same cover of forbs (6%) during late brood-rearing, but the lakebed/meadow habitat had greater (p < .05) forb cover (14%). There was more (;; < .05) cover of medium and tall shrubs in big sagebrush stands compared with low sage- brush (Table 3). Mean home range sizes at Hart Mountain, were 800 and 100 ha for the early and late periods, respectively, whereas at Jackass Creek mean home ranges were 2100 and 5100 ha, respectively. Home range size was smaller (p = .02) in the late period than the eaily peri- od at Hart Mountain, whereas home range size increased (p < .01) during the late period at Jackass Creek. Home range size was small- er (p < .01) at Hart Mountain than at Jackass Creek during both periods. Discussion Sage Crouse hens with broods displayed similar use of cover types on the two study areas. The change in cover- type use of suc- cessfully nesting hens from big sagebrush to low sagebrush during the first 6 weeks after hatching was unique to this study. Perhaps availability of foods partially accounted for this change in use of cover types. Klebenow (1969), Peterson (1970), Wallestad (1971), Aut- enrieth (1981), and Dunn and Braun (1986) reported relationships between habitat use by broods and food availability. Return to use of big sagebrush during weeks 7-12 after hatch- ing was similar to findings elsewhere. Canopy cover and shrub height at brood sites in Mon- tana changed from 6% and a range of 15-30 cm, respectively, in June to 12% and 30—45 cm in August (Peterson 1970). Pyrah (1971) and Wallestad (1971) noted sagebrush height was greater in cover t\'pes used by broods during 174 Great Basin Naturalist [Volume 54 P 55 5 bC > i JIU « II 3 J' _>^ -^ ^ o o lo in oa CO CO ^o o o in -^ o V S 00 — 3 A -S 52 3 2 oi = o u ^ en S H tlH O CAl go M ■»-: '^ X CO -a w S CO _:^ < 0; H 4J U 3 " aj CO ■^ II _>■ in oj in 3 i' ^.00 ^ II o o ^ o o SS3!- oo^ tD CO, CO CO -H 05 oa in t^ -r CO 05 o 00 in CO t-- CO -^ CO CD, ci d- S CO ' of Nevada-Reno, Reno, Nevada 89512. 2U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Reno, Nevada 89512. 177 178 Great Basin Natuiulist [Volume 54 species depending!; on location in the Great Basin (Tausch and Tueller 1988). Develop- ment and refinement of techniques to acquire reliable mensurational data will only enhance the further study, understanding, and proper management of pinyon-juniper woodlands. The objective of this study was to apply evolved methods of estimating whole-tree needle biomass to singleleaf pinyon on a third Great Basin site. Methods Study Area Research was conducted on USDI Bureau of Land Management pinyon-juniper wood- lands 32 km SE of Reno, Nevada (39°17'30"N, 119°42'30"W). The study site lies at 1963 m elevation in a near level east-west saddle formed between a basalt plug and the west- facing slope of the Virginia Mountains. Soils are 0-10 cm deep and poorly developed from decomposed Cretaceous granodiorite and Pliocene-Pleistocene volcanics. This area receives approximately 336 mm of precipita- tion annually (Desert Research Institute 1991), mostly as snow. Some Utah juniper occurs in the area, but singleleaf pinyon is the dominant tree species. Field Techniques Tree selection procedures were based on the five maturity classes described in Black- burn and Tueller (1970). Ten trees from each maturity class were randomly selected for potential measurement within relationships of accessibility associated with placing potome- ters for tree water-use studies (De Rocher 1992). Five trees from each maturity class were randomly selected for a total of 25 trees to be measured and harvested. Total tree height (cm), canopy height (cm), maximum canopy diameter and canopy diameter per- pendicular to the maximum (cm), and trunk circumference (cm) at approximately 15 cm above the root crown were determined for all trees. Past studies have shown that including esti- mates of canopy density improves predictabil- ity of needle biomass (Miller et al. 1981), but visual estimates used were not easily compa- rable between studies. Canopy needle bio- mass density decreases as voids develop with- in the canopy with increasing age. Error can be introduced in needle biomass estimation based on crown volume calculated fi^om sim- ple canopy dimensions without an estimate of the voided space. This study incorporated a grid method of determining average canopy density similar to the method proposed by Belanger and Anderson (1989). The procedure involved viewing each sampled tree canopy through a 6 X 60-cm Plexiglas sheet that had a 3-cm2 grid transposed onto it. A perspective was chosen that approximated an average of canopy fullness that was sufficiently distant to visually contain the entire tree height within the vertically held grid when viewed at arm's length. First, the left upright edge of the grid was then aligned with the trunk and the canopy height centered within the grid. This placed the grid over the right side of the tree. Within the grid a smooth canopy border was imagined as if a ribbon were stretched from the top along the outside edge of the canopy to the base. Next, grid squares more than half- way within this perimeter line were counted for the maximum area covered by the canopy. Last, squares within the perimeter covered by >50% foliage (versus open space, trunk, or branches) were counted. Dividing the number of foliage-covered squares by the total number of squares within the canopy perimeter deter- mined a ratio of relative canopy density. This procedure was then repeated by placing the grid over the left side of the tree canopy. Following crown measurements and esti- mation of canopy density, all green foliage was harvested by cutting off branches and placing them in feed bags. After being air-dried, the needles were dried at 80°C for 24 h to achieve consistency and then weighed. A trunk cross section was also cut approximately 15 cm above the root crown. These cross sections were measured for sapwood area (cm^) using a paper trace, which was cut into small pieces and run through a model LI-3100 leaf area meter (LI-COR, Inc., Lincoln, Nebraska). Analysis Techniques Longest crown diameter, diameter peipen- dicular to it, and crown height for each tree were used to compute the crown volume (ni"^) based on the formula for one-half of an ellip- soid (Tausch 1980, Beyer 1984). Crown vol- ume was also adjusted for canopy density by 1994] SiNGLELEAF PiNYON BlOMASS EQUATIONS 179 Table 1. Singleleaf pinyon foliar mass prediction models using sapwood area and crown volume approximated as one- half of an ellipse. Relation Equation R2 Standard error of estimate Needle vs. sapwood Mass (g) Area (cm^) (All maturity classes) Needle vs. sapwood Mass (g) Area (cm^) (Trees <40cm2 sapwood) Needle vs. sapwood Mass (g) Area (cm^) (Trees >40cm2 sapwood) Needle vs. sapwood Mass (g) Area (cm-) (All maturity classes) Needle vs. sapwood Mass (g) Area (cm-) (Trees <40cm2 sapwood) Needle vs. sapwood Mass (g) Area (cm^) (Trees >40cm^ sapwood) Needle vs. sapwood Mass (g) Area (cm^) (All maturity classes) Needle vs. sapwood Mass (g) Area (cm^) (Trees <40cm2 sapwood) Needle vs. sapwood Mass (g) Area (cm^) (Trees >40cm2 sapwood) Needle vs. ellipsoid Mass (g) Volume (m^) (All maturity classes) Y = Linear regression 265.3 + 122. IX Y = -142.5 + 106.4X Y = 3234.7+ 117.4X InY Log transformed linear regression : 35.31 + 1.24lnX Inv = 33.9 + 1.37lnX InY = 186.5 + 0.94lnX .977 .947 .971 .782 .973 .970 Y = Nonlinear regression 181.4X0-94 Y = 20.1X1-497 Y = 194.4X0929 Y = 2.34X0-58 Needle vs. ellipsoid Y = 1.36X0-62 Mass (g) Volume (m-'^) (Adjusted for canopy density) (All maturity classes) .995 .970 .976 .977 4059 227 6822 16,886 194 6453 3847 83 6444 5643 5459 multiplying the calculated volume by the canopy density ratio. Two types of regression analyses were used to evaluate relationships between oven-dry foliage weight and sapwood area and crown volume. Prediction equations utilizing any form of data transformation to linearize the data for least-squares analysis were repeated using nonlinear regression with the original untransformed data (Tausch and Tueller 1988), and the best fit results are reported. Comparisons among regression results were made using highest coefficients of determination (R^) and lowest standard error of the estimate values, to a level of sig- nificance of p < .01 for the least-squares analyses results. 180 Great Basin Naturalist [Volume 54 Results and Discussion The amount of foliage supported per unit area of conducting tissue ranged from 37.4 g cm-2 for the smallest tree to 157.2 g cm"^ for one of the largest, and averaged 88.8 g cm~2 for all sampled trees. Using a nonlinear regression techni(|ue that iteratively approxi- mates optimal fit without data transformation provided as tight a fit to the full data set as lin- ear regression utilizing log-transformed data in relating dried needle mass and sapwood area. Reapplication of untransformed data to linear models decreased the relationship (Table 1). For trees with <40 cm^ sapwood area (maturity classes 1-3), this relationship was definitely nonlinear, as was observed by Tiiusch and Tueller (1989). The slope of the linear regression line for these western Neva- da data is nearly identical to the slope for pinyon from southwestern Utah (Tausch 1980), suggesting a potential singular relation- ship across the Great Basin. For both these data and the Tausch 1980 southwestern Utah data, the entire foliage was harvested from each tree. A similar relationship was not found between data from Tausch 1980 and Tausch and Tueller 1989. Tausch and Tueller (1989) utilized a foliage subsampling technique to estimate total needle biomass, which may have underestimated the needle biomass to sapwood area ratio. Correlation between needle biomass and the calculated elliptical crown volume was also significant by nonlinear regression (Table 1). When adjusted for percent canopy density, the linear elliptical volume relationship was improved (R^ = .98), and the standard error of the estimate simultaneously was reduced from 5643.1 to 5458.9. Conclusions Prediction of foliar mass using sapwood area of singleleaf pinyon on the Virginia Mountains, Nevada, was equivalent in preci- sion to previously reported results for this species conducted in southwestern Utah (Tausch 1980). Elliptical crown voliuue calcu- lated from canopy widths and crown height again proved to be a significant predictor of dry weight phytomass. Adding an estimate for variations in canopy density further improved the relationships. As previously reported (Tausch and Tueller 1989), application of non- linear regression analysis produced the best fit between phytomass and all tree dimensions, as reflected by increased coefficient of deter- mination values and reduced standard errors of the estimate over other regression methods. Estimates of singleleaf pinyon phytomass for hx'drological and ecological studies of pin- yon-juniper woodlands can be most accurately obtained from a minimum of 10 sapwood area measurements. Canopy dimensions and an assessment of foliage density can also be used to reliably estimate whole-tree phytomass. This study, conducted on the western edge of the Great Basin, achieved needle biomass regressions based on sapwood area that were nearly identical to those from work performed in southwestern Utah. Future research should concentrate on comparing numerous isolated studies across the Great Basin of whole-tree foliar mass hanest so that a regional biomass equation may be developed. Acknowledgments This research was supported in part by ftmds provided by the Intermountain Research Station, Forest Service, U.S. Department of Agriculture. Literature Cited Balda, R. p., and N. Masters. 1980. Avian communities in the pinyon-juniper woodland; a descriptixe analy- sis. Management of western forests and grasslands for non-game birds. US DA General Technical Report INT-86. Belanger, R. p., and R. L. Anderson. 1989. A guide for visually assessing crown densities of loblolly and shortleaf pines. USDA Forest Service, Southeastern Forest Experiment Station, Research Note SE-352. 2 pp. Beyer, W. H., ed. 1984. CRC standard mathematical tables. 26th edition. CRC Press, Boca Raton, Florida. 618 pp. Blackburn, W. H., and P. T. Tueller. 1970. Pin>on and juniper in black sagebrush communities in east-cen- tral Nevada. Ecolog>' 51: 841-848. Budv, J. D., R. O. Meeuwig, and E. L. Miller. 1979. Aboveground biomass of singleleaf pinyon and Utah juniper. Pages 943-952 in Proceedings — forest resource inventories workshop, Colorado State Uni- versity, Fort Collins. BuDY, J. D., and J. A. Young. 1979. Historical uses of Nevada's pinyon-juniper woodlands. Journal ot For- est Histor)' 23: 112-121. Chojn.\CKY, D. C. 1986. Pin\on-juniper site quality and volume growth equations for Ne\'ada. USDA Forest Ser\'ice, Intermountain Forest and Range Research Station, Research Paper INT-372. 22 pp. 1994] SiNGLELEAF PiNYON BlOMASS EQUATIONS 181 Cochran, P. H. 1982. Estimating wood volumes for Doug- las fir and white fir from outside bark measure- ments. Forest Science 28: 172-174. De Rocher, T. R. 1992. Measuring water use of single- leaf pinyon utilizing allometric, potometric and porometric techniques. Unpublislied master's thesis, University' of Nevada-Reno. 101 pp. Desert Research Institute. 1991. Weather patterns and precipitation database. Division of Atmospheric Science, Reno, Nevada. Doughty, J. W. 1987. The problems with custodial man- agement of pinyon-juniper woodlands. Pages 29-33 in R. L. Everett, ed.. Proceedings — pinyon-juniper conference. USDA Forest Service, Intermountain Research Station General Technical Report INT- 215. Fogg, G. G. 1966. The pinyon pines and man. Economic Botany 20: 103-105. Grier, C. C., and S. W. Runninc;. 1977. Leaf area of mature northwestern coniferous forests: relation to site water balance. Ecology 58: 893-899. Grier, C. C., and R. H. Waring. 1974. Conifer foliage mass related to sapwood area. Forest Science 20; 205-206. Hatchell, G. E., C. R. Berry, .and H. D. Muse. 1985. Nondestructive indices related to aboveground bio- mass of young loblolly and sand pines on ectomycor- rhizal and fertilizer plots. Forest Science 31: 419-427. Kaufmann, M. R., and C. A. Troendle. 1981. The rela- tionship of leaf area and foliage biomass to sapwood area in four subalpine forest tree species. Forest Sci- ence 27: 477-482. Long, J. N., F. W. S.mith, and D. M. Scott. 1981. The role of Douglas fir stem sapwood and heartwood in the mechanical and physiological support of crowns and development of stem form. Canadian Journal of Forest Research 11: 459-464. Marchand, p. J. 1984. Sapwood area as an estimator of foliage biomass and projected leaf area for Abies bal- samea and Picea nibens. Canadian Journal of Forest Research 14: 85-87. McCuLLOCH, C. Y. 1969. Some effect of wildfire on deer habitat in pinyon-juniper woodland. Journal of Wildlife Management 33: 778-784. Meeuvvtg, R. O., and J. D. Budy. 1979. Pinyon growth characteristics in the Sweetwater Mountains. USDA Forest Service, Intermountain Forest and Range Research Station, Research Paper INT-227. 26 pp. Miller, E. L., R. O. Meeuwig, and J. D. Budy. 1981. Biomass of singleleaf pinyon and Utah juniper. USDA Forest Service, Intermountain Forest and Range Research Station, Research Paper INT-273. 18 pp. Miller, R. F., L. E. Eddleman, and R. F. Angell. 1987. Relationship of western juniper stem conduct- ing tissue and basal circumference to leaf area and biomass. Great Basin Natin-alist 47: 349-354. Nemani, R. R., and S. W. Running. 1989. Testing a theo- retical climate-soil-leaf area hydrologic equilibrium of forests using satellite data and ecosystem simula- tion. Agricultural and Forest Meteorology 44: 245-260. Short, H. L., and C. L. McCulloch. 1977. Managing pinyon-juniper ranges for wildlife. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Research Station, General Technical Report RM-47. 47 pp. Tausch, R. J. 1980. Allometric analysis of plant growth in woodland communities. Unpublished doctoral dis- sertation, Utah State University, Logan. 142 pp. Tausch, R. J., and P. T. Tueller. 1988. Comparison of regression methods for predicting singleleaf pinyon phytomass. Great Basin Naturalist 48: 39-45. . 1989. E\aluation of pinyon sapwood to phytomass relationships over different site conditions. Journal of Range Management 42: 209-212. 1990 Foliage biomass and cover relationships between tree- and shrub-dominated communities in pinyon-juniper woodlands. Great Basin Naturalist 50: 121-134. Waring, R. H. 1983. Estimating forest growth and effi- ciency in relation to canopy leaf area. Advances in Ecological Research 13: 327-354. Waring, R. H., and S. W. Running. 1978. Water and plant life, problems and modem approaches. Plant Cell and Environment 1: 131-140. Whitehead, D. 1978. The estimation of foliage area from sapwood basal area in Scots pine. Forestry 51: 137-149. Received 24 March 1993 Accepted 4 November 1993 Great Basin Naturalist 54(2), ©1994, pp. 182-188 SOME PHYSIOLOGICAL VARIATIONS OF AGROPY RON SMITHII RYDB. (WESTERN WHEATGRASS) AT DIFFERENT SALINITY LEVELS Rengen Uengl-, Ivo E. Lindauer^-^, and Warren R. Russ^ Abstract. — The purpose of this study was to determine the physiological responses of Agropyron smithii Rydb. to various saline environments as evaluated in the laboratory. Agropyron smithii Rydb. (Rosana) seeds were genninated, transplanted into nutrient solutions with NaCl concentrations of 0, .50, 100, 150, and 200 miVf, and grown for 80 days in a growth chamber. Results indicated that leaf water potential, relative water content of leaf tissue, and concentrations of Na, K, and Cl in plant tissue were significantly affected by increasing NaCl concentration. However, leaf chlorophyll concentration and concentrations of Ca and Mg in plant tissue were not significantly affected by the presence of NaCl. Key words: Agropyron smithii, salinity, physiology, chlorophyll, sodium, potassium, calcium, magnesium, water potential, chloride. Agropyron smithii Rydb. (western wheat- grass) is a strongly rhizomatous, glaucous, often glabrous, long-lived perennial grass. It is palatable, cures well on the ground, and is a native grass of the northern Great Plains (Judd 1962, Schultz and Kinch 1976, Stubbendieck et al. 1986). Because A. smithii is a valuable grazing species in the arid West, it is often sought out for revegetation of these soils. Research has shown that salt stress has an impact on chlorophyll concentrations of leaves, leaf osmotic potentials, and mineral uptake and transport in many plants. Seemann and Critchley (1985) reported that chloiphyll concentration per unit area of Phaseohis vul- garis L. was reduced considerably by NaCl stress. Macler (1988) also reported that in Gelidium coulteri (red alga) the content of chlorphyll was altered with a change in NaCl concentration in the growth media. However, Antlfinger (1981) found that concentrations of chlorophyll a, chlorophyll b, and total chloro- phyll of Borrichia frutescens were not signifi- cantly influenced by salinity. Clipson et al. (1985) used a dew-point hydrometer to measure leaf osmotic potential of Suaeda maritima L. Dum. seedlings grown under different salinities. They found that leaf osmotic potentials were lower (more negative) for those grown under higher salt concentra- tions. Black (1960) reported that, under saline conditions, the osmotic potential of Atriplex vesicaria leaves changed in the direction that maintained a constant water potential gradient between leaf and soil. Evidence is growing that salt stress inhibits the uptake and transport of mineral nutrients in some plants. In Hordeiim vulgare L. (l^arley) seedlings, the uptake and transport of nitrogen (Aslam et al. 1984), phosphate (Maas et al. 1979), K (Lynch and Lauchli 1984), and Ca (Lynch and Lauchli 1985) were reduced by salinity. The transport of K and Ca in Gossypi- um hirsutum L. was also disrupted by high Na"*" concentrations (Cramer et al. 1987). In Sal- icornea europa the uptake of K"*", Mg^"*", and Ca-"*" was also reduced by Na+ (Austenfeld 1974). A study of the physiological responses of A. smithii in saline environments may determine how this species adapts to saline soils and pro- vide additional information for breeding salt- tolerant species. The purpose of this study was to measure leaf chlorophyll concentration, leaf water potential, relative water content of leaf tissue, and mineral content of A. smithii grown under different saline water culture condi- tions. Materials and Methods Agropyron smithii Rydb. seeds (Rosana) were germinated in moist vermiculite (tem- perature alternated between 15 °C for 20 h 'Department of Biology, University of Northern Colorado, Greeley, Colorado 80639. ^Present address: 28-1, Chyau-Ay 4th Road. Hsintein 231, Taiwan. ROC. ''Address correspondence and reprint requests to this author. 182 1994] Physiology of Agropyron Smithii and Salinit\' 183 and 30 °C for 4 h) in complete darkness for 5 days. Thereafter, the seeds were held in dark- ness (15 °C for 3 days), then at an alternating temperature (28 ± 2°C for a 12-h day and 15 °C for the 12-h night) as recommended by Toole (1976). A nutrient solution modified from Arnon and Hoagland (1940) was used in this study (Table 1). One-liter plastic containers were used to hold the experimental plants and nutrient solutions. Plants were supported in square cardboard covers impregnated with paraffin. Both covers and containers were sterilized with a 5% Clorox solution before use. When A. smithii seedlings were 15 days old (2-3 cm long), they were placed through holes in the cover and held in place with loose wads of cotton wrapped around each stem. Four plants were placed in each container. Each container was aerated for 30 min each 24-h period. For the first 9 days after transplanting, damaged or infected seedlings were replaced with fresh ones. Salinization of the medium began 9 days after seedlings had been transferred to the nutrient solution. This was done by increasing NaCl concentration in the culture solutions at the rate of 25 mM every 4 days to the final concentrations of 50, 100, 150, and 200 niM. Plants grown in unsalted cultures were used as controls. The nutrient solution was changed eveiy 12 days during the experiment. Experiments were carried out in a Sherer Gillette Plant Growth Chamber (Model 512 GEL) set for a 12-h day at 28 ± 2°G with a humidity of 40 ± 5% and a 12-h night at 15°G with a humidity of 60 ± 5%. Light was sup- plied by 12 cool white VHO fluorescent bulbs. Eighty plants (20 containers) were used for each treatment, and all plants in each treat- ment were numbered. From these 80 plants, 10 were randomly selected for measuring for each treatment. If a plant had more than one culm, the longest was chosen for the data measurement of leaf chlorophyll, leaf water potential, and relative water content of leaf tissue. Leaf chlorophyll concentration was deter- mined by extracting the chlorophyll in ace- tone (80% v/v in water) from the second and third leaves collected from a randomly select- ed plant when the plants were 93 days old. Absorbance of chlorophyll was measured at Table 1. Composition of the nutrient solution nioditied from Arnon and Hoagland (1940). Salt R/L KNO3 L02 Ca(N03), . 4 H,0 0.71 NH4H,P04 0.23 MgS04 0.24 Salt 111 g/L H3BO3 2.86 MnCl, • 4 H,,0 LSI CUSO4 • 5 HoO 0.08 ZnS04 • 7 H2O 0.22 H,Mo04 . H2O 0.09 FeS04 • 7 H2O 5 g/L 1 0.6 mL/L & tartaric acid 4 g/L J (every 4 days) 645 and 663 nm following the procedure of Withametal. (1986). Leaf water potential was evaluated by Ghardokov's procedure as described by With- am et al. (1986) when the plants were 89 days old. All leaves collected from a randomly cho- sen plant were cut into pieces, mixed, and then equally divided into 17 test tubes con- taining sucrose solutions with concentrations ranging from 0.1 M to 0.5 M with increments of 0.025 molarity. The relationship between sucrose concentration and leaf water potential is shown in Witham et al. (1986). Relative water content (RWG) of leaf tissue was determined by using the following rela- tionship modified from Vassey and Sharkey (1989) when the plants were 84 days old: % RWG = {(FW - DW)/(SAT - DW)} x 100 where: FW = leaf fresh weight in grams, DW = oven-dried (at 105 °G for 72 h) weight in grams, and SAT = weight (in grams) of the tissue after soaking in water for 3 h. All leaf tissue from a randomly selected plant was used for determining relative water content (RWG). Plants selected for mineral analysis (K, Na, Ga, Mg, and Gl) were harvested when they were 80 days old. Ten randomly selected plants were washed with distilled water and then placed in an oven at 105 °G for 72 h. Dry material from each plant was weighed, ground, and then transferred into a crucible that was ashed in a muffle furnace at 500 °G for 4.5 h. The ash was dissolved in 20% HNO3 and filtered for the determination of K, Na, 184 Great Basin Naturalist [Volume 54 Ca, and Mg concentrations. Concentrations of K and Na were determined with a flame emis- sion spectrophotometer (Perkin-Elmer, Model 403), and concentrations of Ca, and Mg were determined with an atomic absorption spec- trophotometer (also Perkin-Elmer, Model 403). The Cl concentration was determined indi- rectly by adding a known excess of silver to the sample solutions, which resulted in pre- cipitation of the Cl as AgCl (Perkin-Elmer Corporation 1971). After the separation of AgCl, the concentration of unreacted silver was determined by atomic absorption. Con- centrations of Cl were calculated using the formula listed at the end of this paragraph. Samples were prepared by placing the oven- dried plants in a muffle furnace at 500 °C for 10 h. The ash was dissolved in distilled water (24-67 xnL) to obtain a solution with a Cl con- centration of 0-1000 /xg/niL Cl and then fil- tered. Ten mL of sample, 2 mL of concentrat- ed nitric acid, and 10 mL of silver nitrate solu- tion (5000 fxgJmL Ag) were transferred into a volumetric flask and diluted with distilled water to a volume of 100 mL. Thirty (30) mL of the mixed solution was centrifuged for 10 min at 2500 rpm. The supernatant was diluted with water 1:100 (v/v). Concentrations of Ag in the diluted supernatant were measured by atomic absolution (Perkin-Elmer, Model 403). Concentrations of Cl~ in the samples were cal- culated as follows (Perkin-Elmer Coiporation 1971): Chloride (mg/niL) = {500 - 100 x (mg/niL Ag in supernatant)} X 3.29 X DF where: DF = dilution factor(s). One-factor ANOVA, following procedures outlined by Kleinbaum et al. (1988), was used to determine statistically significant differ- ences (a = .05) among treatments. Results Leaf water potential, relative water content of leaf tissue (RWC), and concentrations of Na, K, and Cl in plant tissue were significant- ly affected by the presence of NaCl, whereas concentrations of leaf chlorophyll, Ca, and Mg in the plant tissue were not (Table 2). Mean values of leaf water potential de- creased (were more negative) as external NaCl concentrations increased. The range was -0.34 MPa to -0.96 MPa for plants grown in nutrient solutions with 0 mM and 200 mM NaCl, respectively. RWC decreased as NaCl concentration in nutrient solutions increased and varied from a maximum of 88% to a mini- mum of 79% for the plants grown in solutions with 0 mM and 200 mM NaCl, respectively (Table 2). Concentrations of Na, Cl, K, and Ca in A. smithii tissue increased with an increase in NaCl concentration. Mean values (n = 10) of Na concentration varied from a minimum of 1.6 mg/g to a niciximum of 58 mg/g (dry weight) for plants grown in solutions with 0 mM and 200 mM NaCl, respectively (Table 2). Plant Cl concentration increased from 2.7 mg/g to 35 mg/g (dry weight) for plants grown in solutions with 0 niM NaCl and 200 mM NaCl, respectively. Potassium concentration increased from 47 mg/g to 59 mg/g (dry weight). Calcium concentration varied from a minimum of 4.5 mg/g to a maximum of 5.1 mg/g (dry weight). No appreciable change in Mg concentrations was observed, although they decreased from a maximum of 2.7 in control plants to 2.4 mg/g (diy weight) in plants grown in solutions with 200 mM NaCl (Table 2). Discussion Concentrations of chlorophyll o, chloro- phyll h, and total chlorophyll in leaf tissue were not significantly affected by NaCl treat- ments (Table 2). This finding agrees with the previous study q{ Borrichia frutescens by Antlfinger (1981) but not with the study of Seemann and Critchley (1985) on Phascohis vulgaris L. This discrepancy might be a result of using different methods of expressing chlorophyll concentrations. The unit used in this study was mg chlorophyll/g fresh tissue, while the unit used by Seemann and Critchley (1985) was g chlorophyll/m^ leaf area. Since water content of A. smithii leaves decreased significantly as salinity increased (Table 2), this might have caused the values of chloro- phyll concentration to be overestimated for plants grown in solutions with high concentra- tions of NaCl because of reduction of leaf fresh weight. If chlorophyll concentrations of A. smithii were expressed as g/unit leaf area, the treatment difference might have been apparent. Further study is needed to confirm this thesis. 1994] Physiology of Agropyron Smithii and Salinity 185 Table 2. Leaf chlorophyll content (mg/g fresh tissne), leaf water potential (MPa, 20 °C), relative water content of leaf tissue (RWC %), and mineral content of a whole plant (mg/g in dry weight) of A. smithii grown in nutrient solutions with five NaCl concentrations. Values represent mean (n = 10) ± 1 SD. Significance of the F-test from the ANOVA is also given. Treatments (mM NaCl) S ignificance Character 0 50 100 1,50 200 level Chlorophyll a 1.3 ±0.24 1.5 ± 0.19 1.4 ± 0.23 1.4 ±0.21 1.4 ±0.24 .2289 Chlorophyll b 0.9 ± 0.16 1.0 ±0.17 1.0 ± 0.15 0.9 ±0.14 0.9 ±0.20 .2919 Total chlorophyll 2.2 ±0.39 2.6 ± 0.35 2.4 ± 0.36 2.4 ±0.34 2.3 ±0.42 .2166 Leaf water potential -0.34 ±0.06 -0.6 ± 0.1 -0.8 ± 0.07 -0.89 ±0.07 -0.96 ±0.07 .0001 RWC 88 ±3.8 87 ± 1.3 83 ± 6.9 81 ± 7.2 79 ±3.6 .0003 Sodium-' 1.6 ±0.15 37 ± 4.8 48 ± 4.9 54 ±5.2 58 ±5.4 .0001 Chloride-' 2.7 ±0.25 7.3 ± 0.76 18 ± 1.4 30 ±2.4 35 ±2.7 .0001 Potassium 47 ±4.4 55 ± 5.3 57 ± 5.8 58 ±5.8 59 ±5.6 .0001 Calcium 4.5 ±0.51 4.9 ± 0.64 4.9 ± 0.64 5 ±0.64 5.1 ±0.7 .2803 Magnesium 2.7 ±0.21 2.6 ± 0.43 2.4 ± 0.35 2.5 ± 0.23 2.4 ±0.26 .4421 ■'Test of significance was perfonned on the transformed data (common logarithn Several environmental factors that influ- ence chlorophyll content in A. smithii have been identified. Lauenroth and Dodd (1981) discovered that when A. smithii was exposed to SO2, chlorophyll a and b concentrations were reduced. Chlorophyll a was found to be more sensitive to SO2 than was chlorophyll b. Bokhari (1976) found that temperature, water stress, and nitrogen fertilizer also influenced the content of chlorophyll a and b. There was a significant change in leaf water potential of A. smithii along the salinity gradi- ent. Leaf water potential became lower (more negative) in response to increased levels of NaCl (Table 2). The decline of leaf water potential caused by salinity has also been found in other plants such as Cochleria ojfici- nalis, Atriplex littoralis, and Limonium vulgare (Stewart and Ahmad 1983). Reduction of tissue water potential induced by the addition of salt into the nutrient solu- tion has several impacts on plants that are similar to those induced by water stress. Examples of these are the inhibition of cell growth, cell wall synthesis, protein synthesis, carbon assimilation, respiration (Glass 1988), photosynthesis (Black and Bliss 1980), and other enzyme activities (Stewart and Ahmad 1983). Reduction of leaf water potential was thought to be a strategy to maintain turgor and avoid desiccation in saline environments (Class 1988). The change of leaf water poten- tial can occur in a variety of ways, such as changing osmotic potential or turgor pressure, or tlie combination of the two. However, studies of angiosperm halophytes by Stewart and Ahmad (1983) have shown that changes in cell osmotic potential are the major components that effect changes in leaf water potential. In leaf tissue of Limonium vulgare, leaf water potential and osmotic potential decreased in a parallel fashion over a change in growth media water potential from near zero to -2.7 MPa. The turgor potential was more or less constant up to -1.8 MPa. When L. vulgare was grown in media having salinity greater than -2.7 MPa, turgor pressure often decreased. It is currently believed that the decrease in tur- gor pressure is the primary event inhibiting growth. Plant cells will grow only when the protoplast exerts a positive pressure on the cell wall. Crop plants were found severely wilted when leaf water potentials were lower than a range of -1.2 to -1.6 MPa (Hanson et al. 1977). Thus, the ability of plants to main- tain their leaf water potentials above the tur- gor loss point in a saline environment may be used as a measure of their salt tolerance. Data from the present study are unable to predict (1) the value of leaf water potential at which A. smithii will lose its turgor or (2) how osmotic potential and turgor pressure respond to water stress induced by saline environment. RWC of leaf tissue is sometimes used to indicate the degree of water deficit. This value in A. smithii was significantly reduced by the presence of NaCl in nutrient solutions (Table 2). Although RWC has a positive relationship with tissue turgor pressure, the correlation between these two parameters varies from species to species. For example, at the same 186 Great Basin NatuR/\list [Volume 54 value of RWC, Dtihautia ciliolata is able to maintain a higher value of turgor pressure than D. scahra (Robiehaux 1984). The ability to maintain higher turgor pressure is thought to be an adaptation to water stress induced by salt. Further study is needed to e.xplore the rela- tionships of RWC, osmotic potential, and turgor pressure of A. smithii in saline environments. The concentration of Na ion in A. smithii increased dramatically as the external NaCl increased (Table 2). An increase in Na ion has also been found in leaf tissue of Phaseolus vul- garis L. grown in saline environments (See- mann and Critchley 1985). The accumulation of high concentrations of Na ion was thought to balance the low water potential of the exter- nal environment in halophytes (Glass 1988). Data from this study suggest that A. smithii may use the same method as other halophytes to maintain a more negative osmotic potential than that of the external medium. Accumulation of Na in A. smithii tissue also accounts for the reduced growth of this plant, because enzymes of all eukaiyotes are sensi- tive to high concentrations of NaCl (Kramer 1984) and high concentrations of Na cause a disruption of membrane integrity by displace- ment of Ca from cell surfaces by Na (Cramer et al. 1985, Lynch and Lauchli'l988). More- over, Cramer et al. (1987) and Jeschke (1984) suggested that a high level of Na in the apoplast could also inhibit the transport of assimilate and K in the phloem, thus reducing growth. For most plants to survive in saline envi- ronments, Na must be excluded from the bulk cytoplasm. In halophytes it has been demon- strated that Na concentrations are relatively low in the cytoplasm compared to the vacuole. In the root cortical cells of the halophyte Suaeda maritima (L.) Dum., Na was found in the vacuoles at four times the concentration in the cytoplasm or cell walls (Hajibagheri and Flowers 1989). Jeschke (1980) suggests that this kind of Na compartmentalization appears to be brought about by selective K ion influx and Na efflux through the plasmalemma and by Na"''/K"'" exchange across the tonoplast. In addition to cellular Na compartmental- ization, plants employ other methods to avoid or minimize toxic effects of Na. For example, Distichlis stricta (Torr.) Rydb., Atriphx haUmus L. (Mozafar and Goodin 1970, Anderson 1974), and members of the families Phunbaginaceae and Frankeniaceae (Helder 1956) have salt- eliminating glands or hairs that are found on the leaves. In Orijza sativa (rice) the salt is translocated to older leaves that then drop from the plant (Yeo and Flowers 1982). Fur- ther study is needed to clarify how A. smithii avoids or minimizes toxic effects of salts. Sodium is not an essential element for A. smithii but is now considered an essential nutrient for plants capable of fixing CO2 via C4 organic acids. This includes C4 and CAM plants (Glass 1988). It is interesting to note that the C4 plants Zea mays (corn) and Sac- charum ojficimirum L. (sugar cane) have not been shown to require Na (Hewitt 1983). The concentration of Cl in A. smithii increased considerably as the external NaCl concentration increased. But overall concen- trations of Cl in A. smithii tissue are lower than those of Na (Table 2). The reason that A. smithii keeps Na concentrations higher than those of Cl in saline environments is unclear. A similar increase in Cl was found in leaf tis- sue of Phaseohis vulgaris L. grown in saline environments (Seemann and Critchley 1985). The accumulation of this element in halophyte tissue is also thought to influence osmotic reg- ulation. Cellular Cl compartmentalization, like that of Na, has been found in Suaeda maritima (L.) Dum. (Hajibagheri and Flowers 1989). Chloride basically has the same toxic effects on plants as Na does. In fact some toxic effects of NaCl may result from a combination of the two ions. In citrus and grapes, Cl has been shown to be the damaging ion (Shannon 1984). Levitt (1980) claimed that Cl injury occurred earlier and was more severe than Na injuiy because Cl was accumulated by plants from NaCl more rapidly than Na was. Since the accumulation of Na and Cl in plant tissue is a common phenomenon found in halo- phytes, the accumulation of these two ions in A. smithii in this study could indicate this species has some level of adaptation to saline environments. The concentration of K ion increased in A. smithii as the external NaCl increased. How- ever, the increment was not as great as that of Na and Cl (Table 2). These results are compa- rable to those found by Antlfinger (1981) in Borrichia frutescens but do not agree with the findings in Gossypium hirsutum L. In G. hir- 1994] Physiology of Agropyron Smithii and Salinity' 187 stittim L., the transport of potassium was dis- rupted by high Na+ concentrations (Cramer et al. 1987). The discrepancy in potassium content might be because G. hirsutum L. is more sensitive to NaCl than are A. smithii and Borrichia friitescens. The increase of the potassium ion content of A. smithii in this study could be due to an altering of the ionic charges within cells resulting from the rise of Cl concentrations. Potassium is known to gen- erate turgor in many non-halophytes and halo- phytes. It is also an enzyme activator with at least 60 enzymes known to be activated by this ion (Glass 1988). Studies of cell cultures of Nicotiana tabacum (Watad et al. 1983), Medicago sativa (Croughan et al. 1979), and Citrus aurantium (Ben-Hayyim et al. 1985) found that a higher level of internal potassium ion could be corre- lated with a higher level of salt tolerance. Since A. smithii is able to maintain a high internal potassium ion concentration when grown in saline environments, this may be a sign of salt tolerance. The concentration of Ca was not signifi- cantly changed by increased external NaCl concentrations (Table 2). This finding does not agree with the studies on Hordeum vulgare L. (barley) seedlings by Lynch and Lauchli (1985) and on Salicornea europa by Austenfeld (1974). The discrepancy in the findings may be due to the possibility that A. smithii is more tolerant to NaCl than Hordeum vulgare L. and Salicornea europa. In Zea mays root protoplast, Ca is known to be displaced from associated cell membranes by high concentrations of internal Na. This dis- placement is correlated with increased leakage of potassium ion (Lynch and Lauchli 1988) because Ca is known to maintain cell mem- brane integrity for plants in saline environ- ments (Poovaiah and Leopold 1976, Leopold and Willing 1984, Cramer et al. 1987). Magnesium concentration did not change appreciably with increasing external NaCl concentrations (Table 2). The findings do not agree with those studies on Salicornea europa conducted by Austenfeld (1974). Magnesium is well known for its participation in the chlorophyll molecule. The unaltered concen- tration of Mg is consistent with the unaltered concentration of chlorophyll in this study. The role of this element in plants responding to external NaCl is not clear at this point. Conclusions The unchanged chlorophyll concentration, reduction of leaf water potential, and accumu- lation of K and Na of A. smithii in this study are signs of adaptation to saline environments. The biomass study of this species (data are not presented in this article) indicates that Agropy- ron smithii prefers an environment with a low concentration of NaCl, although it can survive in a habitat with a higher concentration of salt. Literature Cited Anderson, C. E. 1974. A review of structure in several North Carolina salt marsh plants. Pages 344-397 in Robert J. Reinonold and William H. Queen, eds., Ecology of halophytes. Academic Press, New York. Antlfinger, a. E. 198L The genetic basis of microdiffer- entiation in natural and experimental populations of Borrichia frutescens in relation to salinity. Evolution 35: 1056-1068. Arnon, D. I., AND D. R. Hoagland. 1940. Crop produc- tion in artificial solutions and in soils with special reference to factors influencing yields and absorp- tion or inorganic nutrients. Soil Science 50: 463. ASLAM, M., R. C. HUFFAKER, AND D. W. Rains. 1984. Early effects of salinity on nitrate assimilation in barley seedling. Plant Physiology 76: 321-325. Austenfeld, F-A. 1974. Correlation of substrate salinity and ion concentration in Salicornea europa L. with special reference to oxalate. Biochemie Physiologic der Pflanzen 164: 303-316. Ben-Hayyim, C, P. Spiegel-Roy, and H. Neumann. 1985. Relation between ion accumulation of salt- sensitive and isolated stable cell lines of Citrus aurantium. Plant Physiologv' 78: 144-148. Black, R. A., and L. C. Bliss. 1980. Reproductive ecolo- gy oi Picea mariana (Mill) B.S.P., at the tree line near Inuvik, Northwest Territories, Canada. Ecolog- ical Monographs 50: 331-354. Black, R. F. 1960. Effects of NaCl on the ion uptake and growth of Atriplex vesicaria Heward. Australian Journal of Biological Sciences 13: 249-266. BOKHARI, U. G. 1976. The influence of stress conditions on chlorophyll content of two range grasses with contrasting photosynthetic pathways. Plant Physiol- ogy 40: 969-979. Clipson, N. J. W., A. D. ToMAS, T. J. Flowers, and W. R. G. Jones. 1985. Salt tolerance in the halophyte Suaeda maritima L. Dum. Planta 165: 392-396. Cramer, G. R., A. Lauchli, and V. S. Polito. 1985. Dis- placement of Ca+2 by Na+ from the plasmalemma of root cells. A primary response to salt stress? Plant Physiology 79: 207-211. Cramer, G. R., J. Lynch, A. Lauchli, and E. Epstein. 1987. Influx of Na+, K+, and Ca+2 into roots of salt-stressed cotton seedlings. Plant Physiology 83: 510-516. Croughan, T. P., S. Y. Stavarek, .^nd D. W. R\ins. 1979. Selection of a NaCl tolerant line of cultured alfalfa cells. Crop Science 18: 959-963. Glass, A. D. M. 1988. Plant nutrition. Jones and Bartlett Publishers, Boston. 188 Great Basin Naturalist [Volume 54 Hajibagheri, M. a., and T. J. Flowers. 1989. X-ray microanalysis of ion distribution within root cortical cells of the halophytes Siiaeda maritima (L.) Dum. Planta 177: 131-134. Hanson, A. D., C. E. Nelsen, and E. H. Everson. 1977. Evaluation of free proline accumulation as an index of drought resistance using two contrasting harley cultivars. Crop Science 17; 720-726. Helder, R. J. 1956. The loss of substances by cells and tissues (salt glands). Pages 468-488 in VV. Ruhland, ed., Handbuch der pflanzenphysiologie. Volume 2. Springer, Berlin. Hewitt, E. J. 1983. Essential and functional metals in plants. Pages 313-315 in D. A. Robb and W. S. Pier- point, eds., Metals and micronutrients: uptake and utilization by plants. Academic Press, New York. Jeschke, W. D. 1980. Roots: cation selectivitv- and com- partmentalization, involvement of protons and regu- lation. Pages 17-28 in R. M. Spanswick, VV. J. Lucas, and J. Dainty, eds.. Plant membrane transport; cur- rent conceptual issues. Elsevier/North-Holland, Amsterdam. . 1984. K+-Na+ e.xchange at cellular membranes, intracellular compartmentalization of cations, and salt tolerance. Pages 59-60 in Richard C. Staples and Gary H. Toenniessen, eds.. Salinity tolerance in plants. John Wiley & Sons, New York. Judd, B. I. 1962. Principal forage plants of southwestern ranges. Rocky Mountain Forest and Range E.xperi- ment Station, Forest Service, U.S. Department of Agriculture. Research paper RM-69. Kleinbaum, D. G., L. L. Kupper, and K. E. Muller. 1988. Applied regression analysis and other multi- variable methods. 2nd edition. PWS-Kent Publish- ing Company, Boston. Kramer, D. 1984. Cytological aspects of salt tolerance in higher plants. Page 3 in Richard C. Staples and Gary H. Toenniessen, eds.. Salinity tolerance in plants. John Wiley & Sons, New York. Lauenroth, W. K., and J. L. Dodd. 1981. Chlorophyll reduction in western wheatgrass {Agropijron smithii Rvdb.) exposed to sulfur dioxide. Water, Air, and Soil Pollution 15; 309-;315. Leopold, A. C, and R. P. Willing. 1984. Evidence for toxicity effects of salt on membranes. Pages 71-73 in Richard C. Staples and Gaiy H. Toenniessen, eds.. Salinity tolerance in plants. John Wiley & Sons, New York. Levitt, J. 1980. Responses of plants to environmental stresses. Volume 2; Water, radiation, salt and other stresses. Academic Press, New York. 372 pp. Lynch, J., and A. Lauchli. 1984. Potassium transport in salt-stressed barley roots. Planta 161; 295-301. . 1985. Salt stress disturbs the Ca nutrition of bar- ley (Hordeum valgare L.). New Phytologist 99; 345-354. . 1988. Salinity affects intracellular calcium in com root protoplasts. Plant Physiology 87; 351-356. Maas, E. v., G. Ogata, and M. H. Finkel. 1979. Salt- induced inhibition of phosphate transport and release of membrane proteins from barley roots. Plant Physiology 64; 139-143. Macler, B. a. 1988. Salinity effects on photosynthesis, carbon allocation, and nitrogen assimilation in the red alga, Gelidium coulteri. Plant Physiology 88: 690-694. Mozafar, a., and J. R. GooDlN. 1970. Vesiculated hairs; a mechanism for salt tolerance in Atriplex halimus L. Plant Physiology 45; 62-65. Perkin-El.mer Corpor\tion. 1971. Analytical methods for atomic absorption spectrophotometrv'. Norwalk, Connecticut. PoovAiAH, B. W., and a. C. Leopold. 1976. Effects of organic salts on tissue permeability. Plant Physiolo- gy 58: 182-185. ROBICHAUX, R. H. 1984. Variation in the tissue water rela- tions of two sympatric Hawaiian Dubautia species and their natural hybrid. Oecologia (Berlin) 61; 75-81. Schultz, Q. E., and R. C. Kinch. 1976. The effect of temperature, light, and growth promoters on seed dormancy in western wheatgrass seed. Journal of Seed Technology 1; 79-85. Seemann, J. R., and C. Critchley. 1985. Effects of salt stress on the growth, ion content, stomatal behavior and photosynthetic capacity' of salt-sensitive species, Phaseolus vulgaris L. Planta 164: 151-162. Shannon, M. C. 1984. Breeding, selection, and the genetics of salt tolerance. Pages 231, 236 in Richard C. Staples and Gar\' H. Toenniessen, eds.. Salinity tolerance in plants. John Wiley & Sons, New York. Stewart, G. R., and I. Ahmad. 1983. Adaptation to salini- t\' in angiosperm halophytes. Pages 33—49 in D. A. Robb and W. S. Pieipoint, eds.. Annual proceedings of the phytochemical society of Europe #21. Metals and micronutrients: uptake and utilization by plants. Academic Press, New York. Stubbendieck, J., S. L. Hatch, and K. J. Hirsch. 1986. North American range plants. 3rd edition. Universi- t>' of Nebraska Press, Lincoln. Toole, V. K. 1976. Light and temperature control of ger- mination in Agropijron smithii seeds. Plant and Cell Physiology 17; 1263-1272. Vassey, T. L., and T. D. Sharkey. 1989. Mild water stress of Phaseolus vulgaris plants leads to reduced starch synthesis and extractable sucrose phosphate syn- thase activity. Plant Physiology 89; 106&-1070. Watad, A. A., L. Reinhold, and H. Lerner. 1983. Com- parison between a stable NaCl-selected Nicotania cell line and the wild type. Plant Physiology 73: 624-629. Witham, F. H., D. F. Blaydes, and R. M. Devlin. 1986. Exercises in plant physiology. 2nd edition. Prindle, Weber & Schmidt, Boston. Yeo, a. R., and T. J. Flowers. 1982. Accumulation and localization of sodium ions within the shoots of rice {Orijza sativa) varieties differing in salinity resis- tance. Physiologia Plantarum 56; 343. Accepted 15 August 1993 Great Basin Naturalist 54(2), ©1994, pp. 189-190 PREVALENCE OF ECTOPARASITE INFESTATION IN NEONATE YARROW'S SPINY LIZARDS, SCELOPORUS JARROVII (PHRYNOSOMATIDAE), FROM ARIZONA Stephen R. Goldberg^ and Charles R. Bursey^ Key words: chigger, Eutrombicula lipovskyana, mite, Geckobiella texana, Sceloporus jarrovii, Phrynosomatidae, neonate, prevalence, intensity. While it is well known that ectoparasites infest lizards (Frank 1981), we know of no reports concerning how quickly newborn (neonate) lizards are infested under natural conditions. Ectoparasites have been shown to cause a diffuse inflammatoiy response in the skin of infected lizards from natural popula- tions (Goldberg and Bursey 1991, Goldberg and Holshuh 1992). The purpose of this inves- tigation is to report the age at which ectopara- site (chigger and mite) infestation first occurs in neonate Yarrow's spiny lizards {Sceloporus jarrovii). This lizard is well suited for deter- mining age at which infestation first occurs since it is a live-bearing lizard in which partu- rition occurs within a short period of time near the end of June each year (Goldberg 1971). This contrasts with egg-laying lizards that may contain eggs for several months (Goldberg 1973), with hatchlings emerging over an extended period. Goldberg and Bursey (1992) reported on prevalence of the nematode Spaidigodon giganticus in neonate S. jarrovii. Methods Thirty-six neonate S. jarrovii were collect- ed by hand or hand-held noose 28-30 June 1991 at Kitt Peak (31°95'N, lir59'W, eleva- tion 1889 m) in the Baboquivari Mountains, 85 km SW of Tucson, Pima County, Arizona. Lizards were measured to the nearest mm snout-vent length (SVL), and ectoparasites were counted at time of capture. Sizes of these wild-caught specimens were compared to 223 S. jarrovii neonates born of 37 female captive lizards in 1967-69 (Goldberg 1970). Specimens were deposited in the herpetology collection of the Los Angeles County Natural History Museum (LACM) (139070-139105). Results and Discussion Lizards in the 1991 sample averaged 30.1 ± 2.0 mm SVL, range 26-36 mm. Eighteen of the 36 (50%) neonate S. jarrovii were infested by ectoparasites (Table 1). Seventeen (47%) were infested by chiggers {Eutrombicula lipovskijana), with a mean intensity of 6.5 ± 6.9 and a range of 1-26 chiggers per lizard. Three (8%) lizards were infested by larval Geckobiella texana, with a mean intensity of 3.0 ± 2.6 and a range of 1-6 mites per lizard. Adult G. texana were not present. Two infect- ed lizards had concurrent infections (£. lipovskijana and G. texana). The sample of 19 male and 17 female lizards contained 11 infested males (58%) and 7 infested females (41%). There was no statistical difference in rate of ectoparasite infestation between males and females (chi square = 1.0, 1 df, P > .05). Likewise, there was no statistical difference in intensity of infestation between male and female lizards (Kruskal-Wallis statistic = 0.46, 1 df, P > .05; E. lipovskijana and G. texana combined). Mean intensities were 5.7 ± 6.3 for infested males and 8.14 ± 9.20 for infested females. Eutrombicula lipovskijana was found most frequently within skin folds on both ventrolat- eral surfaces of the neck (the mite pockets of Arnold 1986), but they were occasionally encountered in other areas of the body. Geck- obiella texana was taken from the hind legs only. Representative specimens were deposited 'Department of Biology, Whittier College, Whittier, California 90608. ^Department of Biology, Pennsylvania State University', Shenango Valley Campus, 147 Shenango Avenue, Sharon, Pennsylvania 16146. 189 190 Great Basin Naturalist [Volume 54 Table 1. Infestation of neonate Sceloponis jarrovii by ectoparasites. # vvitli # with Eutrombicitla lipoiskiiana Geckohiella tcxaiui (#, intensity of chiggers (#, intensity of SVL N per lizard) mites per lizard) 26 1 0 0 27 2 1(26) 0 28 7 4 (5, 2, 1, 1) 0 29 3 1(3) 1(1) 30 8 4 (13, 8, 7, 2) 1(2) 31 7 3 (6, 3, 2) 0 32 5 2 (2, 1) 0 33 2 1(16) 0 34 0 — — 35 0 — — 36 1 1 (13) 1(6) in the U.S. National Parasite Collection (Beltsville, Maiyland 20705) as U.S. National Helminthological Collection Nos. 81992 and 82077 for E. lipovskyana and G. texana, respectively. Neonates born in captivity averaged 28.03 ± 0.98 mm SVL and ranged from 26 to 30 mm (Goldberg 1970). Thus, we estimate our field-collected sample to range from 1 day (those of 26-30 mm SVL) to 2 weeks (36 mm SVL) of age. It would appear that infestation can occur during the first few days of life, indeed, perhaps even on the day of birth (Table 1). To our knowledge, this is the only report indicating when ectoparasitic infesta- tion may first occur in the life history of lizards. The correlation coefficient (R) between SVL and number of mites was 0.16, suggesting to us that infestation of neonates by mites is opportunistic and can occur at any time after birth. Loomis and Stephens (1973) noted that hatchling Uta stansburiana from Joshua Tree National Monument, California, had very few chiggers attached but acquired more mites as they grew. They gave no estimate of age when infestation might first occur. Sceloporus jarrovii neonates grow rapidly, many of them reaching sexual maturity by autumn when they are 5 months of age and average 54 mm SVL (Ballinger 1973). We can- not speculate on the infestation of older juve- nile S. jarrovii since seasonal occurrence and abundance of E. lipovskyana are yet to be determined. Acknowledgment We thank M. L. Goff, University of Hawaii, Manoa, for identification of ectoparasites. Literature Cited Arnold, E. N. 1986. Mite pockets of lizards, a possible means of reducing damage by ectoparasites. Biolog- ical Journal of the Linnean Society 29: 1-21. Balllnger, R. E. 1973. Comparative demography of two viviparous iguanid lizards {Sceloporus jarrovi and Sceloponis poinsetti). Ecology 54: 269-283. Frank, W. 1981. Ectoparasites. Pages 359-383 in J. E. Cooper and O. F. Jackson, eds., Diseases of the Reptilia. Vol. 1. Academic Press, London, England. Goldberg, S. R. 1970. Ovarian cycle of the mountain spiny lizard, Sceloporus jarrovi Cope. Unpublished doctoral dissertation, University of Arizona, Tucson. 115 pp. , 1971. Reproductive cycle of the ovoviviparous iguanid lizard Sceloponis jarrovi Cope. Herpetolog- ica27: 123-131. . 1973. Ovarian cycle of the western fence lizard, Sceloporus occidentalis. Herpetologica 29: 284-289. Goldberg, S. R., and C. R. Bursey. 1991. Integumental lesions caused by ectoparasites in a wild population of the side-blotched lizard (Uta stansburiana). Jour- nal of Wildlife Diseases 27: 68-73. . 1992. Prevalence of the nematode Spauligodon giganticus (Oxyurida: Pharyngodonidae) in neonatal Yarrow's spiny lizards, Sceloporus jarrovii (Sauria: Iguanidae). Journal of Parasitology 78: 539-541. Goldberg, S. R., and H. J. Holshuh. 1992. Ectopara- site-induced lesions in mite pockets of the Yarrow's spiny lizard, Sceloponis jarrovii (Phrynosomatidae). Journal of Wildlife Diseases 28: 537-541. Loo.Mis, R. B., and R. C. Stephens. 1973. The chiggers (Acarina, Trombiculidae) parasitizing the side- blotched lizard (Uta stansburina) and other lizards in Joshua Tree National Monument, California. Bul- letin of the Southern California Academy of Sci- ences 72: 78-89. Received 9 November 1992 Accepted 13 July 1993 Great Basin Naturalist 54(2), ©1994, pp. 191-192 SEASONAL VARIATION AND DIET SELECTION FROM PELLET REMAINS OF SHORT-EARED OWLS {ASIO FLAMMEUS) IN WYOMING Eric Stoned Jocelyn Smith^-, and Polly Thornton^ Key words: Short-eared Otvh, Asio flammeus, diet selection, predators, Wyoming. Short-eared Owls {Asio flammeus) are medium-sized predators of open country, sage flats, grasslands, and roadsides. Often active well after sunrise, they are more diurnal than other owls in northwest Wyoming (Karulus and Eckert 1974, Clark 1985). Their foraging areas significantly overlap those of both small- er and larger owls, namely Great Horned Owls {Bubo virginianus) and Burrowing Owls {Speotyto cunicularia) (E. Stone unpublished data, Karulus and Eckert 1974). Prey sources, including small mammals, birds, and insects, are diverse and overlap those used by Great Horned Owls. This study examines shifts in prey sources through the breeding season by identifying prey remains in Short-eared Owl pellets from wild birds. Shifts in prey sources may be the result of changes in prey abundance or avail- ability, or competition with other owl species for the same resources. Additionally, shifts may result from changes in dietary require- ments of adults or their developing dependent offspring. In this study we sought to describe whether shifts in diet occurred and, if so, what types. This descriptive study may serve as a useful baseline of data upon which future studies can be based. Methods and Study Area The Short-eared Owl study area was an old irrigation ditch located approximately 2.2 km southwest of the Teton Science School in Grand Teton National Park, Wyoming. Short- eared Owl pellets were located by searching on the ground and at the base of willows {Salix sp.), mountain alder {Alnus tenuifolio), aspen {Populus tremuloides), and narrowleaf cotton- wood {Populus angustifolia). One active nest was located within 20 m of the ditch and another within 2 km. We observed as many as four owls roosting along the ditch, either on the ground, in the shade of trees, or perched on the lower branches. At the end of each month (March-October) all pellets were collected at the study site. Thus, each group of pellets collected and their contents could be assumed to have originated during that month. Short-eared Owls were no longer seen in the study area in late October and were presumed to have migrated to areas with ample winter prey, shallower snowjjack, or both. Owls were first seen using the roost site in early March. To assure large enough sample sizes, we combined sample months into the following seasonal groups: spring (March, April, and May), summer (June and July), and fall (August, September, and Octo- ber). Prey items were identified using skaill and teeth parts found in individual pellets. Pellet remains of Microtus montanus and M. longi- caudus were not distinguishable by skull or teeth parts and were combined into a prey category hereafter referred to as M. mont-long. Results Short-eared Owl pellets contained 11 dif- ferent prey items. Of these, the 6 most com- mon prey types constituted 94.42% of the diet. A significant decline in sage voles {Lagurus curtatus) in the fall diet of Short-eared Owls was augmented by an increase in the propor- tions of northern pocket gophers {Thijmomys talpoides) and southern red-backed voles {Clethhonomijs gapped) (Table 1). The com- plete disappearance of L. curtatus in fall, with increases in both the number and proportion 'Teton Science School, Box 68, Kelly, Wyoming 83011. ^Present address: 454 Route 32 North, New Paltz, New York 12561. 191 192 Great Basin Naturalist [Volume 54 Table 1. Seasonal percentages of prey items foinid in Short-eared Owl pellets Prey type Mar-Apr-Ma\- Jnn-July Aug-Sep-Oct Totals Microtiis mont-lonf!, 41.37 46.15 42.28 42.92 Thijmomijs talpokles 10.34 17.31 27.64* 21.03 Peromijsciis manicuhittis 24.14 17.31 16.26 18.45 Lagunis curtatus 13.79 13.46 0.00* 6.44 Cleihrionomys happen 1.72 3.85 8.13* 5.58 Sorex sp. 3.45 0.00 2.44 2.15 ZapiLS princeps 3.45 1.92 0.00 1.29 Tamias minimus ().()() 0.00 0.81 0.43 Microtiis pemisylvaniciis 0.00 0.00 0.81 0.43 Unknown bird 0.00 0.00 0.81 0.43 Unidentified beetle 0.00 0.00 0.81 0.43 Number of prey items 58 52 123 233 *Siniiifkaiit iiR-reasf or decrease in diet (P < .0.5. clii-sqiiare post-hoc cell contributions). of T. talpoides and C. gapped in the diet of Short-eared Owls, represents a significant sea- sonal change in overall diet selection or forag- ing locations. Discussion Short-eared Owl's significant seasonal vari- ation in prey selection may be reflective of changes in the availability of their prey. Sage- brush voles {Lagunis curtatus) are reported to become inactive during dry periods corre- sponding to late summer and fall in western Wyoming (Clark and Stromberg 1987:177). Declines in prey such as L. curtatus, found in open areas containing sage or grassland habi- tats, may indicate that Short-eared Owls forage more in forest edges or under tree canopies during the latter part of the summer. These habitats are where M. montanus, M. longi- cauclus, and C. gapperi are found. T. talpoides, which also increased in the diet later in the season, is found in a variety of habitats with loose soil (Clark and Stromberg 1987). In Grand Teton National Park and else- where, there is strong evidence that small mammal prey availability is dependent on environmental factors and climate (Pinter 1988). In 1993, one year later, a continuation of this study was planned. However, a sudden and prolonged period of warm temperatures resulted in rapid snowmelts and subsequent flooding of the subnivean environment (per- sonal observation). Population studies of small mammals being conducted in the same area found that 1993 summer populations were the lowest recorded in 25 years of monitoring (A. Pinter personal communication). In 1993 Short-eared Owls were first seen at the study area on 15 March but were absent for the duration of the summer. It was assumed that the owls moved their foraging and breeding activities to areas that were not affected by the subnivean flooding and depression of small mammal populations. These observations and the results of more normal years suggest that Short-eared Owls possess the flexibility to shift diets and forag- ing areas with changing seasonal or annual prey availability. Acknowledgments The manuscript benefited from comments made by Robert C. Whitmore and an anony- mous reviewer. Steve Cain and Rick Wallen of Grand Teton National Park offered insights and supported our efforts throughout the study. Literature Cited Clark, R. J. 1985. A field study of the Short-eared Owl, Asio flammeiis (Pontoppidan), in North America. Wildlife Mongraph 47. Clark, T. W., and M. R. Stromberg. 1987. Mammals in Wyoming. University Press of Kansas, Lawrence. Karulus, K. E., and a. W. Eckert. 1974. The owls of North America. Doubledav, Garden City, New York. Pinter, A. J. 1988. Multiannnal fluctuations in precipita- tion and population dynamics of the montane vole, Microtus montamis. Canadian Journal of Zoology 66: 2128-2132. Received 19 March 1993 Accepted 28 September 1993 INFORMATION FOR AUTHORS The Great Basin Naturalist welcomes previously unpublished manuscripts pertaining to the biologi- cal natural history of western North America. Pref- erence will be given to concise manuscripts of up to 12,000 words. SUBMIT MANUSCRIPTS to Richard W. Baumann, Editor, Great Basin Naturalist, 290 MLBM, PO Box 20200, Brigham Young University, Provo, UT 84602-0200. A cover letter accompanying the man- uscript nuist include phone number(s) of the author submitting the manuscript, and FAX number and E-mail address when applicable; the letter must also provide information describing the extent to which data, text, or illustrations have been used in other papers or books that are published, in press, submitted, or soon to be submitted elsewhere. Authors should adhere to the following guidelines; manuscripts not so prepared may be returned for revision. MANUSCRIPT PREPARATION. Consult Vol. 51, No. 2 of this journal for specific instructions on style and format. These instructions, Guidelines for Manuscripts Submitted to the Great Basin Naturalist, supply greater detail than those pre- sented here in condensed form. The guidelines are printed at the back of the issue; additional copies are available from the editor upon request. Also, check the most recent issue of the Great Basin Nat- uralist for changes and refer to the CBE Style Man- ual 5th edition (Council of Biology Editors, One Illinois Center, Suite 200, 111 East Wacker Drive, Chicago, IL 60601-4298 USA, piione: (312) 616- 0800, F.\X: (.312) 616-0226; $24)._ TYPE AND DOUBLE SPACE all materials, including literature cited, table headings, and figure legends. Avoid hyphenated words at the right-hand margins. Underline words to be printed in italics. Use stan- dard bond (22x28 cm), leaving 2.5-cm margins on all sides. SUBMIT 3 COPIES of the manuscript and the origi- nal on a 5.25- or 3.5-inch disk utilizing WordPerfect 4.2 or above. Number all pages and assemble each copy separately: title page, abstract and key words, text, acknowledgments, literature cited, appen- dices, tables, figure legends, figures. TITLE PAGE includes an informative title no longer than 15 words, names and addresses of authors, a running head of fewer than 40 letters and spaces, footnotes to indicate change of address and author to whom correspondence should be addressed if other than the first author ABSTRACT states the purpose, methods, results, and conclusions of the research. It is followed by 6-12 key words, listed in order of decreasing importance, to be used for indexing. TEXT has centered main headings printed in all capital letters; second-level headings are centered in upper- and lowercase letters; third-level head- ings begin paragraphs. VOUCHER SPECIMENS. Authors are encouraged to designate, properly prepare, label, and deposit high-quality voucher specimens and cultures docu- menting their research in an established permanent collection, and to cite the repository in publication. REFERENCES IN THE TEXT are cited by author and date: e.g., Martin (1989) or (Martin 1989). Multiple citations should be separated by commas and listed in chronological order. Use "et al." after name of first author for citations having more than two authors. ACKNOWLEDGMENTS, under a centered main heading, include special publication numbers when appropriate. LITERATURE CITED, also under a centered main heading, lists references alphabetically in the fol- lowing formats: Mack, G. D., and L. D. Flake. 1980. Habitat rela- tionships of waterfowl broods on South Dakota stock ponds. Journal of Wildlife Management 44: 695-700. Sousa, W. P 1985. Disturbance and patch dynamics on rocky intertidal shores. Pages 101-124 in S. T A. Pickett and R S. White, eds.. The ecolo- gy of natural disturbance and patch dynamics. Academic Press, New York. Coulson, R. N., and J. A. Witter. 1984. Forest ento- mology: ecology and management. John Wiley and Sons, Inc., New York. 669 pp. TABLES are double spaced on separate sheets and designed to fit the width of either a single column or a page. Use lowercase letters to indicate foot- notes. PHOTOCOPIES OF FIGURES are submitted initially with the manuscript; editors may suggest changes. Lettering on figures should be large enough to withstand reduction to one- or two-column width. Originals must be no larger than 22x28 cm. NOTES. If the manuscript would be more appro- priate as a short communication or note, follow the above instructions but do not include an abstract. A CHARGE of $50 per page is made for articles published; the rate for individual subscribers will be $35 per page. However, manuscripts with com- plex tables and/or numerous half-tones will be assessed an additional charge. Reprints may be pur- chased at the time of publication (an order form is sent with the proofs). FINAL CHECK: • Cover letter explaining any duplication of information and providing phone number(s), FAX number, and E-mail address • 3 copies of the manuscript and WordPerfect diskette • Conformity with instructions • Photocopies of illustrations (ISSN 001 7-361 4) GREAT BASI N NATU RALIST voi 54 no 2 ap n 1 994 CONTENTS Articles Colony isolation and isozyme variability of the western seep fritillary, Speyeria nokomis apacheana (Nymphalidae), in the western Great Basin Hugh B. Britten, Peter E Brussard, Dennis D. Murphy, and George T. Austin 97 Influence of fine sediment on macroinvertebrate colonization of surface and hyporheic stream substrates Carl Richards and Keniiit L. Bacon 106 Resource overlap between mountain goats and bighorn sheep John W. Laundre 114 Predation of artificial Sage Grouse nests in treated and untreated sagebrush Mark E. Ritchie, Michael L. Wolfe, and Rick Danvir 1 22 Timing, distribution, and abundance of kokanees spawning in a Lake Tahoe tributary David A. Beauchamp, Phaedra E. Budy, Brant C. Allen, and Jeffrey M. Godfrey 1 30 Eull-glacial shoreline vegetation during the maximum highstand at Owens Lake, California Peter A. Koehler and R. Scott Anderson 1 42 Redband trout response to hypoxia in a natural environment Mark Vinson and Steve Levesque 1 50 Field study of plant survival as affected by amendments to bentonite spoil Daniel W. Uresk and Teruo Yamamoto 156 Population structure and ecological effects of the crayfish Pacifastacus lenius- culus in Castle Lake, California James J. Elser, Christopher Junge, and Charles R. Goldman 1 62 Brood habitat use by Sage Grouse in Oregon Martin S. Drut, John A. Crawford, and Michael A. Gregg 1 70 Needle biomass equations for singleleaf pinyon on the Virginia Range, Nevada... T. R. De Rocher and R. J. Tausch 1 77 Some physiological variations o( Agropyron smithii Rydb. (western wheatgrass) at different salinity levels Rengen Ueng, Ivo E. Lindauer, and Warren R. Buss 1 82 Notes Prevalence of ectoparasite infestation in neonate Yarrow's spiny lizards, Sceloporus jarrovii (Phrynosomatidae), from Arizona Stephen R. Goldberg and Charles R. Bursey 1 89 Seasonal variation and diet selection from pellet remains of Short-eared Owls {Asio flammeus) in Wyoming Eric Stone, Jocelyn Smith, and Polly Thornton 191 MCZ LIBRARY H E 11 1^ SEP 2 7 1994 GREAT BASIH NATURALIST HARVARD ERSITY VOLUME 54 N2 3 — JULY 1994 BRIGHAM YOUNG UNIVERSITY GREAT BASIN NATURALIST Editor Assistant Editor Richard W. Baumann Nathan M. Smith 290MLBM 190MLBM PO Box 20200 PO Box 26879 Brighani Young University Brigham Young University Provo, UT 84602-0200 ' Provo, UT 84602-6879 ' 801-378-5053 801-378-6688 FAX 801-378-3733 E-mail: NMS@HBLL1.BYU.EDU Associate Editors Mic:iiAEL A. Bovvers Paul C. Marsh Blandy Experimental Farm, Univer.sity of Center for Environmental Studies, Arizona Virginia, Bo.x 175, Boyce, VA 22620 State University Tempe, AZ 85287 J. R. Callaha.n Stanley D. Smith Museum of Southwestern Biology, University of Department of Biology New Mexico, Albuquerque, NM University of Nevada-Las Vegas Mailing address: Box 3140, Hemet, CA 92546 Las Vegas, NV 89154-4004 Jeffrey ]. Johansen Paul T. Tueller Department of Biology, John Carroll University Department of Environmental Resource Sciences University Heights, OH 441 18 University of Nevada-Reno, 1000 Valley Road Reno, NV 89512 Boris C. Kondratieff Department of Entomology, Colorado State Robert C. Whitmore University, Fort Collins, CO 80523 Division of Forestry, Box 6125, West Virginia University, Morgantown, WV 26506-6125 Editorial Board. Jerran T. Flinders, Chairman, Botany and Range Science; Duke S. Rogers, Zoology; William Hess, Botany and Range Science; H. Duane Smith, Zoology. All are at Brigham Young University. Ex OflPicio Editorial Board members include Steven L. Taylor, College of Biology and Agriculture; Stanley L. Welsh, Director, Monte L. Bean Life Science Museum; Richard W. Baumann, Editor, Great Basin Naturalist. The Great Basin Naturalist, founded in 1939, is published quarterly by Brigham Young University. Unpublished manuscripts that further our biological understanding of the Great Basin and surrounding areas in western North America are accepted for publication. Subscriptions. Annual subscriptions to the Great Basin Naturalist for 1994 are $25 for individual sub- scribers ($30 outside the United States) and $50 for institutions. The price of single issues is $12. All back issues are in print and available for sale. All matters pertaining to subscriptions, back issues, or other busi- ness should be directed to the Editor, Great Basin Naturalist, 290 MLBM, PO Box 20200, Brigham Young University Provo, UT 84602-0200. Scholarly Exchanges. Libraries or other organizations interested in obtaining the Great Basin Naturalist through a continuing exchange of scholarly publications should contact the Exchange Librarian, 6385 HBLL, PO Box 26889, Brigham Young University Provo, UT 84602-6889. Editorial Production Staff JoAnne Abel Technical Editor Jan Spencer Assistant to the Editor Copyright © 1994 by Brigham Young University ISSN 0017-3614 Official publication date: 29 August 1994 8-94 750 11055 The Great Basin Naturalist Published at Provo, Utah, by Brigham Young University ISSN 0017-3614 Volume 54 31 July 1994 No. 3 Great Basin Naturalist 54(3), © 1994, pp. 193-203 HABITAT REQUIREMENTS FOR ERIGERON KACHINENSIS, A RARE ENDEMIC OF THE COLORADO PLATEAU Loreen Allphin^ and Kimball T. Harper^ Abstr.\ct. — Erigeron kachinensis is a rare endemic of the Colorado Plateau in southeastern Utah. This perennial composite grows in small, isolated populations at seeps and alcoves arising along canyon walls in Cedar Mesa Sandstone suhstrates. Characteristics of si.x Erigeron kachinensis sites in Natural Bridges National Monument, San Juan County, Utah, were studied to detennine habitat requirements for this species. Sites were analyzed with respect to geology, soil chemistry, physical properties, and vegetational characteristics. The alcoves studied were veiy saline, often with soil sur- faces covered with a white crust of salt. Living cover was enhanced by perennially moist soils, diminished amounts of solar radiation, soil salinity, and above-average amounts of available soil phosphorus. Kachina daisy vegetative growth appears to be favored by these same abiotic factors. The most commonly associated plant species on E. kachinensis sites were Aquilegia micrantha, Calamagrostis scopulorum, Zigacleniis vaginatiis, and Erigeron kachinensis. These species and the daisy accounted for more than 75% of the total living cover in the alcoves studied. A principal components analysis procedure was developed for evaluating site suitability for Erigeron kachinensis. This daisy has been successfully intro- duced to a site selected using that model. Key words: Erigeron kachinensis, Colorado Plateau, critical habitat. The Kachina daisy {Erigeron kachinensis ommendation to "threatened" status for the Welsh & Moore) was discovered and named in species (U.S. Department of Interior 1988). 1968 (Welsh and Moore 1968). It is a rare Currently, the Kachina daisy is listed by the perennial composite of the Colorado Plateau U.S. Fish and Wildlife Service as a category 2 region of Utah and Colorado. The species species, i.e., a species for which more informa- grows in small, isolated populations at seeps tion is needed before assigning final designa- and alcoves arising along the edges of deep tion (endangered, threatened, or sensitive), canyons in these areas. No discussions of the In 1984 an effort was launched to document habitat requirements of this species have locations of all populations of the Kachina daisy appeared in the literature. The Kachina daisy and tlieir size in Natinal Bridges National Mon- was proposed as "endangered" by the U.S. ument (NBNM), San Juan County, Utah. In that Fish and Wildlife Service on 16 June 1976 study eight populations were recorded (Fagan (U.S. Department of Interior 1975, 1976). 1984). In 1988 and 1989 the National Park Later proposals downgraded the original rec- Sei-vice selected four of the eight populations 1 Department of Biology, University of Utah, Salt Lake Cit); Ut;Ji 84112. ^Department of Botany and Range Science. Brigham Young Uni\ersity, Provo, Utah 84602. Please address all reprint requests to this author. 193 194 Great Basin Naturalist [Volume 54 for a 2-year monitoring program. Relative cover of Kachina daisy plants in these popula- tions was reported to vary by almost 60% between years (Belnap 1988, 1989). Belnap's report to the National Park Service called for protection and further research to determine whether the observed declines in population size were due to natural fluctuations in tem- perature and precipitation or to more perma- nent stresses in the physicochemical environ- ment. It was also considered possible that the species was poorly adapted to the associated environmental conditions. Since this species appears to be restricted to seeps and alcoves, typically along a single seepline within a canyon, there is concern for its preservation. If drought caused seeplines occupied by this species to dry up, many popu- lations could be eliminated. As tourism increas- es in the canyon regions of southeastern Utah, so does the threat of human impacts on rare species. Tourists hiking in canyon bottoms look to shady alcoves as refuge from the hot summer sun. Many of the seeps and alcoves contain small, prehistoric Anasazi Indian ruins that also attract tourists. At the outset of this study, the Kachina daisy was known only from NBNM and a pop- ulation in Montrose County, Colorado. With but few known populations confined to un- common site situations, resource managers were concerned that the species might be vul- nerable to extinction. These concerns resulted in the research presented in this manuscript. Our objectives were to determine habitat requirements for the Kachina daisy and to develop a management strategy for preserva- tion of the species. If habitat requirements can be established accurately, potentially occupi- able sites for the species could be identified and new populations established. Study Area and Methods The Kachina daisy grows in seeps and nat- ural alcoves in the Cedar Mesa Sandstone of White and Armstrong canyons in NBNM between 1680 and 1890 m elevation. The Cedar Mesa Formation, deposited in Permian time (Hintze 1975), is a coarse-grained, porous sandstone that stores considerable water within its massive deposits. In thickness the Cedar Mesa Formation varies from 500 to 1200 m. Its depth is at a maximum in the Elk Ridge and NBNM area. Alcoves where the daisy grows are developed in Cedar Mesa Sandstone immediately above the Halgaito Formation. The Halgaito Formation lies below the Cedar Mesa Sandstone. It is a part of the Cutler group, as is the Cedar Mesa Sandstone. The Halgaito Formation consists of mostly reddish brown siltstone, sandstone, and thin beds of limestone. It varies from 400 to 500 m in thickness (Hintze 1975). Where percolating water encounters finer-textured lamina in the Cedar Mesa Sandstone, it accumulates and moves laterally until it reaches the faces of verti- cal walls of deep, narrow drainage channels that cross outcrops of the Cedar Mesa Formation. A survey of canyons surrounding NBNM for Erigeron kachinensis was initiated under a contract with the Bureau of Land Manage- ment. In addition to populations known in NBNM (Welsh et al. 1987), the Kachina daisy has now been found on lands administered by the Bureau of Land Management (Allphin and Haiper 1991). Populations are known in Fish, Arch, White, and Birch canyons; from Dark Canyon Primitive Area in San Juan County, Utah; and from the lower portion of Coyote Wash, Montrose County, Colorado (Welsh et al. 1987, Allphin and Haiper 1991). Six of the eight populations in NBNM were chosen for detailed study based on accessibili- ty and contrasting site characteristics (Fig. 1). The six study sites vaiy in respect to aspect, soil moisture, and soil salinity. These sites include two west-facing, two south-facing, and two north-facing alcoves. All study sites are located along the 1768-m (5800-ft) contour line. Study sites were monitored in June 1990, 1991, and 1992. It is pertinent to note that all sites occur in alcoves or seeps except site 2, which occupies a slope kept moist by seepage from an alcove directly above. At each site individual daisy plants were selected for study using the point center quar- ter method (Cottam and Curtis 1956) along a 100-m transect. These individuals were marked with numbered aluminum tags held at their bases with galvanized nails. The tags facilitated later relocation of individual plants. Several abiotic characteristics were de- scribed by taking composite soil samples from the six study populations. A minimum of five composite soil samples were taken per alcove. From those soil samples, soil water content. 1994] ERIGERON KACHINENSIS HaBITAT 195 Fig. 1. Location of Kachina daisy sites monitored in this study and a site where Kachina daisy plants have been transplanted in Natural Bridges National Monument, San Juan County, Utah (# I — north-facing alcove A, # 2 = north-facing alcove B, # 3 = south-facing alcove A, # 4 = south-facing alcove B, # 5 = west-facing alcove A, # 6 = west-facing alcove B, and # 1 — transplant site). percentage of coarse material (diameter >2 mm), soil texture, percentage of sand, percent- age of organic matter, pH, electrical conduc- tivity, and concentrations of the biogenic ele- ments (I^ K, Ca, Mg, Na, Cu, Fe, Mn, and Zn) were estimated. All soil analyses were made by personnel of the Soil and Plant Analysis Laboratory, Department of Agronomy and Horticulture, Brigham Young University, and all analytical methods were based on those recommended by Black et al. (1965). Soil tex- ture was determined with a hydrometer. Reaction (pH) of soil was taken with a glass electrode on a saturated soil-water paste. Organic matter was quantitated by digestion with 1.0 N potassium dichromate. Phosphorus was determined with the iron-TCA-molybdate method on a soil extract taken with .2 N acetic acid. Exchangeable bases were freed from the soil with 1.0 N ammonium chloride. Ion con- centrations in extract solutions were estimated by atomic absorption. Vegetational data were also collected at each study site in NBNM. Frequency and cover of individual species and percentage of total living plant cover were determined for each study site using a lOO-cm^ quadrat. In addi- tion, vegetative and floral characteristics of 600 tagged plants were recorded at each visit. Variables evaluated for each plant include number of rosettes, crown diameter, leaf length (average for two longest leaves), num- ber of leaves, number of flower heads, number of flowers per head, and number of filled seeds per flower head. The number of filled seeds per head was determined by harvesting mature heads in June and observing whether seeds were filled with mature embryos. Soil data from the various study sites were analyzed by one-way analysis of variance (AN OVA) for significant differences between sites. The least significant difference multiple range test of Snedecor and Cochran (1967) was used to determine significant differences among individual means for each site. Vegetational data were analyzed using a two- way analysis of variance for both site and year. Once again the multiple range test was used to identify significant differences among means for all measured vegetative characteristics among the six alcoves. ANOVAs were per- formed using the STATA statistical package (Computing Resource Center 1992). To determine the effect of the abiotic envi- ronment on vegetative growth in alcoves con- taining the Kachina daisy, abiotic environmen- tal variables were regressed against total living cover, species richness, and Kachina daisy characteristics. Since no significant regression correlations could be found between abiotic environmental and individual daisy character- istics, a whole-plant response index was deter- mined using standardized values for various daisy characteristics among the six alcoves. Vegetative daisy characteristics were stan- dardized by setting the largest value for each variable at 100 and expressing the value for 196 Great Basin Naturalist [Volume 54 that parameter at each study site as a percent- age of the maximum vahie. These standardized values were summed for each of the alcoves to give an integrated estimate of overall plant response. A reproductive daisy response was also measured as number of filled seeds based upon total number of heads produced by indi- vidual daisy plants and number of flowers per head. Both simple and multiple linear regres- sions were performed to evaluate the response of various plant variables to the following abi- otic factors: moisture content, salinity, and available phosphorus in the soil. All regres- sions were performed using the STATA statis- tical package (Computing Resource Center 1992). Composite soil samples were also taken from populations found on lands administered by the Bureau of Land Management. Species found in association with the Kachina daisy were recorded for each population. An alcove found near Sipapu Bridge within NBNM con- tained no Kachina daisies but did harbor sev- eral of the species commonly associated with the daisy (Fig. 1). This location was considered a potential introduction site for the Kachina daisy. A composite soil sample from the site was analyzed to further assess its suitability for this daisy. Mean soil characteristics from sites on BLM lands and the potential transplant site were analyzed by multiple range comparison with the six study alcoves. Principal compo- nents analysis (Pielou 1984)) was used to finther evaluate the suitability of the Sipapu Bridge site for the Kachina daisy using all environ- mental variables in a multivariate analysis. Soil characteristics considered in that analysis included phosphorus (mg/kg), electrical con- ductivity, average soil moisture (June), and average soil temperature (June). Total species richness at each site was also used in the prin- cipal components analysis. The principal com- ponents analysis was conducted using the SAS statistical package (SAS Institute, Inc. 1993 ). Demography and reproductive biology of the Kachina daisy are addressed in greater detail in another manuscript. Results present- ed in this paper will deal only with informa- tion concerning habitat of the species. Results Physical Environment Soils in alcoves occupied by Kachina daisy are sandy loams ranging from 70.4 to 89.6% sand (Table 1). They are very alkaline with pH values of 7.8-9.1. All alcoves studied develop a crusty layer of white salt on the soil surface during the drier part of the year. Data demon- strate that E. kachinensis is very tolerant of saline conditions. Electroconductivit)' of soils from the study sites ranges from 6 to 31 mmhos/cm. The average electrical conductivi- ty value for soils from the six alcoves was 13.6 mmhos/cm. Conductivity values over 8 mmhos/cm are considered high enough to restrict vield of most crop plants (Richards 1954). Soil temperatures (at ~1 cm depth) at sites occupied by the daisy are cool, ranging from 13.2 to 15.8° C in June (Table 1). Soils in alcoves are always cooler than surrounding soils because they receive less sunlight and stay reasonably moist. Soil temperatures in June vaiy only slightly among sites, with high- er temperatures occurring at sites receiving greater amounts of direct sunlight. Percent soil water content in occupied alcoves varies from 5.8 to 25.6% of diy weight in mid-June (Table 1). NBNM receives most of its precipitation from November through March (Fig. 2; Brough et al. 1987). Water accumulates in the sandstone during these months but is available for plant growth in all seasons because of the large reservoir of water held in the sandstone. Alcoves 3 and 6 have the wettest soils (Table 1). This is perhaps related to the fact that these alcoves receive only about 1 h of direct sun- light per day in June. Site 2 has the driest soil. It occurs on a slope outside and directly below an alcove seep. In addition, it receives the most sunlight daily and is farther from a seepline than other study sites. It faces north and is exposed to sunlight only during after- noon hours. All alcoves considered had similar levels of phosphorus (7.5-11.4 mg/kg). They also had high values for potassium, calcium, magne- sium, and sodium (Table 1). ANOVA and mul- tiple range tests demonstrate that soils at alcove 3 differ significantly from those of other alcoves. Alcove 3 soils have significantly high- er electrical conductivity, a greater percentage of Na saturation, and significantly higher levels of potassium, magnesium, and sodium. This alcove also has significantly lower levels of manganese and calcium than the other alcoves studied. These differences may be related to 1994] ERIGERON KACHINENSIS HaBITAT 197 Table 1. Chemical and physical characteristics of soil from alcoves that support Erigeron kachinensis at Natural Bridges National Monument (1-6), San Juan Count); Utah. Each value represents an average of five samples from each alcove. Alcove 7 represents potential transplant site in Natural Bridges National Monument. Column 8 represents a mean of five alcoves in surrounding canyons on BLM land. Means followed by the same letter do not differ at the p < .05 level of significance. Soil temperature and moisture values represent average conditions for mid- June in summers of 1990-92. 1 3 Alcove no. 4 Elevation (m) Aspect (deg. ) Hr. direct sun- light/day in June Texture 1768 1768 1768 1768 1768 1768 302(\VNVV) 328(nn\v) 160(sse) 197(ssw) 252(\vs\v) 225(s\v) 5.0 5.75 l.O 4.,5-5.25 4.6 0-1.5 idv Im loamv snd sandv Ir idvl sandv Im sandy im 8 1768 1859 250(\vsw) 191(ss\v) dy Im sandy Im sandy Im sandy Im Sand (%) pH EC (mmhos/cm) Ave. soil temp (C) Soil moisture (%) Organic matter (%) Skeletal mtl. (% by wt.) 89.6 a 8.6 c 76.7 a 8.8 cd 13.1 be 17.1c 80.3 a 9.1 d 31.1 d 14.1a 12.3 b 1.8 b 12.1 ab 15.8 be 13.9 a 5.8 a 1.1a 15.1b 18.4 c 2.0 be 17.2 b 70.4 a 8.6 c 7.6 ab 15.1b 12.2 b 2.2 c 6.8 a 80.1a 7.9 ab 6.9 a 13.9 a 12.7 b 0.8 a 7.2 ab 71.7 a 73.3 a 66.8 a 7.8 a 8.6 c 8.2 b 6.0 a 1.6 b 7.2 ab 9.8 b 13.2 a 13.5 a 16.4 c 25.6 d 13.2 b 12.3 b P (mg/kg) K (mg/kg) Ca (mg/kg) Mg (mg/kg) Na (mg/kg) Cu (mg/kg) Fe (mg/kg) Mn(mg/kg) Zn (mg/kg) Na saturation (%) of ex- changeable bases 11.4 a 10.7 a 197.0 ab 229.5 b 9590 c 2498 a 649.0 ab 0.48 ab 1.8 b 6.8 c 0.60 b 2.5 b 7886 b 2754 a 822.4 b 0.36 a 0.96 a 6.0 be 0.44 a 3.5 be Essential elements 9.8 a 10.9 a 640.2 c 153.0 a 6758 a 10508 d 11904 b 2030 a 2535.0 c 322.6 ab 0.56 b 0.32 a 1.4 ab 1.8 a 0.72 b 6.0 c 1.9 b 6.6 c 0.88 c 1.2 ab 7.5 a 80.4 a 10884 d 870 a 126.7 a 0.36 a 1.9 b 2.4 a 0.96 c 0.5 a 8.0 a 82.0 a 10282 cd 730 a 127.4 a 0.60 b 2.2 b 4.8 b 0.72 b 0.6 a 9.3 a 10.2 a 198 Great Basin Naturalist [Volume 54 Natural Bridges Nat. Mon. (Elev. = 1768 m) Mean Ann. Temp. = 10.47 C Mean Ann. Precip. = 316.8 mm Fig. 2. 1987-91. Climatic diagram for Natural Bridges National Monument. Data represent average values for the period the fact that the floor of alcove 3 is developed from a reddish brown siltstone/sandstone probably of the Halgaito Formation rather than the Cedar Mesa Sandstone. Even though alcove 3 soils differ from those of other alcoves, the daisy population is thriving and seems unaffected by soil differences. Kachina Daisy Characteristics Characteristics of the average daisy in each of the alcoves studied intensively are noted in Table 2. The average number of rosettes per individual varied from 1.4 to 2.5 among the six sites. Statistical differences among the sites with respect to the number of rosettes per daisy plant are noted in Table 2. Alcoves 3 and 4 supported the largest plants, while the smallest plants grew in alcoves 2 and 5. Average clump diameter for the Kachina daisy ranged from 3.6 to 7.3 cm among the study sites. Alcove 2 plants had the smallest diameters, which perhaps reflects the fact that alcove 2 had the driest soil observed among the alcoves (Table 2). Average leaf lengths ranged from 1.9 to 3.6 cm among the six study sites. Alcove 2 plants had the smallest leaves observed. Average number of leaves per plant ranged from 9.9 to 23.6 among the study sites. As might have been expected, the number of leaves per plant was smallest for alcove 2 (Table 2). Statistical differences seen among charac- teristics of plants from these alcoves reflect the fact that alcoves were chosen for apparent differences in moisture, aspect, and direct sunlight. Plants from alcoves that receive little direct sunlight (sites 3, 4, and 6; Table 1) had larger leaves and clump diameters than plants from alcoves 1 and 5, which received more sunlight. The amount of sunlight could also be expected to affect soil moisture, which should also produce differences in vegetative charac- teristics. Biotic Associates Ecological and vegetational characteristics showed no significant differences among the years of study (1990-92); thus, data given in 1994] Erigeron kachinensis Habitat 199 Table 2. Ecological and vegetational characteristics and average characteristics of 100 randomly selected Kachina daisy plants at each of the six alcoves shown in Table 1. All values are based on data taken during field seasons 1990-92. Means followed by the same letter do not differ significantly at p < .05. Alcove no. 1 2 3 4 5 6 Total li\'ing cover (%) 42.5 c 17.8 a 45.5 c 62.4 d 18.1a 37.7 b Species richness* 6.3 a 5.3 b 4.7 c 9.0 d 4.3 c 7.0 e Ave. # spec./ 100-cni^ quad. 2.8 d 1.8 b 1.6 ab 2.6 d 1.4 a 2.1c Characteristics of E. kachitietisis No. rosettes/plant 1.6 a 1.4 a 2.5 c 2.1 h 1.6 a 1.9 b Plant diam. (cm) 4.8 b 3.6 a 6.2 c 7.3 d 4.4 b 6.6 c Leaf length (cm) 2.7 b 1.9 a 2.5 b 3.6 c 2.1a 3.4 c No. leaves/plant 13.2 b 9.9 a 23.6 d 22.2 d 13.5 b 17.3 c Plant response indices** 492 (veg.) 225 (rep.) 355 (veg.) 202 (rep.) 540 (veg.) 253 (rep.) 605 (veg.) 290 (rep.) 387 (veg.) 226 (rep.) 604 (veg.) 226 (rep.) Erigeron kachinensis Characteristics of prevalent species frequency (%)*** 66.7 d 44.0 b 38.7 a 54.7 c 53.3 c 72.0 e Ave. cover (%) 6.3 b 2.9 a 7.5 c 7.0 c 3.4 a 6.0 b No. plants/m^ 15.1 12.7 6.1 15.7 7.9 27.4 Aquilegia micmntha frequency (%)*** 73.3 d 56.0 c 57.3 c 26.7 b 8.0 a 56.0 c Ave. cover (%) 18.2 d 8.7 b 16.2 c 2.9 a 2.5 a 17.3 cd Calamagrostis scopiilorum fi-equency (%)*** 40.0 d 16.0 b 32.0 c 40.0 d 9.3 a 41.3 d Ave. cover (%) 4.6 b 2.7 a 16.8 d 15.0 d 2.2 a 8.1c Zigadenus vaginatus frequency (%)*** 53.3 c 33.3 b 30.7 b 61.3 d 68.0 d 2.7 a Ave. cover (%) 10.6 c 2.97 b 4.7 b 22.4 d 10.0 c 0.05 a *Species richness = number of species per 25-m transect. **Plant response indices = summed, relativized values for both vegetative (veg.) and reproductive (rep.) characteristics of the Kachina daisy (see text for details). ***Frequency = percentage of lOO-cm" sampling quadrats placed per 25-m transect that included this species. Table 1 are means based on results for the three summers. Total living cover at locations that support E. kachinensis ranged from 17.8 to 62.4% in the six alcoves studied intensively (Table 2). Most living cover was contributed by E. kachinensis, Aquilegia micrantha, Calamagrostis scopulorum, and Zigadenus vaginatus. Frequency and cover values for those species are reported in Table 2. Several species were found to occur regu- larly with the Kachina daisy in both Natural Bridges National Monument and lands man- aged by the Bureau of Land Management in that part of San Juan County, Utah (Table 3). Considering all known locations for the Kachina daisy in San Juan County, Utah, only four species were found to coexist with it at 75% or more of the occupied sites. Those species were Aquilegia micrantha, Calamagrostis scop- ulorum, Cirsium calcareum, and Zigadenus vaginatus. These species are thus good indica- tors of habitat suitable for the Kachina daisy. Three abiotic variables were found to have a significant effect on total living cover in the 200 Great Basin Naturalist [Volume 54 Table 3. Species associated with Erigeron kachinensis at all (100 + ) sites of occupancy on Bureau of Land Management Lands and the six study sites at Natural Bridges National Monument, San Juan Count>; Utah. Aquilegia inicrantha**** Calamagrosiis scopidonim **** Cirsium calcareum**** Zigaden us vaginatus **** Epipactu.s gigantus*** Helicmthella micrucephcda*** Hi'terothcca vdlosci*** Adiatuiii capdlufi var. veneris** Curex aurea** Castilleja linariifolia * * Cirsiuin n/dbergii** alia suhiiiida** Hijinenopapiisjilif alius var. cinereus** Miinulus ecistwoodii * * Senecio nudtilobatus** Suei~tia radiata** Gidium tnultiflorum var. coloradoense* Gilia congesta var. palmifrons* Hahenaria sparsiflora var. sparsiflora* Juncus arcticus* Leptodactylon pungens* Solidago sparsiflora* Aster chilensis— Comandra unbellata~ Penstemon waisonii~ Ranunculus cymhalaria — Trees and shrubs Pinus edulis*** Wuunnus hetulaefolia*** Cercocarpus intricatus** Juniperus osteospenna * Pinus ponderosa* Amelanchier utahensis— Clematis ligusticifolia — Mahonia freinontii— Populus freinontii — Salix exigua — ****>759c of Kachina daisy populations. ***.50-75% of Kachina daisy populations. **2.5-50% of Kachina daisy populations. *5-2.5% of Kachina daisy populations. ~<5% of Kachina daisy populations. six alcoves: soil water content, soil salinity, and available phosphorus. Results of regression analyses between these three important envi- ronmental variables and total living plant cover, integrated vegetative response of the Kachina daisy, and species richness are report- ed in Table 4. Salinity was negatively i"elated to total living cover and species richness in simple regres- sion analyses. Total available phosphorus was significantly related to total living cover in the alcoves. When soil salinity and soil water con- tent were considered together in multiple regression, they were significantly correlated with total living plant cover Partial correlation coefficients show that soil water content accounts for significantly more of the variation in total living plant cover in the model than does soil salinity [% total cover = 3.6(% water content) +.015 (EC)]. The combined effects of soil moisture, soil salinity, and available phos- phorus gave the strongest multiple correlation coefficient obtained with total living cover In this study increased soil moisture appeared to reduce the effect of salinity on plants in the alcoves and permitted good growth of most adapted species in soils of high salinit\'. Vegetative response of the daisy was found to be significantly influenced by percent soil moisture (Table 4), but salinity and phospho- rus were not significantly correlated with veg- etative response. When salinity and moisture are combined in multiple linear regression, a positive, but not significant, influence is shown for vegetative growth. By adding phosphorus to salinity and moisture in multiple regression analysis, the correlation coefficient became statistically significant. Not only do these three abiotic variables affect total cover of all species, but they also appear to be important for vegetative growth of the Kachina daisy. In contrast, reproductive response of the daisy was not found to be significantly influenced by any abiotic variable or combination of vari- ables. Kachina daisy reproduction by seed thus appears to be tolerant of the differences observed for important abiotic environmental variables in this study. The Sipapu Bridge site (site 7, which did not support a population of the Kachina daisy) was found to be environmentally similar to the six Kachina daisy study sites and sites that supported the daisy on lands administered by the BLM (Table 1). Results of a principal com- ponents analysis that considered all environ- mental variables simultaneously confirm this finding (Fig. 3). The Sipapu Bridge site falls among the alcoves that supported Kachina daisy in the diagram depicting the principal components analysis results. Discussion Characteristics of habitat suitable for the Kachina daisy can be predicted from the results of this study. Sites favorable for daisy growth have sandy loam soils, saline soils (13-14 mmhos/cm on average), and a perennial 1994] ERIGERON KACHINENSIS HaBITAT 201 Table 4. Influence of three important environmental variables (soil water content, soil salinity, and available phospho- rus) on total living plant cover, integrated vegetative response, and species richness. Vegetative daisy response is based upon sunmied relativized values for leaf length, crown diameter, number of rosettes per plant, number of leaves per plant, total daisy cover, daisy density, and daisy frequency. Reproductive daisy response represents number of filled seeds per flower head, considering total number of heads per plant and total number of flowers per head. Significance level is given ne.xt to each r-value. Independent variable(s) Dependent variables % living cover Daisy response Vegetat. Reprod. Species richness **Regression significant atp< .05 level of significance. *Regression significant at p < .10 level of significance. NS Regression not significant. r-value for simple regression Soil water content (%) .15 NS .85** .50 NS .HNS Salinity (EC ) -.79* .09 NS .42 NS -.75* Available P (mg/kg) .75* .05 NS .03 NS Multiple r-value for multiple regression .14 N Soil water content and salinity (EC) .95** .75 NS .43 NS .79 NS Soil water content, salinib,' (EC), and available P (mg/kg) .99** .98** .60 NS .86 NS source of water. Occupied soils are alkaline to strongly alkaline. Kachina daisies require perennially moist, cool soils along seeplines. However, these areas typically have large overhangs that sometimes provide too much shade for adequate growth. Therefore, daisies commonly grow on the outside perimeter of the alcove or sometimes directly below the alcove, as seen at site 2. Vegetative features of the plant at sites that typically receive little direct sunlight are typical of those of shade plants: large, thin leaves and elongated stems. Soil water content appears to buffer the effect of high salinity on vegetative giowth within the alcoves (Table 4). Above-average levels of soil phosphorus appear to enhance growth of the Kachina daisy (Table 4). Available soil phosphorus averaged about 9.5 mg/kg at sites occupied by Kachina daisy in this study. Habitat occupied by Kachina daisy also sup- ports several other characteristic species. The most common species associated with the daisy include Aquilegia tnicrantho, Calamagros- tis scopulorum, Cirsium calcareum, and Zigadenus vaginatiis (Table 3). Three abiotic variables appear to have the greatest influence on total living plant cover in alcoves and vegetative Kachina daisy growth, i.e., soil water content, soil salinity, and avail- able soil phosphorus (Table 4). These variables obviously act in concert to influence plant growth in the alcoves. For instance, soil water content has a strong, positive effect on vegeta- tive growth of the daisy, but the positive effect is amplified when both water and soil phos- phorus are considered in the analysis (Table 4). Although these variables are important to veg- etative growth in the alcoves, they appear to have minimal influence on reproduction in the Kachina daisy. Since we observed no significant variation in Kachina daisy populations in the six alcoves studied over a 3-year period, it is questionable whether the large year-to-year variations in relative cover of this species reported by Belnap (1988, 1989) are real. Belnap's results may be attributable to the taking of samples at different locations within the alcoves in differ- ent years. Our study included three of the same alcoves as those sampled by Belnap (our alcoves 1, 5, and 6), but we could detect no significant differences in either water flow or plant community composition among years of observation. The discrepancy between Belnap's results and ours probably relates to sampling design. Our samples were always taken at the 202 Great Basin Naturalist [Volume 54 ■9 2 -r 1.5 1 -- ■8 .10 0.5 ■2 -1.5 -1 -0.5 -0.5 -1 ■5 -1.5 0.5 1 1.5 2 PCI Fig. 3. Plot of the first two components of soil samples from the si.x Kachina daisy study sites in NBNM (1-6), the Sipapu Bridge transplant site (7), and Kachina daisy sites on ELM land (8-10, three site 8 alcoves in Fish Canyon, site 9 in White Canyon, site 10 in Arch Canyon). See Table 1 and text for explanation. same permanently marked points in each year of study, while Belnap sampled at random within the 2 years of analysis. The U.S. Park Service has encouraged attempts to transplant the Kachina daisy into what may be suitable habitat in NBNM, but until the results reported here became avail- able there were no criteria for evaluating the suitability of a site for the species. Since an al- cove at the Sipapu Bridge at NBNM appeared to fall within the range of environmental suit- ability for this daisy (Table 1, Fig. 3), we made experimental transplants of the species into that alcove. Presently, no known natural popu- lations of Kachina daisy occur near this site. The site is a west-facing alcove along the 1768-m (5800-ft) contour level. The habitat resembles other Kachina daisy sites in that salinity is high and many species occur on the site that are known to consistently grow with the daisy. Six daisies were transplanted in the alcove in summer 1991. Four of tlie six suwived to summer 1992. Approximately 200 plants were taken from a site to be destroyed by road construction in 1992 and transplanted at the Sipapu alcove; at least 100 of these plants sur- vived and some even flowered in 1993. Based on these favorable results, the Park Service is 1994] Erigeron kachinensis Habitat 203 now considering seeding the Kachina daisy in other suitable alcoves within NBNM. The validity of the suite of characters given above as descriptors of suitable Kachina daisy habitat is further supported by the fact that many new locations for the species were locat- ed on BLM lands in the region by searching for habitats and associated species closely sim- ilar to the combination of characteristics noted above. We note that others have collected an Erigeron from rock crevices on Elk Ridge, San Juan County, Utah, that is morphologically indistinguishable from populations in this report. However, the habitats associated with specimens from Elk Ridge rock faces and those associated with alcove daisies of the deep canyons of NBNM are extremely differ- ent in respect to elevation, soil moisture, solar radiation, soil salinity, and associated species. Genetic analysis will be required to determine how closely allied these two taxa are. The populations found on BLM land as a result of this study demonstrate that the Kachina daisy is more common than once believed. The species is well adapted to alcoves of the Cedar Mesa Formation, but it is apparently unable to occupy any other habitats in the deep canyon environments. It is imper- ative that managers protect alcoves with peren- nial seep lines in the Cedar Mesa Formation from disturbance. Trails should not be permit- ted to intrude into the alcoves, and camping within the alcoves should be forbidden. Acknowledgments This study was supported in part by grants from the Bureau of Land Management, the Utah Native Plant Society, and Brigham Young University. The U.S. Park Service provided housing and living expenses for the primary author during periods of field research. We extend special thanks to Dr. Lee G. Caldwell for his assistance with computer analyses and statistical designs used for evaluation of data on which this article is based. Literature Cited Allphin, L., and K. T. Harper. 1991. Survey of the Grand Giilch Primitive Area for Eriaerun kachinensis. Report submitted to Bureau of Land Management, U.S. Department of Interior, October 1991. Belnap, J. 1988. Kachina daisy survey. Report submitted to Resource Management Division, Canyonlands National Park, U.S. Department of Interior, 13 July 1988. . 1989. Kachina daisy survey. Report submitted to Resource Management Division, Canyonlands National Park, U.S. Department of Interior, summer 1989. Black, C. A., D. D. Evans, J. L. White, L. E. Ensminger, and E E. Clark, eds. 1965. Methods of soil analysis, part 1: physical and mineralogical properties, including statistics of measurement and sampling. American Society of Agronomy, Inc., Madison, Wisconsin. 770 pp. Brough, C. R., D. L. Jones, and D. J. Stevens. 1987. Utah's comprehensive weather almanac. Publishers Press, Salt Lake City, Utah. 401 pp. Computing Resource Center. 1992. STATA reference manual: release 3. 5th edition. Santa Monica, California. Cottam, C, and J. T. Curtis. 1956. The use of distance measures in phytosociological sampling. Ecology 37: 451-460. Fagan, D. 1984. Report on status of Kachina daisies in Natural Bridges National Monument. Submitted to Resource Management Division, Canyonlands National Park, U.S. Department of Interior, 19 May 1984. HiNTZE, L. E 1975. Geologic history of Utah. BYU Geology Studies, Vol. 3, Pt. 3. Brigham Young University Press, Provo, Utah. 162 pp. PlELOU, E. C. 1984. The interpretation of ecological data. John Wiley and Sons, New York. 263 pp. Richards, L. A., ed. 1954. Diagnosis and improvement of saline and alkali soils. USDA Agricultural Handbook No. 60. 160 pp. SAS Institute, Inc. 1993. SAS user's guide: statistics. SAS Institute, Inc., Gary, North Carolina. Snedecor, G. W, and W G. Cochran. 1967. Statistical methods. 6th edition. Iowa State University Press, Ames. 593 pp. U.S. Department of Interior, Eish and Wildlife Service. 1975. Review oi Erigeron kachinensis as an endangered species. Eederal Register 40: 27880. . 1976. Proposal that Erigeron kachinensis be listed as endangered. Eederal Register 41: 24524-24572. . 1988. Endangered Species Act of 1973 as amend- ed through the 100th Congress. U.S. Eish and Wildlife Sei-vice, Washington, D.C. 45 pp. Welsh, S. L., and G. Moore. 1968. Plants of Natural Bridges National Monument. Proceedings of the Utah Academy of Sciences, Arts, and Letters 45: 220-248. Welsh, S. L., N. D. Atwood, L. C. Higgins, and S. Goodrich. 1987. A Utali flora. Great Basin Naturalist Memoirs No. 9. Brigham Young University Press, Provo, Utah. 894 pp. Received 28 January 1993 Accepted 1 March 1994 Great Basin Naturalist 54(3), © 1994, pp. 204-21 1 COORDINATION OF BRANCH ORIENTATION AND PHOTOSYNTHETIC PHYSIOLOGY IN THE JOSHUA TREE {YUCCA BREVIFOLIA) Kaylie E. Rasinuson', Jay E. Anderson', and Nancy Huntly' Abstiwct. — Despite the profusion of liglit in deserts, moiphological adaptations to increase light interception are common among desert plants. We studied branch orientation and related physiological parameters in the Mojave Desert Joshua tree. Yucca Irrevifolia (Agavaceae). Azimuth and inchnation were measured on all leaf rosettes of 44 Y. brevifolia trees. Interception of solar radiation was modeled for leaves in hypothetical rosettes facing due south and due north in December, March, and June. Carbon isotope discrimination, nitrogen content, and conductance to water vapor were measured in leaves from north- and south-facing rosettes. Rosette azimuths were nonrandom; rosettes predominandy faced southeast. North-facing rosettes were more steeply inclined than those facing south. The preponderance of south- facing rosettes reduces self-shading and increases interception of solar radiation during the winter-spring growth period. Stomatal conductance was higher for leaves in south-facing than in north-facing rosettes. Nevertheless, discrimination against '•^C was less in leaves of south-facing rosettes, indicating that average intercellular CO2 concentration was also lower. South-facing whorls had higher leaf nitrogen content. Greater allocation of nitrogen to leaves in south-facing whorls probably results in those leaves having a greater photosynthetic capacity than their north-facing counterparts. Orientation of rosettes to increase interception of sunlight during the period most favorable for photosynthesis, coupled with allocation of nutrients to maintain a higher photosynthetic capacity' in those rosettes, should significantly increase whole-plant carbon gain in Y. brevifolia. Key words: Yucca brevifolia, Joshua tree, carbon isotope discrimination, photosynthetic capacity, branch orientation, rosette azimutli, morphological adaptations. Morphological adaptations to the light environment are common among desert plants (Ehleringer and Werk 1986). The angle and inclination of leaves (Ehleringer 1988, Neufeld et al. 1988) or cladodes (Nobel 1980, 1981, 1982) may increase interception of solar radia- tion when air temperatures and evaporative gradients are moderate (e.g., early in the day or during winter months) and reduce incident solar radiation during hotter parts of the day or year. Neufeld et al. (1988) reported that foliage clusters in creosote bush {Larrea tri- dentata), a long-lived evergreen shrub of the Mojave and Chihuahuan deserts, are inclined from 33° to 71° and oriented predominantly toward the southeast. They suggested that such architecture would tend to minimize self- shading and maximize carbon gain during periods most favorable for photosynthesis, which could result in improved water-use effi- ciency. We wondered if similar morphological adaptations might be found in the Joshua tree {Yucca brevifolia), a long-lived arborescent monocot with evergreen leaves. Yucca brevifolia is restricted to the Mojave Desert, where it often occurs with L. tridentata. Its tough, fibrous leaves grow in symmetrical whorls fonning cylindrical rosettes at the end of branches (Fig. lA). The axis of newly expanding leaves at the top of a rosette is par- allel with that of the rosette. As leaves mature and become photosynthetically active, they reflex away from the branch axis so that the adaxial surfaces of the youngest fully expand- ed leaves are at about 55° from the rosette axis. This angle gradually increases along the rosette axis so that the oldest photosyntheti- cally active leaves are nearly peqDendicular to the rosette axis (J. Anderson unpublished data; cf Smith et al. 1983). Rosettes vary from about 0.2 to >1.5 m in length and typically contain 200->1000 leaves. Young trees possess a single vertical rosette of leaves. Older trees have multiple branches that result from dichotomous branching at the apices of rosettes (Fig. IB). Old trees can have over 100 branches and grow to >5 m in height (J. Anderson and N. Huntly unpublished data). 'Department of Biological Sciences and Center for Ecological Research and Education, Ul;dio State University, Pocatello, Idalio 83209. 204 1994] Branch Orientation in Yucca brevifolia 205 Fig. 1. (A) Cylindrical rosette of Yucca brevifolia leaves consisting of a sequence of whorls in which adjacent whorls are non-overlapping; (B) mature individual of Y. brevifolia widi multiple branches. Yucca brevifolia is a C3 species with modest photosynthetic rates (Smith et al. 1983). Photosynthesis is hght saturated at a relatively low photosynthetic photon flux density (PFD) of 400-600 /xmol m~2 s"l; the nonoverlapping leaf arrangement results in a relatively even distribution of light throughout the rosette. Smith et al. (1983) found that stomatal con- ductances are highest during the winter- spring wet season and predicted that 80% of the annual photosynthetic productivity would occur from January through May. During the dry season, conductances are reduced to a very modest peak early in the morning. Thus, like L. tridentata and most aridland plants, Joshua trees live in an environment where opportunities for carbon gain are constrained, both seasonally and diurnally. We tested the hypothesis that leaf rosettes are distributed nonrandomly within crowns of Y. brevifolia. After documenting a strong ten- dency for southeasterly orientation of rosettes, we compared leaf nitrogen content and carbon isotope discrimination (A) of rosettes on south vs. north sides of trees. Finallv, we used porometry to explore the significance of differ- ences in nitrogen allocation and A. We show that nonrandom orientation of photosynthetic leaf rosettes in Y. brevifolia is closely integrat- ed with physiology. Study Area The study was conducted at Lytle Ranch Preserve, 48 km west of St. George, Utah (37°9'N, 114° I'W, elevation 850 m), during March of 1989 and 1991. Lyde Ranch is in the northeastern Mojave Desert near the northeiTi distributional limit of Y. brevifolia. Extensive Yucca woodlands occur on benches adjacent to Beaver Dam Wash. Other common species on the benches are Coleogijne ramosissima. Ambrosia dumosa, Larrea tridentata, Tham- nosma montana, and Krameria graiji. Average annual temperature and precipitation at St. George are 16.5 °C and 209 mm, respectively. Methods Orientation of rosettes was assessed for 44 trees chosen systematicallv at 15-m intervals 206 Great Basin Naturalist [Volume 54 along permanently marked transects on two benches. Each tree had fewer than 60 brandies, and we measured azimuth and inclination of all branches on every tree. Azimuth was mea- sured clockwise from true north to the nearest degree with a compass by sighting along the longitudinal axis of each rosette. Inclination of the rosette a.xis from horizontal was deter- mined using an angle gauge with a built-in level. Carbon isotope composition and total leaf nitrogen were measured in leaves collected in March 1989 from eight trees on the perma- nent transects. On each tree one fully expand- ed young leaf was taken from a rosette point- ing due south, and a paired leaf sample was taken from a rosette pointing due north. Samples were dried and ground and then sub- mitted to the Stable Isotope Research Facility for Environmental Research at the University of Utah for determination of carbon isotope ratios. Total leaf nitrogen was determined on subsamples of the paired leaf samples with a LECO C-H-N analyzer at the Holm Research Center, University of Idaho. Carbon isotope discrimination (A) was cal- culated from the carbon isotope ratios accord- ing to Farquhar and Richards (1984), assuming that the isotopic composition of the air was -7.8^/oo. Carbon isotope discrimination is related linearly to the intercellular concentra- tion of CO2 (cj): A = a -h (b - a)(cj/cj (1) where a is the discrimination against ^'^COq relative to ^^C02 associated with diffusion in air (4.4°/oo), b is discrimination against the heavy isotope by ribulose-l,5-bisphosphate carboxylase/oxygenase (Rubisco) (27°/oo), and c.^ is the concentration of CO2 in the atmos- phere (about 350 ^t,L L~^). Because b, a, and c,^ are usually constant, variation in A reflects variation in Cj/c,j, which results from variation in stomatal conductance and in demand for CO2 by the photosynthetic apparatus (Farquhar et al. 1982). Equation 1 was used to estimate Cj for leaves in north- and south-facing rosettes. We measured leaf conductance to water vapor (g^) on 21 and 22 March 1991 with a LI-COR 1600 steady-state porometer. On both days a high cloud cover was present from dawn until dusk, which blocked direct sun- light and created uniform light conditions on all sides of the trees. Nine trees in the vicinity of the permanent transects, each having at least two north-facing and two south-facing rosettes, were chosen for sampling. A fully expanded leaf near the apex of the rosette was sampled in each of two rosettes on the north and south sides of each tree. Means of the two measurements were used for statistical analy- ses. A preliminary sample indicated that con- ductances of abaxial and adaxial leaf surfaces were similar, as reported by Smith et al. (1983); so, for convenience in holding the porometer in place, we sampled only adaxial surfaces. Faired measurements were made from north- and south-facing rosettes on the nine trees between 0900 and 1200 h (MDT). Photosynthetic photon flux density (FED) was measured with a LI-COR 170 quantum sensor (LI-COR, Inc., Lincoln, Nebraska, USA). FED varied from 200 to 800 Atmol m-2 s"! during the measurement periods, but for the paired sample on each tree FED was essen- tially constant and equal on both sides of the tree. Because of the complexity of the architec- ture of Y. brevifolia (Fig. IB), it was beyond the scope of this study to model light intercep- tion of whole trees, taking into consideration shading by other branches and self- shading within rosettes. Instead, we predicted the in- cident FED for leaves in rosettes on the south and north sides of trees for 22 December, 21 March, and 21 June. Solar azimuths and incli- nations for those dates at the latitude of Lytle Ranch were calculated according to Ehleringer (1989b). We calculated the cosine of the angle of incidence (cos i), the fraction of the direct beam of solar radiation that is intercepted by a leaf, for leaves in hypothetical rosettes facing either due south or due north and having an inclination of 60°. Excluding trees with only one branch, this was the mean inclination of the population. Cos i was calculated for four leaves per hypothetical rosette: leaves on the top and bottom, with their axes peipendicular to the ground, and leaves on both sides, with their axes parallel to the ground. Because of the symmetry of whorls of leaves in rosettes (Fig. lA), estimates of incident radiation for those four leaves should be proportional to that for an entire rosette. We assumed that leaves were planar and that each leaf was inserted at an angle of 55° from the rosette axis, the mean value for the youngest fully 1994] Branch Orientation in Yucca brevifolia 207 expanded leaves on several trees. Cos i was calculated for both surfaces of each leaf (i.e., when cos i was negative, indicating that light would strike the abaxial side of a leaf, the abso- lute value was used). Thus, estimates of inci- dent PFD include both abaxial and adaxial surfaces of the four leaves. Direct solar beam PFD values were predicted from a polynomial regression equation based on actual PFD measurements made throughout a clear day in mid-March at Lytic Ranch. Direct-beam PFD estimates were multiplied by cos i of each hypothetical leaf for a given date/time to esti- mate incident PFD. Those values were aver- aged for the four leaves as an index of incident PFD for the hypothetical rosettes. Statistics were calculated according to Zar (1984). Mean azimuths and inclinations were calculated trigonometrically for individual trees, and those means were used to calculate a grand mean for all trees. Uniformity of leaf whorl azimuths within each tree was tested with the Watson U2 statistic. A chi-square goodness-of-fit statistic was used to test for random distribution of whorl azimuths pooled for all trees sampled using twelve 30° classes. Association between azimuth and inclination was tested using a two-way contingency table, with inclinations and azimuths grouped in 30° classes. Paired t tests were used to determine whether A, total leaf nitrogen, or g^^, differed between the south and north sides of the trees. Results Rosette azimuths within trees having more than four branches (the minimum required for the Watson U2 test) were not distributed ran- domly (P < .001 in all 23 cases). Azimuths for individual branches (pooled for all trees) fell predominantly between 90° and 270° (Fig. 2A); this distribution was also nonrandom (X^ = 78.13, d.f = 11, P < .001). All mean rosette azimuths for trees with two or more branches fell between 90° and 280°, with one exception that had a mean azimuth of 10° (Fig. 2B). Mean azimuths of individual trees were tightly clus- tered around the grand mean of the popula- tion, 163° (angular deviation s = 42, n = 44; Fig. 2B). The mean inclination of rosettes on trees having two or more branches fell between 42° and 82° from the horizontal. Trees with two to 15 45 75 105 135 165 195 225 255 285 315 345 Rosette Azimuth (degrees) 22.5 67.5 112.5 157.5 202.5 247.5 292.5 337.5 Mean Rosette Azimuth (degrees) Fig. 2. Frequency distributions of leaf rosette azimuths for 44 Yucca brevifolia trees at Lytle Ranch Preserve, Utah: (A) distribution for all branches; and (B) distribution for the means of individual trees. five branches had more steeply inclined rosettes than did trees with six or more branches. Inclination was associated with azimuth (X2 = 39.45, d.f. = 22, F < .025); rosettes having northerly azimuths (170° -90°) were more steeply inclined than those with southerly azimuths (Fig. 3). Simulation of light interception shows that leaves in rosettes facing south would intercept substantially more direct sunlight than those in rosettes facing north at all times of year, but the difference is much larger in winter and spring than in summer (Fig. 4). When sun 208 Great Basin Naturalist [Volume 54 100 80 - 40 - >N 20 - U A P P - P - ^ P - ^ - ^ 0-17 18-36 37-54 55-72 73-90 Inclination (degrees) Fig. 3. Frequency distributions for rosette inclinations for 33 Yucca brevifolia trees having more than two branches at Lytle Ranch Preserve, Utah: (A) chstribution for rosettes in the southern 180° arc (azimuths from 90° to 270°) from the tree's trunk; (B) distribution for those in the northern 180° semicircle. angles are low, little direct sunlight is inter- cepted by abaxial surfaces of leaves in rosettes facing southeast and adaxial surfaces of leaves in rosettes facing north; those surfaces con- tribute more to total interception with increas- ing sun angles. Morning and afternoon peaks for leaves in the north-facing rosette in March are a consequence of the insertion angle (55°) for the two horizontally opposed leaves. For the analysis shown in Figure 4, it is assumed that the foin- modeled leaves would be exposed to direct-beam solar radiation throughout the day. Clearly, that assumption would not hold for all leaves at all times of day. Because of the low density of Joshua tree stands and low stature of other plants in the en 1500 1000 O E Q Li_ D_ — I ' 1 ' 1 ' r June 21 / /o°°°°°o\ ^ J , I 1 L I I 1 , I ^ 600 800 1000 1200 1400 1600 1800 Solar Time (h) Fig. 4. Predicted diurnal patterns of direct-beam photo- synthetic photon flu.x density (PFD) incident on leaves in hypothetical rosettes of Yucca brevifolia trees located at Lytle Ranch Preserve, Utah. Incident PFD was estimated for adaxial and abaxial surfaces of leaves in four positions in the rosette (see Methods) for 22 December, 21 March, and 21 June. Closed circles represent rosettes facing due south; open circles represent those facing due north. community, shading by other individuals occurs rarely. However, self-shading occurs among leaves within rosettes and among branches within trees. Both would be mini- mized for branches having southerly aspects (Geller and Nobel 1986). Smith et al. (1983) asserted that the nonoverlapping arrangement of leaves of Y. brevifolia resulted in effective penetration of light into a rosette from the top. Leaves in rosettes having northerly azimuths would receive more shade from the rosette in which they occur, from other branches on the same tree, and from the main trunk of the tree, based on the patterns of shadows cast by 1994] Branch Orientation in Yucca brevifolia 209 Table 1. Carbon isotope discrimination (A), leaf nitro- gen content (N), and leaf conductance to water vapor (g^) for leaves in north-facing and south-facing rosettes of Yucca brevifolia at Lytle Ranch Preserve, Utah. Fig. 5. Direction and length of shadows cast from early morning until late afternoon by an object 1 unit in height (1 unit = distance between concentric dashed circles) at the latitude of Lytle Ranch Preserve, Utah, for (1) 22 December, (2) 21 March, and (3) 21 June. an object at Lytle Ranch in December, March, and June (Fig. 5). Carbon isotope discrimination was lower in leaves of rosettes that faced south than in leaves of rosettes that faced north (Table 1). Estimates of Cj based on these A values were 141 fxh L~^ for leaves in rosettes on the south side of trees and 156 ^tL L"^ for leaves in rosettes on the north side. The corresponding ratios of Cj/c^, assuming c^ was 350 fih L"^, would be 0.40 and 0.45, respectively. Total nitrogen content of leaves in rosettes facing south was higher than that in rosettes fac- ing north (Table 1). For seven of tlie eight paired samples, the estimate of total leaf nitrogen was higher for the leaf from the south-facing rosette. Leaves in rosettes on the south side of trees had higher conductances to water vapor than did leaves on the north side (Table 1). Mean g^ was higher on the south side in each of the nine trees tested. Discussion The distribution of rosette azimuths of Y. brevifolia is clearly not random; rosettes point predominantly in southerly directions. Further- more, branches on the north side of trees tend to be more steeply inclined than those on the Parameter North-facing South-facing rosettes rosettes A (o/oo) N(%) g^^. (mol ni"- s"l) 14.7 0.91 0.12.5 13.7 1.02 0.155 8 .0005 8 .015 9 .001 Probabilitv that the mean of the paired differences equals zero based on results of paired-sample ( test. south, which would tend to elevate the rosettes and reduce self-shading. Rosettes on the south side intercept substantially more direct solar radiation throughout the year, and the difference is especially pronounced in winter and early spring when sun angles are relatively low (Fig. 4). It is at this time that the bulk of annual carbon gain occurs (Smith et al. 1983). Shading of leaves in north-facing rosettes also would be greatest during the winter and early spring. Self-shading would magnify differences in incident direct-beam radiation between leaves in south- vs. north- facing rosettes shown in Figure 4. Thus, a sec- ond advantage of positioning rosettes on the south side of trees is to minimize self-shading. The nonrandom branch orientation in Joshua trees appears to be closely coordinated with parameters related to photosynthetic capacity. Lower A in leaves on the south vs. north side of trees indicates that leaves on the south operate at a lower average Cj. As shown in equation 1, A is a time-integrated measure of Cj reflecting the importance of both stomatal limitation to diffusion of CO2 and capacity of the mesophyll to fix CO2 (Farquhar et al. 1982). Numerous studies have shown that instantaneous gas exchange measurements are related to A as predicted by the theory (Hubick et al. 1986, Ehleringer et al. 1992). We found that g^^ was higher in leaves of south-facing whorls, indicating that lower Cj was associated with higher g^^.. Because we measured g^ under conditions when both sides of trees were equally illuminated, we assume that observed differences in g^^ reflect intrinsic differences related to photosynthetic capacity. These results imply that photosyn- thetic capacity was higher in leaves of south- facing rosettes. Photosynthetic capacity' and Rubisco activi- ty often are positively correlated with leaf nitrogen content (Wong et al. 1985, Field and 210 Great Basin Naturalist [Volume 54 Mooney 1986, Evans 1989). Field (1983) pre- dicted that net photosynthesis would he maxi- mized if nitrogen were allocated preferentially to leaves that receive more light. This is pre- cisely what we observed; leaves in south-fac- ing rosettes had higher nitrogen concentra- tions than those from rosettes on the north side of trees. Relatively low leaf nitrogen con- tents of Y. brevifolia were in a range where any increase in nitrogen would be expected to increase photosynthetic capacity. One might expect lower A and lower Cj in leaves of south-facing rosettes to be a conse- quence of lower stomatal conductance. However, observation of the pattern found here is not without precedence. Korner et al. (1988) reported that A decreased in plants with increasing altitude while carboxylation efficiency' and stomatal conductance increased. Leaf nitrogen content also increased with alti- tude, which contributed to an increased photo- synthetic capacity (Korner et al. 1988). Lower A is associated with higher photosynthetic capacit)' in peanut cultivars (Hubick et al. 1986, Wright et al. 1993) and sunflower (G. Farquhar personal communication). Other factors could contribute to the observed difference in A between leaves on the north vs. south side of trees. Maximum stomatal conductance may occur at light levels somewhat below light saturation for photosyn- thesis, as observed in other species (e.g., Anderson 1982). Thus, leaves on the north side might receive sufficient diffuse radiation to open stomata but not saturate the photosyn- thetic apparatus, which could result in higher average Cj than would occur if photosynthetic tissues were light saturated. Also, shading typ- ically results in near instantaneous reductions in photosynthesis, whereas conductance changes more slowly (Anderson 1982, Knapp and Smith 1987). Therefore, Cj of leaves on the north side might be higher in comparison to those on the south because those on the north experience shading more frequently, particu- larly during the winter-spring growing season when sun angles are relatively low. Although differential light levels and inter- mittent shading may contribute directly to observed differences in Cj, the coincidence of lower Cj, higher g^, and higher nitrogen con- tent in leaves of south-facing rosettes provides strong evidence that the lower Cj is primarily a consequence of higher photosynthetic capacity. We conclude that differential allocation of nitrogen to leaxes on the south side of Joshua trees results in substantially higher photosyn- thetic capacities in those leaves. This, coupled with orientation of rosettes to increase inter- ception of sunlight during the period most favorable for photosynthesis, would enhance productivity of whole trees for a given level of nitrogen availability The A values for Y. brevifolia are among the lowest reported for C3 plants (cf Ehleringer 1989a, Korner et al. 1991, Ehleringer et al. 1992). Ehleringer (1989a) reported carbon iso- tope ratios for desert C3 plants corresponding to A values ranging from 13^/oo to 23^/oo, but values for Y brevifolia are generally much lower than those reported by Ehleringer et al. (1992) and Schuster et al. (1992) for desert shrubs such as Ambrosia diimosa, Larrea tri- dentata, and Coleogyne ramosissima that often occur with Joshua trees. A is negatively corre- lated with water-use efficiency (Farquhar et al. 1989). Low A values and corresponding esti- mates of Cj indicate that Y. brevifolia leaves have very low stomatal conductances relative to their photosynthetic capacities. This would translate to high water-use efficiency com- pared to co-occurring C3 plants, assuming they were subjected to comparable leaf-air vapor pressure deficits. Acknowledgments This research was supported by the Pat Kolbet Undergraduate Research Fund and the Department of Biological Sciences at Idaho State University. We thank the personnel at Lytle Ranch Presei-ve and the Monte L. Bean Life Sciences Museum at Brigham Young University for their cooperation. We also thank Marjorie Daly, Kelly Green, Mike Haslett, and Jeff Hemy for help with data col- lection and Drs. Graham Farquhar and Christopher Field for helpful comments on an earlier version of the manuscript. Literature Cited Anderson, J. E. 1982. Factors controlling transpiration and photosynthesis in Tamarix chinensis Lonr. Ecology 63: 48-56. Ehleringer, J. R. 1988. Changes in leaf characteristics of species along elevational gradients in the Wasatch Front, Utah. American Journal of Botany 75: 680-689. 1994] Branch Orientation in Yucca brevifolia 211 . 1989a. Carbon isotope ratios and physiological processes in aridland plants. Pages 41-54 in P W. Rundel, J. R. Ehleringer, and K. A. Nagy, eds., Stable isotopes in ecological research. Springer- Verlag, New York. . 1989b. Temperature and energy budgets. Pages 117-135 in R. W. Pearcy, J. R. Ehleringer, H. A. Mooney, and P W. Rundel, eds.. Plant physiological ecology: field methods and instrumentation. Chapman and Hall, New York. Ehleringer, J. R., a.nd K. S. Werk. 1986. Modifications of solar-radiation absorption patterns and implica- tions for carbon gain at the leaf level. Pages 57-82 in T. J. Givnish, ed.. On the economy of plant fonn and function. Cambridge University Press, New York. Ehleringer, J. R., S. L. Phillips, and J. P Comstock. 1992. Seasonal variation in the carbon isotopic com- position of desert plants. Functional Ecology 6: 396-404. Evans, J. R. 1989. Photosynthesis and nitrogen relation- ships in leaves of C3 plants. Oecologia 78: 9-19. Farquhar, G. D., and R. A. Richards. 1984. Isotopic composition of plant carbon correlates with water- use efficiency of wheat genotypes. Australian Journal of Plant Physiology 11: 539-552. Farquhar, G. D., M. H. O'Leary, and J. A. Berry. 1982. On the relationship between carbon isotope discrim- ination and the intercellular carbon dioxide concen- tration in leaves. Australian Journal of Plant Physiology 9; 121-137. Farquhar, G. D., K. T. Hurick, A. G. Condon, and R. A. Richards. 1989. Carbon isotope fractionation and plant water-use efficiency. Pages 21-40 in P. W. Rundel, J. R. Ehleringer, and K. A. Nagy, eds., Stable isotopes in ecological research. Springer- Verlag, New York. Field, C. 1983. Allocating leaf nitrogen for the maximiza- tion of carbon gain: leaf age as the control on the allocation program. Oecologia 56: 340-347. Field, C, and H. A. Mooney. 1986. The photosynthesis- nitrogen relationship in wild plants. Pages 25-55 in T. J. Givnish, ed.. On the economy of plant form and function. Cambridge University Press, New York. Geller, G. N., and P S. Nobel. 1986. Branching patterns of columnar cacti: influences on PAR interception and CO2 uptake. American Journal of Botany 73: 1193-1200. Hurick, K. T, G. D. Farquhar, and R. Shorter. 1986. Correlation between water-use efficiency and car- bon isotope discrimination in diverse peanut {Arachis) germplasm. Australian Journal of Plant Physiology 13: 803-816. Knapp, A. K., and W. K. S.mith. 1987. Stomatal and photo- synthetic responses during sun/shade transitions in subalpine plants: influence on water-use-efficiency. Oecologia 74: 62-67. Korner, C, G. D. Farquhar, and Z. Roksandic. 1988. A global survey of carbon isotope discrimination in plants from high altitude. Oecologia 74: 623-632. Korner, C, G. D. Farquhar, and S.-C. Wong. 1991. Carbon isotope discrimination by plants follows lati- tudinal and altitudinal trends. Oecologia 88: 30-40. Neufeld, H. S., F C. Meinzer, C. S. Wisdom, M. R. Sharifi, p. W. Rundel, M. S. Neufeld, Y. Goldring, and G. L. Cunningham. 1988. Canopy architecture of Larrea tridentata (DC.) Gov., a desert shrub: foliage orientation and direct beam radiation interception. Oecologia 75: 54-60. Nobel, P S. 1980. Interception of photosynthetically active radiation by cacti of different morphology. Oecologia 45: 160-166. . 1981. Influences of photosynthetically active radi- ation on cladode orientation, stem tilting, and height of cacti. Ecology 62: 982-990. 1982. Orientations of terminal cladodes of platy- opuntias. Botanical Gazette 143; 219-224. Schuster, W S. E, D. R. Sandquist, S. L. Phillips, and J. R. Ehleringer. 1992. Comparisons of carbon iso- tope discrimination in populations of aridland plant species differing in lifespan. Oecologia 91: 332-337. Smith, S. D., T. L. Hartsock, and P S. Nobel. 1983. Ecophysiology of Yucca brevifolia, an arborescent monocot of the Mojave Desert. Oecologia 60: 10-17. Wong, S.-C, I. R. Cowan, and G. D. Farquhar. 1985. Leaf conductance in relation to rate of CO2 assimila- tion. I. Influence of nitrogen nutrition, phosphorus nutrition, photon flux density, and ambient partial pressure of CO.9 during ontogeny. Plant Phvsiology 78:821-825. Wright, G. C, K. T Hubick, G. D. Farquhar, and R. C. N. Rao. 1993. Genetic and environmental variation in transpiration efficiency and its correlation with car- bon isotope discrimination and specific leaf area in peanut. Pages 247-267 in J. R. Ehleringer, A. E. Hall, and G. D. Farquhar, eds.. Stable isotopes and plant carbon-water relations. Academic Press, Inc., San Diego, California. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall, Inc., Englewood Cliffs, New Jersey. Received 22 May 1994 Accepted 8 July 1994 Great Basin Naturalist 54(3), © 1994, pp. 212-227 VARIANCE AND REPLENISHMENT OF NECTAR IN WILD AND GREENHOUSE POPULATIONS OF MIMULUS Robert K. Vickery, Jr.^ and Steven D. Siitherland- Abstr.'VCT — We compared nectar production in wild populations and greenhouse-grown populations of the monkey flower species of section Erythranthc of the genus Miimiln,s. Nectar was sampled fiom over 1000 flowers. For each flower the volume of nectar was measured with a calibrated micropipette and the percentage of sugar with a hand refractometer. Percentage of sugar varied little from flower to flower in both field and greenhouse studies, but volume varied markedly from flower to flower in field studies and even more in greenhouse studies. This high variance in nectar volume appears to be intrinsic. The amount of nectar in greenhouse populations tended to increase with time in the absence of pollinators. The amount of nectar in field populations tended to remain the same with time despite withdrawals by pollinators. Thus, nectar appears to be replenished Ijoth with time and as a response to pollinator withdrawals. The latter conclusion was corroborated by sampling nectar at 2-h intervals all day and comparing the total volume produced by a flower to the volume of nectar produced in control flowers sampled only at the end of the day Key words: Mimulus, nectar, nectar volume, nectar variaiice, nectar replenishment, pollinator reward. Nectar is the primary reward for the princi- pal pollinators, hummingbirds and bumble- bees, of flowers such as the monkey flowers of section Erythranthe of the genus Mimulus (Scrophulariaceae), based on our own observa- tions and as suggested by Free (1970), Faegri and Van Der Fiji (1979), and Baker (1983). Pollen is a secondary reward, particularly for bumblebees, but the analysis of its effect on attracting pollinators was beyond the scope of this study. Here we concentrate on the nectar rewards of the species of the section. Five of the six species of section Erythranthe — Mimulus cardinalis, M. east- woodiae, M. nelsonii, M. rupestris, and M. ver- henaceus — have red, tubular, hummingbird- pollinated flowers, whereas the two races of the sixth species, M. lewisii, have light laven- der pink or deep magenta pink, open, bumble- bee-pollinated flowers (Hiesey et al. 1971, Vickery 1978, Vickeiy and Wullstein 1987). All the species are self-compatible but usually do not self-pollinate. So, pollinators are required for normal seed set (Vickeiy 1990). To characterize the nectar rewards of this group, we examined (1) the standing crop of nectar present in flowers of wild and green- house-grown populations of each species and (2) the ability of flowers to replenish their nec- tar levels. Methods For field studies, flowers of a population of each species and race (Table 1) were analyzed in the wild for their nectar characteristics. Nectar volume was measured with a calibrat- ed micropipette. Percentage of sugar in the nectar was determined with a hand refrac- tometer. Measurements were made on differ- ent fresh flowers at 2-h intervals all day from dawn to dusk (Appendix 1). Flowers were sampled destructively inasmuch as we found that merely probing the flower with a micropipette failed to remove the occasionally sizeable remainder of nectar (Table 2). Greenhouse studies were undertaken to avoid the variable of unequal numbers of pol- linator visits and variations of climate observed in studies of wild populations. Different fresh flowers of greenhouse-grown populations of each species and race (Table 1) were sampled at 2-h intervals from bud stage (bumblebees often probe and rob buds) until flowers fell or shriveled (Appendix 2). Again, flowers were sampled destructively to be comparable to field studies as well as to obtain as complete measurements of the volume and as accurate measurements of the percentage of sugar of the nectar as possible. 'Department of Biolog>', University of U tali. Salt Lake Cit>, Utali 84112 2The Nature Conservancy, 1504 \V. Front Ave., Columbus, Ohio 4.3212. 212 1994] Variance and Replenishment of Mimulus Nectar 213 Table 1. Localities of populations used in the study by species, population number, habitat, locality, elevation, and collector. Vouchers are in the GaiTet Herbarium (UT) of the University of Utah, Salt Lake City. Mimulus cardinaUs Douglas 6651 Growing by stream. Bear Wallow picnic area, Santa Catalina Mtns., Pima Co., Arizona, elev. 2130 ni. Collected by Charles T Mason, Jr., 2143. 7113 Los Trancos Creek, San Mateo Co., California, elev. 40 m. Collected by Malcom Nobs Februar\' 1958. 7120 South face of San Antonio Peak, Los Angeles Co., California, elev. 2250 m. Collected by Verne Grant 9760. 13106 Growing by spring, Aguage Vargas, Cedros Island, Baja California del Norte, Mexico, elev. 600 m. Collected by Steven Sutherland 25 October 1981. 13486 Growing along road to the Pacific Ocean, ca 2 miles west of turnoff from El Camino Real, Santo Tomas, Baja California del Norte, Me.xico, elev. ca 500 m. Collected by Steven Sutherland 20 February 1984. Mimulus easttvoodiae Rydberg 6079 Growing in seeps under overhanging sandstone cliffs. Bluff San Juan Co., Utah, elev. 1415 m. Collected by R. K. Vickerv', Jr, 800. 13514 Growing in seeps in sandstone shelter caves near Anasazi ruins, south side of river. Bluff San Juan Co., Utah, elev. 1375 m. Collected by Steven Sutherland May 1985. Mimulus lewisii Pursh 5875 Growing along small stream where Patsy Morley ski trail crosses Albion Basin Road, Alta, Salt Lake Co., Utah, elev. 2680 m. Collected by R. K. Vickery, Jr., 2723. 6103 Growing along effluent stream from Ice Lake, Soda Springs, Placer Co., California, elev. 2000 m. Collected by R. K. Vickerv', Jr, 1361. 13515 Smoky Jack campground, Yosemite National Park, California, elev. ca 1800 m. Collected by Steven Sutherland 24 March 1986. Mimulus nelsonii Grant 6271 Growing in and by a small brook in the pine forest on Devil's Backbone, Sierra Madre Occidental, Durango, Mexico, elev. 2555 m. Collected by R. K. Vickeiy, Jr, 2614. Mimulus rupestris Greene 9102 Growing on moist, conglomerate cliff ca 100 m below the Tepozteco Temple, Tepoztlan, Morelos, Mexico, elev. 2300 m. Collected by R. K. Vickery, Jr, 2738. Mimtdus verbenaceus 5924 Growing by Bright Angel Creek near Phantom Ranch, Grand Canyon, Arizona, elev. 612 m. Collected by Earl Jackson November 1954. 13518B Growing near stream, Oak Creek Canyon, Coccino Co., Arizona, elev. ca 1800 m. Collected by Steven Sutherland April 1985. 13547 Growing by spring emerging from a talus slope at base of red sandstone canyon wall, Vassey's Paradise, below Lee's Ferry, Grand Canyon, Arizona, elev. 1015 m. Collected by Steven Sutherland 20 April 1986. Wild and greenhouse nectar studies sug- was repeatedly sampled and the other used as gested to us that nectar replacement might the control wherever possible. Occasionally, occur in response to removal of nectar by pol- fluctuating asymmetry between members of a linators. So, nectar volume and percentage of pair led to one flower developing more rapidly sugar were measured repeatedly on flowers of than the other. Usually the flowers developed greenhouse-grown populations. Flowers were synchronously and to the same size as M0ller gently probed (not destructively sampled) and Pomiankowski (in press) suggest for devel- with micropipettes. Each flower was probed opmentally stable, pollinator-visited flowers every 2 h from 0800 to 1600 h, nectar charac- such as Mimulus. It was important to use flow- teristics recorded, and nectar volumes ers of the same size and developmental stage summed (Appendix 3). At 1600 h previously inasmuch as they produce more nectar than unsampled control flowers were gently probed smaller flowers of pairs exhibiting fluctuating in the same manner and nectar characteristics asymmetry (M0ller and Pomiankowski in recorded for comparison to the repeatedly press) sampled flowers (Appendix 3). Mimulus flow- For the statistical analyses two tests were ers of section Enjthranthe typically develop in employed. F-tests were used to compare vari- pairs at each node of the flower stem, except ances of the pairs of wild populations and in M. easttvoodiae and M. rupestris. One flower greenhouse populations for nectar volumes 214 Great Basin Naturalist [Volume 54 Table 2. Comparison of nectar volume obtained by probing the flower vs. volumes obtained by destructively sampling the flower. Flowers were probed with a micropipette and then destnictively sampled to obtain the remainder of nectar present. Greenhouse-grown plants were used. #1 flower #2 #3 #4 #5 Volume Vol. % V^^l, % Vol. % Vol, % Vol. % X M. cardinalis-7113 initial probe 7,0 24.5 3.0 24.5 0.0 0.0 2.0 21.4 5.5 16.0 8.4 remainder 0.8 29.0 0.5 20.6 0.0 0.0 0,0 0.0 0.5 14.0 0.4 M. cardin(ilis-712() initial probe 7.0 17,1 9,0 19.1 6.5 18.8 8.5 17.0 9.0 7.0 8.0 remainder 12.0 17.2 26,0 19.4 17.0 19.0 14.0 17.1 26.0 7.7 19.0 M. cardinalis-6651 initial probe 6.0 14.0 6.0 18.6 3.0 20.2 0.2 10.0 — — 3.8 remainder 5.5 14.0 8.5 18.6 0.2 17.0 4.0 20.5 — — 4.5 M. eastwoodiae-6079 initial probe 0.0 0.0 1.0 25.0 0.5 34.0 7.0 17.0 7.0 14.0 3,1 remainder 0.8 21.2 0.0 0.0 1.0 27.4 3.5 17.0 5.0 14.0 2,0 M. leivisii-6103 initial probe 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 remainder 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 M. lewmi-5875 initial probe 0.0 0.0 0.7 15.3 0.7 10.6 0.8 8.0 0.8 29.2 0.6 remainder 0.1 26.0 0.3 14.0 0.0 0.0 0.0 0.0 0.2 29.0 0.1 M. nelsonii-6271 initial probe 7.0 16.3 4.7 18.8 8.5 16.0 1.2 14.0 — — 5.4 remainder 12.5 16.2 1.6 19.0 1.0 16.0 1.4 13.2 — — 4,1 M. rupestjis-9102 initial probe 0.0 0.0 1.0 34.0 2.2 25.2 0.0 0.0 0.0 0.0 0,6 remainder 0.0 0.0 0.8 29.2 1.5 25.6 3.0 25.8 5.5 28.1 2,1 M. verbenaceus-5924 initial probe 7.0 24.5 3.0 24.5 0.0 0.0 2.0 21.4 5.5 16.0 3.5 remainder 0.8 29.0 0.5 20,6 0.0 0.0 0.0 0.0 0.5 14.0 0.4 and sugar concentrations (Tables 3, 4). The F- test is particularly suitable to test variances (Sokal and Rohlf 1981). The null hypothesis was that observed variances of the wild and greenhouse populations sampled the same sta- tistical population. The F-test was also used to compare nectar volumes and sugar concentra- tions at successive 2-h intervals during die day. The Tukey-Kramer procedure (Lehman et al. 1989) was used to compare nectar volumes and sugar concentrations of greenhouse- grown representatives of the various species and races to each other (Table 5). This method uses average sample sizes and is to be pre- ferred to the T' -method or GT-2 method for comparisons of unequal sample sizes, accord- ing to Sokal and Rohlf (1981). Results and Discussion Wild Populations Observations of the standing crop of nectar in flowers of wild populations revealed signifi- cant differences in nectar volumes but not sugar concentrations among some populations but not others (Tables 3, 4, 5). Mimulus lewisii (both races), M. rupestris, and M. eastwoodiae formed one group with low nectar volumes that were insignificantly different from each other. Mimulus nelsonii and M. cardinalis formed a second group with significantly high- er volumes. Mimulus verhemiceus bridged the two groups with intermediate nectar volumes. In general, the more tubular and brighter red the flowers, the greater the volume of nectar and the more frequent the visits by humming- birds, although this varied from population to population and locality to locality. Conversely, the more open and pinker the flowers, the less the volume of nectar and the more frequent the visits of bumble or carpenter bees. Despite general trends, actual numbers of pollinator visits to flowers of wild populations varied markedly. Specifically, in an average of 3 h of observation each of M. rupestris (9102), M. cardinalis (13106), and M. eastwoodiae (6079), no pollinator visits were observed at all. In 2 h of observations each of M. verbe- naceus (13518B) and M. nelsonii (6271) 3 and 7 visits, respectively, by hummingbirds were 1994] Variance and Replenishment of Mimulus Nectar 215 Table 3. Summary of nectar volumes (standing crops) produced by flowers of wild populations and greenhouse- grown populations of the species of section Enjthranthe. f = significantly higher variance and | = significantly lower variance in the greenhouse-grown population tlian in the wild population. Wild populations Greenhouse populations^ Variances equal Population Sample Mean Standard Population Sample Mean Standard number size volume deviation number size volume deviation (Table 1) (n) (/aI) (n) (jA) F ratio Probability M. cardinalis 13486 80 12.08 ±1.34 ±11.00 M. eastwoodiae 13514 1.10±0.17 ±1.16 M. lewisii — Sierra Nevada race 13515 69 0.60 ±0.21 ±0.63 M. lewisii — Rocky Mountain race 5875 121 0.97 ±0.07 ±0.74 M. nelsonii 6271 155 16. 10 ±1.16 ±15.39 M. rupestris 9102 13 0.99 ±0.78 ±1.54 M. verhenaceus 13518B 65 7.27 ±1.21 ±6.05 13106 6079 6103 6271 9102 13547 40 50.78 ±1.89 ±13.71 27 6.41 ±0.31 ±2.58 22 2.29 ± 0.38 ± 3.52 5S75 127 1.54 ±0.07 ±0.81 38 19.26 ± 2.34 ± 9.44 55 5.42 ± 0.38 ± 3.02 43 42.49 ±1.49 ±13.67 1.7330 .1906 45.1157 .0000 1 41.5997 .0000 T 0.7317 .3932 4.5126 .0349 f 7.0458 .0099 i 29.4422 .0000 j ^Values for nectar volumes in unopened buds (see .Appendix 2) are omitted from these data. recorded but no bee visits. In the populations of M. lewisii, 7 hummingbird and 232 bumble- bee visits were observed in 13 1/2 h of obser- vations of population 13513 of the Sierra Nevada race, and 2 hummingbird and 12 bumblebee visits were observed in 4 1/4 h of observation of population 5875 of the Rocky Mountain race. The highest number of polli- nator visits was observed in the Santo Tomas population (13486) of M. cardinalis with 600+ visits by hummingbirds and 70+ visits by bumblebees in the course of 4 1/2 h of obser- vation. All observations were made for 15-min periods scattered from dawn to dusk. Each population had at least 200 flowers in bloom. The number of pollinator visits to a population depends strongly on the guild of pollinators in that area at that time. For example, the Santa Tomas area was alive with pollinators, whereas Cedros Island lacked them almost completely. Nectar volume varied so much from flower to flower (Appendix 1) that pollinators would have to visit each flower in order to ascertain its nectar reward. Actually, pollinators appear to be cueing in on shapes and/or colors that promise an acceptable reward, on average, but not necessarily from each flower visited. Variances were so high for nectar volumes that one standard deviation approached the popu- lation mean in magnitude in all populations (Table 3). In contrast, variation in sugar con- centration was far less. It was less than one- fifth the magnitude of the mean on average (Table 4). High variances in nectar volume could be due to unequal visits by pollinators; variations in soil moisture; climatic factors such as wind, dew, or rain; or microclimatic variations in humidity around the flowers (Cruden and Hermann 1983, Wyatt et al. 1992). As a day progressed, from as early as 0600 to as late as 2000 (Appendix 1), the mean vol- ume of nectar in flowers of wild populations changed little despite withdrawals by pollina- tors, evaporation, dilution, stimulation by cli- matic factors (Table 6), or, possibly, by reabsorp- tion of nectar by the nectaries (Burquez and Corbet 1991). Specifically, the nectar volume remained essentially unchanged in flowers of four populations: M. eastwoodiae, M. nelsonii, M. rupestris, and M. verhenaceus. It decreased, as would have been anticipated for all popula- tions, if replenishment were not occurring, in only two populations, M. cardinalis and the Sierra Nevada race of M. lewisii. It actually rose in one population, the Rock\' Mountain race of M. lewisii. The increase in volume was 216 Great Basin Naturalist [Volume 54 Table 4. Summary of nectar sugar concentrations (standing crops) produced by flowers of wild populations and greenhouse-grown populations of the species of section fJn/^/jra/j^/ie. f = significantly higher variance and i = sig- nificantly lower \ariance in the greenhouse-grown population than in the wild population. Wild populations Greenhouse populations'^ Variances ecjual Population Sample Mean Standard Population Sample Mean Standard number size concentration de\iation number size concentration deviation (Table 1) (n) (%) (n) (ji\) F ratio Probability' M. cardinalvi 13486 80 12.86 + 0.42 ±3.82 13106 M. eastwoodiae 13514 88 16.14 ±0.91 ±8.90 6079 M. lewisii — Sierra Nevada race 13515 69 12.07 + 0.68 ±4.46 6703 M. lewisii — Rocky Mountain race .5875 121 16.97 ±0.93 ±8.32 5H75 M. nehonii 6271 155 19.94 ±0.33 ±3.97 6271 M. nipestris 9102 13 18.98 ±2..54 ±14.09 9i02 M. verhenaceiis 13518B 65 14.42 ±0.45 ±4.48 13547 40 20.78 ±0.59 ±3.54 0.31.36 .5766 27 18.97 ±1.65 ±7.27 3.4170 .0671 22 13.72 ±1.21 ±8.53 18.6158 .0000 f 127 33.05 ±0.90 ±11.74 12.3031 .0005 J, 38 17.92 ±0.66 ±4.55 0.4615 .4977 55 17.53 ±1.23 ±7.64 5.8744 .01811 43 17.32 ±0.55 ±1.61 26.4864 .0000 1 "Values for sugar concentrations in unopened buds (See Appendix 2) are omitted from tliese data. not due to dilution inasmuch as there was no corresponding decrease in sugar concentration (Table 6). The only species showing a decrease in sugar concentration was M. rupestris, which, however, showed no significant rise in nectar volume. These observations suggest to us that flowers are producing additional nectar both as the day advances and/or as pollinators remove it. Greenhouse-grown Populations Flowers of the greenhouse-grown popula- tions had, as an overall average, more than three times the volume of nectar found in flowers of wild populations, but essentially the same levels of sugar concentration. In three populations, 5875 of the Rocky Mountain M. lewisii, 6271 of M. nelsonii, and 9102 of M. rupestris (Table 1), direct comparisons could be made between greenhouse-grown plants and plants in wild populations because green- house plants were either transplants or grown from seeds collected from the same wild pop- ulations. These greenhouse plants exhibited about twice the volume of nectar recorded for corresponding wild plants. In the other four populations only indirect comparisons were possible. In these cases wild populations came from similar habitats but different localities than the greenhouse-grown populations of the same species or race (Table 1). Greenhouse- grown plants exhibited over four times the volume of nectar found in their wild counter- parts. Presumably the increase in nectar vol- ume in both groups of populations when grown in the greenhouse reflects lack of nec- tar withdrawals in the greenhouse due to absence of pollinators and to more standard- ized and more consistently favorable climatic, soil moisture, and humidity conditions in the greenhouse. Higher relative humidity has been shown to lead to higher nectar produc- tion in Ascelpias syriaca (Wyatt et al. 1992). The increased nectar was more dilute in Ascelpias in contrast to the Mimiilus nectar, which remained at essentially the same sugar concentration. Relative humidity in our green- house was typically 65%, but ranged up or down by 15%. Relative humiditx' at Moab, the closest station to our locality at Bluff, averaged 19%, with ranges of 11-80% on average (Utah Climate Center 1993). This was during July and August (1993), the Mimulus flowering sea- son. It is small wonder that nectar production for that desert population rose significantly higher, nearly sixfold, in our humid greenhouse 1994] Variance and Replenishment of Mimulus Nectar 217 Table 5. Comparisons of mean nectar volumes and mean nectar sugar concentration of the species of Mimulus of sec- tion Enjthranfhe using the Tukey-Kramer test (Sokal and Rohlf 1981). Positive vakies show pairs of means that are sig- nificantK different. Voknne nehonii cardinalis verhenaceus eastwoodiae rupestris lewisii R. Mtn. lewisii Sierras nehonii -.5.2141 cardinalis -2.336.5 -7.1398 verbenaceiis 1.8260 -2.8556 -9.0312 eastwoodiae 7.9089 3.1797 -3.0400 -8.2443 nipestris 9.5366 4.6160 -1.7667 -6.9066 -5.6009 lewisii — 8.0589 3.3297 -2.8900 -8.0943 -7.0387 -8.2443 R. Mtn. leivisii — 8.8322 4.0631 -2.1927 -7.3825 -6.2735 -7.5325 -7.6327 Sierras Sugar concentration nehonii nipestris lewhii R. Mtn eastwoodiae verhenaceus cardinalis lewisii Sierras nehonii -8.0453 nipestris -7.3070 -8.6421 lewisii — -7.8996 -9.1731 -12.7208 R. Mtn. eastwoodiae -5.9996 -7.2731 -10.8208 -12.7208 verhenaceus -5.5712 -6.8299 -10.2783 -12.1783 -13.9349 cardinalis -2.4944 -3.7911 -7.4909 -9..3909 -11.21,58 -11.0165 lewisii — -2.1016 -3.3875 -7.0176 -8.9176 -10.7241 -10.5711 -11.7772 Sierras (Table 3). Relative humidity at Park City, the closest station to our Alta locality, during the July-August flowering season for Mimulus averaged 46% with ranges of 17-85% on aver- age (Utah Climate Center 1993). Nectar pro- duction in the greenhouse was slightly, but insignificantly, higher than nectar production in the wild for this pair of populations. Relative humidity appears to help set the limit on how much of a flower's potential for nectar production is realized. There was no indication of nectar reabsorption. Greenhouse populations exhibited much the same groupings of nectar volume produc- ers as did wild populations. That is, M. east- woodiae, M. lewisii (both races), and M. nipestris, were the low producers; M. cardi- nalis and M. verhenaceus were the high pro- ducers; and M. nelsonii was the intermediate producer. In all but two cases variance in nectar vol- umes increased significantly in greenhouse- grown populations compared to wild popula- tions (Table 6). This occurred despite lack of pollinators. Variability in the standing crop of nectar appears to be intrinsic and not simply due to uneven nectar withdrawal by pollinators. Variability in nectar volume might function as a strategy to insure pollinator visits to many flowers of a population (Wiens personal com- munication); that is, the psychological principle of intemiittent rewards would seem to be oper- ating (Edward Cook personal communication). In flowers of greenhouse-grown populations sugar concentrations varied insignificantly. They tended to remain in the range of 12-20% (Appendix 2). Nectar Replenishment Nectar replenishment is indicated by the general maintenance of nectar volumes despite nectar removal in wild populations and by the tendency of nectar volumes to increase in the absence of pollinators in green- house-grown populations (Table 6). Comparison of nectar volumes produced when flowers were probed with a micro- pipette every 2 h until the late afternoon — like a pollinator removing nectar — with flowers that were not probed at all until the late after- noon demonstrated that repeatedly probed flowers produced at least twice as much nectar as flowers that were probed only once (Appen- dix 3). While nectar volume apparently 218 Gre.\t B.\.sin Naturalist [Volume 54 Table 6. Changes in (loral nectar volume (^1) and percent (%) .sugar witli time, during the course of a day (Appendices 1, 2). t equals a significant increase with time and I equals a significant decrease with time. Species Wild populations Greenhouse populal ;ioiis Mean F ratio P .Mean F ratio V M. cardinalis Santo Tomas 12.0812^1 12.8637% 27.5578 0.4307 .oooo i .5136 — — — Cedros Island — — — 50.7675 /il 20.7850% 1.2138 16.9257 .2775 .0002 1 M. eastivoodiae San Juan 1.1068^1 16.1454% 3.3306 0.8536 .0715 .3581 — — — BluflT — — — 6.2462 /il 19.2392% 20.1851 0.1186 .0001 1 .7334 M. leicisii — Sierra Yosemite 0.6028 All 12.0753% 4.0774 0.0679 .0475 i .7953 — — — Ice Lake — — — 2.2909^1 13.7272% 2.6352 4.8804 .1202 .0390 i M. leicisii — Rock\' Mtns. Alta 0.9719 yul 16.9743% 5.1894 2.1025 .0245 1 .1497 6.2464/11 19.2392% 20.1851 0.1186 .0001 t .7334 M. nehonii Sierra Madre 16.1045^1 19.9412% 0.0568 1.4890 .8119 .2243 19.2657 yLtl 17.9263% 0.3331 5.9106 .5674 .0202 t M. rupestris Tepozteco 0.9923^1 18.9846% 0.0159 8.0010 .9018 .01641 5.4163 jLtl 17.5345% 2.7497 1.1780 .1032 .2827 M. verhenaceus Oak Creek 7.2333 /il 14.3757 % 0.8416 3.9155 .3624 .0521 — — — Grand Canyon — — — 46.3830/11 17.8630% 12.0601 19.7.339 .0009 1 .0000 1 increases with time alone (see above), volume increases more rapidly with repeated removals. The amount of nectar produced by flowers in successive 2-h periods tended to decrease in M. cardinalis, M. eastivoodiae, M. nelsonii, and M. verhenaceus (Appendix 3). The per- centage of sugar dropped in only two cases, the 7120 population of M. cardinalis and the 6271 population of M. nelsonii. Apparently, production of additional nectar is not achieved, with these possible exceptions, by dilution, but reflects the actual synthesis of more nectar. Consequently, calculations of the amount of sugar produced by a flower depend not only on volume of nectar and percent sugar at the time of sampling (Bolten et al. 1979, Sutherland and Vickeiy 1993), but also on the amount of sampling and hence the amount of replenishment of nectar in that flower. Acknowledgments We thank the National Science Foundation for support— Grant No. BSR-8306997. We thank Jeanette Stubbe for much typing. Literature Cited B.\KER, H. G. 1983. An outline of the history of anthology, or pollination biology. Pages 7-28 in Leslie Real, ed.. Pollination biology. Academic Press, New York. 1994] Variance and Replenishment of Mimulus Nectar 219 Bolton, A. B., P Feinsinger, H. G. Baker, and I. Baker. 1979. On the calculation of sugar concentration in flower nectar. Oecologia 41: 301-304. Burquez, a., and S. a. Corbet. 1991. Do flowers reab- sorb nectar? Functional Ecology' 5: 369-379. Cruden, R. W, and S. M. Her.mann. 1983. Studying nec- tar? Some observations on the art. In: B. Bentley and T. Elias, eds.. The biology of nectaries. Columbia Universit)' Press, New York. 259 pp. Faegri, K., and L. Van Der Pijl. 1979. The principles of pollination ecology. 3rd ed. Pergamen Press, Oxford. 244 pp. Free, J. B. 1970. Insect pollination of crops. Academic Press, London. 544 pp. HiESEY, W. M., M. A. Nobs, and O. Bjorkman. 1971. Experimental studies on the nature of species. V. Biosystematics, genetics, and physiological ecology of the Erijthranthe section of Mimulus. Carnegie Institution of Washington, Washington, D.C. Publi- cation No. 628. 213 pp. Lehman A., J. Salt, and M. Cole. 1989. JMP user's guide. SAS Institute, Gary, North Carolina. 464 pp. M0LLER, A. E, AND A. PoMlANKOWSKL In press. Fluctua- ting asymmetry and sexual selection. Genetica. SOKAL, R. R., AND F J. RoHLE 1981. Biometry. 2nd ed. W. H. Freeman and Co., New York. 859 pp. Sutherland, S. D., and R. K. Vickery, Jr. 1993. On the relative importance of floral color, shape, and nectar rewards in attracting pollinators to Mimulus. Great Basin Naturalist 53; 107-117. Utah Climate Center. 1993. Station reports for Moab and Park City, July and August 1993. Utah State University, Logan. Vickery, R. K., Jr. 1978. Case studies in the evolution of species complexes in Mimulus. Evolutionary Biology 11:404-506. . 1990. Pollination experiments in the Mimulus car- dimilis-M. leivisii complex. Great Basin Naturalist 50: 155-159. Vickery, R. K., Jr., and B. M. Wullstein. 1987. Comparison of six approaches to the classification of Mimulus sect. Erijthranthe (Scrophulariaceae). Systematic Botany 12: 339-364. Wyatt, R., S. B. Broyles, and G. S. Derda. 1992. Environmental influence on nectar production in milkweeds (Asclepias sijriaca and A. exaltata). American Journal of Botany 79: 636-642. Received 17 June 1993 Accepted 19 January 1994 (Appendices 1-3 follow on pages 220-227." 220 Great Basin Naturalist [Volume 54 Appendix 1. Standing crop of nectar in flowers of wild populations of the species and races of section Erythranthe of the genus Mimulus at different times of day. Time is given in terms of a 24-hour clock, volume of nectar is in microliters (;u.l), and sugar concentra- tion of the nectar in percent sugar (%). Data were gathered by Steven Sutherland in 1986-87. M. cardinalis— -Santo Tomas, Baja California, del Norte !, Mexico 0600 0800 1000 1200 1400 1600 1800 2000 Vol. % Vol. 9f \bl. % Vol. 7. Vol. % Vol. % Vol. % Vol. % 8.5 18.0 10.0 8.0 13.5 12.2 4,0 11.6 3.0 9,0 8.0 14.8 1.0 8,0 3.0 13.0 43.5 16.8 6.5 6.6 18.0 11.2 49.5 14.2 19.0 19,2 3.5 14.8 1.5 12,4 3.0 12.4 46.0 16.2 16.0 12.0 7.5 14.8 8.5 20.4 8.0 9,6 1.5 9.0 1.5 16,2 22.0 13.6 22.5 10.0 25.0 13.2 6.5 9.6 21.5 14,3 5.5 13.0 1.0 7.4 2.0 21,0 9.0 15.4 12.0 15.6 4.5 12.2 4.0 14.0 32.5 12,8 4.0 7.6 1.0 10.4 4.0 14,2 11.5 17.2 34.0 16.6 11.0 11.4 6.5 10.4 13.5 14.4 22.5 14.8 8.5 11.2 5.0 6,8 14.0 19.2 13.0 12.4 35.0 14.0 7.0 9.0 3.0 11.0 21.0 23.8 7.0 8.0 4.0 5,0 6.5 17.2 18.5 9.0 40.0 16.4 11.5 6.4 13.0 12.4 6.5 13.0 3.5 13.4 4.0 7,4 5.0 14.6 29.0 14.0 16.0 12.2 10.0 11.2 7.0 14.2 8.0 16.8 3.5 15.4 6.0 13,4 10.5 16.4 20.5 14.2 13.0 9.6 23.5 14.8 5,5 12,2 10.0 7.2 3.5 7.4 7,0 9,0 7,0 21,4 x= 24.8 14.3 17.7 11.6 10,8 11.4 15.8 13.8 10.8 13.4 4.1 11.2 3,6 11,3 9,2 16,0 M. eastwoodiae — Blu iff, Utah 0800 1000 1200 1400 1600 1800 Vol. % Vol. % Vol. % Vol. % Vol. % Vol, % 7.3 17.6 0.3 13.8 0.2 21.6 0.3 27.2 1.0 14.4 1,2 17.2 2.0 18.2 0.9 33.0 1.3 12.4 0.4 17.0 1.2 11.8 1,2 12.8 1.0 15.4 2,5 31.2 1.8 33.0 0.3 26.6 0.6 3.8 1,0 6.8 0.6 7.0 2.3 11.4 0.7 18.2 0.2 5.4 0.3 2.0 2.5 12.0 0.8 14.0 0.5 5.6 0.5 23.8 1.0 10.4 0.7 7.4 1.2 12.8 1.3 5.2 1.4 11.2 0.6 22.4 0.3 25.2 0.6 9.2 1.3 18.6 0.6 10.0 0.8 7.2 4.5 28.0 0.4 22.0 2.6 25.0 0.8 26.0 0.3 10.4 1.2 8.4 2.4 17.4 1.8 9.0 0.2 7.6 0.1 9.0 0.6 6.8 0.6 5.2 4.3 33.0 0.5 21.8 1.2 10.0 2.5 8.8 1.3 20.2 1.2 33.0 0.4 19.0 l.S 11.2 0,6 23.0 x= 1,6 11,6 1,2 1,7 24,3 0,5 18,3 1.0 10.2 1.2 14.7 M. lewisii — Yosemite, California 0600 0800 1000 1200 1400 1600 1800 Vol. % Vol, % Vol. % Vol. % Vol, % Vol, % Vol, % 0.1 6.0 0,3 10,0 0,7 9,4 0.4 14.0 0,8 11,4 0.7 9,0 0,5 10,6 3.2 10.8 1,2 10,6 1,3 10,2 1.2 14.6 0,1 7,4 0,4 9,0 1,0 12.2 0.2 5.0 0,8 10,0 1.2 29,4 0.2 16.4 0,8 10.8 0.2 15,2 0,5 12.2 1.4 17.4 2,0 22.6 0.5 7,0 0.1 16.2 0,2 13.2 0.1 6,0 0,3 16,8 0.5 16.4 0,2 6.0 0.2 8,4 0.2 12.6 0,1 11.4 0.1 10,8 0,1 9,4 0.8 20.2 0,3 10.2 1.6 26,8 0.1 14.0 0,3 12.2 0.1 4,0 0,2 15,2 1.0 13.2 0.3 12.4 0.2 10,8 0.6 13,6 0,2 11.0 0.2 11,2 0,1 12.8 1.0 5.2 0.3 8,6 0.4 11,0 0.1 14,0 2,7 16.8 0.1 16,0 1,5 10.8 0.3 11.6 0.3 10,4 0.2 9,4 0,1 13,6 0.2 9.0 0.6 12,8 X = 0.9 12,0 0.7 11,3 0.8 13.1 0,3 13,9 0.5 11.1 0.2 10,3 0,6 12.6 M. letvisii — All ta, Utah 0600 0800 1000 1200 1400 1600 Vol. % Vol. 7c Vol. % Vol. % Vol, % Vol, % 0,3 15,1 0.2 3,6 0,8 14.4 0.5 14.8 0.4 21.2 2,4 20,2 1,1 9,2 0.1 7,6 0,8 27.2 2.2 15.2 1.7 11,0 1,4 20,4 0,8 10,8 0.2 13,8 1,4 24.2 0.5 19.6 0.8 11,8 0,4 19,4 2,8 15.4 0.2 7.6 0.2 9.0 0.2 6.0 1,3 10,2 1,2 20,8 3.0 12.1 2.2 8.6 0.3 20.2 0.8 10.2 0,1 11,0 3,2 20,8 0.7 25.8 1.0 27.2 0.9 25.0 0.5 33,0 0,2 29,2 2,0 20,8 0.4 17.6 0.6 7.4 2.2 9.0 0.2 10,2 1,6 8,2 0,8 14,4 0.8 20.0 0.2 12.8 2.3 16.4 1.0 15,2 0,3 33,0 1,3 21,2 0.8 12,6 0,2 6.8 0.4 33.0 1.0 7,0 1,2 15,2 1,2 21,4 0.6 4,0 0.3 3.3 0,9 19.2 1.5 15,0 0,6 16,0 0,4 11,0 x= 1.1 14.2 9.9 1.0 19.8 0.8 14.6 0.8 16.' 1.4 19.0 1994] Variance and Replenishment of Mimulus Nectar 221 Appendix 1. Continued. M. nekonii— -Sierra Madre Occidental. , Sinaloa 0600 0800 1000 1200 1400 1600 1800 2000 Vol. %■ Vol. 9c Vol. % Vol. % Vol. 9c Vol. % Vol. "c Vol, % 12.3 18.4 32.3 21.4 12.4 16.8 10.8 16.8 2.1 17.0 5.2 20.2 12.3 18.4 32.3 21.4 7.3 15.6 6.1 17.0 30.2 19.0 44.0 18.0 2,8 17.0 4.4 21.2 7.3 15.6 6.1 17.0 22.0 30.4 15.0 70.2 10.6 17.0 23.6 15.2 12.0 17.6 6.2 25.4 22.0 30.7 15.0 20.2 5.5 20.0 15.4 18.8 14.0 15.4 32.4 18.2 2.0 17.2 11.3 23.8 5.5 20.0 15.4 18.8 2.9 16.4 27.6 16.0 11.3 15.8 22.6 16.6 2.1 25.4 2.6 20.2 2.9 16.4 27,6 16.0 3.7 15.6 53.5 17.8 5.7 19.8 4.7 15.0 2.5 21.0 2.7 19.0 3.7 15.6 53,5 17.8 14.6 21.6 12.7 20.0 5.8 16.8 7.3 19.4 33.2 23.4 3.4 18.2 14,6 21.6 12.7 20,0 27.5 27,0 1.0 25.2 2.2 16.6 6.9 17.6 4.3 20.8 2.5 15.8 27.5 27.0 1.0 25,2 10.8 21.2 10.0 18.3 2.4 16.0 12.7 17.6 14.5 29.2 3.2 23.2 10.8 21.2 10,0 18.2 5.0 21.6 1.7 18.4 2.4 16.2 4.2 29.0 2.7 20.2 5.0 21,6 1,7 18.4 x= 11.2 20.8 17.5 19.3 11.3 17.0 16.7 17.0 8.0 21.8 4.4 20.7 11.2 20,8 17,5 19.3 M. rupestris — Tepozteco, Morelos 1000 1200 1400 Vol. % Vol, % Vol, % 0.3 20.2 0,7 14.2 5,6 22,2 0.3 32.0 2.6 18,4 0,2 6,0 1.3 55,2 0.3 6,0 0,3 15,2 0.4 32,0 0.1 6,0 0,2 0,6 11,4 8,0 x= 0.6 .34,8 0.9 11,1 1,4 12,6 M. verbenaceus — Oak Creek Canyon, Arizona 0800 0900 1200 1400 14,6 5,1 12,9 9,8 13,4 3,9 14,0 1600 Vol, % Vol. % Vol. % Vol, 9c Vol, % 13,3 13.0 1.0 13.0 19.5 13.2 11,7 12.2 2,6 22,2 4.3 15.0 0.2 7.0 9.7 13.2 0,7 21,2 11,4 16.8 2.4 14.0 0.3 6.0 5.4 11.2 0,7 12,2 6,3 16.4 14.0 17.2 4.3 11.8 18.3 12,4 2.5 12,2 2,0 12.2 3.6 16.8 0.5 9.8 9.4 8,4 4.0 12,2 17,0 15.8 1.8 24.4 7.4 17.4 7.8 12,4 2,4 24.0 0.7 6.8 9.2 17.0 8.5 12,8 3,1 18.2 1.5 12.0 5.0 16.2 10.6 25.8 10,1 10.2 7.0 13.0 10.2 14.6 7.3 12.4 3,0 21.4 5.2 13.2 13.4 16.4 1.6 12,4 26.8 16.8 8.5 17.4 222 Great Basin Naturalist [Volume 54 % a o — o o a, c 2 ^ .S o "? <= s <=> < ^ Z >. C-) (M O) (M o) IN ic lo lo ■^ CO C-) -* to 00 o )0)oqcio-HO— ]05 -rcocDoooooooj'M'n>-rto o] p c-i 00 eg Cvj CD C<) ;Do6ci f^ ^jj m tn -H oq " rt cq (M O CM -r CD 00 O (M -^ (M -H o] — I oq -T C-] O ao CD -r (M CO og oi -H — . -H -H I I I I I I I ]-<—( -H -H (M-f-rcocCoooooooqrN-T"^ -r to O -f t !£) IN cocsit-^aJ-roiriio Cn1_( _,,-h— i^— <,-lrtCl C-I CO -H jDiDcocooooicg-r-t" (N(N(N(MCOCOCOCOCOCO — t(MOJ-H -HtN-H— Hr^-^oq o]'^0;0oooooooo-i O'M'l'-rtCiCDOOOOOO'MCM -H Ol -H — -4 CM rt ooootNOoqppp'^op lOt^O— <0(NO-HOt^OCO -H -H (N j^io;C^O O'*mcotoco-*"-HO-5<-rco M-Tfin'^iO'i'ioio OOOlMtOOOOO -flOlOlClOLO^CO ICCOtN'^COTfcOCOCOlMTt'cOCO -rtD;D0000OOC^]' E Reese^ and John VV. Connelly^ Ab.stract. — We translocated 196 Sage Grouse {Centrocercus urophasianus) into Sawtooth Valley, Idaho, during March-April 1986-87 to augment a small resident population. Forty-four grouse equipped with radio transmitters were monitored through spring and summer. Nest sites (n = 6) had greater (P = .032) horizontal cover than did independent random plots {n — 7). During summer, grouse used sites (n = 50) with taller live and dead shnib heights, greater shrub canopy cover, and more ground litter (P < .009) than were foimd on dependent random plots [n — 50) 50-300 m from use sites. Distance to edge and mountain big sagebrush {Artemisia tridentata vaseyana) density best separated use sites from independent random plots in logistic regression analysis and conectly classified 64% of the use sites and 78% of the independent random plots. Sage Grouse used sites that had narrower frequency distributions for many variables than did independent plots (P < .04), suggesting selection for uniform habitat. Key words: Centrocercus urophasianus, dispersal, habitat use. hntiie raii3200 m) and to the east by the White Cloud Mountains (>3500 m). The periphery of Saw- tooth Valley is composed of rolling glacial moraines with slopes >10°. The valley floor is composed of glacial and alluvial deposits with slopes 0-5° (Tuhy 1981). Average annual precipitation is 26 cm and average annual temperature is 6.5° C. The val- ley averages 2.5 m of snow, which accounts for 85% of the annual precipitation (Tuhy 1981). Sagebrush cover dominates approximateK' 125 km2 (75%) of Sawtooth Valley Mountain big sagebrush/Idaho fescue {Festuca idahoensis) is the major habitat type (Tuhy 1981). 'Department of Fish and Wildlife Resources, University of Idaho, Moscow, Idaho 83843. ^Present address: Idaho Department of Fish and Game, 868 East Main Street, Jerome, Idaho 83339. 3ldaho Department of Fish and Game, 134.5 Barton Road, Pocatello, Idaho 83204. 228 1994] Habitat of Translocated Grouse 229 Wet meadows and riparian areas cover 19 km- (11%) of the valley, irrigated pastnres 19 km^ (11%), and isolated stands of lodgepole pine {Pinus contorta) 3 km^ (2%; Musil 1989). Methods During late March and early April 1986 and 1987, we captured 196 Sage Grouse (46 adult females, 19 yearling females, 115 adult males, 16 yearling males) by spotlight trapping (Giesen et al. 1982) on 1 1 leks from nonmigra- tory populations (J. W. Connelly personal observation) in southeastern Idaho. Capture areas were at similar elevations approximately 144 km from Sawtooth Valley. Grouse were classified to age and sex (Dalke et al. 1963) and leg-banded at the capture site. Males were transported in wooden crates and females were moved individually in modified card- board boxes to reduce head-scalping and other injuries (Patterson 1952). Birds were trans- ported by truck to Sawtooth Valley each morn- ing after capture and moved by snowmobiles to the release site adjacent to the last active lek. Releases occurred from 19 March to 6 April 1986 and 25 March to 1 April 1987. We equipped 44 (22%) grouse (31 females, 13 males) with solar-powered radio-transmit- ters (Musil 1989, Musil et al. 1993) attached to ponchos (Amstrup 1980). Fifteen grouse (8 females, 7 males) were marked with radios in 1986 and 29 (18 females, 11 males) in 1987. Weight of telemetry packages (<25 g) was <2.2% of the mean body weight of female grouse. We located birds at least twice per week, equally dividing locations during the day among three periods (Dunn and Braun 1986). We tracked radio-marked birds from the ground using a hand-held 4-element Yagi antenna and receiver (Mech 1983). Radio-locations were obtained by walking a 15-30-m-radius circle around the signal (Musil et al. 1993). We plotted radio-locations on aerial photographs and 7.5-minute U.S. Geological Survey orthoquadrangle topo- graphic maps overlaid with the Universal Transverse Mercator (UTM) grid system (Lancia 1974) scaled to 100 m^/grid. Habitat Characteristics Nest sites. — Nests of translocated Sage Grouse were located by telemetry and inci- dental sighting. Nest site characteristics were measured after nesting efforts ceased. At each nest the number of shrubs in contact with the nest bowl was counted. Height of the shrub over the bowl and area (length X width) of the shrub mass surrounding the nest were mea- sured. Density of shrubs <40 cm and >40 cm tall was measured within a 2-m radius of the nest. A cover board (Jones 1968) was placed in the nest, and horizontal cover was estimated at 2 m from the nest at 0° and 45°. The board was also placed flat in the nest and cover at 90° was measured. Four 20-m transects were positioned at cardinal directions intersecting the nest, and shrub cover was measured using the line-intercept method (Canfield 1941). Shrub and grass heights were measured at 5-m intervals along the transects. To determine whether Sage Grouse were selecting nest sites based on stand characteris- tics, we established a dependent random plot in 1987 at a random direction and distance 50-300 m from each nest site. A correspond- ing independent random plot was located by randomly selecting two 5-digit numbers corre- sponding to the last five numbers of the east and north UTM coordinates covering the study area (167 km^). To find the independent random plot, we paced the distance along a compass line from the nearest landmark to the point. Only points in sagebrush habitat were used for independent random plots because this is the only habitat used for nesting by this species (Patterson 1952, Petersen 1980, Wakkinen 1990, Connelly et al. 1991). Daily use sites. — Vegetative and topo- graphic variables were measured at sites used by radio-marked Sage Grouse during May- July 1987. Use plots were centered at radio- locations and selected uniformly among daily use patterns of Sage Grouse (Dunn and Braun 1986). Habitat characteristics were also mea- sured at dependent and independent random plots as described for nest sites. At each use site we measured vegetation along two parallel 15-m transects placed 8 m apart. Transects were positioned peipendicular to the contour of the slope and centered with- in a 60-m-radius circle for use sites. Shrub canopy cover was measured by line-intercept (Canfield 1941). Shrub density (plants/m^) within 0.5 m of each side of the transect was measured, and a clinometer was used to record slope at each vegetation site. 230 Great Basin Naturalist [Volume 54 We estimated understory cover with modi- fied Daubenmire (1959) 4 X 5-dm plots at 1.5-m intervals (20 frames/site) along the transects (Mosley et al. 1986). At each Daubenmire plot, heights of the closest live and dead shrub < 1 m from the transect were measured. Locations of vegetation sampling sites were plotted on 7.5-minute orthoquadrangle topo- graphic maps, elevations recorded, and the distance to the nearest change in cover type (i.e., pasture, riparian, wet meadow, or timber) measured with an electronic planimeter. Vegetation analysis. — Depending on nor- mality, univariate parametric or nonparametric statistical tests were used for comparing equal- ity of both means and variances between use and random sites. Separate analyses were con- ducted for use vs. dependent random sites (matched pairs) and between use and indepen- dent random sites using SAS (SAS Institute, Inc. 1985) and Statistix II (Analytical Software, Box 130204, St. Paul, Minnesota 55113) com- puter programs. We used logistic regression (Harrell 1985) to identify variables that best distinguished Sage Grouse use from independent random sites. Maximum-likelihood estimates were computed to determine coefficients for vari- ables in the predictive model. The significance level to enter and stay in the logistic regres- sion model was set at .10, and addition of vari- ables to the model was stopped once the X^ test of the residual variables was no longer sig- nificant. Nonparametric tests were used to compare nests with random plots because of the small sample of nests (n = 6). Wilcoxon s signed rank test (Conover 1980) was used to compare height of the shrub covering the nest and average height of live shrubs along the tran- sects surrounding the nest. We did not intentionally flush radio- marked grouse; thus flock composition was largely unknown. Occasionally, mixed-sex flocks were flushed, which suggested that plots used for habitat sampling were not represented by one sex. Therefore, we did not compare habi- tat use by male and female grouse. Results Nest Sites At least one translocated Sage Grouse nest- ed in 1986 and six nested in 1987. Two of the grouse that nested in 1987 were birds released in 1986; the others were released during spring 1987. Vegetation at nest sites {n = 6) did not differ (F > .10) from dependent ran- dom plots (Wilcoxon signed rank test, P > .249 for all values). Although average height of shrubs covering nests (x = 50.7 ± 6.7 cm) was greater {P = .04) than average shrub height surrounding nests (x = 27.3 ± 4.0), there were no differences in shrub height or cover at nest sites compared with dependent or independent random sites. Grouse nested at sites^with greater {P = .03) horizontal cover at a 45° angle to the nest (x = 86.0 ± 12.5) than at independent random sites (x = 66.9 ± 16.5). Daily Use Sites Between 22 May and 23 July 1987, 50 use sites were sampled for 15 (3 males, 12 females) radio-marked grouse, with an equal number of dependent and independent random sites. Dependent random sites averaged 163 ± 16 m from use sites. Grouse used sites with more shrub canopy cover (P < .01), greater litter cover (P < .01), and taller live and dead shrubs (P = .00) than at dependent random sites (Table 1). Variance tests indicated few dif- ferences in frequency distributions between Sage Grouse use and dependent random sites (Table 1). Sage Grouse used areas with flatter slopes (P < .01), farther from habitat edges (P = .01), with more litter cover (P = .00), less bare ground (P = .00), and greater density of moun- tain big sagebrush (P = .04) (Table 1) than at independent sites. Variance tests indicated that grouse used narrower frequency distribu- tions of slope, elevation, live shrub canopy cover, bare ground, density of shrubs other than sagebrush, and live shrub height (P = .00) but wider distributions of distances to edge (P = .00), dead shrub canopy cover (P < .01), total shrub density (P = .03), and dead shrub height (P = .00) (Table 1). Two variables were identified by logistic regression to best separate use sites from independent random sites. Distance from edge and mountain big sagebrush density cor- rectly classified 64% of the use sites and 78% of the independent random sites. The proba- bility that a site would be classed as a use site increased as distance from habitat edge and density of mountain big sagebrush increased. 1994] Habitat of Translocated Grouse 231 o = o m CO t^ o CD ^ O lO CO ^ -H e>^ CO t- -f csi ^ t- 00 LD O t-~; -H oq 00 p ;b CO oi -i- od CO o o o o o o -H ^ o o (M 05 O CO o pp'i^cOcOLO UO'^IOCN t-^ c-^ — < -t< t-^ t^ o-i ^ o ^ p p -r t^ CO o 't -t O 1- lO o p CO p lO t-- o> CD CD ^ 'tt CC «D -^ o 05 -r O O O O O -H 05 — ( — I CD CD t^ GO 00 oq t--;t^coLocoo5 oqic OOOOOO CDCO (N CO ro t^ C^ CM in lo -H CO CD C75 CD CO m in CD CM CO -H -^ O O ^ CM CO in 5 1 CD 05 t^ 00 in CO 05 -H o o ^ CD CD CD CO CO in CD t^ CO O -ii C I- t- O -T CO CO TT CD 05 ^ oq CM CO —1 ^ O O -H CM CD CO CO on W 5 p 4J -C t/) „ ij "2 ^ — 2 i .^ Sji — Sf - ^ - -■ •' pa U CO C/2 §- a; 232 Great Basin Naturalist [Volume 54 Discussion Nest Sites All nests were under sagebrush, similar to findings for many established populations (Patterson 1952, Klebenow 1969, Wallestad and Pyrah 1974, and Petersen 1980) but some- what different from Sage Grouse nesting in southeastern Idaho (Connelly et al. 1991). No differences were detected between nest sites and dependent plots in the same stand of sagebrush, but hens did nest under shrubs that were taller than shrubs surrounding the nest. These findings are similar to those reported by Wallestad and Pyrah (1974) and Petersen (1980) for established populations. However, Wakkinen (1990) reported that Sage Grouse in southeastern Idaho nested under taller shrubs with a larger area than shrubs in the same stand. Hens may select tall plants and clumps of shrubs for nest sites because these provide more visual obstruction to predators. We detected few differences in vegetation between nest sites and independent random sites. Wakkinen (1990) reported similar find- ings and suggested this indicated an abun- dance of suitable nesting habitat. Daily Use Sites Translocated Sage Grouse in Sawtooth Valley used sites with greater physical obstruction than at dependent random sites, and these may have provided more conceal- ment from predators. Grouse use sites also had greater litter cover, which may be related to shrub cover and live shrub height as well as insect abundance (Patterson 1952, Johnson and Boyce 1990). In a comparison of summer use sites with independent random plots, grouse used flatter sites near the center of the valley rather than the rolling glacial moraines along the perime- ter. The central part of the valley has extensive stands of mountain big sagebrush, whereas mixtures of sagebrush and antelope bitter- brush {Purshia tridentata) occur on the moraines. Areas used by grouse had little interspersion of habitat edges when compared to sagebrush along the perimeter of the valley. The perimeter had narrow peninsulas of sage- brush on steeper slopes that extended into lodgepole pine timber. These sites were not used by radio-marked Sage Grouse. Ratti et al. (1984:1193) tested variances between Spruce Grouse {Dendragapiis canaden- sis) use sites and random plots and stated that "these differences indicated a preference for sites having habitat characteristics with less variation than the general environment. " Similarly, translocated Sage Grouse used nar- rower ranges for several microhabitat charac- teristics, both topographic and structural, compared with habitat available throughout the study site. However, within a stand of sagebrush, translocated grouse selected habi- tat with greater-than-average values rather than narrower frequency distributions. Translocated Sage Grouse were not associ- ated with edges of cover types as was reported for grouse in Colorado (Dunn and Braun 1986). Grouse in Sawtooth Valley were associated with greater-than-average structural charac- teristics of sagebrush within a stand (i.e., taller brush and greater canopy cover). This sug- gests that variability in habitat structure not only among but also within stands of sage- brush is important to Sage Grouse by provid- ing adequate habitat during different seasons and for diurnal uses (Dunn and Braun 1986). Characteristics of nesting and summer habitats used by translocated grouse within Saw1:ooth Valley were generally similar to those reported for established Sage Grouse popula- tions in many parts of the species' range. This similarity suggests that translocations of Sage Grouse, if carefully planned, are a feasible method of augmenting or reestablishing Sage Grouse populations (Musil et al. 1993). Patterson (1952) concluded that restoration of relatively small Sage Grouse populations by translocation was not effective because of the birds' tendency to disperse from the release site. Contrary to Patterson's (1952) findings. Sage Grouse translocated into the Sawtooth Valley remained near the release site (Musil et al. 1993). Dispersal of these birds may have been greatly reduced because they were released during the breeding season, into the relatively insular and isolated Sawtooth Valley, and, perhaps most importantly into an area with adequate spring and summer habitat. Ac KNOWLE DGM E NTS We thank T L. Parker, G. D. Power, M. D. Scott, and J. E. Wenger for help with field- work. We also appreciate the help of G. W 1994] Habitat of Translocated Grouse 233 Gadwa and B. R. Snider during releases. R. D. Guse entered data, and R. L. Crabtree and G. D. Hayward assisted in data analysis. Reviews by C. E. Braun, T. P Hemker, and an anonymous referee improved earlier drafts of this manuscript. This is Gontribution No. 707 of the Idaho Forest, Wildlife and Range Experiment Station and Federal Aid in Wildlife Restoration Project W-160-R-15. Literature Gited Allred, W. J. 1946. Sage Grouse trapping and transplant- ing. Proceedings of the Western Association of State Game and Fish Commissioners 26: 143-146. Amstrup, S. C. 1980. A radio-collar for game birds. Journal of Wildlife Management 44: 214-217. Autenrieth, R. E. 1981. Sage Grouse management in Idaho. Wildlife Bulletin 9. Idaho Department of Fish and Game, Boise. 238 pp. Batterson, W. M., a.nd W. B. Morse. 1948. Oregon Sage Grouse. Oregon Game Commission, Federal Aid in Wildlife Restoration Project 6-R. 29 pp. Ca.nfield, R. H. 1941. Application of the line interception method in sampling range vegetation. Journal of Forestr>' ,39: 388-394. Connelly, J. W, W L. Wakkinen, A. D. Apa, and K. P Reese. 1991. Sage Grouse use of nest sites in south- eastern Idaho. Journal of Wildlife Management 55: 521-524. CONOVER, W J. 1980. Practical nonparametric statistics. 2nd edition. John Wiley and Sons, New York. 493 pp. Dalke, R D., D. B. Pyrah, D. C. Stanton, J. E. Crawford, and E. F Schlatterer. 1963. Ecology, productivity, and management of Sage Grouse in Idaho. Journal of Wildlife Management 27; 811-841. Daubenmire, R. F. 1959. A canopy coverage method of vegetational analysis. Northwest Science 33: 43-64. Dunn, R O., and C. E. Braun. 1986. Summer habitat use by adult female and juvenile Sage Grouse. Journal of Wildlife Management .50: 228-235. GlESEN, K. M., T. J. SCHOENBERG, AND C. E. BraUN. 1982. Methods for trapping Sage Grouse in Colorado. Wildlife Society Bulletin 10: 224-231. Hamerstrom, F, and F Hamerstrom. 1961. Status and problems of North American Grouse. Wilson Bulletin 73: 284-294. Harrell, F E., Jr. 1985. The logist procedure. Pages 269-293 in P. S. Reinhardt, ed., SUGI supplemental library user's guide, version 5 edition. SAS Institute Inc., Gary, North Carolina. Johnson, G. D., and M. S. Boyce. 1990. Feeding trials with insects in the diet of Sage Grouse chicks. Journal of Wildlife Management 54: 89-91. Jones, R. E. 1968. A board to measure cover used by Prairie Grouse. Journal of Wildlife Management 32: 28-31. Klebenow, D. a. 1969. Sage Grouse nesting and brood habitat in Idaho. Journal of Wildlife Management 33: 649-662. Lu-KNCLA, R. A. 1974. A universal grid system for map loca- tions. Wildlife Societv' Bulletin 2: 72. Mech, L. D. 1983. Handbook of animal radio tracking. University of Minnesota Press, Minneapolis. 197 pp. Mosley, J. C, S. C. Bunting, and M. Hironaka. 1986. DeteiTuining range condition from frequency data in mountain meadows of central Idaho. Journal of Range Management 39: 561-565. Musil, D. D. 1989. Movements, survival and habitat use of Sage Grouse translocated into the Sawtooth Valley, Idaho. Unpublished master's thesis. University of Idaho, Moscow. 72 pp. Musil, D. D., J. W. Connelly, and K. P Reese. 1993. Movements, survival, and reproduction of Sage Grouse translocated into central Idaho. Journal of Wildlife Management 57: 8.5-91. Oakleae R. J. 1971. The relationship of Sage Grouse to upland meadows in Nevada. Unpublished master's thesis. University of Nevada, Reno. 73 pp. Patterson, R. L. 1952. The Sage Grouse of Wyoming. Sage Books Inc., Denver, Colorado. 341 pp. Petersen, B. E. 1980. Breeding and nesting ecology of female Sage Grouse in North Park, Colorado. Unpub- lished master's thesis, Colorado State University, Fort Collins. 86 pp. Ratti, J. T, D. L. Mackay, and J. R. Alldredge. 1984. Analysis of Spruce Grouse habitat in north-central Washington. Journal of Wildlife Management 48: 1188-1196. SAS Institue, Inc. 1985. SAS user's guide: statistics. SAS Institute, Inc., Gary, North Carolina. 956 pp. Schoenberg, T. J. 1982. Sage Grouse movement and habitat selection in North Park, Colorado. Unpub- lished master's thesis, Colorado State University, Fort Collins. 86 pp. Thompson, W. K. 1946. Live trapping and transplanting Ringnecked Pheasants and Sage Grouse. Proceedings of the Western Association of State Game and Fish Commissioners 26: 133-137. TUHY, J. S. 1981. Stream bottom community classification for the Sawtooth Valley, Idaho. Unpublished mas- ter's thesis. University of Idaho, Moscow. 178 pp. Wakkinen, W. L. 1990. Nest site characteristics and spring-summer movements of migratorv' Sage Grouse in southeastern Idaho. Unpublished master's thesis, University of Idaho, Moscow. .55 pp. Wallestad, R. O., and D. Pyr\h. 1974. Movement and nesting of Sage Grouse hens in central Montana. Journal of Wildhfe Management 38: 630-633. Received 7 October 1992 Accepted IS January 1994 Great Basin Naturalist 54(3), © 1994, pp. 234-247 VEGETATION ZONES AND SOIL CHARACTERISTICS IN VERNAL POOLS IN THE CHANNELED SCABLAND OF EASTERN WASHINGTON Elizabeth A. Crowe', Alan J. Busacca-, John P Reganold^, and Benjamin A. Zamora'^ Abstract. — Vernal pools are seasonal pools occurring in Mediterranean-tyiJe climates within which grow concentric zones of vegetation. We studied two vernal pools that lie within an Artemisia tridentata/Festuca idahoensis shrub-steppe landscape in the Channeled Scabland of eastern Washington to determine the relationship between vegetation zonation and soil characteristics. Abimdant plant species in the pools include Ehjmus cinereus, Poa scabrella, Lornatiwn graiji, Allium geyeri, Eleocharis palti.sths, Epilobium mintttuin, Myosunis aristatis, Deschainpsia danthonioides, and Psilocarphus oregonus. We surveyed topography, measured plant species fiequency and cover to describe the vegetation zones, and used Sorenson's inde.x of percent similarity to verify our designation of plant zones as communities. In one pool we described soil profiles and sampled soils throughout the growing season according to plant commimities. We analyzed soils for pH; electrical conductivity; sodium, calcium, and magnesium ions; sodium adsorption ratio; particle size; organ- ic carbon; and water matric potential. ANOVA tests of soil characteristics and topography among plant communities showed that only differences in topography are statistically significant. There are, however, trends in particle size, some soil chemical parameters, and soil moisture potential among plant communities along the topographic gradient. Electrical conductivity decreased with increasing diyness of the soil through the spring and summer. Seasonal changes in soil moisture potential showed that shallower soils in the centers of pools are wetter during the wet season and drier during the dr\' season than are deeper soils. These changes in moisture may be the most important influence on vegeta- tion distribution within the vernal pools. Key words: vernal pool, vegetation zones, soil charaeteristics, eastern Washington. Vernal pools occur in grasslands, parklands, and forests where Mediterranean-type rainfall patterns prevail. These biotic systems are geo- graphically widespread and are among the casualties of the widespread modification of natural landscapes. Vernal pools are typically formed in shallow depressions where soils have impermeable hardpans or are underlain by impermeable bedrock. Vernal pools fill with water from wdnter rains (and snowmelt in colder climates) and gradually dr\' during late spring and early summer through evapotranspiration. Vegetation within the pools is different from that of the surrounding landscape and often forms a pattern of more or less concentric zones of different species groupings. These unique natural sites are excellent for studying ecological processes in relatively self-contained ecosystems. Zonal vegetation patterns of vernal pools have attracted many researchers in California, where vernal pools are numerous. Scientists have approached the study of pools by examin- ing aspects of seasonal hydrology and soil phys- ical and chemical characteristics. One theoiy is that seasonal duration of standing water directly affects distribution of plant species according to their ability to germinate and/or grow either under water or within the short- ened growing season after evaporation of the pool (Purer 1939, Lin 1970, Zedler 1987). Other researchers have found trends in soil particle size, available nitrogen and phosphorus, ex- changeable magnesium and sodium, electrical conductivity, pH, and unsaturated soil moisture potential that correlate to position along the gradient from outside to inside the pool (Lathrop 1976, Bauder 1987). Thus, a second theory is that soil chemical and physical fac- tors influence plant distribution. Researchers have taken different views as to whether there are spatially discrete zones of species groupings (plant communities) or whether the distribution of species is continu- ously varialile, with overlapping growth ranges. Some argue the latter case and maintain that 'Area 3 Ecology Program, U.S. Forest Service, WallowaAV'liitiiian National Forest, Box 907, Baker City, (Oregon 97S14. ^Department of Crop and Soil Sciences, Washington State University; Pullman, Washington 99164, ■'Department of Natural Resource Sciences, Washington State University, Pullman, Washington 99164. 234 1994] Vernal Pools of Eastern Washington 235 only temporal groupings of species occur (Purer 1939, Zedler 1987). The alternate view has been supported by several studies whose authors typically have delineated three to five zones, one of which is the surrounding grass- land vegetation (Lin 1970, Kopecko and Lathrop 1975, Macdonald 1976). Pools of the Pacific Northwest, specifically those in central and eastern Oregon and Washington, have been little studied. We chose two vernal pools in eastern Washington to examine the validity of vegetation zonation and to determine a possible relationship between vegetation zones and soil properties. Study Area We chose to study vernal pools in the Marcellus Shioib Steppe Natural Area Preserve (47°14'N, 118°24'E, Sec. 15, T20N, R35E, WBM, in Adams County, Washington) because the site has not been seriously degraded by past grazing (Schuller 1984) and because it has been fenced as a preserve since 1986. The preserve covers approximately 290 ha (Fig. 1) and is at the northeast end of a larger tract of uncultivated scabland, surrounded by wheat fields, that extends west-southwest for approx- imately 6.5 km along Rocky Coulee, part of the Channeled Scabland, an enormous land- scape in eastern Washington formed by repeated cataclysmic glacial outburst floods during the Pleistocene (Fig. 1; Baker 1978). Surrounding the Channeled Scabland and in some areas within it are deposits of Pleisto- cene-age loess (windblown silt) many meters thick (Busacca 1991) that overlie the Miocene Columbia River Basalt. The last glacial floods about 15,000 years ago removed most of the older loess from the Telford-Crab Creek Scabland tract, including the project site, so that the thin cover of loess presently on the site has accumulated, and soils have devel- oped, only since the last floods. Deeper soils of the site, approximately 40-150 cm deep, are on the loess mounds. Soils within the venial pool basins are only approximately 10-30 cm deep. Most vernal pools are part of an inter- rupted channel system running through the study site (Fig. 1). Solitaiy pools and those in the intermittent channel system may have been formed by a combination of cataclysmic flood scour, variable loess deposition, and local slope erosion. Rainfall distribution of the Columbia Plateau is a Mediterranean type. Forty years of average temperatures, precipitation, and evapotran- spiration were used to produce a climate dia- gram (Thornthwaite 1948, NOAA 1988-89), which indicated that November through March are the months of greatest precipitation (approximately 60% of the yearly total) and soil water storage, whereas April through October bring little precipitation and high evapotran- spiration. Methods Average monthly precipitation and temper- ature values recorded at the Ritzville ISE weather station (NOAA 1988-89) were used to compute average monthly evapotranspiration for the months of pool filling and plant growth in the study year (Palmer and Havens 1958). Precipitation values for November 1988 through March 1989, the months typically receiving the majority of precipitation, and April through June of 1989, the months during which most vernal pool species grow vegeta- tiveh' and flower, were compared with average values for these time periods. We chose two vernal pools for this study and called them South Pool and North Pool. South Pool is 109 m long and 70 m wide, and North Pool (Fig. 2) is 57 m long and 34 m wide. A 25-m-wide swath through South Pool (Fig. 3) and all of North Pool were surveyed using a Leitz Set2 Total Station Electronic Distance Meter and SDR22 data collector to produce topographic maps. Within the two pools, groups of plant species fomied concentric zones from the outsides to the centers of the pools. Although zones were uneven in width and some were even absent in some places, the sequence in which they appeared was con- sistent. Six vegetation zones were identified in each pool. Boundaries of the vegetation zones were marked along transect lines across each pool. We marked four sites for plant and soil sampling within each vegetation zone, two on either side of each pool center (Fig. 3). Vegetation (in both pools) and soils (in South Pool only) were sampled on 23 April, 10 May, 24 May, 11 June, 29 June, and 13 July 1989. Repeated sampling was done to monitor maturation of vegetation. When plants within each zone reached their seasonal maturity, we measured frequency and coverage for each 236 Great Basin Naturalist [Volume 54 _ ,.^^ Columbia Rivor Fig. 1. Map of the Marcellus Shrub Steppe Natural Area in eastern Washington state with inset map showing general- ized topography, natural area boundai-y, and principal vernal pools. Triangles are locations of the sagebrush (SBl), wildrye (WRl), and vernal pool (VPl) soil profile sampling sites; North Pool and South Pool are labeled. species in forty 20 x 50-cm plots along con- tours. We used nested frequency plot sizes of 5 X 5 cm, 12.5 x 20 cm, 20 x 25 cm, 20 x 37.5 cm, and 20 x 50 cm and cover classes of 0-5%, 5-25%, 25-50%, 50-75%, 75-95%, and 95-100% (Daubenmire 1970). Frequency and cover values were averaged over the forty plots for analysis. Soil samples were collected in South Pool with a 2-cm-diameter probe. The first sets of soil samples were collected at site stakes. On each successive date, samples were taken approximately 10 cm from the previous sample at the same elevation as the site stake. Sampling depth increments were 2-10 cm, 10-30 cm, 30-60 cm, 60-90 cm, etc., until we reached basalt bedrock (0-2 cm consisted of ash from the 1980 eruption of Mount St. Helens). We measured the matric potential of soil moisture in soil samples collected on all dates using the filter paper equilibration method (Campbell and Gee 1986, Campbell 1988). We produced a moisture characteristic curve specifically for the filter paper (Whatman #42) used for determination of water potential from water content. Organic carbon and parti- cle-size distribution were measured using wet combustion (Nelson and Sommers 1975) and the hydrometer metliod (Gee and Bander 1986), respectively. The only pretreatment used in the particle-size analysis was sodium hexameta- phosphate. We analyzed the 2-10-cm samples collected on 24 May fi-om sites S1-S12. Saturation extracts (using distilled H^O) were collected from all samples from the 23 April and 28 June sampling dates (early and late in the growing season in the pools) and fiom all moiphological description site samples. Electrical conductivity (EC) and pH were measured on the extracts, as well as soluble Na+, Ca+2, and Mg+2 to determine the sodi- um adsorption ratio (SAR). All ion concentra- tions were measured on an atomic absoiption spectrophotometer. To observe moiphological properties of the soils (e.g., thickness of horizons and presence 1994] Vernal Pools of Eastern Washington 237 ^** , • -. • ;i>fe«n. i. "i!'-' ,*> «^« ^H*»« ,,"'>^-:-^ L V'.'^* ,^^. ^^>^3, ■^', . J._ i Fig. 2. Low-angle oblique aerial photograph of North Pool taken in May 1990. View is to the west; the pool is 57 m in length. A. trklentata is the dominant shi"ub in the plant zone 1 area surrounding the pool. of structure that would indicate certain pedo- logical processes), we excavated three soil pits: one in an upland position surrounding the pools (SBl), one in a pool rim (WRl), and one in a pool basin (VPl). To avoid damaging pools of the preserve, we located pits in the adjacent scabland tract (Sec. 16, T20N, R35E; Fig. 1), where soils were similar to those of the study site. Morphological features were described, block descriptions written, and horizons sam- pled for the three profiles. Block descriptions are in Crowe (1990). The three profiles were classified according to Keys to Soil Taxonomy (Soil Survey Staff 1990). We also analyzed horizons for particle size and SAR according to methods discussed above. Vegetation, soil, and topographic data were analyzed using descriptive and inferential sta- tistics. From the vegetation tallies we compared different vegetation zones (e.g., zone 1 vs. zone 2, zone 1 vs. zone 3, etc.) using Sorensen's similarity index for frequency (combining all frequency plot sizes) and for frequency weight- ed by cover (Barbour et al. 1980). Zonal eleva- tions, particle-size classes, and organic matter were statistically compared among zones using a one-way analysis of variance (ANOVA). A repeated-measures ANOVA was used to ana- lyze differences in soil matric potential among the vegetation zones over all five sample dates. The repeated-measures ANOVA was also used to analyze EC, pH, and SAR among all plant zones and between the 23 April and 28 June sampling dates. Results Precipitation Comparison of the long-term average monthly precipitation with monthly totals for November 1988 through March 1989 revealed that total precipitation for that period was almost identical to the long-term average (169 mm during 1988-89 vs. 168 mm on average). April through June precipitation totals (58 mm) were also quite similar to the long-term aver- age (63 mm); however, only 1.52 mm of rain fell in June compared to an average of 23 mm, and temperature and evapotranspiration were slightly higher than average. Matric potential values measured during April and May thus were probably indicative of the moisture nor- mally available to vernal pool species at that time of year, and those measured in June were 238 Great Basin Naturalist [Volume 54 Fig. 3. Topographic cross section through South Pool showing selected soil sampling locations, plant zones, and soil profile moiphology. Note ca lOX vertical exaggeration. Shaded areas indicate plant zones. Zones are numbered and cor- respond to te.\t. S2-S24 are selected soil sampling locations. probably lower than normal. The pools proba- bly dried faster than they do in some years, which would shorten life cycles of annuals that normally flower during June or early July. Vegetation We identified six vegetation zones (zones 1 through 6) in South Pool (Fig. 3), five similar zones (zones 1, 3, 4, 5, and 6) in North Pool, and an additional zone (zone 7) of rush (£. palustris) that occurs only in North Pool. Total species cover and frequency in the 25% cover class for all zones in the South and North pools are shown in Table 1. Zone 1 was an example of a shrub-steppe community as described by Daubenmire (1970). It was dom- inated by F. idahoensis with a scattered shrub overstory of A. tridentata. More abundant forbs included Fleet riciis maeroeera. Flantago patagonica, and Draha verna. Shrub height was about 70 cm and herb height about 40 cm. Zone 2 was dominated by E. cinereus, with a shrub overstory of Chysothamnus naiiseosus and C. vicidiflorus. The more abundant forbs included Senecio interrigemiis, F. macrocera, D. verna, and AchUIea nullefoliiun. Average height of vegetation in this community was about the same as in zone 1. Zones 3-7 had no shrub overstory. In zone 3 Foa seahrella (40 cm in height) dominated, with Lomatiwn grayi, A. geyeri, and Montia linearis scattered throughout. Deschampsia dantho- nioides had very high cover in zones 4, 5, and 7. Other forbs and graminoids in zones 4 and 5 were E. palustris, Agrostis diegoensis, and Agoseris lieterophylla. Forbs that distinguished zone 4 were A. geyeri and E. minutum; those that distinguished zone 5 were Navarettia intertexta, Grindelia nana, and Myosuris arista- tis. Zone 7 had an abundance of E. palustris and well-distributed M. aristatus, Alopechurus genieulatus, and E. minutum. Vegetation 1994] Vernal Pools of Eastern Washington 239 height in these zones averaged 30 cm. Zone 6 was dominated by the low-growing annual forbs N. intertexta and Plagiobothrys scoideri (5-10 cm). Other forbs included Psilocarphus oregonus and G. nana. We calculated similarity indices fi-om our comparisons of plant communities of different zones within each of the tsvo pools (Table 2). Results from comparisons using absolute fre- quency of species alone resulted in higher similarity indices between zones, whereas the addition of total canopy cover of species re- duced most indices substantially. In the latter comparison only the comparison of zones 4 and 5 in South Pool and zones 5 and 7 in North Pool produced similarity indices greater than 50%. Species that were fairly well established (i.e., they have high cover and/or frequency percentages) in two or more zones were most- ly annual forbs and graminoids: D. verna, A. heterophylla, E. minutum, D. danthonioides, A. diegoensis, N. intertexta, and M. aristatus. In addition, three perennial herbs, E. palustris, G. nana, and A. geyeri, had a strong presence in more than one zone. Species that were noticeably unique to one zone were F. ida- hoensis and A. tridentata (zone 1), E. cinereus and Chrysothamnus spp. (zone 2), P. scabrella (zone 3), and Boisduvalia stricta (zone 6). Soils Microtopography, soil sampling points, and soil moiphology are depicted for South Pool in Figure 3. An ANOVA of zonal elevation means demonstrated strongly significant differences among zones in each pool (P < .0001). Soil profiles examined in the soil profile pits off the Marcellus site confirmed moipho- logical characteristics of the soils surrounding (SBl), on the margins of (WRl), and within (VPl) vernal pools of the area, including those of North and South Pool (full soil profile descriptions in Crowe 1990). We produced a cross section showing the soil morphology of South Pool (Fig. 3) based on features seen in the offsite soil profile pits combined with soil properties and depths to bedrock measured during the repeated soil sampling of South Pool. SBl was typical of the deeper soils in the shiTib-steppe zone surrounding the pools. It is classified as a Xerollic Camborthid. VPl was typical of the shallow, stony soils within the pools and is classified as a Lithic Camborthid. The two soils differed primarily in that VPl was shallower to bedrock (33 cm in VPl, 116 cm in SBl). Neither of these soils had strongly developed soil profile features. They consisted principally of a dark, organic matter-rich mol- lic epipedon or topsoil horizon typical of steppe soils, and a blocky brown cambic sub- soil horizon. WRl was typical of the pool rim or margin landscape position that supported the wildrye zone (zone 2) . It is classified as a Lithic Natrix- eroll and is distinguished by its natric horizon (a clay-enriched or argillic horizon caused by a high exchangeable sodium percentage), shal- low depth to bedrock, and mollic epipedon. The SAR (a measure of the dominance of sodi- um on the exchange complex of clay colloids) of the natric horizon of WRl was 9.7, which is less than the value of 13, i.e., the lower limit set for the definition of the natric horizon (Soil Survey Staff 1990); however, other features typical of the influence of high exchangeable sodium were present, including columnar structure in the natric horizon, dark organic colloid stains on these columns, and an overly- ing eluvial horizon. Clay content of the natric horizon of WRl was 17.0% compared to 11.0% in the overlying eluvial horizon. ANOVAs performed on soil physical and chemical properties among zones were not statistically significant, but there are recogniz- able trends in some properties that may have ecological significance. Sand, silt, and clay con- tents differed with respect to South Pool topo- graphic positions (Fig. 4). Sand percentage was greater and silt percentage was less in zone 3 than in other zones. In zone 3, at the bottom of the pool "rim," erosion and deposition may have caused a winnowing of fines from and an accumulation of sand in the soils. Clay content increased modestly from the outer zone to the center of South Pool. Organic carbon values were higher in tlie plant zones dominated by A. tridentata, E. cinereus, and P. scabrella (zones 1, 2, and 3) than in the remaining zones of South Pool (Fig. 4). Soil-water matric potentials were similar, zone by zone, in the 2-10-cm and 10-30-cm depth increments on the first sampling date, 24 May (Fig. 5). On successive dates the 2-10- cm increment dried more than did the 10—30- cm increment as plants extracted soil moisture from the near-surface zone and soil moisture evaporated from the soil surface. On 11 June in 240 Great Basin Naturalist [Volume 54 ;,. I ' — 1/ " N 5 Z CD 1; 1 Z CC 1; ■^ o o N C/5 lO 1; 5 1 Z IC JZ Zj ■^ o o N vA3 §7 N ^ t; »0 lO CD CO Csi CO in 00 d -H o ^ o in d 2 ^ si CD ^ <>J O 0; "t! in si in O N o 05 CO Z d CO in in in p p t^ (>i <>j p in -f 1—1 in -H CO d d ci d d> o o in o o in cj ^ iij c4 ^ d ci d in Jo -t< ^ CD f^ d o\ d d CO d .~ ^ cj 3 2 5 ff c S *= ~ .s 3 .SP S ~ '^ Cj ^ ^ o o --■■ Zi ll^SSfS O 1^ C2 S Q Q ^ c*: ■^ < ^ uj CJ^ 1^ 5^ Si ^ CO ^ in l>; ;o — 1 CO c4 CO CO a5 05 ci d o-i ^^ in -^ o S lo 5 5 -J SJ '-: o "-; o o in So -: 05 <^ <^ <^ iX o _j; — I o ^ (N CD ^ CO m o in m ^ o o oi d <>i p-i i2 d d -r CO o fM -H _; o o in 2 fU in in in p p p oj oi d d in CO 5^ CD t^ eg in CO cc in in — : oj -r CO -H »n c 5 o o o o d> d i6 d CO -* in CD CO CO CO 05 ^ Tfi CO CO in 5^ CD ^ O t^ o ° == in in yj :t «j ^ 2 oj oi 05 oi -r ^ ^ o ^H c (M c _^ _; P in '^ o o 2 c s w = ~ ^ ~ ^ a. •2 ''-^ (i, (i, < Q ' - _2 - S ~ -Is [^;§ a: ;^ ^ s .2 3 §b i^ 3 .* =1 5 B g ^ ^ s: .3 cS cS Q cs 5= cr, .2 5 .2 s ti^ I < s CJ 3 o t>^ s ^ o ? i^ >- ^ o 5 CQ Q tli -< U CQ 242 Great Basin Naturalist [Volume 54 Table 2. Similarity index comparison of plant zones. Similarit) index (frecjnencN' onK) South Pool Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6 Zone 1 *** 57.9 42.1 25.0 26.3 7.0 Siniilaritv Zone 2 26.6 *** 42.1 35.0 21.1 0.0 index Zone 3 11.9 12.2 *** 65.0 68.4 41.4 (weighted by Zone 4 6.9 8.4 19.5 *** 80.0 51.6 cover) Zone 5 2.7 2.3 9.8 66.4 *** 62.1 Zone 6 0.3 0.0 6.9 Similarit\- index 11.3 (frequency only) 35.5 *** North Pool Zone 1 Zone 3 Zone 4 Zone 5 Zone 6 Zone 7 Zone 1 *** 50.0 34.3 28.6 11.1 21.6 Similarity Zone 3 15.6 *** 51.9 47.1 21.4 27.6 index Zone 4 4.4 36.4 *** 62.1 43.5 50.0 (weighted by Zone 5 2.6 11.5 46.2 *** 66.7 83.9 cover) Zone 6 0.6 0.4 9.1 23.2 *** 72.0 Zone 7 1.4 5.9 17.9 60.7 31.1 *** zone 2, for example, the average matric poten- tial was -56 MPa in the 2-10-cm increment, -27 MPa in the 10-30-cm increment, and -2 MPa in the 30-60-cm increment. On 23 April, when soils were moist, ECs were consistently higher in all plant zones and in both depth increments than they were on 28 June when soils had dried considerably (Fig. 6). Zone 2, the pool rim, had the highest ECs of any plant zone in the 10-30-cm incre- ment on both the 23 April and 28 June sam- pling dates (Fig. 6). All mean ECs of saturation extracts, however, were below 2 dS/m, the minimum value for designation of a saline soil (Bohn et al. 1985). In the 2-18-cm soil incre- ment of profile pit VPl and the 30-48-cm increment of profile pit WRl, EC values were 2.20 and 2.62, respectively, indicating the potential for at least modest salinity in soils of the vernal pools. The range of mean pH values of the 2-10- cm samples for each zone increased from 23 April (6.2 to 7.5) to 28 June (7.7 to 8.1; Fig. 6). pH values of the 10-30-cm, 30-60-cm, and 60-90-cm depth increments in zone 2 aver- aged 8.4, 9.1, and 9.3 over both dates, respec- tively. pH values of the WRl soil profile pit, which has natric features, ranged from 7.3 to 7.9 among its horizons. Sodium adsorption ratios (SAR) in soils of zone 2 of South Pool met the definition of sodic soil. Average SARs were 15.5 and 14.0 in the 10-30-cm depth increment on 23 April and 28 June, respectively (Fig. 6); 11.8 and 27.2 in the 30-60-cm increments on the earlier and later dates; and 27.4 in the 60-90-cm increment on 28 June. High SAR values in zone 2 of South Pool were in concert with moderately high SAR values of 9.7 in the 26-30-cm increment and 9.5 in the 30-48-cm increment of WRl. The SAR in the 2-10-cm increment of zone 3 was slightly higher than in other zones on 28 June. Discussion Our study examined vegetation patterns and soil characteristics of vernal pools on the Marcellus site. Zedler (1987) suggested that the cycle of regional weather patterns is reflected in patterns of species germination and distribution in vernal pools. Monthly averages of precipitation, temperature, and evapotranspiration during our study season were fairly typical of the approximately 100-year climatic record in the Ritzville area; therefore, although we do not have data for many seasons, our data represent the vegetation structure and composition as expressed during a typical annual climatic patteiTi. Cover-weighted similarity indices with two exceptions support the vegetation zone divi- sions we made at the outset of the study. It is not surprising that similarity indices based on frequency alone are inconclusive in that there were many species present in several zones. Cover values are a much better indicator of the strength of that presence. For example, although D. clanfhonioides had a high cover value in three different zones, each zone had a unique combination of co-occurring species: 1994] Vernal Pools of Eastern Washington 243 Co 1.5 3 4 Vegetation Zones Fig. 4. Particle size and organic carbon content of the 2-10-cm depth of soils by vegetation zone in South Pool. In zone 4, A. geyeri, A diegoensis, and E. min- utum had highest cover vakies after D. dantho- nioides. In zone 5 those niches were filled by N. intertexta and £. palustris, and in zone 7 by E. palustris and M. aristatus. Higher similarity indices between zone 4 and zone 5 (Table 2) reflect their position in the transition between drier upland-type com- munities and the pool basin area that contains standing water and/or has saturated soils more consistently and for a longer time period. We feel that species unique to each of these zones are ubiquitous and abundant enough to classi- fy them as separate communities. We found a greater ratio of annual to peren- nial species in wetter zones (zones 4-7) of the pools than in drier ones (zones 1-3). This ten- dency was also found in vernal pools in California (Holland and Jain 1977) and in a study of a seasonally flooded river marsh in Zimbabwe in which vegetation zonation was also reported (Cole 1973). It is difficult for perennials to withstand large changes in microenvironmental factors throughout the year, whereas the short life cycles of annual species may be completed within only one set of edaphic conditions. Annuals also produce abundant seeds, thus ensuring some survival due to variation in adaptability and the oppor- tunity to delay germination until conditions are favorable. Morphology and chemistry of the soils allow us to reconstruct the genesis of the pool sys- tem soils. Natric horizons have been docu- mented as occurring in several situations in loessial soils across parts of the Columbia Plateau region where the mean annual precip- itation is low to moderate (Peterson 1961). One specific situation is on the flanks of low mounds in areas of mound-and- swale micro- relief where shallow soils are underlain by an impervious hardpan or bedrock. Calcium ion is presumed to become tied up as precipitated CaC03 in the calcic horizons of these soils over time. Apparently, water has moved later- ally down the slope gradient within the mound soil during its genesis, transporting sodium ions and gradually concentrating them 244 Great Basin Naturalist [Volume 54 (0 (1 CL 'Z S >-^ CO ^_ s (0 *- ^^ C o 0) en o Q. -10000 -1000 ■ Zonel ! Zone 2 i Zone 3 D Zone 4 Zones Zone 6 2-10cm -10000 ? -1000 •z ^ -100 CO — S .2 -10 i«-« = c O 0) -1 ^ o Q. -0.1 10-30cm -0.01 April 23 May 10 May 28 June 11 June 28 July 13 Sampling Dates Fig. 5. Soil inatrie potential at 2-10-cni and 10-30-cni depth intei"vals by vegetation zone and sampling date in South Pool. on the exchange complex of soil clays where the slope levels. Exchangeable sodium must accumulate to a critical level to effect the dis- persion of clay colloids and organic matter, which in turn leads to the development of the typical natric soil moiphology with a pale elu- vial horizon overlying the clay-enriched, columnar-structured natric horizon. We ob- sei"ved a large variation in degree of expression of this morphology in pool-rim soils across the Marcellus site. A natric horizon can have several adverse effects on the growth of plants. (1) When satu- rated, sodic soil horizons become dispersed and disaggregated, thereby clogging pores and preventing the flow of oxygen to plant roots. (2) When they dry, natric soils can form dense surface crusts that can prevent seedling emer- gence. (3) The availability of calcium, magne- sium, and potassium can decrease in the soil due to preferential replacement of these ions by sodium on the exchange sites of clays and organic matter. (4) Sodium salts can create osmotic stress or be toxic to plants by interfer- ing with their phvsiological processes (Black 1968). Soil in zone 2 on the rim of the pools was not saturated during the part of the year when plants are physiologically most active, and the concentration of exchangeable calcium and magnesium was actually higher in zone 2 than in zone 1 (perhaps due to a higher cation ex- change capacity in zone 2 soil). Osmotic stress and sodiiun salt toxicit)' appear to be the most likely conditions excluding zone 1 plants from zone 2, especially A. tridentata, whose roots 1994] Vernal Pools of Eastern Washington 245 23 April ^ 2-10 cm 10-30 cm 28 June C3 IIDn 2-10 cm 10-30 cm 10 Vegetation Zones Fig. 6. Electrical conductivity (EC), pH, and sodium adsoiption ratio (SAR) at 2-10-cni and 10-30-cm depth intervals by vegetation zone and sampling date in South Pool. would penetrate to the most saline and sodic soil horizons. These conditions favor zone 2 plants such as E. cinereus. Choudhuri (1968) tested effects of soil salinity on E. cinereus and A. tridentata and found that E. cinereus was very tolerant of sodium salts and that A. tri- dentata was not. Surface crusts, high bulk density, and potassium deficiencies may also adversely affect zone 1 plants. Although none of the ECs in South Pool were greater than 2 dS/m, soil salinity may have subtler influences on plant distribution. The 2 dS/m salinity limit is based on studies of crop plants, which generally are not native to the area in which they are grown. Limits of tolerance of native plants, such as E. cinereus as mentioned above, can be higher than in crop plants. Unfortunately, no specific studies exist of salinity tolerance of the species or genera found in vernal pool basins, either in California or Washington. Choudhuri (1968) found that all species he tested had a certain degree of self- adaptive capacity to increase their tolerance to salinity if the increase was gradual. Because plant tolerances to salinity can vary with envi- ronment (Levitt 1980), the interaction of 246 Great Basin Naturalist [Volume 54 slightly different salinity levels with changing moisture conditions may produce varying responses among the pool species. Some species will take up soluble ions through the cell walls of their roots, but the ions either will not pass through the cell membranes or will be stored inside cell vacuoles and thus not interfere with physiological processes in the plant (Levitt 1980). The decrease in soil ECs from the 23 April sampling date to the 28 June sampling date may be the result of plants pref- erentially taking up soluble ions to decrease their water potentials as soil water potential decreased. Increased pH values from the first to the last sampling dates may indicate precipitation of alkaline salts. Again, specific responses of plants to this phenomenon can vaiy, and con- trolled studies of individual species would be necessaiy to make any further conclusions. Our particle-size results coincide with those of California researchers (Lathrop 1976, Bander 1987), who also found that finer particle sizes increased and sand fractions decreased from outside to the interior of the pool basins. Slightly more clay occurred in zone 4 soils than in zones 5 and 6 (the center of South Pool; Fig. 4), which is similar to the pattern in vernal pools on Kearny Mesa in California (Zedler 1987). A higher percentage of silt- sized particles can give a soil greater water- holding capacity under unsaturated condi- tions. The lack of a statistically significant dif- ference in soil moisture potentials from the first to the last sample dates is difficult to understand. Given the generally large spatial variability in properties such as soil moisture potential, analysis of a greater number of sam- ples would perhaps result in the finding of sig- nificant differences between dates and might also show significant differences among zones on particular sampling dates. It is interesting to note that some species were still sui-viving and in some cases photosynthesizing in soil matric potentials far less than -15MPa, com- monly considered the permanent wilting point for plants. This seems to indicate an adaptabil- ity of vernal pool species to seemingly unfa- vorable moisture conditions. Zedler (1987) stated that duration of stand- ing water is the crucial factor in structuring plant distribution of vernal pools. The highly significant statistical difference we found between topographic elevation of the vegeta- tion zones should be most closely related to the location and extent of above- and below- ground free water. We believe that soil mois- ture potential under unsaturated conditions probably also has a large effect on plant growth in the various vegetation zones. For example, different vernal pool species may have different types of root systems to take advantage of moisture in different parts of the soil profile, as was found in a study of several eastern Washington grasses (Harris and Wilson 1970). Also, maximum physiological activity can occur at different water potentials for different species (Wieland and Bazazz 1975). The duration of free-standing water and changes in unsaturated soil water potential through the season are probably both impor- tant to plant distribution in the pools. Vernal pools can help us understand the physiological ecology of self-contained, water- controlled terrestrial ecosystems. We need to learn more about specific ecological processes in the vernal pool system to understand larger functions described so far Futm-e studies might include (1) examination of rooting systems and their relationship to water use and availability, (2) determination of the minimum physiologi- cally detrimental salinity (especially of various sodium salts) and optimal pH levels for indi- vidual species, and (3) tolerance of individual species to vaiying durations of standing water and levels of unsaturated soil moisture poten- tials. Acknowledgments We thank The Nature Conservancy for allowing us to conduct this study on the Marcellus Shrub Steppe Natural Area. We also extend thanks to Gaylon S. Campbell for his help in measuring the matric potential of our soil samples and Jim Harsh for helpful dis- cussion about saturation extracts and sodium adsorption ratio. R. Alan Black and Jonathan Halvorson reviewed an early draft of this man- uscript and offered many useful suggestions. This manuscript forms contribution number 9407-14 of the Department of Crop and Soil Sciences, Washington State University. Literature Cited Bakkr, V. R. 1978. Quaternan,- geolog)' of the Channeled Scabland and adjacent areas. Pages 17-35 in V R. Baker iind D. Nuniniedahl, eds., The Channeled Scab- 1994] Vernal Pools of Eastern Washington 247 land — a guide to the geonioiphologv' of the ColumlMa Basin, Wasliington. Comparative Planetary Geology Field Conference, 5-8 June 1978. NASA Planetary Geology Program. Barbour, M. G., J. H. Burk, and W. D. Pitts. 1980. Terrestrial plant ecology. The Benjamin/Cummings Publishing Company, Inc., Menlo Park, California. 604 pp. Bauder, E. T 1987. Species assortment along a small- scale gradient in San Diego vernal pools. Unpublished dissertation, University' of California, Davis. 297 pp. Black, C. A. 1968. Soil-plant relationships. Robert E. Krieger Publishing Company, Inc., Malabar, Florida. 792 pp. BoHN, H. L., B. L. McNeal, and G. A. O'Conner. 1985. Soil chemistry. 2nd edition. John Wiley and Sons, New York. 329 pp. BusACCA, Alan J. 1991. Loess deposits and soils of the Palouse and vicinit>'. Pages 216-228 in V. R. Baker et al.. The Columbia Plateau, Chapter 8 in R. B. Morrison ed.. Quaternary non-glacial geology of the United States. Geology of North America. Vol. K-2. Geological Society of America. Campbell, G. S. 1988. Soil water potential measurement: an overview. Irrigation Science 9: 265—273. Campbell, G. S., and G. W. Gee. 1986. Water potential: miscellaneous methods. Pages 619-633 ;/i A. Klute, ed.. Methods of soil analysis, part 1. Physical and mineralogical methods. 2nd edition. American Society of Agronomy-Soil Science Society of America, Inc., Madison, Wisconsin. Choudhuri, G. N. 1968. Effect of soil salinity on genni- nation and survival of some steppe plants in Wash- ington. Ecology 49: 465-471. Cole, N. H. A. 1973. Soil conditions, zonation and species diversity in a seasonally flooded tropical grass-herb swamp in Sierra Leone. Journal of Ecology 61: 831-847. Crowe, E. A. 1990. Soil and vegetation relationships of vernal pools in the Channeled Scabland of eastern Washington. Unpublished thesis, Washington State University, Pullman. 100 pp. Daubenmire, R. F 1970. Steppe vegetation of Washington. Washington Agricultural Experiment Station Technical Bulletin 63. 131 pp. Gee, G. W, and J. W. Bauder. 1986. Particle-size analy- sis. Pages 383-411 in A. Klute, ed.. Methods of soil analysis, part 1. Physical and mineralogical methods. 2nd edition. American Society of Agronomy-Soil Science Society of America, Inc., Madison, Wisconsin. Harris, G. A., and A. M. Wilson. 1970. Competition for moisture among seedlings of annual and perennial grasses as influenced by root elongation at low tem- perature. Ecology 51; 530-.534. Holl\nd, R. F, and S. K. Jain. 1977. Vernal pools. Pages 515-533 in M. G. Barbour and J. Major, eds.. Terrestrial vegetation of California. John Wiley and Sons, New York. KOPECKO, K. R J., AND E. W Lathrop. 1975. Vegetation zonatioTi in a venial marsh on the Santa Rosa Plateau of Riverside Count)', California. Alsio 8: 281-288. Lathrop, E. W 1976. Vernal pools of the Santa Rosa Plateau, Riverside County, California. Pages 22-27 in S. Jain, ed.. Vernal pools; their ecology and con- servation. Publication No. 9, Institute of Ecology, University of California, Davis. Levitt, J. 1980. Responses of plants to environmental stresses. Vol. II. Academic Press, New York. 607 pp. Lin, J. W Y. 1970. Floristics and plant succession in ver- nal pools. Unpublished thesis, San Francisco State University, San Francisco, California. 99 pp. Macdonald, R. 1976. Vegetation of the Phoenix Park ver- nal pools on the American River bluffs, Sacramento County, California. Pages 69-76 in S. Jain, ed., Vernal pools: their ecology and conservation. Publication No. 9, Institute of Ecology, University of California, Davis. Nelson, D. W, and L. E. Sommers. 1975. A rapid and accurate procedure for estimation of organic carbon in soil. Proceedings of the Indiana Academy of Science 84; 456-462. NOAA (National Oceanographic and Atmospheric Administration). 1988-89. Climate data, Washington 92(9-12) and 93(1-8). Palmer, W. C, and A. V. Havens. 1958. A graphical method for determining evapotranspiration by the Thomthwaite method. Monthly Weather Review 86: 123-128. Peterson, F F 1961. Solodized solonetz soils occurring on the uplands of the Palouse loess. Lhipublished dis- sertation, Washington State Universit); Pullman. 278 pp. Purer, E. A. 1939. Ecological study of vernal pools, San Diego County Ecology 20: 217-229. SCHULLER, R. 1984. Flora, vegetation and management of Marcellus Shrub Steppe Preserve, Adams County, Washington. Unpublished report for The Nature Conservancy, Olympia, Washington. Soil Survey Stafe 1990. Keys to soil taxonomy. 4th edi- tion. SMSS Technical Monograph No. 6. Virginia Polytechnic Institute, Blacksburg. 422 pp. Thornthwaite, C. W 1948. An approach toward a ratio- nal classification of climate. Geographical Review 38: 55-94. Wieland, N. K., and F a. Bazazz. 1975. Physiological ecology of three codominant successional annuals. Ecology 56; 681-688. Zedler, R H. 1987. The ecology of southern California vernal pools; a community profile. U.S. Fish Wildlife Sei-vice Biolog>' Report 85(7.11). 136 pp. Received 8 October 1993 Accepted 10 January 1994 Great Basin Naturalist 54(3), © 1994, pp. 248-255 GOLDEN EAGLE {AQUILA CHRYSAETOS) POPULATION ECOLOGY IN EASTERN UTAH J. William Bates ^ and Miles O. Moretti^ Abstract. — Golden Eagle population ecology was studied from 1982 to 1992 in eastern Utah where over 47% of 233 teiritories monitored during the study period were active. Golden Eagle use of four habitat types was compared. Talus territories were used less often than expected; valley, aspen-conifer, and pinyon-juniper temtories were used as expect- ed. Number of young produced per territory averaged 0.612 and was correlated with rabbit abundance. Observations on the impacts of coal mining at two locations are discussed. Key words: Aquila chrysaetos. Golden Eagle, population, habitat use. prey relationships. The Golden Eagle {Aquila chrysaetos) is a year-round resident of eastern Utah but is most common during the nesting season. Golden Eagle nests in the area are found at elevations of 1546 m (5070 ft) to 3000 m (9800 ft). iMost are located on cliffs, while others are located in cottonwood {Populus freinontii) and Douglas fir {Pseudotsuga menziesii) trees. Golden Eagle eyries are found in riparian areas, shadscale-clay hills, pinyon-juniper hills with sandstone cliffs, steep talus slopes with large cliffs, and aspen-conifer areas in trees or on smaller cliffs (Jensen and Borchert 1981). Many nests are located on prominent escarpments found in the Castle Valley area. These escarpments are part of the Castle Gate and Hiawatha formations, which are rich in coal deposits (McGregor 1985). Coal mining is a major industry in the area, and mining activi- ties have the potential to impact nesting Golden Eagles. As a result, federal land-management agencies have required mining companies to monitor eagle nests on their properties. The primary objective of this project was to monitor Golden Eagle and eagle prey popula- tions in a variety of habitats in eastern Utah. The secondary objective was to summarize data collected by mining companies required to monitor raptor nests. Study Area Golden Eagle nests monitored during this study were located in Carbon and Emery counties in eastern Utah (Fig. 1). The study area includes territories from Scofield and Emma Park south to Quitchipah Creek, and from Horse Canyon on the east to Huntington Canyon on the west. Elevations in the study area range from 1546 m (5070 ft) to 3000 m (9800 ft). Vegetative zones include riparian, saltbush {Atriplex sp.), sagebrush {Artemisia sp.), pinyon-juniper {Pinus edulis, Jurnperus osteospenna), and mixed aspen-conifer. The study area was classified into four habi- tat types that typify eagle use in the area: (1) valley territories, located on saltbush flats, on clay hills, or along riparian areas, with nests in cottonwood trees, on conglomerate pinnacles, or on clay ledges; (2) pinyon-juniper territories, with nests found on sandstone cliffs; (3) talus territories, where eyries were located on thick sandstone cliffs; and (4) aspen-conifer territo- ries, where one nest was located in a Douglas fir and all others were on sandstone cliffs. Methods The U.S. Fish and W^ildlife Service, in cooperation with the Utah Division of Wildlife Resources (UDWR), conducted extensive heli- copter surveys in 1981 and 1982 to locate Golden Eagle nests in the area. Over 250 nests were located and monitored during these sur- veys. Beginning in 1986 several mining com- panies were required to monitor approximate- ly 26 territories within a 10-mile radius of the areas affected by mining to assess the impacts of coal mining on the local Golden Eagle pop- ulation. In 1990 the UDWR began monitoring 'Utali Division of Wildlife Resources, 455 West Railroad A\enue. Price, Utali 84501. 248 1994] Golden Eagle Population Ecology 249 *Fairview SANPETE A N 0 DUCHESNE I Emma Park Scofield Res. * Helper * Price CARBON *Huntington * Castle Dale * Perron Horse/ Canyon EMERY uitchlpah Creek UTAH * General location of study area Fig. 1. Map showing Golden Eagle study area. an additional 13 territories beyond the 10-mile radius impact area. A total of 39 territories were monitored in 1992. A Bell Jet Ranger helicopter with a pilot and two observers was used to check all known nests in the area affected by mining. Previously unknown nests occasionally were found and recorded during these flights. Normally, the helicopter was able to fly close enough to allow direct observation of the nest. Adult eagles usually would remain in the nest as the helicopter passed, although occasionally they flushed. Adult eagles also left the nest area when they were viewed from the ground. Eyries in nonimpacted areas were observed from a distance to determine whether eagles were present. If adult eagles, greenery, or fiesh mutes were present, the nest site was classi- fied as occupied. If young or eggs were pres- ent, it was classified as active. The nest was classified as inactive if no sign of eagle use was present. If eggs were present but failed to hatch, or if all nestlings were observed to die before fledging, it was classified as failed. Due to commitments to other projects, we had in- sufficient time to return to each territory to determine the number of successfully fledged young. Therefore, these data cannot be inter- preted to indicate Golden Eagle recruitment or nesting success. Rabbit populations were monitored in the area to determine prey base trends during 1986-91. Eleven 5-mile transects were com- pleted each year in the study area. Transects were conducted just after dusk or just before dawn by mounting a spotlight on a vehicle and recording all rabbits seen on one side of the road. Transects were completed in desert shrub, pinyon-juniper, sagebrush, and aspen- conifer habitat types. Data were analyzed using descriptive sta- tistics, contingency table analysis, and linear regression in the Number Cruncher Statistical System (Hintze 1990). The Bonferroni Z test (Neu et al. 1974) was used to analyze utiliza- tion data. Results Habitat Use Of 233 Golden Eagle territories checked from 1981 to 1992 (average/year = 26), 109 (47%) were active and produced young. Almost 78% of the territories were occupied. The year with the most active territories (56%) was 1990 (Fig. 2). In that same year 94% of monitored territories were occupied. The year with the fewest known active territories (33%) was 1988. 250 Great Basin Naturalist [Volume 54 1981 1982 1986 1987 1988 ^989 1990 1991 1992 n=l8 n=14 n=19 n=24 n=15 n=24 n=36 n=38 n=39 ^ Active :::::;^ Uccupjed InactJve WM Failed Fig. 2. Status of Golden Eagle territories in eastern Utah. Of 185 territories checked in consecutive years, over 28% (52) were active. Five territo- ries were observed to produce young for 3 consecutive years. One territory was active 4 consecutive years, while another produced young 5 consecutive years. One territory failed 3 years in a row. Generally, eagles use differ- ent nest sites within the same territoiy in con- secutive years, but in our study eagles used the same nest as the previous year 11 times (21%). Golden Eagle nesting activity was analyzed by habitat type. A significant difference was found between the four habitat types (chi- square = 20.6, P < .015). The number of active territories in each habitat type was com- pared to the expected number active using the Bonferroni Z statistic (Neu et at. 1974). Talus territories were active less frequently than expected, accounting for almost 37% of avail- able habitat, but only 24% of active territories (Table 1, Fig. 3). The number of active nests in valley, pinyon-juniper, and aspen-conifer terri- tories did not differ significantly from the number expected. Talus eyries had their highest incidence of use in 1982, 1987, 1990, and 1991, with over 40% of territories active. In 1989 only one of nine talus territories was active. Over 75% were active in 1986, 1987, 1989, 1991, and 1992. Six of nine were active in 1990, seven of nine in 1991 (although two eyries failed), and seven of nine in 1992. Two or fewer valley ter- ritories were checked in 1981, 1982, and 1988. Over 57% of aspen-conifer territories were active each year, with the exception of 1982, 1986, and 1992, when only one of three, one of four, and tliree of nine, respectively, were active. Nesting was relatively late in 1991 because of an unusually wet and cold spring; precipita- tion was 4.34 cm (1.71 in) greater than normal and temperatures were 1.65 °C (3°F) cooler than the 30-year average at the Hiawatha weather station. Golden Eagles also showed a shift in habitat use in 1991. All known valley tree nests were active (n = 9). Talus territories were used less than expected and were initiat- ed up to 4 weeks later in 1991 than in 1990. In spite of the cool spring, all four known aspen- conifer territories over 2400 m in elevation near Joe s Valley Reservoir were active and began incubation earlier than lower talus ter- ritories and close to the time incubation began at this elevation in previous years. 1994] Golden Eagle Population Ecology 251 Table 1. Active Golden Eagle eyries by habitat type in eastern Utah, 1982-92. Habitat t\pe Sample points Proportion ol habitat Territories acti\e Expected acti\e Prop, of active teiTitories 95% confidence intei-val \^ille> Pin\ on/juniper Talus Aspen/conifer .51 41 85 56 0.219 0.176 0.364 0.241 32 22 26 29 24 19 40 26 0.294 0.202 0.239 0.266 .196 < /; < .392 .116 < 1981 1982 1986 1987 1988 1989 1990 1991 1992 Fig. 4. Average number of young per territory in eastern Utah. 1994] Golden Eagle Population Ecology 253 Linear regression was used to determine if there was a relationship between number of rabbits seen per mile during rabbit transects in 1986-91 and number of eaglets per territory. A weak relationship was found {R^ = .33, P = .24), indicating that part of the variability in Golden Eagle productivity was explained by rabbit population levels. The data indicated a lag effect, with productivity higher the year after rabbit populations increased (Fig. 5). By using linear regression to test this hypothesis, we found a near-significant relationship be- tween number of rabbits the previous year and number of eaglets per territory (R^ = .63, P = .058; Fig. 6). A significant relationship was also found between number of rabbits/mile and number of young produced per active ter- ritory in the same year, indicating higher pro- duction in years when rabbits were more abundant (R2 = .83, P = .01; Fig. 7). These data demonstrate that Golden Eagles produce more young in the same year that rabbit popu- lations increase, but a higher proportion of territories are active the year following an increase in rabbits (Fig. 5). Discussion Number of young produced per territory and proportion of active territories in south- eastern Utah were similar to those of other studies. Phillips et al. (1990) found 0.78 young produced per occupied territory in Montana and Wyoming from 1975 to 1985, compared to 0.82 in this study. They also found 1.46 young produced per successful territory, compared to 1.39 per active eyrie in this study. Results from southeastern Utah are inflated as the Phillips study was based on number of fledged birds and this study recorded only the number present in nests. However, most eaglets in this study were approaching fledging age when observed. Murphy (1975) found 0.69 young fledged per occupied territory in central Utah. Number of eaglets produced was associated with rabbit population densities in the study area. Although other prey, such as white-tailed prairie dogs, are available, correlations with rabbit populations were quite high. High rabbit populations seemed to influence Golden Eagle nesting in two ways. First, num- ber of young produced per active nest was 1986 1987 1988 1989 1990 1991 Eaglets/Territory Rabbits/Mile Fig. 5. Golden Eagle production and rabbit population trends. 254 Great Basin Naturalist [Volume 54 0.8 0.75- o 0.7 0 0.65 (D n (f) 0.6 +-J Q) O) 03 0.55 LU ) . jj ^ y=0.42x+0.40 R**2=0.63 P=0.058 » ^ ^^ ^ ^^ ^ ^^ B 1 1 1 1 1 1 1 0.5 0.45 0.1 0.3 0.5 0.7 0.2 0.4 0.6 Rabbits per Mile the Previous Year Fig. 6. Eaglets per territory as a function of rabbits the previous year. 0.9 0.8 affected by number of rabbits in the area that year; i.e., more eaglets were produced in years with higher rabbit populations. This relation- ship has also been found in other studies (Murphy 1975, Phillips et al. 1990). Second, there appeared to be a lag effect on number of eagles that attempted to nest. There was a sig- nificant correlation between number of young produced per territoiy and number of rabbits the previous year. High rabbit populations may have allowed more pairs in the area to nest, or enticed more eagles into the area, resulting in an increased number of active territories. Use of valley territories increased in years with higher rabbit populations. Golden Eagles may have selected nest location to minimize the energy required to obtain food. In years with higher rabbit populations, eagles may have spent more time hunting in valley loca- tions. The 2 years with the fewest active talus eyries, 1988 and 1989, were years of relatively high rabbit abundance. Eagles possibly avoid- ed talus eyries in years of high rabbit popula- tions because they were too far from an abun- dant food source. In years with fewer cotton- tail and jackrabbits they may have used these territories to take advantage of other prey, such as snowshoe hares or woodrats. Data on mining impacts caused by cliff spaulings are too few to draw empirical con- clusions. However, we offer some observa- tions. When ample suitable habitat is nearby, there appeared to be no net loss in produc- tion. The territory at Star Point was active 2 of 4 years before and after the escarpment fail- ure. Although the pair at Newberry Canyon did not re-nest in the canyon for 3 years after the original nests fell, they may have been using alternate nests of adjoining pairs. The five territories in the area averaged 2.25 pairs active/year before and 3 active/year after the escarpment failure. In consideration of these obsei"vations, we offer several recommendations to protect against loss of birds or territories. First, if spaul- ing can be controlled, it should be done in the nonnesting season. Othei^wise, physically fenc- ing may help prevent loss of nestlings. The two fenced nests were not used; however, the pair built a new nest below a fenced nest on a cliff that was failing. The pair did not attempt to raise young in that nest. Second, there must be ample suitable nesting habitat to allow other nests to be built. In Newbeny Canyon a sheer wall was the result of escaipment failure and may not provide suitable nesting struc- ture. This pair built a new nest 150 m east of a 1994] Golden Eagle Population Ecology 255 0) LU > '■4-> o < CL W O) to LU 1.8 1.7 1.6 1.5^ 1.4 1.3 1.2 1.1 0.1 y=0.84x+0.97 R**2=0.824 P=0.012 0.5 0.4 0.6 Number of Rabbits per Mile 0.7 0.9 0.8 Fig. 7. Eaglets per active eyrie as a function of rabbits, eastern Utah. fallen nest on a ledge that did not fail. Loss of nesting structure could be a consideration in areas with limited cliff habitat where the whole face fails. Acknowledgments We thank Energy West Mining and Cyprus Plateau Mining for providing helicopter flight time to obtain much of these data. We also extend special thanks to V Payne and B. Grimes for their interest and support. J. Bingham of Helicopter Services proved to be a skilled pilot and a valuable observer. Appreciation is also given to E Howe and J. Felice for review- ing this manuscript. Literature Cited Bates, J. W. 1989. Utah furbearer harvest report. Utah Division of Wildlife Resources Publication 89-7. Salt Lake City, Utah. 27 pp. HlNTZE, J. L. 1990. Number cruncher statistical system, version 5.03. Kaysville, Utah. 442 pp. Jensen, E. H., and J. W. Borchert. 1981. Soil survey of Carbon area, Utah. U.S. Geological Survey Publica- tion. 294 pp. McGregor, E. E. 1985. Engineering geology of surficial materials in the Price 30' X 60' quadrangle. Carbon, Duchesne, Wasatch, and Utah counties, Utah. U.S. Geological Survey Open-file Report 85-356. Mitchell, D. L., and J. A. Roberson. 1992. Utah upland game annual report. Utah Division of Wildlife Resources Publication 92-11. Salt Lake City, Utah. 223 pp. Murphy, J. R. 1975. Status of a Golden Eagle population in central Utah. Pages 91-96 !/i J. R. Murphy C. M. White, and B. E. Harrell, eds.. Population status of raptors. Proceedings of the Conference on Raptor Consei^vation Technical Research Report 3. Neu, C. W, C. R Byers, and J. M. Peek. 1974. A tech- nique for analysis of utilization-availability data. Journal of Wildlife Management 38: 541-545. Phillips, R. L., A. H. Wheeler, J. M. Lockhart, T. R McEneaney, and N. C. Forrester. 1990. Nesting ecology of Golden Eagles and other raptors in south- eastern Montana and northern Wyoming. U.S. Fish and Wildlife Service Technical Report 26. Washington, D.C. 13 pp. Received 8 March 1993 Accepted 18 January 1994 Great Basin Naturalist 54(3), © 1994, pp. 256-271 IDENTIFICATION OF PURSHIA SUBINTEGRA (ROSACEAE) Frank W. Reichenhacherl Abstract. — Populations of Purshia in central Arizona are intermediate in some characters between Purshia subinte- gra, an endangered species, and Purshia stanshuriana, the common cliffrose. These intermediates may represent forms derived from a history of hybridization and introgression between the putative parent species. Morphological data were obtained from 216 pressed specimens of P. stibmtegra, P. stanshuriana, and introgressed forms. Over 50 separate dis- criminant function analyses (DFA) and principal components analyses (PCA) were nm on numerous combinations of raw and log-transformed data. The best variable suite, providing the clearest discrimination between groups, used log- transformed data on 15 morphological characters, but DFA post-hoc identifications were 90-100% correct with only 7 characters using raw data. DFA distinguished four separate nodes of variation. Two groups consisting of 122 P. subinte- gra and 29 P. stanslniriana were easily discriminated in DFA and were distinguished in PCA as well. Introgressed forms were consistently identified in two much less well-defined groups of 46 and 19 specimens. Introgressed forms are not intermediate between the two supposed parents in some characters, appearing most similar to P. stanshuriana in most measured characteristics. Principal distinguishing characteristics of the four groups are as follows; P. suhintegra — usual- ly eglandular, has 0-2 leaf lobes and short hypanthia-pedicels; P. stanshuriana — always abundantly glandular, has 4 leaf lobes and short hypanthia-pedicels; the introgressed fonn "Tonto" is usually eglandular, has 4 leaf lobes and long hypan- thia-pedicels; the introgressed fonn "Verde " is usually glandular, has 4 leaf lobes and slightly shorter hypanthia-pedicels. Key words: Purshia subintegra, Purshia stanshuriana, Arizona clijfrose, clijfrose, endangered species, morphometries, introgression, taxonomy. Purshia suhintegra (Kearney) Henrickson (Arizona cliffrose) is protected under federal law as an endangered species (USFWS 1984). For a federally endangered species like P. suhintegra it is important, indeed vital, to know the taxonomic identity of every individual plant in a given population because the protec- tive measures of the Endangered Species Act are available to species (including forms that exhibit characteristics of introgression with other species), but not to their early generation hybrids. Purshia suhintegra is found in four widely scattered locations from northwestern to southeastern Arizona (Table 1, Fig. 1). The first collection was made by Darrow and Benson in 1938 (Kearney 1943, Schaack 1987a) near Burro Creek in Mohave County, Arizona. A second population was documented in a col- lection by Pinkava, Keil, and Lehto in 1968 (Pinkava et al. 1970) almost 300 km from Burro Creek, near Bylas in Craham County, Arizona. Anderson (1986) found a third popu- lation on bluffs overlooking the upper Verde River near the town of Cottonwood (referred to as the Verde Valley area) and reported on Barbara G. Phillips' 1984 discovery of the fourdi locality for P. suhintegra near Horseshoe Dam along the lower Verde River. At the four locations cited above, Purshia suhintegra is restricted to outcrops of Tertiary deposits of limy lacustrine rock formations (Anderson 1986). Soils derived from these ancient lake basin rocks are characterized by low nitrogen and phosphorus levels, which limit, or preclude, typically Sonoran Desert species that are common on nearby sites with soils derived from igneous and metamorphic rocks (Anderson 1986). Purshia subintegra is a species of the northern and eastern perimeter of the Sonoran Desert; all four sites supporting the species are at or below 1000 m elevation. Purshia stanshuriana, the common cliffrose of the Southwest, is not a Sonoran Desert species and consequently is not sympatric with P. suhintegra at three of the four P. suhin- tegra sites known. In the Verde River Valley of eastern Yavapai County, Arizona, the highest- elevation P. suhintegra site, the two species occur in close enough proximity that gene exchange may occur, at least occasionally. Scattered populations and individuals of 'Southwestern Field Biologists, 8230 East Broadway Boulevard, Suite W8, Tucson, Arizona 85710-4002. 256 1994] Identification of Purshia subintegra 257 Table 1. Locations oi Purshia spp. collections sampled in multivariate morphometric analysis; 216 specimens collected for morphological analysis in Arizona. Region and Letter Elev. collection site N County designation (m) Biotic community Substrate Purshia subintegra 1. Bylas 20 Graham A 850 Sonoran desertscrub Tertiary lacustrine 2. Burro Creek 20 Mohave B 790 Sonoran desertscrub Tertiary lacustrine 3. Horseshoe Lake 42 Maricopa C 640 Sonoran desertscrub Tertiary lacustrine 4. Verde Valley 8 Yavapai D 1025 Semidesert grassland Verde Formation 5. Verde Valley 24 Yavapai E 1050 Semidesert grassland Verde Formation 6. Verde Valley 8 Yavapai F 1065 'Verde" Semidesert grassland Verde Formation Verde Valley 7. DKWellRd. 3 Yavapai a 1185 Semidesert grassland Verde Formation 8. Seventeen Tank Rd. 6 Yavapai b 1020 Semidesert grassland Verde Formation 9. Mesa Blanca 3 Yavapai c 1210 Semidesert grassland Verde Formation 10. DKWellRd. 1 Yavapai d 1175 Semidesert grassland Verde Formation 11. DKWellRd. 3 Yavapai e 1155 Semidesert grassland Verde Formation 12. Seventeen Tank Rd. 3 Yavapai f 1035 Semidesert grassland Verde Formation 13. Cottonwood Hwy. 1 Yavapai g 1030 Semidesert grassland Verde Formation 14. Comville 2 Yavapai h 1110 Semidesert grassland Verde Formation 15. Black Mtn.Rd. 6 Yavapai i 1385 Piiion-juniper woodland Supai Formation 16. Cherry Rd. 3 Yavapai J 1070 Semidesert grassland Verde Formation 17. Cherry Rd. 3 Yavapai k 1230 Interior chaparral Verde Formation 18. Cherry Rd. 6 Yavapai 1 1270 Interior chaparral Verde Formation 19. Cherry Rd. 6 Yavapai m 1341 "Tonto" Interior chaparral Verde Formation Verde Valley 20. Camp Verde 2 Yavapai n 990 Interior chaparral Verde Formation South of Globe 21. Dripping Springs Rd. 6 Gila 0 990 Semidesert grassland Limestone? Tonto Basin 22. Pinal Creek 3 Gila P 944 Semidesert grassland Tertiary lacustrine 23. Punkin Center 6 Gila q 725 Sonoran desertscrub Tertiary lacustrine 24. Beeline Hwy. 2 Gila r 1015 Interior chaparral Tertiar}' lacustrine Purshia stansburiana 25. Jerome 16 Yavapai X 1770 Interior chaparral Tertiary basalt 26. Skull Valley 3 Yavapai Y 1270 Semidesert grassland Weathered volcanics 27. Sonoita 10 Santa Cruz Z 1435 Plains grassland Quaternary alluvium Purshia in central Arizona exhibit what appear to be intermediate characteristics between P. subintegra and P. stansburiana. Numerous botanists (Schaack and Morefield 1985, Schaack 1987a, 1987b, Anderson 1986, 1993, Henrickson personal communication 1988) have assumed that intermediate forms arose from past hybridization and subsequent intro- gression with one or both of P. subintegra and P. stansburiana, and that hybridization and introgression may still be occurring in some locations. Throughout this paper the term "introgressed form," recognizing that the origin of the intermediates is still unclear, is applied to forms that are intermediate in some charac- ters between P. subintegra and P. stansburiana. Related Taxa The systematics of Purshia is under investi- gation by Dr. James Henrickson for the upcoming Chihuahuan Desert Flora and the revised Arizona Flora. Henrickson (1986) pub- lished a brief note recombining species previ- ously placed in Cowania to Purshia, a move generally agreed upon by botanists. Thus, the genus Purshia now consists of seven species: Purshia ericifolia (Torr. ex Gray) Henrickson, P. glandulosa (Curran), P. mexicana (D. Don) Henrickson, P. plicata (D. Don in Sweet) Hen- rickson, P. stansburiana (Torr in Stansb.) Hen- rickson, P. subintegra (Kearney) Henrickson, and P. tridentata (Pursh) DC. Although the 258 Great Basin Natur.\list [Volume 54 O P- subintegra V P- stansbur " Introgressed Form V2 Fig. 1. Distribution of sampling sites, species, and introgressed forms. See Figure 2 for an e.xpanded view of Verde Valley. Letters identify sites illustrated in graphs of multivariate analyses. relationship of P. subintegra to P. stanshuriana is the subject of this study, that o{P. subintegra to the moiphologically similar P. ericifolia is of interest as well. All other Purshia taxa have lobed leaves except P. ericifolia. Leaves of P. ericifolia are about 6 mm long, simple, acute, linear, and eglandular The species is restricted to limestone outcrops in the Texas Big Bend region. It has been speculated that P. subintegra may have evolved from some ancient series of crosses and backcrosses involving P. ericifolia and some other Purshia, perhaps P. stansburi- ana (McArthur et al. 1983). Phylogenetic investigation of the whole genus in relation to closely related genera such as Fallugia would be valuable in interpreting P. subintegra and should be the logical next step for future work. Relation to Purshia stansburiana Schaack (1987a, 1987b) suggested that the basionym P. subintegra was based on material of hybrid origin, resulting from past hybridiza- tion of P. stansburiana and a previously unnamed central Arizona Purshia. He conse- quently published a new species name, Purshia pinkavae Schaack, to include a very pure concept of what had been included in P. subintegra, ". . . restricted to late Tertiary cal- careous, lacustrine deposits ca. 16-21 km north- west of Bylas, Graham County, Arizona. Few botanists have adopted Schaack's tax- onomy, but most recognize that variation is present and concede that it may have resulted from some form or degree of past hybridiza- tion or introgression involving P. subintegra and P. stansburiana. McArthur et al. (1983) reported n = 9 and 2n = 18 for P. subintegra and P. stansburiana, respectively. Phillips et al. (1988) used starch gel electrophoresis to investigate isozymes at 14 loci in three populations of Purshia stans- buriana and four populations of P. subintegra. The material used in this study was collected by the author and presei^ved in liquid nitrogen from the same plants that provided dried specimens for this study. Phillips et al. (1988) were unable to discern differential patterns of variability useful in identification of taxonomic groups; between-groups similarities ranged from 0.925 to 0.992 (Nei [1978] unbiased genetic identity). Fitts et al. (1992) studied Purshia subinte- gra in Verde Valley, reporting on many impor- tant, but heretofore unknown, aspects of the pollination biology of the species. They found that flowers may be pollinated anytime in the first three days after anthesis, the plants are partially self-compatible, and native and intro- duced bees are primaiy pollinators. Reciprocal crossing experiments between P. subintegra and what was believed to be P. stansburiana were also conducted by Fitts et al. (1992). As is discussed in the concluding section of this report, plants from which the P. stansburiana was taken are actually introgressed forms. The purpose of this study was to analyze moiphological character variation in species of Purshia in order to identify the range of mor- phological variation in P. subintegra and to develop a means of discriminating between Purshia subintegra, an endangered species, and other non-endangered Purshia taxa with which P. subintegra is most likely to be con- fused. This study was undertaken solely to address the need of natural resource managers to have a means of determining which individ- uals and populations of Purshia are protected under the Endangered Species Act. The methods used in this study were carefully cho- sen to obtain this result. 1994] Identification of Purshia subintegra 259 Methods A total of 216 Purshia plants were sampled at 27 widely scattered sites from southeastern to northwestern Arizona for measurement and analysis of morphometric characters (Fig. 1). Much attention was given to sampling Purshia in Verde Valley, the only location where it is believed P. subintegra and P. stanshuriana are in close enough proximity that gene exchange might currently be occurring. It was hoped that if hybridization were occurring between the two taxa, it would be possible to isolate characters useful in discriminating between P. subintegra, P. stanshuriana, and the introgressed forms. In determining where to sample, it was important to have some firsthand notion of where the introgressed forms might be: this turned out to be more difficult than it might seem given the disparate views of several researchers. Table 1 lists three separate collec- tion sites for P. subintegra in Verde Valley (labeled D, E, and F), in addition to several locations for introgressed forms and one loca- tion for P. stanshuriana on mountains over- looking the valley (Jerome, labeled X). Figure 2 shows Verde Valley and collection sites in the valley. The author made three separate collecting trips to Verde Valley— 1987, 1989, and 1992 — each time expanding the scope of the sampling effort to try to obtain a more rep- resentative sample of character variation. Morphometric samples from each of the 216 sampled plants consisted of two to four 20-40-cm-long branches dried in a standard herbarium press. Samples were collected in April 1987, 1989, and 1992 from each of 4 P subintegra populations, 3 P. stanshuriana pop- ulations, and 20 sites of introgressed forms. Table 1 shows the locations and sample sizes of each collection site. A rigorous stratified- random sampling method was employed at the P. subintegra and P. stanshuriana sites, and at least 10 specimens were collected at each site. Collections at the introgressed fonn sites were made much more subjectively and the sample sizes were much smaller, only one to six speci- mens. Data on 15 characters judged to be poten- tially useful in taxonomic differentiation between the 27 groups were obtained from the pressed specimens. Floral characters were heavily relied on, and characters that could be used in field identification of unknown speci- mens were also employed. Table 2 shows the character palette developed for the morpho- metric analysis. The list of characters indicates that a mix of binary, categorical, and continuous data was used. This was taken into account in subsequent statistical analyses. All measure- ments and counts were made under a binocu- lar dissecting microscope with a micrometer disk or electronic calipers. Scoring procedures are described in Table 2. SYSTAT version 4.0 was used to subject the data to more than 50 separate discriminant function analyses (DFA) and principal compo- nents analyses (PCA) to identify moqDhologi- cal groups and to determine which characters could be most confidently used to separate the groups. Numerous combinations of characters were used to group like data (binary, categori- cal, and continuous) and to examine the effects of including ratios as characters (hypanthium- pedicel length/width, sepal length/width, petal length/width) in the data set. Initial analyses using PCA were run on several combinations of characters to identify characters responsible for within-group similarity. A priori assignments of plants to groups required for DFA involved grouping collec- tion sites in several combinations by morpho- logical, geographical, and ecological criteria. Most DFAs were run with the following groupings: (1) 27-group analysis, all 27 collec- tion sites coded as separate groups; (2) 3- group analysis, 4 P. subintegra sites, 3 P. stans- huriana sites, and all introgressed forms in one group; (3) 4-group analysis, 4 P. subintegra sites, 3 P. stanshuriana sites, and the intro- gressed forms separated into two groups iden- tified as "Tonto" and "Verde." Results Purshia subintegra can be differentiated from P. stanshuriana and introgressed forms by leaf glandularity and leaf lobing. The mean score of leaf glandularity in P. subintegra is less than 0.4, and the mean number of lobes/ leaf is 2.5 or less. All others are more glandular or have more leaf lobes. A population of what I initially believed to be introgressed forms at site a (Fig. 2) in Verde Valley possesses glan- dularity and lobing characteristics of P. subin- tegra and, based on this and the results of the multivariate analyses, should probably be clas- sified as P. subintegra. 260 Great Basin Naturalist [Volume 54 1994] Identification of Purshia subintegra 261 Table 2. Characters measured for analysis of character variation in Purshia stansburiana, introgressed forms, and Purshia subintegra. Descriptions of the character measurements used in multivariate analyses and acronyms (in paren- theses) used as variable labels in Tables 4 and 5. Most hypanthia-pedicel, petal, sepal, pistil, and stamen measurements and counts were from the same five flowers from each plant. 1. Leaf pubescence. (LEFPUB) The adaxial surface of Purshia leaves is densely tomentose, though the midvein is often bare. The abaxial surface ranges fi-om completely glabrous to completely obscured by long arachnoid hairs. Twenty leaves from each of the 216 plant specimens were scored on an index of leaf pubescence density. Only the dorsal (abaxial) surface was scored. The scale ranged from 1 (completely to nearly completely glabrous) to 5 (densely pubescent). 2. Leaf glands. (LEAFGLAN) Ten leaves from each of the 216 plants were examined and scored for presence or absence of impressed-punctate glands. 3. Hypanthium-pedicel glands. (HYPGLAN) Five hypanthia (with pedicels) from each plant were examined and scored for presence or absence of stipitate glands. 4. Leaf lobes. (LOBES) The number of lobes on each of 20 leaves from each of the 216 plants was counted. The leaf tip was not counted. Lobes var- ied in distinctness, i.e., from much longer than wide to mere bumps on the edge of the leaf Even the most minor lobes were scored. Figure 3 illus- trates variation in leaf lobing among P. subintegra, the "Tonto" and "Verde" introgressed forms, and P. stansburiana, and provides an example of lobe scoring. 5-7. Hypanthium-pedicel dimensions. (HYPLGTH, HYPWDTH) Length and maximum width of five hypanthia-pedicels from each of the 216 plants were measured under a binocular dissecting microscope using a micrometer disk or electronic calipers. The length/width ratio was also entered as a character variable (HYPRAT). 8-10. Sepal dimensions. (SEPLGTH, SEPWDTH) All sepals (usually 5) from five flowers from each of the 216 plants were dissected and measured under a microscope using a micrometer disk or electronic calipers. Maximum (basal) width and length of the sepals were recorded. The length/ width ratio was also entered as a character vari- able (SEPRAT). 11-13. Petal dimensions. (PETLGTH, PETWDTH) Afl petals (usually 5) from five flowers from each of the 216 plants were dissected and measured under a microscope using a micrometer disk or electronic calipers. Maximum width and length of the petals were recorded. The length/width ratio was also entered as a character variable (PETRAT). 14. Pistil number. (PSLSFLR) Flowers normally con- tained 2-A pistils. Aborted pistils were easily dis- tinguished by their small size (<1.25 mm long) and brown to dark brown color. Viable pistils were pale yellow with silvery-white achene hairs and were nearly always longer than 1.25 mm. The total number of pistils per flower was counted on each of the 216 specimens, as well as the number of viable and aborted pistils. 15. Number of stamens. (STMNS) Stamens were counted in five flowers from each of 216 plants. Table 3 lists mean values and standard errors obtained for each of the 15 characters included in the analysis for each of the four identified groups of Purshia spp. Figure 4 illustrates the distribution of variation in sepal and hypanthium-pedicel dimensions. Note that sepal length and width are highly and positively correlated among groups, while hypanthium-pedicel width and length are not. P. subintegra plants have shorter, narrower sepals, while P. stansburiana have longer, wider sepals; introgressed forms are intermediate. "Tonto" forms have very long, wide hypanthia- pedicels, while "Verde" forms have slightly shorter, but much narrower, hypanthia-pedicels; neither of the two introgressed forms is inter- mediate in hypanthium-pedicel dimensions between the supposed parent species, P. subintegra and P. stansburiana. Principal Components Analysis Rotated factor scores derived from three PCAs are listed by character in Table 4 and are graphed in Figure 5. The first three factor axes together account for 73-87% of variance in the data. Horizontal relationships on the FACTOR(2)/FACTOR(l) graphs for each analysis (x-axis, Figs. 5A-5C) are primarily based on leaf lobing, while vertical relation- ships are based on glandularity. The horizontal relationship is again based on glandularity in the FACTOR(3)/FACTOR(2) graphs (y-axis, Figs. 5A-5C), but the vertical relationship (z- axis. Figs. 5A-5C) is mostly influenced by hypanthium-pedicel length. PCAs illustrate similarities of the three groups of Purshia spp. to each other, but graphs must be interpreted carefully. It appears that 262 Great Basin Naturalist [Volume 54 P. subintegra n8 0 0 0 28 9 32 2 3 46 0 4 16 0 20 2 0 0 120 122 39 36 18 123 216 not be just as prevalent as those with Q. veluti- na, argue instead that the problem lies in not properly delineating the range of variation in putative parents, putative hybrids, and intro- gressed forms. They found characters with non-intermediate dimensions, but in every case these were closer to one putative parent than the other and not, as described here for hypanthium-pedicel dimensions, quite differ- ent from both putative parents. This could be accounted for by several factors including nat- ural variation, heterosis, and linked gene con- trols of simple characters. No breeding experi- ments involving P. subintegra and P. stansburi- ana have yet been carried out so that we may describe the quantitative characteristics of an actual hybrid. It has only been assumed so far that a hybrid should possess a morphology roughly intermediate between the parents. The close proximity of some "Verde" plants to P. subintegra suggests the opportunity for hybridization exists now, although data pre- sented here do not indicate which of the sam- pled plants, if any, represent F^ hybrids. Results obtained in this study of Purshia indicate that while it is reasonably simple to determine what is P. subintegra, it is not always possible to distinguish what have been referred to as 'introgressed forms' from P. stans- buriana. Results of Anderson and Harrison (1979), in a situation similar to that described here, might suggest that we consider intro- gressed forms a part of the natural variation inherent in locally adapted P. stansburiana genomes, and not, as has been assumed by many, a result from a predominance of back- crosses with P. stansburiana. Why then are forms as similar to P. stansburiana, as described above, found in such proximity to P. subintegra at Dead Horse Ranch, and how is it that P. subintegra in Verde Valley has remained morphologically distinct despite what appears to be ample opportunity for extensive hy- bridization? 270 Great Basin Natur.\list [Volume 54 Table 7. Reliahk' characters iiscliil in cliscriiiiiiiatiiiK hctween P. suhintCfira, P. staii.shiiriana, and introgressed fomis. Quantities in parentheses are mean \ahies taken from Tabic 3. Other data are based on author's observations and Anderson (1986). Purshia Piirshiii Character suhintcura "Verde" "Tonto" stunsburiana Distribution 640-1065 m elevation, 1020-1385 m elevation, 725-1015 m elevation. Statewide (except south- northwest of Bylas, Verde Valley, near Camp Verde, western quarter). Graham Co.; north Yavapai Co. Yavapai Co.; generally above 1500 m of Burro Cn, Mohave Tonto Basin and elevation Co.; Horseshoe L., south of Globe, Gila Co. .Maricopa and Yavapai COS.; east of Cottonwood, Yavapai Co. Ecology Restricted to limey soils derived from weathered Tertiaiy lakebed limestones Mostly restricted to limey soils derived from weathered Tertiary lakebed limestones, occasionally on Supai Formation sandstones Limey soils of weathered lakebed lime- stones; other limestone fomiations, soils derived from volcanics and alluvial materials Various Growth tonii Shrub, up to 1 m tall, stems widely and sparingly branching from the base Shrub, 1-2 m tall when mature, widely and sparingly branched from the base Shrub, 1-2, or some- times 3 m tall when mature, widely and sparingly branched from the base Shrub, 3-5 m tall when mature, stems erect, branching from the base Leaf shape Entire or 1-2 lobes (0.7 lobes per leaf) 3-5 lobes (3.7 lobes per leaf 3-5 lobes (4.0 lobes per leaf) 3—5 lobes (3.7 lobes per leaf) Leaf vestiture Very densely pubescent on lower sui'face, less pubescent on upper surface Veiy densely pubescent on lower surface, less pubescent to bare on upper surface Very densely pubescent on lower surface, less pubescent to bare on upper surface Very densely pubescent on lower surface, less pubescent to bare on upper surface Leaf glandularity Usually none, rarely with impressed punctate glands (2% glandular) Usually glandular (86% glandular) Usually not glandular (20% glandular) .-^bundantK glandular (100% glandular) Hypanthia- pedicels Short, usually not glandular, rarely with stalked glands (.5.1 mm long, 2.4 mm wide, 11% glandular) Long, usually with stalked glands (9.2 mm long, 2.2 mm wide, 89% glandular) Long, usually without stalked glands (10.1 mm long, .3.2 mm wide, 40% glandular) Short, with abundant stalked glands (6.6 mm long, 2.9 mm wide, 100% glandular) Sepals Short and narrow (3.6 Long and narrow (4.1 mm long, 2.8 mm wide) mm long, .3.3 mm wide) Short and wide (4.0 mm long, 3.5 mm wide) Long and wide (4.9 mm long, 4.3 mm v\ide) Petals Short and very narrow Long and wide (9.9 mm (8.5 mm long, 5.7 mm long, 8.4 mm wide) wide) Long and wide (10.7 mm long, 8.2 mm wide) Very long and ver\' wide (11.3 mm long, 10.3 mm wide) Pistils Stamens 3-4 (3.5 per flower) (48.6 per flower) 4—6 (5.4 per flower) (67.4 per flower) 4-6 (5.2 per flower) (66.6 per flower) 4—6 (5.5 per flower) (89,0 per flower) Deciding whether the introgressed forms, "Verde " and "Tonto, " should be considered as such, or, alternatively as manifestations of a broader concept off? stanslniriana, is not vital to the need that prompted this study; namely, a means of accurately identifying P. subintegra for the purpose of determining what is and is not to be considered endangered under fed- eral law. No nomenclatural revisions or additions to the classification of Purshia are proposed; plants reported here as introgressed forms 1994] Identification of Purshia subintegra 271 should be regarded as such unless carefully controlled crossing and backcrossing experi- ments are conducted that clearly show result- ing progeny are essentially identical to "Verde" and "Tonto" plants in the field. I recommend- ed in 1986 that most fruitful results for the inteipretation of moiphological variation in P. subintegra would be obtained from a 'common garden' experiment. Under more carefully controlled conditions than may be found in the field, it should be possible to trace genetic bases for morphological characters relied on so heavily in this paper. Had such a study been initiated at that time, we may very well be enjoying the lucidity provided by early results. Hybrid formulae for nothotaxa have not yet been validly published for the introgressed forms and should not be until their origins are firmly established. Protective measures of the Endangered Species Act should be applied to those forms that conform to the P. subintegra character list in Table 7, but not to plants con- forming to the characteristics listed for the introgressed forms or to P. stansburiana. Acknowledgments This study was financed, in part, by U.S. Bureau of Reclamation, PO. 7-PG-32-16590, and U.S. Fish and Wildlife Sewice, PO. 20181- 1-1358. Many thanks to Tom Gatz (Reclamation) and Sue Rutman (Fish and Wildlife) for the opportunit\' to pursue the study in my own way. John Anderson, U.S. Bureau of Land Manage- ment, provided good company and many excellent ideas on several of the collecting trips. J. Mark Porter (University of Arizona) showed me how to set up the multivariate analyses and, more importantly, helped me to interpret the results. Thanks also to Jim Hen- rickson, who gave me the benefit of a biosys- tematist's insights into this sticky problem. Literature Cited Anderson, E. 1949. Introgressive hybridization. John Wiley, New York. 109 pp. Anderson, J. L. 1986. Biogeographical analysis of Cowania subintegra Kearney (Rosaceae), an Arizona Sonoran Desert endemic. Unpublished thesis, Arizona State University', Tempe. 118 pp. . 1993. A synthetic analysis of a rare Arizona species, Purshia subintegra (Rosaceae). Pages 20.5-220 in R. M. Sivinski and K. L. Lightfoot, eds., Proceedings of the Southwestern Rare and Endangered Plant Confer- ence, New Mexico Depailment of Natural Resources, Santa Fe. Anderson, R. C, and T. Harrison. 1979. A limitation of the hybrid inde.x using leaf characters. Southwestern Naturalist 24; 463-473. COOPERRIDER, M. 1957. Introgressive hybridization between Quercus marilandica and Querciis velutina in Iowa. American Journal of Botany 44: 804-810. FiTTS, R. D., V J. Teppedino, and T. L. Griswold. 1992. The pollination of the Arizona cliffrose [Purshia subintegra), including a report on experimental hybridization with its sympatric congener P. stans- buriana (Rosaceae). Pages 359-368 in R. M. Sivinski and K. L. Lightfoot, eds., Proceedings of the Southwestern Rare and Endangered Plant Conference. New Mexico Department of Natural Resources, Santa Fe. Henrickson, J. 1986. Notes on Rosaceae. Phytologia 60: 468. Kearney, T. H. 1943. A new cliff-rose from Arizona. Madroilo 7: 15-18. McArthur, E. D., H. C. Stutz, and S. C. Sanderson. 1983. Taxonomy, distribution, evaluation, and cyto- genetics of Purshia, Cowania, and Fallugia (Rosoideae, Rosaceae). Pages 4—24 in K. L. Johnson and A. R. Tiedemann, eds.. Proceedings — Research and management of bitterbrush and cliffrose in western North America. USDA Forest Service General Technical Report INT-152. Intermountain Forest and Range Experiment Station, Ogden, Utah. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89: 583-590. Phillips, B. G., A. M. Phillips III, ,\nd D. J. Howard. 1988. Final report. Starch gel electrophoresis of Purshia subintegra and Purshia stansburiana. Museum of Northern Arizona report to U.S. Bureau of Reclamation, Phoenix, Arizona. 42 pp. PiNKAVA, D. J., E. Lehto, and D. Keil. 1970. Plants new to the Arizona flora-II. Journal of the Arizona Academy Science 6: 134. ScHAACK, C. G. 1987a. Lectotypification of Cowania subintegra Kearney, basionym of Purshia subintegra (Kearnev) Henrickson (Rosaceae). Taxon 36: 4,52-454.' . 1987b. A new Arizona Purshia (Rosaceae). Phy- tologia 63: .301-302. SCHAACK, C. G., AND J. D. MoREFlELD. 1985. Field survey for Cowania subintegra Kearney, Coconino National Forest. Unpublished report to Coconino National Forest, Flagstaff, Arizona. 29 pp. U.S. Fish and Wildlife Service (USFWS). 1984. Final rule to determine Cowania subintegra (Arizona cliffrose) to be an endangered species. Federal Register 49: 22326-22329. Received 4 November 1993 Accepted 14 March 1994 Great Basin Naturalist 54(3), © 1994, pp. 272-286 OBSERVATIONS ON DOUBLE-CRESTED CORMORANTS {PHALACROCORAX AURITUS) AT SPORTFISHING WATERS IN SOUTHWESTERN UTAH Michael J. Ottenbacher^, Dale K. Hepworth^ and Louis N. Berg^ Abstract. — Counts of Double-crested Cormorants {Phalacrocorax aiiritiis) were made at 13 reservoirs and lakes in southwestern Utah during 1989-91 to detennine current abundance of that species. Food habits of cormorants were studied at three of the reservoirs in 1989. Data were also collected on trout abundance during standardized gill-netting to make comparisons between comiorant numbers and trout abundance. Cormorants were observed at all waters studied except one and were generally most numerous during the spring as they migrated through the area. Estimated cormorant abundance ranged from 0 to 34 bird-days per ha and was highest at the larger, lower-elevation reservoirs. Cormorants were summer residents at several of the larger reservoirs and nested successfully at Piute Reservoir. Trout accounted for 24—81% of the diet of cormorants, with Utah chubs constituting most of the remainder of the diet. Estimates of the annual consumption of fish by cormorants ranged from 0 to 15.8 kg per ha. The index of trout abundance was inversely related to cormorant abundance [P < .01) at the waters studied. Cormorants apparently have increased in numbers and extend- ed their range in southwestern Utah during the past decade. This change may be the result of factors that have led to similar changes throughout North America as well as some factors unique to Utah. Methods to mitigate the impact of predation by piscivorous birds on sportfisheries are discussed. The Utah Division of Wildlife Resources has initiated a new management plan at Miners ville Reservoir that incorporates piscivorous birds into sportfish management at that reservoir. Key words: cormorants, Phalacrocorax auritus, trout, abundance, food habits, predation, management, sport fishing, reservoirs, Utah. Various factors influencing survival of stocked trout were examined at Minersville Reservoir, Utah, in 1985-88 (Hepworth and Duffield 1991, Wasowicz 1991). During that study we observed an increase in the number of Double-crested Cormorants {Phalacrocorax auritus, hereafter referred to as cormorants) at Minersville Reservoir compared with previous years. An apparent increase in the abundance of cormorants at several other reservoirs was also noted, and we received reports of cor- morants at some waters where they previously had not been reported (Walters and Sorenson 1983). Apparent changes in abundance and distribution of this species in Utah coincided with reported increases in the number of cor- morants in many parts of North America (Price and Weseloh 1986, Christie et al. 1987, Campo et al. 1988, Findholt 1988). As the rela- tive abundance of cormorants has increased, there have been conflicting reports concern- ing their impact on recreational fisheries. A number of authors have concluded that cor- morants take considerable numbers of game fish and potentially impact important fisheries (Ayles et al. 1976, Myers and Peterka 1976, Christie et al. 1987, Campo et al. 1988). Others have felt that cormorants have had lit- tle impact on economically valuable species of fish (Baillie 1947, Carroll 1988, Findholt 1988). To evaluate the potential impact of cor- morants on fisheries in southwestern Utah, we continued to document the number of cor- morants at Minersville Reservoir and 12 addi- tional waters. We also collected data on trout abundance at these waters during standard- ized annual gill-netting and initiated a study of the food habits of cormorants at three of the larger reservoirs. Based on these observations, we determined current abundance of cor- morants at local waters, compared estimates of cormorant abundance to indices of trout abun- dance, and estimated annual consumption of fish by cormorants. Study Area Data on distribution, relative abundance, and seasonal occurrence of cormorants were ^Utah Division of Wildlife Resources, 622 N. Main, Box 606, Cedar City, Utah 84720. 272 1994] Cormorant Observations in Southwestern Utah 273 collected at 13 reservoirs in southwestern Utah (Table 1). Reservoirs ranged in size from 36 to 1020 ha, and elevations from 910 to 2695 m above MSL. Most reservoirs were originally constructed for irrigation storage and have water levels that fluctuate substantially on an annual basis. Highest water levels occurred in late winter and spring, with minimum levels in the fall following the irrigation season. Fish Lake and Panguitch Lake are natural lakes where storage has been increased by the addi- tion of small dams. All the reservoirs except Quail Creek and Gunlock had ice cover for a period of 2-5 months during winter and spring. Sportfishing is a major activity at all of the waters since they are open year-round to angling by the general public with various re- strictions (State of Utah 1992). Sportfisheries at all reservoirs except Gunlock are managed, at least in part, as put-grow-and-take trout fisheries. Various sizes and numbers of rain- bow trout {Oncorhynchus mykiss) were stocked annually at the different reservoirs. Fingerling rainbow (76 mm total length [TL]) were stocked at waters where numbers of competing species were low and predation was not a concern. Larger rainbow (127-178 mm TL) were stocked at reservoirs where sur- vival of small trout was poor because of com- petition with nongame species and/or preda- tion. Limited numbers of other species of trout were stocked at some waters to provide variety in fishing opportunity. Recruitment from spawning in tributaries associated with reservoirs also provided a small number of trout in addition to those stocked at some of the study waters. Stocked trout were harvest- ed by anglers after they reached a catchable size (>230 mm TL), generally after they had been in the reservoirs for 7-11 months. Few rainbow trout survive longer than 2 years fol- lowing stocking (Stuber et al. 1985, Hume and Tsumura 1992). Most reservoirs contained few fish species other than trout, and five con- tained primarily stocked trout (Enterprise, Kolob, Koosharem, Lower Bowns, and New- castle). Three of the reservoirs contained only stocked trout and Utah chubs (Gila atraria; Miners ville. Otter Creek, and Panguitch). Two were primarily warm-water fisheries where trout abundance was not evaluated (Gunlock and Quail Creek). The remaining three waters (Fish Lake, Johnson and Piute reservoirs) con- tained more than two other fish species besides trout. Only two or three of these other species were abundant, while the rest were of minor occurrence. A number of the waters in which Utah chubs and Utah suckers {Catastomus ardens) occurred were periodically treated with rotenone to remove all fish when those non- game species became abundant. When recla- mation projects were conducted, chubs and suckers often outnumbered trout by hundreds to one. Following treatments, trout were the Table 1. Description of waters in southwestern Utah where scheduled counts of Double-crested Cormorants were conducted, 1989-91. Maximum Fish species present'* surface Water Location Elevation (m) area (ha) Common Uncommon Enterprise Reservoir T38S R18W, Washington Co. 1755 200 RT Fish Lake T26S R2E, Sevier Co. 2695 1012 RT LT SR US, UC, YP RS, MS Gunlock Reservoir T40S R17W, Washington Co. 1092 108 CC, LB, GS BC, BG Johnson Reservoir T25S R2E, Sevier Co. 2688 285 RT CT UC, US YRRS Kolob Reservoir T38S RllW, Washington Co. 2474 136 RT CT BK Koosharem Reservoir T25S RIE, Sevier Co. 2132 125 RT CT BK Lower Bowns Reservoir T31S R6E, Garfield Co. 2271 36 RT CT BK Minersville Reservoir T29S R8W, Beaver Co. 1677 401 RT, CT UC BN Newcastle Reservoir T36S R15W, Iron Co. 1659 66 RT SB Otter Creek Reservoir T29S R2W, Piute Co. 1942 1020 RTUC CTBN Panguitch Lake T35S, R7W, Garfield Co. 2502 505 RT BK, BN, CT UC Piute Reservoir T28S, R2W, Piute Co. 1828 1015 RT UC, US BN, CT RS, SB Quail Creek Reservoir T42S, R14W, Washington Co. 910 239 RT LB, BC, BG BB ^Fish species: RT = rainbow trout, GS = green sunfish, LT = lake trout, SP = splake trout, US = Utah sucker, UC = Utah chub, YP = yellow perch, MS = mottled sculpin, CC = channel catfish, LB = largemouth bass. BC = black crappie, BG = bluegill. CT = cutthroat trout, BK = brook trout, RS = redside shin- er, BN = brown trout, SB = smallmouth bass, BB = black bullhead. 274 Great Basin Naturalist [Volume 54 predominant species for at least a \'ear or two. In situations where undesirable nongame species of fish could not he completely removed from a drainage, Utah chubs and Utah suckers would gradually increase and eventualK' return to pre-reclamation densities. In addition to cormorants, other piscivorous birds observed at the study waters included Common Loons {Gavia immer). Western Grebes {AechmopJiorous occidcniulis), American White Pelicans {Pelecanus crfhrorychos). Mergansers {Mergus merganser and M. serra- tor), and Great Blue Herons {Ardea herodias treganzai). Methods Counts of cormorants were made at 1- to 3- week intervals at 11 resei^voirs during 1989. In 1990 we made biweekly counts at four of the larger reservoirs. Counts were made again in 1991 at the four reservoirs surveyed in 1990, as well as three additional ones. Counts gener- ally began following ice-out at each reservoir and continued through November at most waters. We discontinued counts early at sever- al reservoirs that were drained during the summer or chemically treated to remove nongame fish. At most locations counts were made from shore using binoculars or a spot- ting scope. At larger resei"voirs we often used a boat to facilitate counting. Technicians mak- ing counts were instructed using a standard training program by the authors. The same one or two technicians counted birds at all waters during any one year of the study. Cormorants were easily identified. Knowledge of the birds' feeding and resting patterns, as well as other behaviors, also aided in making accurate counts. An annual estimate of cormorant abun- dance (bird-days per ha per year) was made for each reservoir studied. The estimate of abundance was calculated using methods commonly employed to estimate sportfishing pressure in creel surveys of anglers (Robson 1960, Lambou 1961). A bird-day was defined as one day spent by one cormorant at a given water. The sampling period was stratified by 3-month intei^vals, March-May, June-August, and September-November. The number of days within a stratum varied among waters, depending on the time of ice-out and whether a given reservoir was drained or treated in the fall. The number of bird-days for a stratum at a given water was estimated using the follow- ing formula: D = K(—^) n Var (D) = n (n-1) n /h D = estimated total bird-days; K = number of days within a sampling stratum; oCj = number of cormorants counted on the ith day; n = number of days sampled within a stratum; Var = estimated variance; N = total cormorants counted per stratum. 95% confidence intei"val = ± 2 yVar (D). The estimate of annual cormorant abundance was the sum of estimated bird-days for strata within a sampling year divided by the mean surface area of the reservoir. Gill-nets were used to estimate trout abun- dance at each reservoir (Bennett 1962, Hubert 1983) during early spring, 2-4 weeks after winter ice cover was completely gone. Net numbers, styles, and locations were based on long sampling histories at each water. Gill-net data have been collected on most of the study waters for 10 years or more. We followed stan- dardized netting practices used by the Utah Division of Wildlife Resources (UDWR). Two to six nets, depending on lake or reservoir size, were set at each water in areas less than 30 ft deep. Nets were set during the afternoon and retrieved the following morning. Each net was 1.8 m deep by 38.1 m long and consisted of five monofilament nvlon panels with bar mesh sizes of 19.1, 24.4, 31.8, 38.1, and 50.8 mm. Data recorded for fish gill-netted at each water included numbers, species, and individ- ual lengths. Gill-net samples generally consist- ed of trout stocked the previous year and a few from stocking 2 years earlier. The trout abundance index used for each reservoir in the study was the mean number of trout col- lected per net, set overnight (trout per net- 1994] Cormorant Observations in Southwestern Utah 275 night). When comparing trout abundance and estimates of cormorant abundance, we paired the trout abundance index for a given water with the estimate of cormorant abundance for the previous year. Because spring gill-net catches consisted primarily of trout stocked the previous year, the relationship between the cormorant abundance estimate and trout abundance index reflected impacts of preda- tion on one cohort of stocked trout over one year Large trout were excluded from the data at two waters when calculating the trout abun- dance index. These larger fish represented older cohorts that were not vulnerable to cor- morant predation during the study period. Large trout occurred at Minersville Reservoir and Fish Lake as the result of unusual circum- stances or the presence of unique trout popu- lations. At Fish Lake a few large lake trout {Salvelinus namaijacush) were not used in the index. One cohort of cutthroat trout {Oncorhyn- chus clarki) at Minersville Reservoir was not used in the trout abundance index for that reservoir. This 1986 cohort grew rapidly to a large size following a chemical renovation in 1985 and comprised a substantial portion of the annual spring gill-net catches through 1991. A simple linear regression was used to compare estimates of cormorant abundance (bird-days) and the trout abundance index (trout per net-night) using both untransformed data and log-transformed data. Data on cormorant diet were collected at three large, lower-elevation reservoirs where birds were relatively abundant. At Minersville and Otter Creek resei'voirs, primaiy potential fish prey species were stocked rainbow trout and Utah chubs, with lesser numbers of cut- throat trout and brown trout {Salmo trutta). At Piute Reservoir primaiy prey species included rainbow trout, Utah chubs, and Utah suckers. Piute Reservoir also contained limited num- bers of redside shiners {Richardsonius haltea- tiis), smallmouth bass {Microptenis dolomieiii), cutthroat trout, and brown trout. We collected 10 cormorants each at Minersville, Otter Creek, and Piute reservoirs (30 birds total). Birds were collected using shotguns during July and August at 1000-1100 h following morning feeding periods. We also used food- habit data collected by Wasowicz (1991) at Minersville Reservoir in April 1988, which included cormorants collected during after- noon hours. Additional food-habit information was obtained from six fledgling cormorants at Piute Reservoir in 1989 by approaching active nests and collecting regurgitated stomach con- tents. In total, diet data were obtained from 52 cormorants, with samples taken in mid-April, late April, late June, late July, early August, and late August. Stomach contents were identified to fish species using flesh color, peritoneum color, fin rays, and pharyngeal teeth as key characteris- tics. We made TL measurements of ingested fish when possible. TL estimates were also based on a measurement from the front of the dorsal fin to the front of the anal fin. Estimates of biomass of ingested fish were made using length-weight relationships for each species (Carlander 1969, Varley and Livesay 1976). Annual trout consumption by cormorants was estimated by multiplying values for bird abundance (bird-days) by a daily biomass con- sumption rate of 465 g per day (after Wasowicz 1991), and by the percentage of trout in the diet (this study, Wasowicz 1991). The daily biomass consumption rate used by Wasowicz (1991) and this study was based on an average adult body weight for cormorants of 1860 g (Ross 1977) and a daily biomass consumption rate of 25% of body weight. Dunn (1975) reported that daily consumption rates for free- living adults and juveniles of several species of cormorants averaged approximately 20-30% of body weight. When information on diet com- position of cormorants was not available for a particular water, we made a conservative esti- mate of the percentage of trout in the diet by determining relative abundance of trout and other forage species in that water. Season-long creel surveys of sport fisher- men and chemical treatment projects to remove undesirable nongame fish were con- ducted at a number of study waters. Although not directly related to this study, data collect- ed during these activities provided a means to validate trout abundance indices and verify relative abundance of different fish species. We estimated total annual trout harvest by anglers and the percent return to the creel of the total numbers of fish stocked (Robson 1960, Lambou 1961) during creel surveys. High and low harvest estimates corresponded widi higli and low trout abundance as measured by standardized gill-netting. Visual inspections following chemical treatments provided anoth- er way of verifying relative fish abundance and 276 Great Basin Naturalist [Volume 54 species composition. Following a chemical treatment, we could be certain that stocked trout dominated a fishery for a year or two. Creel surveys were conducted at Fish Lake in 1989, Johnson Reservoir in 1984 and 1989, Kolob Reservoir in 1991, Lower Bowns Reservoir in 1991, Minersville Reservoir in 1986 and 1988, Newcastle Reservoir in 1991, and Otter Creek Reservoir in 1985. Chemical treatments were conducted at Johnson Reservoir in 1986, Kolob Reservoir in 1985, Koosharem Reservoir in 1985, Minersville Reservoir in 1984 and 1991, Otter Creek Reservoir in 1989, Panguitch Lake in 1991, and Piute Reservoir in 1985 and 1990. Results Cormorant Distribution and Abundance Eight to 35 counts were made at each of the 13 reservoirs (Table 2). Individual counts of cormorants ranged from 0 to 264 birds. Cormorants were observed early in the year (2 February 1989) at Quail Creek Reservoir, which was the lowest in elevation and most southern reservoir studied. At most other waters, cormorants were first observed soon after ice-out, usually in March. Numbers of cormorants were generally highest in spring or early summer. At lower-elevation waters, cormorants were often absent during midsum- mer but were observed again in late summer or fall. At some higher-elevation waters, high- est counts occurred in midsummer They were present throughout the summer at several of the larger reservoirs. Cormorants were observed at all waters surveyed except one. Lower Bowns Reservoir, the smallest and most easterly located. Cormorants attempted to nest at 2 of the 13 locations studied. In 1988 and 1989 nesting was initiated at Minersville Reservoir. Cormorants constructed nests in a flooded grove of Cottonwood trees in the shallow north end of the reservoir. The nests were aban- doned, however, in late spring when the water level receded beyond the nesting trees. Water levels at Minersville Reservoir remained low during the spring of 1990 and 1991. The area in which nesting had been attempted the pre- vious 2 years remained some distance above the shoreline, and cormorants made no further attempts to nest. Cormorants did nest success- fully at Piute Reservoir in 1989 and 1990. On 26 June 1989, 45 fledgling cormorants were observed in nests in flooded cottonwood trees in the south end of that reservoir. In 1990, 55 pairs of nesting birds were observed in the same area on 11 April. Young cormorants were observed in 16 of the nests on 26 May 1990, in spite of rapidly dropping water levels that had left nesting trees well above the shoreline. Piute Reservoir was drained in the fall of 1990 causing water levels to remain low in 1991 and exposing the ground below trees used for nesting the previous 2 years. No nesting activ- ity was observed at any of the locations stud- ied in 1991. Estimates of cormorant abundance at the 13 reservoirs ranged from 0 bird-days at Lower Bowns Reservoir in 1991 to 20,329 bird-days at Otter Creek Reservoir in 1989 (Tables 2 and 3). When we accounted for the size of various waters surveyed, cormorant abundance was highest at Minersville Reservoir where we estimated 34 bird-days per ha for 1989 (Table 4). Cormorant abundance was low at most of the higher-elevation waters, such as Kolob Reservoir, Johnson Reservoir, and Fish Lake. Trout Abundance Stocking rates ranged from 186 to 669 trout per ha per year at the waters studied, except at Gunlock Reservoir, which was managed only for warm- water species. In general, num- bers and sizes of trout stocked at each reser- voir or lake were considered sufficient to pro- duce high numbers of catchable-size trout providing that survival was adequate. Trout abundance indices at the waters surveyed ranged from 1 to 91 trout per net-night (Table 4). Our past experience indicates that trout abundance indices of at least 25-30 fish per net-night yield a population of trout that will produce good fishing during the year. Rainbow trout accounted for the majority of the gill-net catch at most waters. The trout abundance index was inversely related to esti- mates of cormorant abundance (P < .01, Fig. 1). Although a log transformation of cormorant abundance data statistically improved the fit of the regression line, the negative relationship was also significant [P < .05) for the original, untransformed data. Trout abundance indices were low when bird abundance was greater than 15 cormorant-days per ha. Both high and low trout abundance indices occurred with low cormorant abundance; however, there 1994] Cormorant Observations in Southwestern Utah 277 Table 2. Statistics fiom cormorant counts at 13 reservoirs in southwestern Utah, 1989-91. Time of year WaterA'ear/Statistic Mar- May Jun-Aug Sep-Nov Total Otter Creek Reservoir, 1989 Total days in intei-val 78 92 52 222 Number of counts 12 7 5 24 Mean birds per count 65 135 56 92 Estimated bird-days 5044 12,394 2891 20,329 Standard error (bird-days) 883 715 947 1479 95% confidence interval ±1765 ±1431 ±1894 ±2958 Otter Creek Reservoir, 1990 Total days in interval 76 92 61 229 Number of counts 6 6 3 15 Mean birds per count 2 11 8 7 Estimated bird-days 139 1043 488 1670 Standard error (liird-days) 53 563 347 663 95% confidence interval ±107 ±1126 ±694 ±1327 Otter Creek Reservoir, 1991 Total days in interval 76 92 61 229 Nimiber of counts 5 6 2 13 Mean birds per count 53 7 0 20 Estimated bird-days 4013 675 0 4688 Standard error (bird-days) 1204 357 0 1256 95% confidence interval ±2407 ±714 ±0 ±2511 Newcasde Reservoir, 1989 Total days in intei-val 92 92 91 275 Number of counts 9 6 7 22 Mean birds per count 6 0 1 2 Estimated bird-days 593 0 65 658 Standard error (bird-days) 114 0 52 125 95% confidence interval ±228 ±0 ±103 ±250 Newcasde Reservoir, 1991 Total days in interval 92 92 31 215 Number of counts 18 18 3 39 Mean birds per count 3 t^ 1 1 Estimated bird-days 240 15 21 276 Standard error (bird-days) 103 8 21 105 95% confidence interval ±206 ±17 ±41 ±210 Minersville Reservoir, 1989 Total days in interval 92 92 56 240 Number of counts 14 12 7 33 Mean birds per count 78 33 11 45 Estimated bird-days 7209 3067 624 10,900 Standard error (bird-days) 1785 475 190 1856 95% confidence interval ±3569 ±949 ±379 ±3712 Minersville Reservoir, 1990 Total days in interval 92 92 61 245 Number of counts 12 11 2 25 Mean birds per count 64 14 3 30 Estimated bird-days 5850 1296 153 7299 Standard enor (bird-days) 1265 317 153 1313 95% confidence interval ±2529 ±634 ±305 ±2625 Minersville Reservoir, 1991 Total days in interval 92 92 30 214 Number of counts 7 6 2 15 Mean birds per count 31 1 0 14 Estimated bird-days 2852 107 0 2959 Standard error (bird-days) 963 77 0 966 95% confidence interval ±1926 ±153 ±0 ±1931 278 Great Basin Natufl\list [Volume 54 Table 2. Continued. Ti )f \ far WaterAVar/Statistic Mar-\la\ Jun-Aug .St'p-Xo\' Total Piute Reservoir, 1989 Total days in intenid Number ol coiuits Mean birds per count Estimated bird-days Standard error (]:)ird-days) 95% confidence interval Piute Reser\'oir, 1990 Total days in interval Number of counts Mean birds per count Estimated bird-days Standard error (bird-days) 95% confidence interval Piute Reservoir, 1991 Total days in inter\'al Number of counts Mean birds per count Estimated bird-days Standard error (bird-days) 95% confidence intei^val Fisli Lake, 1989 Total days in interval Number of counts Mean birds per count Estimated bird-days Standard error (bird-days) 95% confidence intei'val Panguitch Reservoir, 1989 Total days in intei"val Number of counts Mean birds per count Estimated bird-days Standard error (bird-days) 95% confidence interval Panguitch Reservoir, 1990 Total days in intei"val Number of counts Mean birds per count Estimated bird-days Standard error (bird-days) 95% confidence interval Panguitch Reservoir, 1991 Total days in interval Number of counts Mean birds per count Estimated bird-days Standard error (bird-days) 95% confidence interval 82 92 11 7 70 65 5702 5967 690 1358 ±1380 ±2715 92 92 7 6 60 29 5559 2683 1395 670 ±2790 ±1341 82 92 5 6 2 5 180 475 160 196 ±320 ±392 82 92 4 6 0 t 0 15 0 15 ±0 ±31 71 92 4 5 1 6 71 570 71 128 ±142 ±256 49 92 7 12 6 25 301 2285 121 175 ±242 ±349 30 92 2 6 0 5 0 429 0 272 ±0 ±543 91 265 8 26 12 48 1081 12,750 605 1639 bl209 ±3277 0 184 0 13 0 45 0 8242 0 1548 ±0 ±3095 174 — 11 — 4 — 655 — 253 — ±506 74 248 5 15 0 t 0 15 0 15 ±0 ±31 73 236 5 14 5 4 365 1006 247 287 ±493 ±573 61 202 3 22 13 17 773 3359 458 505 ±917 ±1010 0 122 0 8 0 4 0 429 0 272 ±0 ±543 1994] Cormorant Observations in Southwestern Utah 279 Table 2. Continued. Time of year WaterA'ear/Statistic Mar-Ma> Jun-Aug Sep-Nov Total Koosharem Resei^voir, 1989 Total da\'S in interval 92 92 Number of counts 10 6 Mean birds per count 0 1 Estimated bird-days 0 46 Standard error (bird-days) 0 21 95% confidence interval ±0 ±41 Johnson Resenoir, 1989 Total da\'s in intenal 31 92 Number of counts 3 6 Mean birds per count 0 t Estimated bird-days 0 31 Standard eiTor (bird-days) 0 31 95% confidence interval ±0 ±61 Enterprise Resen'oir, 1989 Total days in interval 71 92 Number of coimts 9 6 Mean birds per count 4 1 Estimated bird-days 252 77 Standard enor (liird-days) 147 50 95% confidence inten'al ±295 ±100 Lower Bowns, 1991 Total days in interval 31 92 Number of counts 6 17 Mean birds per count 0 0 Estimated bird-days 0 0 Standard enor (bird-days) 0 0 95% confidence interval ±0 ±0 Kolob Reservoir, 1991 Total days in interval 31 92 Number ot counts 3 18 Mean birds per count t t Estimated bird-days 10 5 Standard error (bird-days) 10 5 95% confidence intei"val ±21 ±10 Gunlock Resei-voir, 1989 Total days in interval 92 92 Number of counts 10 7 Mean birds per count 8 0 Estimated bird-days 727 0 Standard eiTor (Ijird-davs) 296 — 95% confidence interval ±592 ±0 Quail Creek Reservoir, 1989 Total days in intenal 92 92 Number of counts 12 7 Mean birds per count 3 0 Estimated bird-days 284 0 Standard error (l:)ird-days) 167 — 95% confidence interval ±334 — 11 195 2 18 1 t 6 52 6 21 ±11 ±43 58 181 4 13 0 t 0 31 0 31 ±0 ±61 88 251 6 21 t 1 29 358 29 158 ±59 ±317 61 184 12 35 0 0 0 0 0 0 ±0 ±0 61 184 14 35 t t 4 19 4 12 ±9 ±25 91 275 6 23 t 3 15 742 15 297 ±30 ±593 91 275 6 25 1 1 91 375 30 170 ±60 ±339 lorants present but mean number of l)ircls per count was less tlian 0.1. 280 Great Basin Naturalist [Volume 54 Table 3. Estimated annual c()n.sinnpti(jn of fish l)y cormorants at 13 reservoirs in southwestern Utah, 1989-91. Estimated'' Total annual fish Percent Annual trout Survey bird-days consumption trout in consumption Water Year period (95% C.l.) (kg) diet (kg) Lower Bowns 91 1 May-31 Oct 0(±0) 0 — 0 Enterprise 89 22 Mar-27 Nov 358 (±317) 166 100^ 166 Fish Lal', Jr, and Steven D. Sutherland 212 Nesting and summer habitat use by translocated Sage Grouse {Centrocercits urophasianus) in central Idaho David D. Musil, Kerry P Reese, and John W. Connelly 228 Vegetation zones and soil characteristics in vernal pools in the Channeled Scab- land of eastern Washington Elizabeth A. Crow^e, Alan J. Busacca, John P Reganold, and Benjamin A. Zamora 234 Golden Eagle {Aquila chrijsaetos) population ecology in eastern Utah J. William Bates and Miles O. Moretti 248 Identification oiPurshia subintegra (Rosaceae) Frank W Reichenbacher 256 Observations on Double-crested Cormorants {Phalacrocorax auritus) at sport- fishing waters in southwestern Utah Michael J. Ottenbacher, Dale K. Hepworth, and Louis N. Berg 272 Notes New mammal record for Fremont Island with an updated checklist of mammals on islands in the Great Salt Lake Kenneth L. Cramer 287 H E GREAT BASIN MTURAUST VOLUME 54 NS 4 — OCTOBER 1994 BRIGHAM YOUNG UNIVERSITY GREAT BASIN NATURALIST Editor Assistant Editor Richard W. Baumann Nathan M. Smith 290 MLBM 190 MLBM PO Box 20200 PO Box 26879 Brigham Young University Brigham Young University Provo, UT 84602-0200 Provo, UT 84602-6879 801-378-5053 801-378-6688 FAX 801-378-3733 E-mail: NMS@HBLL1.BYU.EDU Associate Editors Michael A. Bowers Paul C. Marsh Blandy Experimental Farm, University of Center for Environmental Studies, Arizona Virginia, Box 175, Boyce, VA 22620 State University Tempe, AZ 85287 J. R. Callahan Stanley D. Smith Museum of Southwestern Biology, University of Department of Biology New Mexico, Albuquerque, NM University of Nevada-Las Vegas Mailing address: Box 3140, Hemet, CA 92546 Las Vegas, NV 89154-4004 Jeffrey J. Johansen Paul T. Tueller Department of Biology, John Carroll University Department of Environmental Resource Sciences University Heights, OH 441 18 University of Nevada-Reno, 1000 Valley Road „ r. ^ Reno, NV 89512 Boris C. Kondratieff Department of Entomology, Colorado State Robert C. Whitmore University, Fort Collins, CO 80523 Division of Forestry, Box 6125, West Virginia University, Morgantown, WV 26506-6125 Editorial Board. Jerran T. Flinders, Chairman, Botany and Range Science; Duke S. Rogers, Zoology; William Hess, Botany and Range Science; H. Duane Smith, Zoology. All are at Brigham Young University. Ex Officio Editorial Board members include Steven L. Taylor, College of Biology and Agriculture; Stanley L. Welsh, Director, Monte L. Bean Life Science Museum; Richard W. Baumann, Editor, Great Basin Naturalist. The Great Basin Naturalist, founded in 1939, is published quarterly by Brigham Young University. Unpublished manuscripts that further our biological understanding of the Great Basin and surrounding areas in western North America are accepted for publication. Subscriptions. Annual subscriptions to the Great Basin Naturalist for 1994 are $25 for individual sub- scribers ($30 outside the United States) and $50 for institutions. The price of single issues is $12. All back issues are in print and available for sale. All matters pertaining to subscriptions, back issues, or other busi- ness should be directed to the Editor, Great Basin Naturalist, 290 MLBM, PO Box 20200, Brigham Young University Provo, UT 84602-0200. Scholarly Exchanges. Libraries or other organizations interested in obtaining the Great Basin Naturalist through a continuing exchange of scholarly publications should contact the Exchange Librarian, 6385 HBLL, PO Box 26889, Brigham Young University Provo, UT 84602-6889. Editorial Production Staff JoAnne Abel Technical Editor Jan Spencer Assistant to the Editor Copyright © 1994 by Brigham Young University ISSN 0017-3614 Official pubUcation date: 25 October 1994 10-94 750 11712 MCZ LIBRARY :0V 3 0 1994 The Great Basin Naturalist Published at Provo, Utah, by Brigham Young Universit\ ISSN 0017-3614 Volume 54 31 October 1994 No. 4 Great Basin Naturalist 54(4), © 1994, pp. 291-300 MYCORRHIZAL COLONIZATION, HYPHAL LENGTHS, AND SOIL MOISTURE ASSOCIATED WITH TWO ARTEM/S/A TRIDENTATA SUBSPECIES James D. Trent^, Tony J. Svejcar^, and Robert R. Blanks Abstract. — Mycorrhizal tungi are thought to benefit associated plant species via enhanced nutrient uptake and/or improved water relations. However, detailed descriptions of the components of mycorrhizal colonization and mycor- rhizal hyphal growth are not availalile for Artemisia tridentata. This species occupies sites characterized by relatively low levels of both soil nutrients and moisture. We studied patterns of vesicular, arbuscular, and hyphal mycorrhizal colo- nization, mycorrhizal hvphal lengths, and soil moisture associated with two subspecies of A. tridentata over a 2-year period. A. tridentata ssp. vaseijana (ATV) is generally associated with more mesic and slightly higher elevation sites compared to A. tridentata ssp. tridentata (ATT). Nearly twice as much precipitation was received the first year compared to the second. In general, there were higher levels of colonization and hyphal lengths associated with ATV than with ATT. The ATV site received slightly more precipitation and was lower in available nutrients than the ATT site. HN^phal lengths and arbuscular colonization appeared more responsive to precipitation than were either vesicular or hyphal col- onization. Hyphal colonization did not necessarily follow the same temporal pattern as hyphal lengths. Thus, mycor- rhizal activity' was greater for the subspecies that received slightly more precipitation and occupied a site lower in avail- able nutrients. Arbuscular colonization and hyphal lengths appeared to be most closely associated with soil moisture and thus plant activit\'. Key words: vesicular-arhuscular nnjcorrhizae, hijphal length, Artemisia tridentata, soil moisture, soil temperature. Understanding ecosystem processes requires 1988). Morphogenesis of arbuscule and vesicle baseline data that describe spatial and tempo- formation should be differentiated when ral variations in microbial mediated processes assessing functionality or dependency of the (Burke et al. 1989). Such information is also plant on VAM on a seasonal basis. The pres- needed to assess the role of mycorrhizae and ence of arbuscules indicates plant-fungal other fungi in native plant communities. In interactions (Hirrel et al. 1978, Allen 1983) native plant communities, vesicular-arbuscu- since arbuscules are the site for P and C trans- lar mycorrhizal (VAM) colonization has been fer between symbionts (Cox and Tinker 1976, shown to vary both seasonally and among Wilcox 1993). Seasonal changes in extramatri- plant species (Read et al. 1976, Rabatin 1979, cal VAM fungal hyphae indicate that plant- Daft et al. 1980, Gay et al. 1982, Allen 1983, fungal interactions are dynamic (Wilcox 1993). Giovannetti 1985, Brundrett and Kendrick Therefore, it is necessary to measure seasonahty lUSDA-ARS, 920 Valle>' Road, Reno, Nevada 89512. ^USDA-ARS, HC 71, 4..51 Hu-\ 205, Bums, Oregon 97720. .\ddress reprint requests and correspondence to this author. 291 292 Great Basin Naturalist [Volume 54 of arbuscular colonization and extraniatrical hyphae to adeqiiateK' assess the changing rela- tionship between symbionts in the field. The Great Basin environment is character- ized by winter precipitation, normally as snow, followed by hot, dry summers (C^omstock and Ehleringer 1992). Root growth is most abun- dant in upper soil horizons in the early spring. Rooting activity diminishes in upper soil hori- zons as root growth follows the soil moisture profile into deeper soil layers (Fernandez and Caldwell 1975). In desert soils, N and P are most abundant in upper soil horizons (West 1991), and their availabilit)' to plants diminishes with decreasing soil moisture. Moisture move- ment from deep roots in moist soils to shallow roots in diy soils could make shallow soil nutri- ents available through the process of hvxlraulic lift (Passioura 1988, Caldwell and Richards 1989, Caldwell et al. 1991). Mycorrhizae could play a role in this process (Richards and Caldwell 1987, Caldwell et al. 1991); however, little is known about seasonal dynamics of mycorrhizae in arid ecosystems. Fitter (1993) suggests that plant root systems evolve in a manner that optimizes the use of plant carbon. Mycorrhizal colonization and the formation of extraniatrical hyphae should also reflect an optimization of plant carbon usage. But to date there is relatively little information on spatial or temporal variation in mycorrhizal activity in the Great Basin. In this study we quantified VAM arbuscular root col- onization, vesicular root colonization, hyphal root colonization, and mycorrhizal hyphal length through the plant growing season for 2 years in Artemisia thdenfata ssp. tridentata and Artemisia tridentata ssp. vaseijana. We have characterized the seasonality of the above parameters and show their relationship to changes in both soil moisture and temperature. Study Site Description An Artemisia tridentata ssp. tridentata (ATT) and an Arte7nisia tridentata ssp. vaseyana (ATV) plant community were chosen for study. The study sites are located approximately 30 miles northwest of Reno, Nevada, and are within 3 miles of each other. The ATT community is at 1555 m elevation and is composed of the fol- lowing vegetation: Artemisia tridentata ssp. tridentata, Chrysothamnus viscidiflorus, Ephedra viridis, Stipa cnmata, S. thnrJjeriana, Oryzopsis hymenoides, Elymus hystrix, and Broinus tectorum. The soil is classified as a coarse-loamy, mixed, mesic Aridic Argixeroll. This is an alluvial fan soil that was mainly derived from granitic rocks. The soil consists of about 40 cm of loamy sand to gravelly loamy sand overlying a subsoil of about 50 cm of sandy loam. The underlying material to over 200 cm is loamy coarse sand. The ATV community is at 1830 m elevation and includes the following vegetation: Artemisia tridentata ssp. vaseyana, Purshia tridentata, Ribes sp., Chrysothatnnus viscidiflorus, Stipa cohnnhiana, S. occidentalis, Elymus hystrix, and Bromits tectorum. The soil is classified as a coarse-loamy, mixed, frigid Ultic Argixeroll. This type of upland soil formed in a residuum from granodiorite and consists of about 60 cm of gravelly coarse sand and loamy coarse sand overlying a subsoil of about 30 cm of loamy coarse sand and sandy loam. The underlying material consists of about 10 cm of weathered granodiorite. Methods Sampling was conducted on seven dates in 1989 and five dates in 1990. Four replicate macroplots (20 X 20 m) were randomly selected at each site. One shiTib within a macroplot was selected for sampling each year. We changed shrubs to ensure that the prior year's sampling did not influence measured parameters. Within each macroplot a CRIO micrologger (Campbell Scientific, Logan, Utah)'^ was equipped to measure soil temperature at 10 cm and soil moisture at 10, 30, and 60 cm, adjacent to the target shrub. Soil temperatures were measured with thermocouples, and soil moisture with gypsum blocks. Leader length was measured as a plant growth indicator (Barker and McKell 1986) to avoid destructive hai-vest of Artemisia shrubs. Five marked lead- ers were measured on the target shrub \\'ithin each macroplot at each sampling date. Soil samples were collected with a spade to a depth of 20 cm from within the dripline of the target shrub in each macroplot on each date. Sample volumes were about 8000 cm'^ (20 X 20 X 20-cm cube). Roots were sieved from soil, and both root and soil samples were placed in plastic bags and kept on ice prior to 'Mfiilion of tradt' names does not indieate endorsement bv USDA. 1994] Mycorrhizae in Subspecies of Sagebrush 293 storage at -2°C. Diirins; the sieving process we discarded non-Artemisia roots, which are easily differentiated by color and morphology. In the laboratory roots were washed and cut into 1-cm segments; they were then cleared with KOH and stained with trypan blue (Phillips and Hayman 1970). We estimated per- cent total, arbuscular, hyphal, and vesicular colonization using the gridline-intersect method with a compound microscope at 160X. Hyphal lengths were quantified using the following modified Bethlenfalvay and Ames (1987) procedui-e: (1) 20 g of soil was added to 100 ml of 0.05% trypan blue solution and boiled for 15 min; (2) samples were cooled and 100 ml of sodium hexametaphosphate was added to each flask; (3) flask contents were added to a Waring blender and blended for about 5 sec; and (4) an aliquot was added to a microscope slide and scored for hyphal lengths at 400X. Hyphal lengths were mea- sured on six different aliquots per soil slurry using an improved Neubauer ultra plane counting chamber with a depth of 0.1 mm. Four randomly selected transects from each aliquot were scanned at 400X. Each scan was 7.9 mm long and 0.395 mm wide. From these dimensions the liquid volume scanned was calculated (3.12 X 10"*^ ml/ scan). Slides were scanned and hyi3hal lengths quantified using an image analysis system. Aliquots were averaged prior to statistical analysis. We tried a 1:50 soil to extractant ratio initially, but found we had to use 1:10 because VAM hvqjhal lengths were low. The 1:10 extraction has been shown by Ingham and Klein (1984) to be adequate for measuring hyphal lengths. Since roots were removed before extraction, VAM hyphal lengths pre- sented do not include rhizoplane hyphae. Inclusion of rhizoplane hyphae would no doubt elevate hyphal lengths; however, exclu- sion of roots allows a more accurate depiction of the hyphae that extend beyond the zone of phosphorus depletion around roots. Criteria for determining VAM fungi were similar to those established by Allen and Allen (1986). Most mycorrhizal hyphae are branched, have a knobby appearance, are aseptate, absorb try- pan blue, and are about 3-10 ^tm in diameter. Soil samples collected from each site were composited each year for soil chemical and physical analyses. Nitrate, NH4+, and 804"^ were extracted using 0.1 M KCl; phosphorus was extracted using sodium bicarbonate (Olsen and Sommers 1982); organic carbon was deter- mined using the Walkley-Black procedure (Nelson and Sommers 1982); and particle-size analysis was determined using standard meth- ods (Gee and Bauder 1982). Data were analyzed by year. Artemisia sites and sampling date effects were assessed using analysis of variance with SAS (Statistical Analysis System). All VAM root colonization data were transformed by taking the arcsine and square root of VAM root intersects per total root intersects prior to conducting analy- sis of variance. Results The ATT site had significantly higher levels of bicarbonate extractable P, KCl extractable SO4--, and KCl extractable NO3- than the ATV site (Table 1). ATT soils had a significant- ly lower proportion of sand and significantly more silt and clay than ATV soils. Organic car- bon and NH4+ were not significantly different between the two sites. Neither soil chemical nor physical characteristics were significantly different between years (Table 1). The signifi- cant Site* Date interaction for extractable P is evident since P increases in the second year at the ATT site, yet decreases at the ATV site. Leader lengths were shorter in both plant communities in 1990 than in 1989 (Fig. 1). In addition, leader length was slightly higher in the ATV community in both years. Decreased leader lengths in 1990 are attributed to lower Table 1. Means and probability values for soil physical and chemical characteristics of the A. tridentata ssp. tri- dentata (ATT) and A. tridentata ssp. vaseijana (ATV) sites in 1989 and 1990. Soil samples used for mycorrhizal char- acterization were used in this analysis, which accounts for the change in textural analysis between years. Inorganic P (bicarbonate extractable), N (KCl extractable), S (KCl extractable), and Walkley-Black soil organic carbon (O.C.) are given in the table {n — 4^). ATT M\ Soil variable 1989 1990 1989 1990 Site % sand 84.2 80.4 91.1 92.3 .006 % silt 11.5 14.8 6.2 5.8 .008 % clav 4..3 4,9 2.8 1.8 .046 P/xgg-1 12.1 14..3 9.8 7.9 .0005 NH4Mgg-i 5.2 5.1 5.6 3.4 .39 N03-Mgg-i 6.6 8.3 4.6 4.0 .03 S04-2/.gg-l 7.8 9.6 4.6 5.2 .001 O.C.mgg-l 9.6 8.8 9.3 10.0 .62 "The effect of year was not significant (P > .05) for any variable, and iiiils in the case of P was the Site and Year interaction significant (P < ,0.5), 294 Great Basin Naturalist [Volume 54 • ATT 89 ATV89 A ATT 90 Q P = A ATV90 -- UJ > s ;A\ I- ? ^ / ^\^ ■■ UJ > UI > // / "^^^ ? // / ^ Ul 4 ^ / y^ UJ / /y/ (- > 0 ^ — 1 1 — —\ 1 ^ 1 1 1 1 1 J J A MONTH Fig. 1. Seasonal change in leader length for Artemisia tridentata ssp. tridentata (ATT) and Artemisia tridentata ssp. vaseijana (ATV) during the 1989 and 1990 growing seasons. Phcnological stage appears at the top of the graph. precipitation received during that year (Fig. 2). Annual precipitation for the growing sea- son was measured beginning in November. In 1989 ATT and ATV communities received 384 and 436 mm of precipitation, respectively. However, during 1990 the ATT and ATV com- munities received only 194 and 206 mm, respectively. The 4-year average precipitation (1984-87) for ATT and ATV was 211 and 256 mm, respectively. The hvo sites were not different from each other in either maximum or minimum temper- ature (Fig. 2). Maximum March temperatures at 10-cm soil depths for both sites and years were 7-13 °C, while minimum temperatures were 2-4 °C. Maximum soil temperatures in July-August at 10 cm for both sites and years were 30-31 °C, while corresponding minimum 35 o o LU 30 - 25 - 20 - 5 - UJ 10 I- 5 - E E, z o 0 -r 100 90 80 70 60 50 a. 40 o UJ 30 a: Q. 20 10 0 J J A S 0 N D J F M A M J 89 DATE 90 Fig. 2. Daily ma.ximum and minimum soil temperatures at the 10-cm soil depth and monthly precipitation for Artemisia tridentata ssp. tridentata (ATT) anJ Artemisia tridentata ssp. vaseijana (ATV) during the 1989 and 1990 grow- ing seasons. 1994] Mycorrhizae in Subspecies of Sagebrush 295 temperatures were 19-21 °C. By mid-October maximum temperatures were 16-19 °C and minimum temperatures were 8-9 °C. During early 1989 percent arbuscular colo- nization was similar for the two Artemisia sub- species but diverged to greater levels in ATV roots by early June (Fig. 3). Arbuscular colo- nization of ATT roots dropped to much lower levels than ATV roots during the midsummer dry period. This was followed by a minimal increase in colonization during fall for both subspecies. In 1990 percent arbuscular colo- nization was consistently lower for ATT than ATV roots (Fig. 4, Table 2). There was also a decrease in colonization from early spring through summer for both subspecies. Overall analysis of variance indicates a significant sea- son and subspecies effect for percent arbuscu- lar colonization in both years (Table 2). In 1989 percent hyphal, vesicular, and total colonization did not significantly change through the season (Fig. 3, statistics in Table 2). However, in 1990 both percent hyphal and total colonization changed significantly through the season (Fig. 4, statistics in Table 2). In both years a significant subspecies effect was ob- served for vesicular and total colonization. Subspecies vaseijana roots had greater levels of colonization than tridentata roots. Mycorrhizal hyphal lengths changed signif- icantly through the season for both years (Fig. 5, statistics in Table 2). In 1989 and 1990 hyphal length more than doubled from March to May for all sites except the ATV site in 1990. Hyphal lengths decreased during summer and remained constant during fall for all sites with the exception of ATV in 1990, which increased slightly during the fall. Mycorrhizal hyphal- lengths were significantly greater at the ATV site when compared to the ATT site for most sampling dates. Soil moistiue depletion data are presented in Figure 6. In both years the ATV site had higli- er soil water potentials for a greater proportion J J A S MONTH J J A MONTH Fig. 3. Monthly changes in percent total, arbuscular, vesicular, and h\phal root colonization for Artemisia tridentata ssp. tridentata (ATT) and Artemisia tridentata ssp. vaseijana (ATV) during the 19S9 growing season. 296 Great Basin Naturalist [Volume 54 J J A MONTH J J A MONTH Fig. 4. Montlily changes in percent total, arhuscular, vesicular, and h\ phal rout colonization for Artemisia tridcntatu ssp. tndentata (ATT) and Artemisia tridenfata ssp. vaseyana (ATV) during the 1990 growing season. Table 2. Site means and results of two-way ANOVAs showing the effect of sagebrush site (Artemisia tridentata ssp. tndentata vs. ssp. vaseyana) and time of year (Date) on mycorrhizal colonization (ARB = arhuscular colonization, VES = vesicular colonization, HYP = hyphal colonization, TOT = total colonization) and mycorrhizal hyphal length (nig-l). Site means'' ATT AT\' ProhahilitN \alues Site Date Site* Date %ARB %VES %HYP %TOT Length 10.1 5.4 3.9 19.4 l.:3 13.1 8.8 4.8 26.7 1.8 - - 1989 .009 .004 .346 .004 .018 .014 .517 .205 .067 .0004 .316 .031 .563 .149 .491 %ARB %VES %HYP %TOT Length 6.1 5.9 3.1 15.1 0.9 13.4 14.6 4.6 32.6 1.7 - - 1990 .0001 .0003 .065 .OOOl .005 .005 .116 .003 .004 .050 .461 .206 .852 .511 .373 ''Site means • averaged over sampling date ior each year (ATI" = Artemisia tridcntatu ssp. tndentata, Xr\' = Artemisia tridentata ssp. vaseyana). 1994] Mycorrhizae in Subspecies of Sagebrush 297 2 — 1 -- Fig. 5. Seasonal changes in mycorrhizal hyphal lengths (m of hyphae/g of soil) in soil adjacent to Artemisia tridentata ssp. tridentata (ATT) and Artemisia tridentata ssp. vaseijana (ATV) during the 1989 and 1990 growing seasons. of the season than did the ATT site. Soil mois- ture dechned more rapidly in 1990 compared to 1989. Soil moisture at 10 and 30 cm fell below -1.5 MPa at both sites by midsummer 1990. In 1989 late summer and early fall precip- itation raised soil water potentials above -0.2 MPa at both sites by October. Discussion The two Artemisia subspecies exhibited consistent differences in both colonization and hyphal density of the associated mycorrhizae. Because the study was observational in nature, we could not separate genetics of two sub- species from the differences in sites they occu- pied. Thus, differences must be attributed to the species/site combination. The ATV site was slightly more mesic than the ATT site, but we did not detect any differences in tempera- ture (Fig. 2). In general, ATV tends to occupy sites of higher elevation or with more summer precipitation than ATT (e.g., Winward 1980). In this study we also found that on a seasonal average the ATV site was lower in available nutrients than the ATT site (Table 1). Nutrient differences may be the result of either plant community differences or the fact that the ATV site had a higher sand content. Higher levels of colonization and greater hyphal lengths at the ATV site compared to the ATT site were thus associated with more available 298 Great Basin Naturalist [Volume 54 J J A S 1989 Fig. 6. Monthly change in soil water potential (-MPa) in Artemisia tridentata ssp. trideiifata (ATT) and Aiiemisia tri- dentafa ssp. vaseijana (ATV) plant communities. Water potentials were measured in 1989 and 1990 at 10-, 30- and 60-cm soil depths with gypsum blocks. soil moisture and lower nutrient availability. These data tend to support the hypothesis that mycorrhizae are more active when moisture is available and nutrients are limiting. Arbuscular colonization is thought to be particularly important to carbon and phospho- rus exchange (Cox and Tinker 1976, Wilcox 1993), and thus this featiue plays a critical role in plant/fungus interaction (Hirrel et al. 1978, Chilvers and Daft 1982, Allen 1983). The two study sites differed in arbuscular colonization during both years. However, the difference was more consistent during 1990 compared to 1989. The difference in P availability between sites was 23% in 1989 and 81% in 1990. Thus, the greatest separation in mycorrhizal activity between sites corresponded to the greatest differences in P availability. Timing of maximal mycorrhizal activity appeared to correspond to aboveground activ- ity during the higher precipitation year of 1989. Previous research has demonstrated that A. tridentata achieves highest photosynthetic rates during late spring when moisture is available and temperatures are generally not limiting (DePuit and Caldwell 1973). The'peak in root growth of A. tridentata also occurs in mid-April to late May (Fernandez and Caldwell 1975). Our measurements of leader growth (Fig. 1) appear to confirm that peak above- ground activity occurred in late spring or early summer during this study. However, during the diy year of 1990 there was no increase in mycorrhizal activity during the spring; rather, activity generally declined. The fact that pre- cipitation and leader growth in 1990 were roughly half of 1989 values indicates that envi- ronmental stress may have limited plant activ- ity in 1990. Plant stress and limited carbon assimilation likely reduced root and/or mycor- rhizal activity. Because mycorrhizal activity can be a rather substantial carbon cost to the 1994] Mycorrhizae in Subspecies of Sagebrush 299 associated plant (Chapin et al. 1987), it is not surprising that mycorrhizal colonization and hyphal length would decline during a drought \ear Hyphal lengths appeared to be quite responsive to precipitation patterns within a year. Increases in hyphal lengths in the spring of 1989 were associated with spring precipita- tion. Hyphal lengths declined during the dry summer period and were stable during fall as both sites began receiving precipitation again. During 1990 hyphal lengths were stable dur- ing spring and declined during summer. In late summer and fall of 1990, the ATV site received precipitation and hyphal lengths increased (Figs. 5, 6). However, the ATT site did not receive significant precipitation and hyphal lengths did not increase. It appears that the pattern of arbuscular colonization more closely follows precipitation and hyphal lengths than does either vesicular or hyphal colonization. Arbuscular colonization and mycorrhizal hyphal lengths are more respon- sive to seasonal and yearly variations in plant growth than are either vesicular or hyphal col- onization levels. We suggest, from a functional standpoint, that measurement of mycorrhizal hyphal length and arbuscular colonization is more relevant than is total colonization. Acknowledgments We thank Dr M. Allen and Dr. C. Friese for reviewing the manuscript. We also thank Jill Holderman for technical assistance. Literature Cited Allen, M. F. 1983. Formation of vesiculai-arbuscular mycorrhizae in Atriplex gardneri (Chenopodiaceae): seasonal response in a cold desert. Mycologia 75: 773-776. Allen, E. B., and M. E Allen. 1986. Water relations of xeric grasses in the field: interactions of mycorrhizas and competition. New Phytologist 104: 559-571. Barker, J. R., and C. M. McKell. 1986. Differences in big sagebrush {Artemisia tridentata) plant stature along soil-water gradients: genetic components. Journal of Range Management 39: 147-151. Bethlenfalvav, G. J., AND R. N. Ames. 1987. Comparison of two methods for quantifying extraradical myceli- um of vesicular-arbuscular mycorrhizal fungi. Soil Science Societv' of America Journal 51: 83-1-837. Brundrett, M. C, and B. Kendrick. 1988. The mycor- rhizal status, root anatomy, and phenology of plants in a sugar maple forest. Canadian Journal of Botany 66: 11,5.3-1173. Burke, I. C, W. A. Reiners, and D. S. Schimel. 1989. Organic matter turnover in a sagebrush steppe land- scape. Biogeochemisti-y 7: 11-31. Caldwell, M. M., and J. H. Richards. 1989. Hydraulic lift: water efflux from upper roots improves effec- tiveness of water uptake by deep roots. Oecologia 79: 1-5. Caldwell, M. M., J. H. Richards, and VV. Beyschlag. 1991. Hydraulic lift: ecological implications of water efflux from roots. Pages 423—436 in D. Atkinson, ed., Plant root growth: an ecological perspective. Blackwell Scientific Publications. Chapln, E S., Ill, A. J. Bloom, C. B. Field, and R. H. Waring. 1987. Plant responses to multiple environ- mental stresses. Bioscience 37: 49-57. Chilvers, M. T, and J. E Daft. 1982. Effect of low tem- peratures on development of the vesicular-arbuscu- lar mycorrhizal association between Glomus cidedo- niitm and Allium cepa. Transactions of the British Mycological Society 79: 153-157. COMSTOCK, J. R, and J. R. Ehleringer. 1992. Plant adap- tation in the Great Basin and Colorado Plateau. Great Basin Naturalist 52: 195-215. Co.x, G., and P B. Tinker. 1976. Translocation and trans- fer of nutrients in vesicular-arbuscular mycorrhizas. I. The arbuscule and phosphorus transfer: a (iuanti- tative ultrastructural study. New Phytologist 77: 371-378. Daft, M. J., M. T Chilvers, and T. H. Nicolson. 1980. Mycorrhizas of the liliiflorae. I. Morphogenesis of Eiidymion nonscriptus (L.) Garcke and its mycor- rhizas in nature. New Phytologist 85: 181-189. DePuit, E. J., and M. M. Caldwell. 1973. Seasonal pat- tern of net photosynthesis of Artemisia tridentata. American Journal of Botany 60: 426-435. Fernandez, O. A., and M. M. Caldwell. 1975. Phenology and dynamics of root growth of three cool semi-desert shrubs under field conditions. Joiunal of Ecology 63: 703-714. Fitter, A. H. 1993. Characteristics and functions of root systems. Pages 3-24 in Y. Waisel, A. Eshel, and U. Kalkafi, eds., Plant roots: the hidden half. Marcel Dekker, Inc., New York, Basel, Hong Kong. Gay, R E., R J. Grubr, and H. J. Hudson. 1982. Seasonal changes in the concentrations of nitrogen, phospho- rus and potassium, and in the density of mycorrhiza, in biennial and matrix-forming perennial species of closed chalkland turf. Journal of Ecology 70: 571-593. Gee, G. W, and J. W Bauder. 1982. Particle size analy- sis. Pages 383^09 ;/i A. Klute, ed.. Methods of soil analysis. Part 1. Physical and mineralogical methods. 2nd edition. Agronomy No. 9, Part 2. American Soil Association and Soil Science Society of America publication. Madison, Wisconsin. GioVANNETTi, M. 1985. Seasonal variations of vesicular- arbuscular mycorrhizas and endogonaceous spores in a maritime sand dune. Transactions of the British Mycological Society 84: 679-684. Hirrel, M. C., H. Mehravaran, and J. W. Gerdemann. 1978. Vesicular-arbuscular mycorrhizae in the Chenopodiaceae and Cruciferae: Do they occur? Canadian Journal of Botany .56: 2813-2817. Ingh.\m, E. R., and D. a. Klein. 1984. Soil fungi: rela- tionships between h>q5hal activity' and staining with fluorescein diacetate. Soil Biolog>' Biochemistiy 16: 273-278. 300 (iHEAT Basin Naturalist [Volume 54 Nelson, D. W, and L. E. Sommers. 1982. Total carbon, organic carbon, and organic matter. Pages .539-577 in A. L. Page. R. H. Miller, and D. R. Keeney, ed.s., Methods of soil analysis. Part 2. Chemical and micro- biological properties. 2nd edition. Agronomy No. 9, Part 2. American Soil Association and Soil Science Society of America publication. Madison, Wisconsin. Olsen, S. R., and L. E. Sommers. 1982. Phosphorus. Pages 403-427 in A. L. Page, R. H. Miller, and D. R. Keeney, eds.. Methods of soil analysis. Part 2. Chemical and microbiological properties. 2nd edi- tion. Agronomy No. 9, Part 2. American Soil Associ- ation and Soil Science Society of America publication. Madison, Wisconsin. Passioura, J. B. 1988. Water transport in and to roots. Annual review. Plant Physiologist 39: 245-265. Phillips, J. M., and D. S. Hayman. 1970. Improved pro- cedures for clearing roots and staining parasitic and vesicular-arbuscular mycorrhizal fungi for rapid assessment of infection. Transactions of the British Mycological Society 55: 158-161 R-VBATlN, S. C. 1979. Seasonal and edaphic variation in vesicular-arbuscular mycorrhizal infection of grasses by Glomus tenuis. New Phytologist 83: 95-102. Read, D. J., H. K. Koucheki, and J. Hodgson. 1976. Vesicular-arbuscular mycorrhiza in natural vegeta- tion systems. I. The occurrence of infection. New Phytologist 77: 641-653. Richards, J. H., and M. M. Caldwell. 1987. Hydraulic lift: substantial nocturnal water transport between soil layers by Artemisia tridentata roots. Oecologia 73: 486-489. West, N. E. 1991. Nutrient cycling in soils of semiarid and arid regions. Pages 295-332 in J. Skujins, ed., Semiarid lands and deserts: soil resource and recla- mation. Marcel Dekker Inc., New York. WiLCO.X, H. E. 1993. Mycorrhizae. Pages 731-767 in Y. Waisel, A. Eshel, and U. Kalkafi, eds.. Plant roots; the hidden half Marcel Dekker, Inc., New York, Basel, Hong Kong. Winward, a. H. 1980. Ta.xonomy and ecology of sage- brush in Oregon. Oregon State University Station Bulletin 642. Received 13 July 1993 Accepted 20 April 1994 Great Basin Naturalist 54(4), © 1994, pp. 301-312 SOIL AND VEGETATION DEVELOPMENT IN AN ABANDONED SHEEP CORRAL ON DEGRADED SUBALPINE RANGELAND James O. Kleniniedsoni and Arthur R. Tiedemann^ Abstract. — Vegetation and soils inside and outside an abandoned sheep corral on degraded subalpine range of the Wasatch Plateau were studied to determine the influence of appro.ximately 37 years' use of the corral on soil and plant development. Vegetal and surface cover were estimated. Herbage, litter, and soils were sampled inside and outside the corral and analyzed for Cg^„, N, P, and S. Soil pH, bulk density, and CO3-C also were measured. Storage (mass/unit area) of C^rg. N, P, and S was determined for each component. Yield and vegetal composition were significantly affected inside the conal boimdary. Herbage yield was 2.2 times greater, litter mass 16 times greater, foliar cover of grasses 2 times greater and forb cover 70% lower inside than outside the corral. Cover of meadow barley {Hordeum brachyantherum), a component of the predisturbance vegetation of the Wasatch Plateau, was nearly 12 times greater inside than outside the corral. These and other vegetal and cover differences reflect inside-outsidc differences in concentration, storage, and availability of soil Co^jt, N, P and S. Concentrations of Cq^„ and total and available N, P and S were greater in the surface 5 cm of soil inside the corral. Available P inside the coiTal was much higher in all soil layers. Because of bulk density dif- ferences, storage was greater inside the corral only for Cy^jr and N at 0-5 cm and for P at 5-15 cm. Lower soil pH inside the corral appears related to soil P distribution and CO3-C storage. Results suggest a need to reexamine earlier conclu- sions that tall forbs are the climax dominants of the Wasatch summer range. Key words: summer range; sod Cy^„, N, P, S, CO^-C, pH, and bulk density; plant composition and cover; biomass yield: litter After 35 years of destructive grazing by cattle and sheep in the late 1800s, the subalpine range of the Wasatch Plateau east of Ephraim, Utah, was in extremely poor condition (Reynolds 1911, Sampson and Weyl 1918, Sampson 1919). Erosion and alteration of vegetal cover reached such severe proportions that most of the soil A horizon was lost to erosion, and mud-rock floods were a common occurrence in the canyons leading to valleys and settle- ments at the base of the Wasatch Front (Reynolds 1911, Croft 1967). In some places only subsoils remained when control of graz- ing was finally achieved with establishment of the Manti National Forest in 1903 (Reynolds 1911, Sampson and Weyl 1918, Ellison 1949). Although condition of the range improved steadily over the next several decades, most of the summer range was still unstable in 1950, and accelerated erosion was continuing but at greatly reduced rates (Ellison 1954, Meeuwig I960).' Under moderate grazing secondary succes- sion occurred from 1903 to about 1940 when it slowed perceptibly (Ellison 1954). Since then succession has been extremely slow (Johnson 1964, Intermountain Research Station, Ogden, Utah, unpublished data). Our observa- tions suggest that soil and vegetal conditions have essentially stabilized since Ellison s last obsei-vations in the mid-1950s. We believe the slow rate of succession and range improve- ment in the Wasatch subalpine since then is directly attributable to extreme amounts of soil loss and relatively low fertility of soils that remained after the period of degradation. Based on examination of numerous soil profiles on the plateau and those of similar soils else- where, we believe at least 50%, and possibly as much as 80 or 90%, of the A horizon was lost from this summer range via accelerated erosion. Such a loss would certainly remove a large portion of the soil's organic matter and nutrient capital and significantly alter produc- tive potential. In recent years we have pursued this hypothesis with several studies. This paper reports the results of a fortuitous observational study designed to demonstrate the effect of organic matter and nutrient additions over time on development of soils and vegetation of the Wasatch subalpine range. During field ^School of Renewable Natural Resources, University of Arizona, Tucson, Arizona 85721. ^Pacific Northwest Research Station, La Grande, Oregon 97850. 301 302 Great Basin NATURy\LisT [Volume 54 studies in 1988, we happened upon ;in aban- doned and dilapidated sheep corral that obvi- ously had not been used in many years (Fig. 1). No remains of manure were present inside the conal; vegetation and litter development were advanced and perennial grasses were abun- dant (Fig 2A). The contrast with vegetation and litter outside of what remained of the cor- ral fence was striking (Fig. 2B). The corral offered an opportimity to docu- ment effects of use of the corral (1936-73) and inputs of organic matter and nutrients via sheep manure during its use to soil and vegetal development inside the corral boundaiy. Study Area The Buck Ridge corral study site (39°15'N, lir26'W) is located about 18 km east of Manti, Utah, on Cheriy Flat adjacent to Buck Ridge Road and about 1.6 km east of Skyline Drive. This location is 5 km south of the Alpine Station and the well-known and stud- ied Watersheds A and B of the Great Basin Experimental Range established by Dr. Arthur W. Sampson in 1912 (Sampson and Weyl 1918, Meeuwig 1960). Cheriy Flat is typical of the crest of the Wasatch Plateau, which is about 3150 m elevation. The plateau is long, narrow, and oriented approximately north and south with riblike ridges extending east and west. The top of the plateau is gently rolling to nearly level. Average annual precipitation is about 840 mm; two-thirds of this falls as snow between November and April. Precipitation averages 173 mm during the summer months (June through September) but varies considerably. Mean annual temperature is about 0°C (Ellison 1954). In the vicinity of the corral, Cherr>' Flat has a gentle 2% slope to the east; microtopography is smooth. Soil parent materials are of the Flagstaff Formation (Stanley and Collinson 1979) that crop out over about 7200 km- in central Utah (Schreiber 1988). Dominant lith- ology is freshwater lacustrine limestone and Fig. 1. Reiiuiiiis of Buck Ridge tonal as it appeared in 1989. 1994] Soil and Vegetation Development on Subalpine Range 303 . " ' > ' * . B-'.. ■ -■-. ^~ 1 .. 1 i^ ■^- '' . ' Vv"'^*"' ■■ ■ • >-,j'J..-»k ' '.^ '■ *;^'- -'*• "V ''*'■-- ■*""'«•• 5 ■-'^''~ ■ y"' ' '^l . , V f r-' ' ■, , •' ■'"'?', '•_■* ■• -V. 'l. 5i 3 r-^i'^'^ :.\\^:*^^ Fig. 2. Close-up view of vegetation and groundcover inside (A) and outside (B) Buck Ridge corral in 1989. calcareous shales with minor interbeds of sand- stone, oil shale, conglomerate, gypsum, and volcanic ash (Weber 1964, Schreiber 1988). Soils in this region of the plateau are mostly fine, mixed Argic Cryoborolls, but lithic, pachic, and vertic Ciyoborolls also are present. They are shallow to moderately deep; the sub- soils are silty clays or clay loams. Thickness of the A horizon averages about 4 cm; the B hori- zon averages approximately 52 cm thickness. Based on typical profile descriptions (H. K. Swenson, Soil Conservation Service, Boise, Idaho, personal communication), these rela- tive horizon thicknesses suggest that much of the original A horizon was lost by wind and water erosion following the period of unre- stricted grazing prior to 1903. Vegetation of the Wasatch Plateau is chiefly herbaceous, but small patches of Engelmann spruce {Picea engelmannii) and subalpine fir {Abies lasiocarpa) occupy steep northerly exposures of east-west ridges and dot the plateau landscape. Because there were no remnants of the original pristine vegetation (Ellison 1949, 1954), opinions differ regarding its exact character. Ellison (1954) describes the original plant community as mixed-upland herb dominated by tall forbs, while Sampson (1919) considered wheatgrasses to be the pri- mary species of the herbaceous climax (i.e., what he referred to as summer range). Based on file records and discussion with former permittees, we determined Buck Ridge corral was built and first used in 1936. It was last used about 1973 (Ed Shoppe, Manti-LaSal National Forest, Ephraim, Utah, personal communication). Hence, the corral was used annually by sheep for about 37 years. During this period undetermined and variable amounts of organic matter and nutrients were added annually via dung and urine, depending on the number and size of bands using the 304 Great Basin Naturalist [Volume 54 corral and frequency of use. For the past 10 years, the Buck Ridge allotment, comprising 4235 ac, has been grazed by sheep at the rate of 3.3 ac/AUM from 1 July to 30 September. Methods As a basis for conducting this observational study, we first assured ourselves that areas inside and outside the corral boundary were initially alike in every respect and that site characteristics had no bearing on the precise location of the corral at the outset or on the findings. There were no differences in topog- raphy (or microtopography) and no evidence that other state factors (climate, biotic factor, parent material) differed within the small study site (about 0.75 ha) of the corral area. To assess the differential effect of nutrient addi- tions inside and outside the corral, we sampled vegetation, litter, and soil in midsummer. Moisture conditions were dry at the time and little grazing had occuned inside or outside die corral. Present condition of the corral fence (Fig. 1) indicates that sheep have had near equal access to both sides of the corral bound- ary for many years, but there is no record of how long fence cross rails have been down. Cover by species (foliar projection), litter, soil, and rock were estimated in 10 randomly located 0.5-m^ plots inside and outside the corral (within 10 m of the corral boundary). Herbage and litter were hai'vested in the same plots, oven-dried (70 °C), and weighed. Six randomly located soil pits were sampled inside and outside the corral. At each pit, soil cores (5.197 cm dia.) were collected from the 0-5-, 5-15-, and 15-30-cm layers. Plants and litter were ground to pass through a 0.425-mm sieve. Soils were air-dried, sieved to remove the >2-mm fraction, and then ground to pass through a 0.150-mm sieve. Plant and soil samples were analyzed for total N by semi-micro-Kjeldahl (Bremner and Mulvaney 1982) and total S by dry combustion (Tiedemann and Anderson 1971) in a LECO high-frequency induction furnace (LECO Corp., St. Joseph, Michigan). Plant and soil samples were analyzed for total C by dry com- bustion (Nelson and Sommers 1982) in the LECO high-frequency induction furnace. Organic C (C^^g) of soils was determined by correcting total C for carbonate-C as deter- mined by a gasometric method (Dreimanis 1962). Total P was determined in plant materi- al using the vanado-molybdo-phosphoric yel- low color method after dry-ashing (Jackson 1958) and in soils using ascorbic acid color development (Olsen and Sommers 1982) fol- lowing hydrofluoric acid digestion (Bowman 1988). Available nutrients in soils were deter- mined as follows: P using ascorbic acid color development following 0.5 M sodium bicar- bonate extraction (Olsen and Sommers 1982), N by steam distillation of 2 N KCl extracts (Keeney and Nelson 1982), and S with 1:1 water extracts, followed by ion chromatogra- phy (Dick and Tabatabai 1979). Tests of significance for difference between inside and outside values for all variables stud- ied were carried out with the t test. We recog- nize the desirability of replicating the inside- outside corral comparison. Unfortunately, that was not a design feature we could control; other abandoned corrals — even less than 37 years age — simply do not exist on this summer range. Results Vegetation Obvious visual differences in vegetation, litter, and soil surface conditions inside and outside the old corral (Figs. 2A, 2B) were con- firmed in the data (Table 1). Herbage yield in- side the con-al was 2.2 times greater than out- side. Although total herbage cover inside and outside the corral was the same (65%), grasses comprised a much greater percentage of foliar cover than did forbs inside than outside the corral. Three perennial grasses {Agropyron trachycaiiliim, Hordeiiin hrachijanthcrwn, and Stipa letiennani) dominated vegetal cover inside the corral (Table 2); outside the corral A. trachijcmdwn and S. leffcrmani were equal- ly important, but H. hrachyanthcrum was unimportant. Forbs were represented by 7 species inside the corral and 11 species outside (Table 2). Taraxacum officinale and Achillea miUefolium were the dominant forbs inside and outside the corral, but their cover outside was much greater than inside. No other forb species con- stituted more than 4% of the herbage compo- sition. Soil surface protection by litter differed markedly inside and outside the corral. Mass of litter inside the corral was 16 times greater 1994] Soil and Vegetation Development on Subalpine Range 305 Table L Influence of long-term corral effects on vegeta- tion, litter, and soil surface characteristics at Buck Ridge. Component and attribute Inside Outside Difference sign, at P < Herbage yield (g m--) L39 ± 2L' 64 ±9 .005 Litter mass (g m~^) 312 ± 78 19 ±2 .005 Foliar cover (%) Grasses 55 ±5 28 ±5 .001 Forbs 11 ±3 37 ± 8 .05 Total 66 ±3 65 ±7 NS Basal cover (%) Litter 71 ±4 18 ±4 .001 Bare groimd + rock 3±2 25 ±4 .005 Table 2. Percentage of species composition (by foliar cover) of vegetation inside and outside Buck Ridge corral. Percentage ".Mean ± standard error; n = 10. than outside, while cover of Htter was 4 times greater (Table 1). This is consistent with the 12-fold difference in bare ground between inside and outside locations. Only 2% of the soil surface was bare inside the conal. Nutrients Concentrations of all nutrients studied were influenced by dung and urine accumulation inside the corral (Table 3). Nitrogen concentra- tion was higher in the herbage, litter, and 0-5- cm soil layer, but lower in the 15-30-cm soil layer inside than outside the coiTal. Concentra- tion of C^j.„ was parallel to that of N for the litter and soil layers, while P concentration was higher inside than outside only for litter and the 0-5- cm soil layer (Table 3). Concentration of S was higher inside than outside only for the upper soil layer. Storage (mass/unit area) of all four nutrients was significantly greater inside than outside the corral for herbage and litter components (Table 4). Amounts of Cqj.„ were greater inside the conal in the surface soil, but lower in the 15-30-cm soil layer. Storage of N was greater inside the corral in the 0-5-cm soil layer, while storage of P was greater inside the cor- ral in the 5-15-cm soil layer. Availability of P was much higher inside than outside the corral in all soil layers (Table 5). Availability of N and S was significantly higher (P < .10 ) inside the corral only in the 0-5-cm soil laven Inside Outside Grasses Agropijron trachijcauhim 8.2 Alopecunis pratensis Brorntis carinatm 4.0 Hordeuin hrachijantherwn 35.6 Poa pratensis Stipa Columbiana 0.4 Stipa lettermani 30.4 Total grasses 78.6 Forbs Achillea millefolium Androsace septentrionalis Artemisia ludoviciana v. imcompta Aster foliaceus v. canhyi Cijmopteris lemmonii Descurania richardsonii Erigeron ursinus Gilia aggregata Lesqucrella utahensis Polygonum gkindulosa Ranunculus inamoeniis Rtmiex mexicanus Taraxacum officinale Viola nuttallii v. nuttallii Total forbs 8.8 0.3 0.9 0.2 0.3 3.5 7.4 21.4 10.3 1.6 3.0 0.7 26.5 42.1 18.0 0.3 2.7 2.4 0.2 0.1 2.6 0.2 2.1 29.2 0.1 57.9 Discussion and Conclusions Development of vegetation and stabiliza- tion of the soil surface at Buck Ridge corral were indeed striking considering the slow pace of secondary succession of the Wasatch summer range from 1903 to 1940 (Ellison 1954, Meeuwig 1960) and the apparent lack of trend since 1940. Values for herbage production, litter mass, and cover data portray control of the soil surface inside the corral and are in marked contrast to conditions outside the corral and the surrounding summer range, which is still relatively unstable and subject to accelerated erosion. We believe the vegetation trend ob- served inside the corral has occurred within a relatively short time — no more than 20 years, assuming vegetal development did not com- mence until after abandonment of the corral. This is not to say that ephemerals did not occupy the corral annually between periods of use, only to be eliminated during use. The large changes in vegetation and soil surface conditions are consistent with changes in nutrient status of the soil-plant-litter system 306 Great Basin Naturalist [Volume 54 Table 3. Conct'Titration ot C,„.j,, \, R and S in components of tlic soil-plant-litter system inside and outside Buck Ridge corral. Difference Nutrient Component Inf iide Outside sign, at P < -gku-'- ^org Herbage 447.4 ± 7.P' 433.0 ± 4.2 NS Litter 420.0 ±8.3 448.9 ± 7.2 .025 Soil, 0-5 cm 134.2 ± 17.6 45.6 ± 2.2 .001 5-15 cm 34.8 ±2.7 39.5 ± 1.4 NS 15-30 cm 21.8 ±2.5 36.4 ± 2.0 .005 N Herbage 21.75 ± 0.75 15.50 ± 0.82 .001 Litter 19.48 ±0.83 11.77 ± 0.77 .001 Soil, 0-5 cm 14.05 ± 1.95 3.39 ± 0.21 .001 5-15 cm 2.99 ±0.23 3.05 ± 0.10 NS 15-30 cm 1.89 ±0.17 2.73 ± 0.17 .005 P Herbage 2,23 ±0.10 2.22 ± 0.21 NS Litter 1.99 ±0.12 1.47 ± 0.08 .001 Soil, 0-5 cm 2.36 ±0.22 1.61 ± 0.06 .01 5-15 cm 1.70 ±0.13 1.44 ± 0.08 NS 15-30 cm 1.36 ±0.13 1.39 ± 0.09 NS S Herbage 1.29 ± 0.04 1.22 ± 0.04 NS Litter 1.10 ± 0.07 0.96 ± 0.10 .05 Soil, 0-5 cm 1.42 ±0.18 0.88 ± 0.08 .025 5-15 cm 0.69 ± 0.22 0.74 ± 0.11 NS 15-30 cm 0.68 ±0.11 0.71 ± 0.05 \S •'Mean ± standard error; n = 6. over the life of the corral and subsequent to its abandonment (Crocker and Major 1955, Olson 1958, Blackmore et al. 1990). Unfortunately, much of the corral history (actual herd use, inputs of dung and urine, and character of vegetation that occupied the corral between periods of use by sheep) was not documented. However, changes in concentration and accu- mulation of nutrients inside the corral seem reasonable, based on what might be expected from traditional use of a subalpine corral by sheep for 37 years, experience from other grazed systems (Blackmore et al. 1990, Scholes 1990), and an understanding of the chemistiy of the elements studied here. We can be confi- dent that concentrations of soil C,j,.j, and N in- side the corral have declined since abandon- ment and change in the biotic factor (Jenny 1941). However, whether a new steady state has been reached yet is conjectural (Jenny 1941, Tiedemann and Klemmedson 1986), even though observed concentrations of soil C,,,.^, and N are within the range for comparable undisturbed soils (Retzer 1956, Youngberg and Dyrness 1964). Because C, N, F, and S are ubiquitous in soil organic matter and its precursors (Stevenson 1986), close association among these elements should be expected in components of the Buck Ridge soil-plant-litter system, especially be- tween Cy,.„ and N because soil N is almost entirely organic (i.e., about 98%). On the other hand, because of certain dissimilarities in the chemistry of these four nutrients and differ- ences among them in physiological separation into dung and mine pathways (C and P entirely via dung, N and S predominately via urine; Floate 1970, O'Connor 1981, Barrow 1987, Sagger, MacKay et al. 1990), certain differences in nutrient accumulation patterns can be expected. Close association of Cg^g and N was appar- ent even in the 15-30-cm soil layer where concentration of C^,,.;, and N and amount of C,„.„ were lower inside than outside the corral. Such an unexpected difference at this depth lacks explanation, certainly none related to corral effects. It is not consistent with the large differences in C^j.^ and N in the litter and 0-5-cm soil layer where effects of the cor- ral would be most expected. Parent material seems the most likely cause of this difference, but samples from the study site were uniform in CO3-C and varied randomly in N, P and S. However, limestone and shales are noted for spatial variation, even within very short 1994] Soil and Vegetation Development on Subalpine Range 307 Table 4. Storage of Co^g, N, F, and S in components of the soil-plant-litter system inside and outside Buck Ridge corral. Difference Nutrient Component Inside Outside sign, at P < kgm-2 --. r Herbage 0.064 ± 0.00?' 0.028 ± 0.004 .001 Litter 0.129 ±0.032 0.009 ± 0.001 .005 Soil, 0-5 cm 3.46 ±0.17 1.84 ±0.13 .001 5-15 cm 4.44 ± 0.61 3.94 ±0.25 NS 15-30 cm 4.30 ± 0.62 6.44 ±0.63 .05 Soil total'^ 12.19 ± 1.29 12.22 ±0.80 NS Total s\'stem 12.34 ± 1.27 12.25 1-2 ±0.80 NS gm 1 -------- N Herbage 3.17 ± 0.27 1.03 ±0.16 .001 Litter 6.23 ± 1.68 0.23 ±0.04 .001 Soil, 0-5 cm 348 ± 16 138 ± 10 .001 5-15 cm 379 ± 50 304 ± 17 NS 15-30 cm 370 ± 46 485 ±50 NS Soil total 1079 ± 96 927 ±62 NS Total system 1115 ± 105 928 ± 152 NS P Herbage 0.33 ± 0.04 0.15 ±0.03 .001 Litter 0.66 ± 0.18 0.03 ± <0.01 .005 Soil, 0-5 cm 63 ±4 64 ±3 NS 5-15 cm 216 ± 30 145 ± 14 .10 15-30 cm 231 ± 43 250 ±32 NS Soil total 551 ± 70 460 ±44 NS Total system 552 ± 70 461 ±44 NS S Herbage 0.19 ± 0.02 0.08 ±0.01 .001 Litter 0.37 ± 0.11 0.02 ± <0.01 .01 Soil, 0-5 cm 37 ±2 35 ±3 NS 5-15 cm 90 ± 16 77 ± 14 NS 15-30 cm 134 ± 25 128 ± 18 NS Soil total 261 ± 42 241 ±33 NS Total system 262 ± 42 241 ±33 NS ■'Mean : ; n = 10 for herbage and litter. 6 lor soil components. distances (C. F. Lohrengel, Department of Geology, Snow College, Ephraim, Utah, per- sonal communication). Of 23 rock samples from the near vicinity (4-km radius) classified bv Schreiber (1988), P concentration ranged 3i-fold, with a C.V. of 1.46. The sample high- est in P content, an organic-rich shale, burned under a match flame. PhosphoiTis is relatively immobile but should accumulate in soils over time where P inputs exceed removal in grazed herbage (Sagger, Hedley et al. 1990). This would characterize the situation in Buck Ridge corral with the large inputs of animal excreta from 1936 to 1973. Moreover, decomposition of this material should facilitate P mobility, especially after pul- verization by hoof action (Bromfield and Jones 1970). Although urine and dung hydrolyze rapidly causing NH4 to accumulate and pH to rise, nitrification quickly takes over and, in the case of urine, within days pH will drop below control levels (Doak 1952, During et al. 1973, Haynes and Williams 1992). Indeed the initial impact of decomposition of most plant materials is an increase in bulk pH (Williams and Gray 1974). However, products of organic decay are predominantly acid; hence, acidification even- tually dominates. Those horizons or soil layers that contain the products of primary decompo- sition, in this case the litter and 0-5-cm layer (Table 6), will show the greatest acidity (Swift et al. 1979) and a tendency for enhanced solu- bility and mobility of P Significantly lower carbonate-C of soil inside than outside the corral (Table 6) manifests increased soil acidity inside the corral. James Clayton (personal communication, Intermountain Research Station, Boise, Idaho) suggests the CO3-C dif- ference inside and outside the corral is reason- able, based on estimated H"*" supplied by nitrification of urea and organic matter decomposition over a period of 37 years. 308 Great Basin Naturalist [Volume 54 1 ABLF 5. Concentration of available soil N, P, and S inside and outside Buck Ridge corral. N P s Soil Layer Inside Outside Diff. sign. P < Inside Outside Dill, sign. P< Inside Outside Diff. sign. P < 0-5 cm 5-15 cm 15-30 cm mgkg^' 105 ± 26" 52 ± 11 29 ± 7 20 ± 3 13 ± 3 13 ± 2 .10 NS NS mg kg-l 158 ± 19 57 ± 6 142 ± 16 22 ± 4 70 ± 1 1 10 ± 3 .001 .05 .001 mg kgr^ 50 ± 13 22 ± 3 24 ± 2 22 ± 1 15 ± <1 14 ± 1 .10 NS NS .standard cnor; » = 6. The P distribution pattern described here is similar to that found by Sagger, MacKay et al. (1990), who closely predicted observed P accumulation in soil of sheep pastures. In areas where sheep camped, 85-90% of P accu- mulated in the upper 15 cm of soil was accounted for l)y animal waste. Williams and Haynes (1992) noted significant increases of P in the top 20 cm of soil in pastures grazed by sheep for 38 years and treated with super- phosphate. Nitrogen and S losses from the corral soil- plant system could have been large. Nitrogen may be lost by volatilization of NH3, leaching and surface runoff of NO3, or denitrification under appropriate conditions (Ball et al. 1979, Floate 1981, O'Connor 1981), while SO4 may be lost by surface runoff and leaching, depend- ing on SO4 retention capacity of soils (Sagger, Hedley et al. 1990). Williams and Haynes (1992) assumed most of the S loss they observed (48-73%) was due to leaching. Thus, whether Buck Ridge corral was devoid of vegetation during much of the year and hence subject to leaching and runoff, or whether its use was intermittent so as to permit vigorous growth of ephemerals and uptake of available nutrients, we would not expect mineralized N and S to accumulate in the soil profile. Similarity in P and S accumulation in the 0-5-cm soil layer, in view of higher concentra- tions of these nutrients, is attributed to lower bulk density of the upper soil layer inside the corral (Table 6). By contrast, bulk density of the 5-15-cm layer was significantly greater inside than outside the corral. These opposite trends in adjacent soil layers appear to be due to compaction of the entire upper 15 cm dur- ing 37 years of use of the corral by sheep, fol- lowed by amelioration of this effect in the absence of trampling after the corral was abandoned, especially in the top 5 cm where organic matter was concentrated (Tables 3, 4). Heavy clay subsoils of this site should compact readily. Sommerfeldt and Chang (1985) found that long-term manure treatments reduced bulk density of the upper 15 cm of a cultivated soil by as much as 39%. Increased availability of nutrients in the upper 5 cm of soil of the corral soil is associat- ed with higher concentration of nutrients and the large pool of organic matter in that layer. Carbon/element ratios of all soil-plant compo- nents (Table 7) indicate that conditions gener- ally more favorable for net mineralization of N and P (Stevenson 1986) prevailed inside than outside the corral at the time of sampling. Greater availability of P in all soil layers inside the corral also can be associated with pH in a range that one might expect maximum avail- ability of the labile inorganic P fraction (Stevenson 1986). Coupled with this is low mobility of P, in contrast to N and S, which allows P to be retained in place. Since abandonment of the corral, organic matter would have continued to accumulate, but fi-om a new source, i.e., autotrophic produc- tion of vegetation that presumably developed soon after corral abandonment. Significant im- port of new nutrients since abandonment is unlikely. Almost all nutrients in post-abandon- ment crops of herbage would have been recy- cled from the soil. Presumably, the level of herbage production inside the corral has exceeded that outside almost since abandon- ment owing to higher fertility status of soils inside the corral. The present condition of the corral fence (Fig. 1) would indicate that it has not been a barrier to sheep for many years. Hence, differential grazing probably has played a minor role in vegetal differences inside and outside the corral boundary. Although we have emphasized the role of nutrients in the observed changes, we cannot dismiss the possibility that improvement in moisture-holding capacity of surface soils 1994] Soil and Vegetation Development on Subalpine Range 309 Table 6. Effects of inside and outside positions of Buck Ridge corral on bulk density, pH, and carbonate-C of soil layers. Soil Difference Soil property layer (cm) Inside Outside sign, at F < Bulk density (mg ni"'^) O-o 0.59 ■+■ O.O.S'' 0.84 ± 0.06 .05 5-15 1.26 + 0.04 1.10 ± 0.02 .005 L5-30 1.39 + 0.06 1.26 ± 0.10 ns pH 0-5 6.58 + 0.11 7.23 ± 0.13 .005 5-15 7.15 + 0.06 7.28 ± 0.09 NS 15-30 7.17 + 0.06 7.37 ± 0.09 .10 Carbonate-C concentration (g kg-*) 0-5 5.20 + 0.82 8.92 ± 1.87 .10 5-15 5.35 -H 1.24 9.23 ± 2.10 NS 15-30 4.13 -H 0.94 10.63 ± 2.19 .05 amount (kg m~2) 0-5 0.15 + 0.04 0.38 ± 0.10 .10 5-15 0.65 Hh 0.13 0.74 ± 0.17 NS 15-30 0.77 + 0.15 1.80 ± 0.36 .025 Soil 1.57 + 0.25 3.05 ± 0.54 .05 ''Mean ± standard error, n = 6. Table 7. Carbon-element ratios of soil-plant-litter components for inside and outside positons of Buck Ridge conal. Ratio Component Inside Outside Difference sign, at F < C/N C/P C/S Herbage 19.6 + 0.8^ 30.7 ± 2.6 Litter 22.4 + 1.0 37.2 ± 3.6 Soil, 0-5 cm 10.0 + 0.2 13.3 ± 0.6 5-15 cm 11.7 + 0.5 13.0 ± 0.5 15-30 cm 11.5 + 0.6 13.4 ± 0.6 Herbage 210 + 15 241 ± 44 Litter 241 + 23 308 ± 15 Soil, 0-5 cm 56 ± 5 29 ±2 5-15 cm 21 ± 1 28 ±2 15-30 cm 16 ± 1 27 ±2 Herbage 346 ± 16 368 ± 17 Litter 406 ± 43 439 ± 55 Soil, 0-5 cm 95 + 5 54 ± 5 5-15 cm 53 + 5 61 ± 11 15-30 cm 34 + 3 53 ± 5 .05 .05 .001 NS .05 NS .05 .001 NS .025 NS NS .001 NS .01 ■'Mean ± standard error; n = 10 for lierliage and litter, 6 for soil components. inside the corral also may have influenced the successional trend following abandonment. Comparison of basal cover data for Buck Ridge corral (inside and outside) with that portraying conditions in 1946 for six "relic nat- ural areas" and four stands on Elk Knoll (Elk Knoll Research Natural Area), as described by Ellison (1954), is revealing (Table 8). Similarity between cover of litter, bare ground, and rock at Elk Knoll (3.2 km west-northwest of Buck Ridge corral) in 1946 and in 1989 and what we found outside Buck Ridge corral supports ob- sei'vations that successional trend has been vir- tually static in the last 40-odd years. Although Elk Knoll had been previously grazed, it has been protected fiom grazing, except for wildlife, from 1903 (Ellison 1954) to the present. According to Ellison, the natural areas, which he claimed had never been grazed by domes- tic livestock or had been grazed only lightly for many years, provided a partial description of pristine vegetation in the Wasatch sub- alpine at that time. Vegetal cover data (Table 1) demonstrate a marked trend toward perennial grasses inside the corral. We interpret this as an upward 310 Great Basin Naturalist [Volume 54 Tabli-: 8. Comparison of cover data for Buck Ridge cor- ral with Ellison s data for "relic natural areas and sites at Elk Knoll. Litter Bi ire ground Location cover + rock Buck Ridge corral - - 9f - - Inside corral 71 3 Outsitle 18 25 "Relic natural areas "•' 11-20 11-50 Elk Knoll 194&' 26-31 29-37 1989'' 27-49 17-39 ■'Sec Ellison (19541. '^Klemniedson ;incl Tieck'iiiann. i mp,il,lisl,.-a data trend in succession; vegetation changes are accompanied by greater herbage production, increased htter mass and cover and stal)ihty of the soil surface. These characteristics have been commonly associated with improvement toward high range (ecological) condition (Laurenroth and Laycock 1989). Interestingly herbage composition inside the corral is in marked contrast to that described by Ellison (1954) for his six relic natural areas. He said that "one of the most striking things about the natural areas is the abundance of perennial forbs ; they constituted 70-88% of the vegetation in relic natural areas in 1946. The trend toward perennial grasses (79% of total foliar cover) we have observed inside the corral is quite the opposite of Ellison's compo- sition data and conesponds more to vegetation development of the summer range described by Sampson (1919). The increase in Hordeum bracJiijantJierum also suggests an upward trend. Bodi Ellison (1954) and Sampson (1919) noted the presence of H. nodosum, a synonym misapplied to H. brachyantherum (Hitchcock 1950, Holmgren and Reveal 1966), on the summer range. But only Sampson (1919) dis- cussed successional status; he described H. nodosum as a shallow-rooted species that occupied space between bunched wheat- grasses, the primary species of the subclimax type. He did not list this plant with types of lower developmental stage. Normally where ecosystem degradation has been as severe as experienced here, with almost complete loss of the A horizon, we would expect successional processes in soil and vege- tation to occur simultaneously (Sampson 1919, Crocker and Major 1955, Olson 1958). But, in this case where livestockmen controlled the input of manure and associated effects for 37 years, ecosystem development was essentially one-sided; soil development advanced rapidly for 37 years before sheep use of the corral ceased and development of vegetation was allowed to proceed. The fact that development in vegetation, litter, and soil surface conditions has advanced so far in just 20 years, far out- pacing comparable development outside the corral, even in Ellison's "relic natural areas, " leads us to conclude that soil fertility has been a key factor controlling succession and im- provement of the Wasatch summer range. Acknowledgments This research was supported by Grants 85- CRSR-2-2717 and 91-38300-6156, Range Research Grant Program, USDA-CSRS. The authors gratefully acknowledge the Shrub Improvement and Revegetation Project, Intermountain Research Station, Provo, Utah, for providing facilities for this research; field assistance of Gary Jorgensen, range techni- cian, Intermountain Research Station, and Dr Clyde Blauer, professor, Snow College, both of Ephraim, Utah; and laboratory analysis by Justine McNeil, University of Arizona. Thanks are also due to Drs. James Clayton, Fred Provenza, and E. Lamar Smith for reviews of the manuscript. Literature Cited Ball, R., D. R. Keeney, E W. Theobald, and P. Nes. 1979. Nitrogen balance in urine-affected areas of a New Zealand pasture. Agronomy Journal 71: 309-314. Barrow, N. J. 1987. Return of nutrients by animals. Pages 181-186 in R. W. Snaydon, ed.. Managed grasslands. Elsevier, Oxford. Blackmore, a. C, M. T. Mentis, and R. J. Scholes. 1990. The origin and extent of nutrient-enriched patches within a nutrient-poor savanna in South Afiica. Journal of Biogeography 17: 463—470. Bowman, R. A. 1988. A rapid method to determine total phosphorus in soils. Soil Science Societ\' of America Journal 52: 1301-1.304. Bremner, J. M., AND C. S. MuLVANEY. 1982. Nitrogen — total. Pages 595-624 in A. L. Page, ed.. Methods of soil analysis. Part 2. 2nd ed. Agronomy Monograph 9. American Society of Agronomy and Soil Science Society of America, Madison, Wisconsin. Bromfield, S. M., and O. L. Jones. 1970. The effect of sheep on the rec\ cling of phosphorus in hayed-off pastures. Australian Journal of Agricultural Research 21:699-711. 1994] Soil and Vegetation Development on Subalpine Range 311 Crocker, R. L., and J. Major. 1955. Soil development in relation to vegetation and surface age at Glacier Bay, Alaska. Jonrnal of Ecology 43: 427-448. Croft, A. R. 1967. Rainstorm debris floods. Arizona Agricultural E.xperiment Station, Tucson. 36 pp. Dick, W. A., and M. A. Tab.'VTABAI. 1979. Ion chromato- graphic determination of sulfate and nitrate in soils. Soil Science Society of America Journal 43: 899-904. DoAK, B. W. 1952. Some chemical changes in the nitroge- nous constituents of urine when voided on pasture. Journal of Agricultural Science 42: 162-171. Dreimanis, A. 1962. Quantitative gasometric determina- tion of calcite and dolomite by using Chittick appa- ratus. Journal of Sedimentary Petrology 32: 520-529. During, C, W. C. Weeda, and E D. Dorofaefe 1973. Some effects of cattle dung on soil properties, pas- ture production, and nutrient uptake. II. Influence of dung and fertilizers on sulphate sorption, pH, cation-e.xchange capacity, and potassium, magne- sium, calcium and nitrogen economy. New Zealand Journal of Agricultural Research 16: 431-438. Ellison, L. 1949. Establishment of vegetation on deplet- ed subalpine range as influenced by microclimate. Ecological Monographs 19: 97-121. . 1954. Subalpine vegetation of the Wasatch Plateau, Utali. Ecological Monographs 24: 89-124. Flo,\te, M. J. S. 1970. Mineralization of nitrogen and phosphorus from organic materials of plant and ani- mal origin and its significance in the nutrient cycle in grazed upland and hill soils. Journal of the British Grasslands Society 25: 295-302. . 1981. Effects of grazing by large herbivores on nitrogen cycling in agricultural ecosystems. Pages 585-601 in E E. Clark and T. R. Rosswell, eds.. Terrestrial nitrogen cycles. Ecological Bulletin (Stockholm) 33. Haynes, R. J., AND R H. Williams. 1992. Changes in soil solution composition and pH in urine-afFected areas of pastures. Journal of Soil Science 43: 323-334. Hitchcock, A. S. 1950. Manual of the grasses of the United States. 2nd ed. Rev. by Agnes Chase. U.S. Government Printing OfBce, Washington D.C. 1051 pp. Holmgren, A. M., and J. L. Reveal. 1966. Checklist of the vascular plants of the Intermountain region. USDA Forest Service Research Paper INT-32. Intermountain Forest and Range Experiment Station, Ogden, Utah. 160 pp. Jackson, M. L. 1958. Soil chemical analysis. Prentice- Hall, Englewood Cliffs, New Jersey. Jenny, H. 1941. Soil formation. McGraw-Hill Book Co., New York. Johnson, H. B. 1964. Changes in vegetation of two re- stricted areas of the Wasatch Plateau, as related to re- duced grazing and complete protection. Unpublished masters thesis, Brigham Young University, Provo, Utah. Keeney, D. R., and D. W Nelson. 1982. Nitrogen — inor- ganic forms. Pages 643-698 m A. L. Page, ed., Methods of soil analysis. Part 2. 2nd ed. Agronomy Monographs 9. American Society of Agronom\' and Soil Science Society of America, Madison, Wisconsin. Laurenroth, W. K., and W. A. Laycock, eds. 1989. Secondary succession and the evaluation of range condition. Westview Press, Boulder, Colorado. Meeuwig, R. O. 1960. Watersheds A and B — a study of surface runoff and erosion in the subalpine zone of central Utah. Journal of Forestiy 58: 556-560. Nelson, D W, and L. E. Sonimers. 1982. Total carbon, organic carbon and organic matter. Pages 539-580 in A. L. Page, ed.. Methods of soil analysis. Part 2. 2nd ed. Agronomy Monographs 9. American Society of Agronomy and Soil Science Society of America, Madison, Wisconsin. O'Connor, K. F 1981. Comments on Dr Floate's paper on grazing effect by large herbivores. Pages 707-714 in F. E. Clark and T. Rosswall, eds.. Terrestrial nitro- gen cycles. Ecological Bulletin (Stockholm) 33. Olsen, S. R., and L. E. Sommers. 1982. Phosphorus. Pages 404-430 in A. L. Page, ed.. Methods of soil analysis. Part 2. 2nd ed. Agronomy Monographs 9. American Society of Agronomy and Soil Science Society of America, Madison, Wisconsin. Olson, J. S. 1958. Rates of succession and soil changes on southern Lake Michigan sand dunes. Botanical Gazette 119: 125-170. Retzer, J. L. 1956. Alpine soils of the Rocky Mountains. JouiTial of Soil Science 7: 22-32. Reynolds, R. V R. 1911. Grazing and floods: a study of conditions in the Manti National Forest, Utah. U.S. Forest Sei^vice Bulletin 91. 16 pp. Sagger, S., M. J. Hedley, A. G. Gillingham, J. S. Rowarth, S. Richardson, N. S. Bolan, and R E. H. Gregg. 1990. Predicting the fate of fertilizer sulphur in grazed hill country pastures by modelling the transfer and accumulation of soil phosphorus. New Zealand Journal of Agricultural Research 33: 129-138. Sagger, S., A. D. MacK-ky, M. J. Hedley, M. G. Lambert, AND D. A. Clark. 1990. A nutrient-transfer model to explain the fate of phosphorus and sulphur in a grazed hill-countiy pasture. Agriculture, Ecosystems and Environment 30: 295-315. Sampson, A. W 1919. Plant succession in relation to range management. U.S. Department of Agriculture Bulletin 791. 76 pp. Sampson, A. W, and L. H. Weyl. 1918. Range preserva- tion and its relation to erosion control on western grazing lands. U.S. Department of Agriculture Bulletin 675. 35 pp. Scholes, R. T. 1990. The influence of soil fertility' on the ecology of southern African diy savannas. Journal of Biogeography 17: 415-419. Schreiber, J. E, Jr. 1988. Final report on the Flagstaff Limestone (Paleocene-Early Eocene) in the Manti- LaSal National Forest, east of Manti-Ephraim, Sanpete County, Utah. Department of Geosciences, University of Arizona, Tucson. Unpublished report. Sommerfeldt, T. G., .\nd C. Chang. 1985. Changes in soil properties under annual applications of feedlot manure and different tillage practices. Soil Science Society of America Journal 49: 983-987. Stanley, K. O., and J. W Collinson. 1979. Depositional history of Paleocene-Lower Eocene Flagstaff Limestone and coeval rocks, central Utah. American Association of Petroleum Geologists Bulletin 63: 311-323. Stevenson, F J. 1986. Cycles of soil; carbon, nitrogen, phosphorus, sulfur, micronutrients. John Wiley and Sons, New York. Swift, M. J., O. W Heal, and J. M. Anderson. 1979. Decomposition in terrestrial ecosystems. Studies in 312 Great Basin Naturalist [Volume 54 Ecology. Vol. 5. University of California Press, WlU.I.WIs, R H., AND R. J. Haynes. 1992. Balance sheet of Berkeley. phosphorus, sulphur and potassium in a long-term TiEDE.MANN, A. R., AND T. D. ANDERSON. 1971. Rapid grazed pasture supplied with superphosphate. analysis of total sulphur in soils and plant materials. Fertilizer Research 31: 51-60. Plant and Soil 3.5; 197-200. Williams, S. T, and T R. G. Gray. 1974. Decomposition TiEDEM.XNN, A. R., AND J. O. Klkmmedson. 1986. Long- of litter on the soil surface. Pages 611-6.32 in C. H. term effects of inesquite removal on soil characteris- Dickinson and G. L. E Pugh, eds.. Biology of plant tics: I. Nutrients and bulk density. Soil Science litter decomposition. Academic Press, New York. Society of America Journal 50: 472-475. Youngberg, C. T, and C. T. Dyrness. 1964. Some physi- Weber, J. N. 1964. Carhon-o.xygen isotopic composition cal and chemical properties of pumice soils in of Elagstaff carbonate rocks and its bearing on the Oregon. Soil Science 97: 391-399. history of Paleocene-Eocene Lake Flagstaff of cen- tral Utah. Geochimica et Cosmochimica Acta 28: Received 4 J anuanj 1994 12 19-1242. Accepted 29 March 1 994 Great Basin Naturalist 54(4), © 1994, pp. 313-328 GEOSTATISTICAL ANALYSIS OF RESOURCE ISLANDS UNDER ARTEMISIA TRIDENTATA IN THE SHRUB-STEPPE Jonathan J. Halvorson^ Hai^vey Bolton, Jr.-, Jeffrey L. Smith'^, and Richard E. Rossi- Abstract. — Desert plants can influence the pattern of resources in soil resultinj^ in small-scale enriched zones. Although conceptually simple, the shape, size, and orientation of these "resource islands ' are difficult to study in detail using conventional sampling regimes. To demonstrate an alternative approach, we sampled soil under and around indi- vidual Artemisia tridentata (sagebrush), a dominant shrub of cool desert environments, and analyzed the samples with univariate statistics and geostatistics. Univariate statistics revealed that soil variables like total inorganic-N, soluble-C, and microbial biomass-C were distributed with highest mean values within about 25 cm of the plant axis and significant- ly lower mean values at distances beyond 60 cm. However, such simple analyses restricted our view of lesource islands to identically sized, symmetrical acciunulations of soil resources under each plant. Geostatistics provided additional information about spatial characteristics of soil variables. Variography revealed that samples separated by a distance of less than about 70 cm were correlated spatially. Over 75% of the sample variance was attributable to spatial variability. We modeled these spatial relationships and used kriging to predict values for unsam- pled locations. Resulting maps indicated that magnitude, size, and spatial distribution of soil resource islands vary behveen individual plants and for different soil properties. Maps, together with cross-variography, further indicate that resource islands under A. tridentata are not always distinguishable from the sunounding soil by shaip transition bound- aries and may be asymmetrically distributed around the plant a.\is. Key words: resource islands, geostatistics, Artemisia tridentata, nutrient availability, kriging, spatial correlation. Recognition that individual plants can sig- nificantly affect the local soil environment dates back to at least the middle of the nine- teenth centuiy (see Zinke 1962) and has been documented for many plant forms including broadleaf and coniferous trees (Zinke 1962, Everett et al. 1986, Doescher et al. 1987, Belsky et al. 1989), bunch grasses (Hook et al. 1991), herbaceous legumes (Halvorson et al. 1991), and, in particular, desert shrubs (e.g., Fireman and Hav^ward 1952, Garcia-Moya and McKell 1970, Nishita and Haug 1973, Earth and Klemmedson 1978, Burke 1989, Burke et al. 1989, Virginia and Jarell 1983, Bolton et al. 1990, 1993). Soil associated with plants typi- cally contains greater concentrations of limit- ing resources (e.g., N, P), contains larger pop- ulations of soil microorganisms, and exhibits higher rates of nutrient cycling processes like mineralization (Charley and West 1977, Bolton et al. 1990) and denitrification (Virginia et al. 1982). These small-scale enriched zones, vari- ously temied "fertile islands" (Gamer and Stein- berger 1989), "isles of fertility" (West 1981, Whitford 1986), "resource islands" (Reynolds et al. 1990), or "ecotessara" (Jenny 1980), are hypothesized to result from several mecha- nisms (Garner and Steinberger 1989), includ- ing litter-fall or stemflow (Zinke 1962), decreased erosion or increased deposition (Coppinger et al. 1991), microclimatological amelioration of the soil (Smith et al. 1987, Pierson and Wight 1991), or inputs of resources via insects, birds, or animals (Davidson and Morton 1984). Detailed knowledge of the size and internal dynamics of resource islands is important for understanding energy flu.x, mass transport, and nutrient cycling processes at the scale of the individual plant. Resource islands may also connote a tier in a progressive, hierarchical mosaic of plant and animal habitats, resource distributions, and biogeochemical processes (i.e., patches sensu Kotliar and Wiens 1990). Estimates of the distribution and numbers of resource islands in the landscape may aid in understanding population level processes and can be used to refine regional estimates of 'Pacific Northwest Laboiaton, Richland, Washington 99352. Direct correspondence to 215 Johnson Hall, Washington State University, Pullman, Washington 99164-6421. ^Pacific Northwest Laborator\', Richland, Washington 99352. "^Land Management and Water Conservation Research Unit, USDA-ARS, Washington State University-, Pullman, Washington 99164. 313 314 Great Basin Naturalist [Volume 54 energy flow and mass transfer. Furthermore, interrelationships of large numbers of individ- ual resource islands may influence ecosystem stiTicture, flinction, and stal)ilit> (Reynolds et al. 1990, Schlesinger et al. 1990, Malvorson et al. 1991). Although conceptually simple, the size, shape, and orientation of resource islands are not easy to evaluate. Previous studies have typically been based on relatively small num- bers of samples collected using a binaiy regime (i.e., samples collected beneath the plant versus samples collected "away " from the plant) or along a transect passing from plant to bare soil. Such an approach cannot be used to provide a detailed spatial analysis of resource concentra- tions or processes in the soil that are likely to exhibit complex responses to landscape and microsite variations (Burke et al. 1989). Additionally, data collected from different locations (or depths) have often been analyzed using inferential statistics such as AN OVA or t tests that assume samples are spatially inde- pendent and identically distributed. However, these assumptions may be dubious, if untested, since ecological phenomena are often spatially or temporally correlated and their frequency distributions are rarely normal (Rossi et al. 1992). Recently, a branch of applied statistics, known as geostatistics, has been demonstrated to be usehil for detennining spatial correlations among ecological data and for estimating values at unsampled locations (e.g., Robertson 1987, Robertson et al. 1988, Rossi et al. 1992). Objectives of this study were (1) to use geo- statistics to describe and model the spatial continuity of soil variables around individual plants, (2) to use this information to produce graphical representations or maps of specific resource islands, and, finally, (3) to quantify the spatial correlation between plants and soil variables. We examined Artemisia tridentata (sagebrush), a prominent shrub of cool desert environments (West 1983) previously known to affect the distribution of resources in the soil. Several workers have measured higher concentrations of resources such as total-C, total-N, inorganic-N, and higher rates of N cycling in soil beneath A. tridentata than in nearby open soil using a binary sampling regime (e.g., Burke 1989, Burke et al. 1989, Bolton et al. 1990, 1993). However, these stud- ies have not accounted for possible spatial auto- correlation of the samples, evaluated resource islands of individual plants, nor quantified the scale of soil heterogeneit>' beneath A. tridentata. Geostatistics has previously been used to describe environmental and soil parameters associated with A. tridentata. For example, Pierson and Wight (1991) used one-dimen- sional geostatistics to analyze spatial and tem- poral variability of soil temperature under A. tridentata. Halvorson et al. (1992) demonstrat- ed that geostatistics was an appropriate means of measuring resource islands at the scale of an individual A. tridentata plant Jackson and Caldwell (1993a) attempted to quantify the scale of nutrient heterogeneity around indi- vidual A. tridentata and Fseudoroegneria spi- cata in a native sagebrush steppe using semi- variograms. They demonstrated increasing autocorrelation of soil nutrients at spatial scales <1 m but did not determine whether small-scale effects were attributable to individ- ual plants or an artifact of the nested sampling design used. More recently, they constructed kriged maps that showed relatively high con- centrations of soil variables like soil organic matter, extractable phosphate, and potassium near Pseiidoruegneria tussocks but not Aiieumia shmbs (Jackson and Caldwell 1993b). However, these kriged maps did not directly quantify spatial covariation between locations of indi- vidual plants and resource islands. Further, Jackson and Caldwell did not obsewe autocor- relation for microbial processes at any scale that was measured. To meet our objectives, we applied geosta- tistics in three steps. First, we characterized and modeled the similarity between samples as a function of their separation distance and direction. Second, we used this relationship to interpolate values at unsampled locations directly under and near individual plants. Finally, we quantified spatial covariation be- tween soil properties and plants. Study Site The study was conducted at the Arid Land Ecology (ALE) Reserve, located on the Hanford Site in south central Washington (see Bolton et al. 1990 for details). There, remnants of the native Artemisia tridentata-Elytrigia spicata association occur on silt loams of the Warden or Ritzville series. This perennial shrub-steppe is the largest grassland-type in 1994] Geostatistical Analysis of Resource Islands 315 North America and covers more than 640,000 km- of the Intermountain Pacific Northwest too chy to support forests (Daubenmire 1970, Rogers and Rickard 1988). In an undisturbed state the A. tridentata-E. spicata association would be composed typically of three layers of vegetation: an overstory shrub {Artemisia tri- dent ata tridentata), a large caespitose perenni- al grass {Eh/trigio spicata [formerly Agropyron spicatiim]), and a small caespitose perennial grass {Poa secunda) growing on soil with a thin cryptogamic crust (Daubenmire 1970). However, following disturbance such as tillage, grazing, or fire, the alien annual grass Bromus tectorum becomes established. Methods Soil Collection and Analysis Cores of surface soil (10.5 cm dia. X 5 cm deep) were collected at 41 specific locations within 2 X 2-m plots centered on mature A. tridentata individuals (Fig. 1). Samples were located so as to minimize the number of data points needed for analysis of spatial character- istics and to avoid preferential clustering. We sampled five identically oriented plots (205 points) in March 1991, when levels of soil moisture and microbial biomass activity were 1.5 -- 1.0 () 0.5 -- N -e^ — e-® 2.0 ^-e — he h Q o o d) o o o o o o o o o o o o o o o O O () o CD O O () 0.0 ^-o — he \ e^ — e-e 0.0 0.5 1.0 1.5 East (m) 2.0 Fig. L Schematic of a typical sampling plot. Each plot (five total) was centered on an Artemisia tridentata plant. Dashed circles show the location of 41 soil cores (10. .5 cm dia. X 5 cm deep). All five plots were oriented as shown. high. All plots were located within approxi- mately 20 m of each other within a flat area with randomly spaced plants. Multiple plots were sampled for two reasons: first, to assess spatial characteristics of resource islands by basing our calculations on several examples rather than a single instance; and, second, to provide replicates in the event that no spatial dependence of soil properties was observed. Data fiom all plots were combined to consider spatial dependence of soil properties around several A. tridentata plants simultaneously. This approach was chosen because it provided a more generalized evaluation of resource islands under individual A. tridentata and greatly increased the number of data pairs at any separation distance. Estimates of plant location were required for cross variography (see below). Therefore, vegetation maps were produced from vertical photographs. Each plot was divided into 5 X 5-cm squares. Each square was classified into one of three groupings — bare, grass species, or A. tridentata — based on predominant cov- erage. For this work no attempt was made to distinguish among grass species. Each soil sample was sieved (5 mm), mixed, and analyzed for a variety of soil variables. For this work we present data only for water-solu- ble forms of C, total inorganic-N (i.e., nitrate + ammonium), and soil microbial biomass-C. Soluble soil C (H2O-C) was extracted with room temperature deionized water and ana- lyzed using an infrared gas analyzer (Ionics Inc., Watertown, Massachusetts). Total inor- ganic nitrogen (TI-N) was extracted within 48 h of collection from 10-g subsamples of soil using 25 ml 2M KCl and analyzed colorimetri- cally (Alpkem Coip., Clackamas, Oregon). Soil microbial biomass-C (SIR-C) was estimated from the respiratoiy response of soil to glucose, a source of C readily utilized by heterotrophic soil microorganisms (Anderson and Donisch 1978). Ten-gram samples of soil were placed in 40-ml glass vials, moistened with deionized H2O, covered with Parafilm, and incubated in the dark at 23.5 ± 0.5 °C for 1 wk. Each sam- ple was then amended with a glucose solution at the rate of 600 mg glucose (240 mg C) kg-1 soil, bringing the final H2O content of the soil to 20-25% (w/w; equivldent to 30-50 kPa). Glass vials were sealed with silicone septa and incubated for 3 h. Soil respiration was mea- sured by gas chromatography and related to 316 Great Basin Natur.\list [Volume 54 estimates of soil microbial biomass-C with equations developed by Anderson and Domsch (1978). Univariate Statistics Univariate statistics were calculated for each soil parameter. For classical inferential statistics, data were also assigned to one of five distance classes depending upon sample loca- tion within a plot. These classes can be envi- sioned as concentric rings located at increas- ing distances from the center of the plot. The first distance class comprised samples collected directly under A. tridentata (distance = 0 cm, n = 1 per plot), followed by the second (approximate distance = 25 cm, n = 8 per plot), third (approximate distance = 60 cm, n = 8 per plot), fourth (approximate distance = 110 cm, n = 12 per plot), and fifth (appro.ximate distance = 130 cm, n = 12 per plot). Average values of soil properties in each distance class were plotted as a function of radial distance from the plant axis. Following variography (see below), logjo transformed samples deemed spatially independent were compared with ANOVA using plot as a blocking factor. Geostatistics Variography. — We evaluated spatial char- acteristics of each soil parameter with the non- ergodic autocorrelation function (Srivastava and Parker 1989) and summarized results graphically as correlograms. Like variograms, correlograms represent the average degree of similarity between samples as a function of their separation distance (lag) and direction. Unlike the variogram, the correlogram filters out the effects of changes in both lag means and lag variances. Each point in a correlogram was calculated from this equation: p*(h) = N(h) V^ Nfh) 2^i=l {[z(xi) - m_iJ[z(xi+h) - m+i,]} S-liS+ii (1) where z(xi) and z(xj -I- h) are two data points separated by the distance (lag) h. Datum z(xj) is the tail and z(xj + h) is the head of the vec- tor, N(h) is the total number of data pairs sepa- rated by lag h, m_j^ and m+j^ are means of the points that correspond to tail and head of the lag, respectively, and S_i, and S+|^ are standard deviations of tail and head values of the lag, respectively. We chose the correlogram because it removes the effects of lag means and standardizes by the lag variances (Rossi et al. 1992). For this work we express correlo- grams in the form of a standardized variogram by subtracting each p*(h) from 1 (Isaaks and Srivastava 1989, Rossi et al. 1992). Correlograms were first calculated solely as a function of lag distance (i.e., the omnidirec- tional case) without considering any differ- ences in spatial continuity with direction (i.e., anisotropy). However, since resource islands need not be symmetric (e.g., Zinke 1962), we also calculated directional correlograms. For each soil property a separate correlogram was calculated for samples oriented 0°, 45°, 90°, and 135° (±15° tolerance) from each other. Since correlograms are symmetric about the origin (i.e., 0° = 180°, 45° = 225°, etc.), 0°, 45°, 90°, and 135° directions correspond to samples aligned along east-west, northeast- southwest, north-south, and northwest-south- east axes. Because the data of each plot were concate- nated during this analysis, local anisotropics (i.e., anisotropics specific to each plot) were in effect combined. Thus, any directional effects we obsei-ved were a composite of the five plots and presumably indicative of overall direction- al trends. To identify directions of maximum and minimum continuity; we estimated the lag distance corresponding to a common value for each directional correlogram (Isaaks and Srivastava 1989). The directional correlogram with the greatest lag associated with a correlo- gram value of 1 was identified as the direction of greatest continuity. The conelogram with the smallest lag corresponding to 1 was deemed the direction of minimum continuity. The empirically determined scatter of data points in each correlogram was fit with models known to produce a positi\'e definite kriging system (i.e., matrices diat provide both a unique solution and a positive estimation variance; Isaaks and Srivastava 1989). Such models typi- cally contain several salient features known as nugget, range, and sill. The nugget is the amount of variance not explained or modeled as spatial correlation. It is the apparent ordi- nate intercept and is due to (1) unsampled cor- relation below the smallest lag and (2) experi- mental error (Rossi et al. 1992). A small nugget relative to the sill indicates that a large pro- 1994] Geostatistical Analysis of Resource Islands 317 portion of the sample variability^ is modeled as spatial dependence. Conversely, a large nugget indicates less sample variability can be mod- eled as spatial dependence. The sill is charac- terized by a leveling off of the correlogram model. If present, it indicates that spatial cor- relation is, on average, constant. However, if spatial correlation continues to change at lags greater than those considered in the correlo- gram, then a sill will not be apparent. The lag value when the correlogram model reaches the sill is known as the range. It represents the maximum separation distance within which samples are spatially correlated. At lags > the range, the sill of the variogram may approach the sample variance (Barnes 1991). Kriging to estimate data at unsampled locations. — Kriging has been likened to "multiple linear regression with a few twists" (Rossi 1989). In classical multiple linear regres- sion, an estimate of tiie dependent variable, Y, is calculated from a weighted linear combina- tion of independent variables where each is measured at about the same location in time or space. Usually, only a single value of Y is estimated. Similarly, in kriging, z'''(Xq), the estimated value of the variable for an unsam- pled location (x^), is calculated as a weighted linear combination of the surrounding sam- pled neighbors: N z^ixj = I gi • z{xi) (2) / = 1 where the z(x;)'s are the sampled values at their respective locations, and the gj's are the weights associated with each sample value. In ordinary kriging, weights used to estimate z*(Xq) are chosen so that the resulting estimate is unbiased and has a minimum estimation variance and sum to unity. Kriging incorpo- rates a model of spatial continuity' (here the cor- relogram model) and accounts for the degree of clustering of nearby samples and their dis- tance to the point being estimated (Isaaks and Srivastava 1989). We used ordinaiy point kiiging to estimate values of soil properties at unsam- pled locations. For each plot we estimated val- ues for the nodes of a 5 X 5-cm- grid. Each predicted value was based on a minimum of 6 and a maximum of 12 neighbors located with- in a 0.8-m circular search radius. Cross-variography. — In addition to spatial characteristics of single soil properties, we also determined how soil properties covaried with plants. We modeled spatial covariation with p*^g(h), the estimated nonergodic cross- correlogram. Like the correlogram, it accounts for both variables' fluctuating lag means and variances (Isaaks and Srivastava 1989, Rossi et al. 1992). Because comparisons between contin- uous variables (i.e., TI-N, SIR-C, and H2O-C data ) and discrete variables (i.e., plant data) might be complicated by a "contact effect " (Luster 1985), we converted TI-N, SIR-C, and H2O-C data to binary variables using an indi- cator transformation (Journel 1983). For this work, continuous data values of TI-N, SIR-C, and H2O-C were coded 1 if they were greater than the local (within-plot) median, or 0 follow- ing Halvorson et al. (in review). Cross-correlo- grams were then calculated for grass species or A. tridentata and indicator transformed TI-N, SIR-C, and H2O-C data using the equation, P*AB(h) = N(h) N(h) N(h) Z X [Ia (Xi,ZA) - mA_J[lB(xk^ZB) - niB^i i=l k= 1 (3) where N(h) is the total number of data pairs separated by vector h, I^(xj,z^) is the coded plant data (equal to 1 if the specified plant type was present or 0 if absent) at some loca- tion (xj), m^i and S^_, are the mean and standard deviation, respectively, for the plant variable at those data locations that are -h away from a soil property data location. Similarly, Ig(xi^,ZB) is the coded soil variable data (equal to 1 if the data value is greater than the local plot median or else 0) at location (xj^), mg , and Sg . are the mean and standard deviation of the son variable indicator calcu- lated for those locations that are +h away from a plant variable data location. Note, when h is 0, equation 3 is equivalent to the Pearson correlation coefficient (Isaaks and Srivastava 1989). Unlike the correlogram, values calculated for the cross-correlogram may not be symmetric about the origin because both the order and direction are switched when variables are re- versed (Isaaks and Srivastava 1989). Conse- quently, we calculated individual cross-correl- ograms for the 0°, 45°, 90M35M80°, 225°, 318 Cheat Basin Naturalist [Volume 54 270°, and 315° directions (±30° tolerance). These correspond to soil samples aligned to the east, northeast, north, northwest, west, southwest, south, or southeast of a plant. Results Summarx' statistics indicated that samples of TI-N and H2O-C were positively skewed, while samples of SIR-C were more normally distributed (Figs. 2A-C). Total inorganic-N ranged from 0.6 mg / kg soil to a maximum of 23.6 mg / kg soil (Fig. 2A). The mean value for TI-N of 3.8 mg / kg soil compared reasonably to the values reported by Bolton et al. (1990) of 4.1 and 4.9 mg / kg soil for open soil crust and A. tridentata soil, respectively. Values obsei^ved for H2O-C ranged widely from 9.8 mg / kg soil to 633.9 mg / kg soil (Fig. 2B). Estimates of SIR-C ranged from less than 200 to over 1800 mg / kg soil (Fig. 2C). The average value for SIR-C, 750 mg/ kg, was within the range reported by Burke et al. (1989) and equivalent to about 980 kg C / ha soil assuming a bulk densit\' of 1.3 (Bolton et al. 1990). Comparative- ly, Smith and Paul (1990) reported average microbial biomass pool size for grassland sys- tems of 1090 kg C / ha. Univariate statistics also indicated how soil properties varied with distance from the A. tridentata axis (i.e., center of the plot; Figs. 3A-C). Concentrations of HgO-C and SIR-C were greatest within 25 cm of the plant axis and lowest at distances beyond 60 cm (Figs. 3B, C). A similar pattern was obsei^ved for TI-N except that mean concentration was low in soil collected from directly beneath the A. triden- tata plant (Fig. 3A) and from distances beyond 60 cm. This somewhat unexpected finding of a resource "hole" in the center of the resource island may be indicative of differences in the cycling of N under sagebrush and grass plants. Variography indicated that samples of TI-N, H2O-C, and SIR-C were spatially correlated (Fig. 4). Correlograms for SIR-C and TI-N exhibited similar ranges of about 0.7 or 0.8 m. The correlogram for H2O-C was similar to the others at small lag distances and equaled the sample variance at a range near 0.7 m. How- ever, at greater lags, correlogram values for H2O-C increased above the sample variance and did not appear to reach a sill until lags were greater than 1 m. A correlogram sill greater than 1 for H.9O-C can occur if the uu - 80 - — 60 - - 40 - - 20 - - 0 - Number 205 Mean 3.8 Standard Deviation 3.9 Skewness 2.7 Kurtosls 11.1 P: T~L 4=^ -h 24 TI-N (mg/kg soil) B 150 -r o 90 -- 60 30 Number 205 Mean 62.6 Standard Deviation 73.0 Skewness 3.9 Kurtosls 24.9 tn^ 300 600 H20-C (mg/kg soil) C 40 -r 30 -- 10 - Number 205 Mean 743 Standard Deviation 281 Skewness 0.9 Kurtosls 4.3 tUd^ 0 400 800 1200 1600 2000 SIR-C (mg/kg soil) Fig. 2. Frequency histograms and some basic summary statistics for (A) total inorganic-N (TI-N), (B) water solu- ble-C (HoO-C), and (C) soil microbial biomass-C (SIR-C). majority of sample values are collected from an area with dimensions equal to or less than the variogram range (Barnes 1991) or if dis- crete regions of high and low concentration occur at lags greater than the maximum lag in 1994] Geostatistical Analysis of Resource Islands 319 «n 7 - o> • - JlC 6 - - en E • ( / V z 5 - - / \ Ui / J L \ _l / \ IS ■ / \ < / \ T a 4 - r/ \ z / \ ^» < I o ■ ( ^ V^\i X 3 - _ T O UJ I _l . •*■ < 1- o o _ 1 1 r 1 1 1 1 »- Z ~* 1 1 1 1 1 c 1000 -1- 900 800 -- 700 ^ 600 \ \ 1 1 \ 1 0 25 50 75 100 125 150 DISTANCE FROM PLANT (cm) Fig. 3. Mean ± standard error for (A) TI-N, (B) H2O-C, and (C) SIR-C. Numbers in parentheses are the number of data points (from five plots combined) that contribute to each estimate. 1.50 -r 1.00 -- 0.75 0.50 -- 0.25 - 0.00 f * ♦■ # Total Inorganic — N O HjO soluble-C V SIR blomass-C 0.00 0.25 0.50 0.75 1.00 1.25 1.50 Lag (m) Fig. 4. Omnidirectional correlograms for TI-N, H2O-C, and SIR-C. Each point shown was calculated from a mini- mum of 253 pairs (range 253-644). the correlogram (i.e., an incompletely mod- eled "hole" effect). The apparent nuggets for all three soil parameters suggested that more than 75% (a correlogram value of 0.25 or less) of total sample variability could be modeled as spatial dependence. Ranges observed in correlograms of soil properties were used to establish the separation distance beyond which correlation between samples could not be distinguished from sam- ple variance. In other words, samples separat- ed by distances greater than the range were candidates for analysis using more traditional statistical techniques that assume indepen- dence such as AN OVA. For our data, samples in the first two distance classes were thus combined and compared to samples from the last two classes because they were separated by more than about 0.8 m. Samples in the third distance class, lying at an intermediate dis- tance between the center and outer boundary of the plot, were excluded from analysis. Analysis of variance of log^Q transformed data using the five plots as blocking factors showed mean concentrations of TI-N, H2O-C, and SIR-C to be significantly greater (P < .001) within 29 cm of the plant axis than values col- lected > 1.07 m away from the plant axis (Table !)• Omnidirectional correlograms character- ized spatial correlation or continuity purely as a function of lag distance. However, by consid- ering the orientation of samples in addition to their lag distance, directional anisotropics were suggested. Directional correlograms showed 320 Great Basin Naturalist [Volume 54 Tablic 1. Summaty statistics and randomized complete block ANOV'As for log],, transformed TI-N, HgO-C, and SIR- C. Data arc summarized into two sample location classes: Near (all measurements collected from within 29 cm of the plant axis; n — 45) and Away (samples collected at distances greater than 107 cm from the plant axis; n — 120). Average separation distance between the two location classes was 99.4 cm. Mean Stand arc! de\ iation Standard error Near Awa\ Near Away Near A\va\ TI-N 0.58 0.39 0.34 0.32 0.05 0.03 H2O-C 2.01 1.49 0.31 0.26 0.05 0.02 SIR-C 2.93 2.82 0.14 0.17 0.02 0.02 Source of variation df MS F P Plot 4 0.59 6.14 <.001 TI-N Sample Location Error 1 159 1.14 0.10 11.81 .001 Plot 4 0.21 2.85 .026 II2O-C Sample Location Error 1 159 8.79 0.07 122.49 <.001 Plot 4 0.06 2.32 .060 SIR-C Sample Location Error 1 159 0.36 0,03 13.63 <.001 differences in both nuggets and ranges (Fig. 5). Generally, the largest apparent nuggets were observed in correlograms oriented in the 0° (east-west) and 45° (northeast-southwest) directions. With the exception of H2O-C, these correlograms had estimated nuggets of 0.4 or more. Conversely, correlograms calcu- lated for the 90" (north-south) and 135" (northwest-southeast) directions generally exhibited nuggets of < 0.2. Directions of maximum and minimum con- tinuity were identified. For TI-N, maximum continuity was observed in the 45° direction while lower but similar ranges of continuity were observed in the other three directional correlograms (Fig. 5, left column). Little anisotropy was observed in directional correl- ograms for H2O-G, suggesting that spatial cor- relation could be adequately modeled with a single isotropic correlogram (Fig. 5, center column). Like TI-N, the direction of maximum continuity observed for SIR-G was 45°, with a direction of minimum continuity in the 135° direction (Fig. 5, right column). The aniso- tropics for TI-N and SIR-G were accounted for in kriging by using a model that evaluated both distance and direction (Table 2). Maps of the estimates generated with these models and ordinary kriging were constructed for each soil property in each plot. Taken together they suggest that generalizations about spatial distribution of resources in the soil beneath A. tridentata can be complicated by the variation obsei-ved between individual plants and specific soil properties. For exam- ple, distinct "islands of TI-N were not always clearly associated with A. tridentata. Instead, in three of five plots highest concentrations of TI-N appeared to be associated with location of grasses (Fig. 6, plots A,B,E). In plots G and D, concentrations of TI-N were highest in the vicinit\' of the A. tridentata canopy. However, vegetation maps of these plots indicate that grass species were present near the A. triden- tata plant. In plot E smallest concentrations of TI-N were predicted to lie under the A. tri- dentata plant. Conversely, highest accumulations of HoO- G were clearly associated with A. tridentata in kriged maps. In each plot high concentrations of H2O-G were localized near the plot center under the plant canopy. Location of grass species did not always appear to coincide strongly with high concentrations of H2O-G (see NW 1/4 of plot A, NE and SE 1/4 of plot B, and NE 1/4 of plot E). However, high con- centrations of both TI-N, and H2O-G did coincide with location of grasses in the NW 1/4 of plot B and SW 1/4 of plot E. Maps of SIR-G indicate that islands of soil biomass-G were present under plants but not apparently specific to any particular type of vegetation. Relatively high concentrations of SIR-G were estimated under A. tridentata plants in all plots. However, SIR-G was also 1994] Geostatistical Analysis of Resource Islands 321 1.5 n 0° Tl-N r H2O-C -, SIR _ -C () • 0.5 -- i • • • • • ;;•••'•• n n () 1 1 1 1.5^45° 1 .0 - 0.5^ • • •• • • • •• • • • •• • • 0.0 - 1 1 1 III 1 . . 1 - 1 .5 1.0 0.5 0.0 90 •; • ® 1 — 1 — I 1 1.5 1.0 0.5 + 0.0 135 • • • • • • •• • • 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 LAG (m) 0.0 0.5 1.0 1.5 Fig. 5. Directional correlograms for TI-N, HoO-C, and SIR-C in the 0°, 45°, 90°, and L35° directions calculated with a tolerance of ±L5°. Each point shown was calculated from a minimum of 47 pairs (range 47-212). Vertical lines corre- spond to a correlogram value of 1 and were used to identify' directions of ma.\imum and minimum spatial continuity. Open symbols are estimates of the apparent nugget (the apparent ordinate) fit by eye. accumulated elsewhere in relation to the loca- tion of grass species. High concentrations of SIR-C were observed in several instances not associated with high concentrations of either TI-N or H2O-C (e.g., NE and SW 1/4 of plot B, NWl/4ofplotD). Cross-variography indicated how TI-N, H2O-C, and SIR-C covaried spatially with respect to A. tridentata and grass species. Indicator transfomied TI-N data were similarly and positively correlated with A. tridentata and grass species at a lag of 0 (equivalent to the Pearson correlation coefficient; Figs. 7A,B). However, conelation varied with increasing lag distance (i.e., showed spatial dependence) in a different manner for each. For A. tridentata highest positive correlations with TI-N were observed in the 45° (northeast) and 90° (north) directions (Fig. 7A). The range over which A. tridentata remained positively correlated with TI-N was longest in die 45° and 90° directions, extending to about I m and 0.75 m, respec- tively. Positive correlations with TI-N were obsei^ved for other directions too but only to a lag of about 0.5 m. In contrast, grass species were positively correlated to above-median TI-N concentrations in the 315° (southeast), 270° (south), and 225° (southwest) directions (Fig. 7B). The range over which grass species were positively correlated to TI-N in these directions was less than that for A. tridentata. In other directions grass species were uncor- related or negatively correlated to TI-N at lags above 0. 322 Great Basin Naturalist [Volume 54 TaBLF 2. Model'" parameters used for ordinan kri^ini ALE soil properties. Soil parameter TIN l-p*45 = .L54 + 0.428 Sph(0.27) + 0.524 Sph(L80) l.p*j3^ = .L54 + 0.428 Sph(0.71) + 0.524 Sph(L20) SIR-C l.p*^5 = .072 + 0.381 Spli(0.23) + 0.717 Sph(1.87) l.p*^35 = .072 + 0.381 Sph(0.74) + 0.717 Sph(0.75) H.,0-C "l-p* = 1.40 Sph( 1.3) 'Models shown for TI-N and SlR-C art- a coniliination ol a nuggi't constant and two spherical models. The spherical model, denoted "Sph, is an autho- rized model commonly used to model variograms. The number that precedes "Sph" can be thought of as the local sill for that model while the number in parentheses is the range at which the local sill is reached (see Isaaks and Srivastava 1989). For a correlogram, the standardized fomi is l-p*(h) = 15 (lag/range) -0.5 (lag/range)'^ if lag ^ = range, else 1 if otherwise. ^NOTE: Geostatisticians often distinguish between the nugget used for diag- nostic purposes, which is the apparent ordinate intercept, and the nugget \'alue. which is used in modeling. Indicator transformed HoO-C data were positiveK' correlated with A. tridentata but not with grass species at a lag of 0 (Figs. 7C,D). As with TI-N, highest correlations with A. triden- tata were observed in the 45° and 90° direc- tions (Fig. 7C). Similar patterns of spatial dependence were obsei'ved for all directions. The distance to which H^O-C remained posi- tively correlated with A. tridentata ranged from a minimum of about 0.5 m to a maximum of near 0.75 m in the 45° and 90° directions. Unlike A. tridentata, H2O-C was not positive- ly correlated with grass species (Fig. 7D). Instead, H2O-C was moderately negatively correlated in the 0°, 45°, 90°, and 135° direc- tions, meaning grass species were more asso- ciated with below-median concentrations of H2O-C. At lags greater than about 0.2 m, little change in cross-correlograms was observed, indicating only a weak spatial dependence. In contrast to H2O-C, cross-correlograms indicated that SIR-C was slightly more corre- lated with grass species than with A. tridenta- ta at a lag of 0 (Figs. 7E,F). Like other soil properties, strongest positive correlation between A. tridentata and SIR-C was observed in the 45° direction, which also remained positively correlated to lags in excess of 1 m (Fig. 7E). Lowest correlations with A. tridentata were to the 270° and 225° directions. Indicator transformed SIR-C data were most correlated to grass species in the 225°, 270°, and 315° directions (Fig. 7F). Spatial dependence of correlations was obsened out to a lag of about 0.5 m. Beyond this, correlations of grass species with SIR-C remained approximately constant. Discussion Geostatistics can be applied to resource island data to provide several useful diagnostic features prior to actual mapping of the land- scape itself. For example, variography can define the presence and extent of spatial cor- relation and alert the researcher to apply with caution classical statistical comparisons that assume samples are independent and from identically distributed populations. For these methods to be more properly applied to spa- tial data, comparisons should probably be lim- ited to those samples separated l^y distances > range of the correlogram (Table 1; Webster 1985, Robertson 1987). This is true for studies that compare samples collected along a con- tinuum such as distance, depth, or concentra- tion (i.e., resource gradient) or as a function of time. Another promising use of variography is to relate spatial continuity of two or more variables at the same site or the same variable at two or more sites by comparing variograms, covario- grams, or correlograms with one another This approach may be useful for comparing the scale of ecological processes or ecosystem boundaries, but should be approached with caution for several reasons. First, each point in a traditional variogram represents the average value of the squared difference between many pairs of data points. While an average value may be appropriate for modeling spatial conti- nuity (as a summary statistic), it does not indi- cate the range of individual squared differ- ences or provide an estimate of the "goodness of fit " about each point in a variogram. The range of the deviation about the average value may be large or small (see Webster and Oliver 1992), complicating comparisons of variograms. Thus, for comparative purposes other more "robust" representations of spatial dependence such as Journel s niAD estimator, Cressie- Hawkins' robust estimator, or the rodogram (see Rossi et al. 1992) may be more appropriate choices. Second, even if variograms for two properties are similar, resultant estimates may yield veiy different maps (e.g., TI-N and SIR-C in this study). This is because estimation of un- known data values by kriging depends not only 1994] Geostatistical Analysis of Resource Islands 323 Plot A Plot B Plot C Plot D Plot E 1 L 7 „• \ 1C ^ 1 o^ - ( * < ♦ ^ 50^ \ 700 1/700 % Fig. 6. 2 X 2-m maps of vegetation and kriged estimates of TI-N, H2O-C, and SIR-C for five ALE plots. Each kriged plot is composed of 1681 points estimated by ordinary point kriging. Vegetation maps indicate vertical projections of A. tridenfata (black) and grass species (crosshatch) as determined from photographs. Soil properties are depicted in mg / kg soil (dw). upon a model of spatial continuity, but ultimate- ly upon degree and configuration of the known sample values in the field. During our analysis of resource island data using geostatistical methods, we made several assumptions or decisions about the data that could have affected our inteipretations. First, we assumed that resource islands under A. tri- denfata could be monitored using a particular configuration of samples located within a 2 X 2-m plot. A different number of samples col- lected from a larger plot with a different shape or in a different pattern might have generated different correlogram models or kriged esti- mates (Webster and Oliver 1992). Second, we collected data at a single time during the year, thereby making the kriged maps "snapshots" in time and space. Data values and spatial continuity undoubtedly vary for some types of environmental variables (e.g., soil moisture or TI-N) on a seasonal or shorter time scale. Other environmental variables such as soil te.xture, total-N, or C might change more slow- ly. Third, we chose to analyze data for the five plots collectively rather than for each plot individually. This choice reflects an inteipreta- tion that correlograms for individual plots were reasonably similar to each other and allowed our calculations to be based on a greater num- ber of paired comparisons. We reasoned that correlogram models of spatial continuity, de- rived fiom concatenated data, would summarize typical patterns of spatial continuity and direc- tional anisotropics. Alternatively, although sin- gle-plot analyses would result in plot-specific models of spatial continuity with greater specificity, they might make generalizations difficult. Finally, we assumed that spatial con- tinuity for II2O-C was reasonably described by a single isotropic model, but we concluded 324 Great Basin Naturalist [Volume 54 w?ih Artemisia O.i -- O 0 • 45^ ■ 225 V 90° A 270 ▼ 135' A 315 3 with □180° T grass species 0.00 1.25 Fig. 7. Directional cross-correlograms showing spatial covariance of indicator transformed TI-N, H2O-C, SIR-C data, and presence/absence (coded as 1 or 0) of A. tridentata and grass species. Correlograms were calculated for 0° (east), 45° (northeast), 90° (north), 135° (northwest), 180° (west), 225° (southwest), 270° (south), and 315° (southeast) directions ±30° tolerance. Each point shown summarizes a minimum of 54 pairs (range 54-158). 1994] Geostatistical Analysis of Resource Islands 325 that directional anisotropics obser\'cd for TI-N and SIR-C were important enough to be accounted for in the estimation process. Results of a geostatistical analysis cannot completely replace "sound ecological reason- ing" or theoiy (Rossi et al. 1992). Thus, the re- searcher must decide whether directional ani- sotropics observed in descriptive variography portray significant spatial patterns or are merely a coincidental result of the number and arrangement of data. The decision to account for spatial anisotropy in the kriging procedure is, in part, related to the desired end product of the geostatistical analysis. For ex- ample, if the goal of an analysis is the most accu- rate representation possible of a particular resource island under a specific A. tridentata, then a highly detailed model of spatial continu- ity would be appropriate regardless of the source of spatial variability. In this case variog- raphy based on a concatenated data set might be less appropriate than analysis based only on the single plot. However, the goal of geo- statistical interpretation of ecological data may not be to produce detailed site maps. Instead, the ecologist may be more interested in pat- terns that are broadly applicable. Anisotropics can often be related to information about the environment such as stratigraphic, meteoro- logical, or hydrogeological patterns (Isaaks and Srivastava 1989) and may suggest linkages be- tween environmental variables. Our decision to account for anisotropics in the kriging process was, in part, influenced by information about another environmental parameter, prevailing wind direction. For this reason we would expect the anisotropics observed for TI-N and SIR-C to be a consistent feature of tlie ALE landscape. Directional correlograms revealed greatest spatial continuity for samples of TI-N and SIR-C in the 45° direction, northeast-south- west. Cross-correlograms, more specifically, indicated that above-median concentrations of soil properties were most correlated to A. tri- dentata in the 45° direction, or northeast. For the ALE site, cumulative records indicate that prevailing local wind direction is from the southwest quadrant (Table 3), corresponding to the downwind direction of greatest spatial continuity. Prevailing wind direction might influence spatial patterns of soil resources by affecting distribution of litter deposition which, for A. tridentata at the ALE site, may exceed 60 kg / ha annually (Mack 1971). Table 3. Frequency of occurrence of wind at the ALE site. Source: H. Bolton, Pacific Northwest Lalioratorv. Source quadrant 1990 1991 Hours Percent Hours Percent 0-90° (NE) 90-180° (SE) 180-270° (SW) 270-360° (N\V) 1161 13.25 1557 17.78 4090 46.70 22-^7 ■^••> 07 1208 13.97 1613 18.66 3974 45.96 1851 21.41 Evidence for the occurrence of resource islands in the ALE landscape was provided by comparing concentrations of soil resources collected near A. tridentata vegetation to those collected away from the plant. However, the specific sampling regime employed to evaluate "near" vs. "away" influenced the par- ticular conclusion reached. For example, Bolton et al. (1990) were unable to conclude that concentrations of TI-N in soil under A. tridentata were significantly greater than con- centrations measured in open soil crust based on six samples drawn at random from each soil type (see also Doescher et al. 1984). In con- trast, we found that evaluating TI-N vs. dis- tance away from the A. tridentata axis resulted in the naive conclusion that significantly high- er concentrations of TI-N would always occur under A. tridentata plants (Fig. 3A, Table 1). Such a conclusion for TI-N and other soil properties would lead to a model of a land- scape composed of identically sized, symmet- rical resource islands centered on each A. tri- dentata individual and would infer some sort of causal relationship between concentration of TI-N and A. tridentata presence. However, kriged maps suggest that greatest concentra- tions of TI-N were not always associated with A. tridentata. Autocorrelation of soil properties was described using variography. The association of soil variables with A. tridentata individuals was supported jointly by kriged maps and cross-correlograms, with the latter showing that soil properties (especially H2O-C) were positively correlated to A. tridentata. Kriging is a means for producing visually satisfying maps of soil properties and provided addition- al insight into characteristics of resource dis- tribution under A. tridentata. However, we relied on these maps primarily as heuristic tools because we recognized that kriged maps 326 Great Basin Naturalist [Volume 54 are models that can be influenced by decisions about the data set (e.g., concatenated vs. single plot), the "art " of variograni modeling, the type of kriging chosen (ordinary kriging is a data "smoother"), the specific search strategy used, and the method of graphical representa- tion. Finally, kriging by itself does not provide a measure of estimate confidence or reliability like nonparametric methods (Journel 1983) or stochastic conditional simulation (Rossi et al. 1993). Kriged maps of HqO-C appeared to be most similar to graphs of summaiy statistics vs. distance fi-om plant axis (Fig. 3A). In each kiiged map (Fig. 6), high concentrations coincided with A. tridentata in a classic resource island pattern. These accumulations might be tied closely to inputs from A. tridentata litter fall representing a source of C that could be accessed by heterotrophic soil microorgan- isms. Alternatively, high concentrations of H2O-C under A. tridentata might not indicate large C inputs. Instead, they might indicate the accretion of soluble, but recalcitrant, forms of C not readily useable by soil micro- organisms. In this case the term "resource island " would be ambiguous. To have ecologi- cal significance, a resource island must be evaluated for resource quantity, resource qual- ity, and presence of alternative resource sub- stitutes. Further, the significance of resource accumulation into islands might change with time in relation to diurnal cycles, growing sea- son, or successional stage (Halvorson et al. 1991). Kriged maps of SIR-C showed accumula- tions of soil microbial biomass in close pro.xim- ity to each A. tridentata individual. However, high concentrations were also observed for locations corresponding to other plant species, demonstrating that resource islands of micro- bial populations or activity can be numerous and are nonspecific to A. tridentata. Addition- ally, a significant amount of SIR-C was esti- mated for locations not associated with any plant. This suggests that while local inputs by plants may stimulate microbial population growth or activity, sufficient resources exist in the environment to support moderate amounts of SIR-C during some times of the year. However, plant location may control the dis- tribution of SIR-C indirectly through influ- ence on microclimatological factors such as soil temperature and evapotranspiration. These factors would become more important during the hot, dry summer months and could limit distribution of SIR-C to locales closer to A. tridentata. Assessing the distribution of soil microbial populations or microbially mediated nutrient- cycling processes such as mineralization or denitrification is complicated by multiple resource requirements and compensatory capabilities of living microorganisms (Smith et al. 1985). For example, microbial population size or activity within a C-substrate resource island might be limited by the availability of soil N. Conversely, the same microbial popula- tion might be limited by the availability of C- substrate despite an N-rich environment. Under such a scenario the greatest population size or activity might occur in a location with low or intermediate quantities of both C and N, and the resource island for soil microorgan- isms or mineralization potential would appear distinct spatially from other resources. Estimation of soil properties like SIR-C that depend on the distribution of one or more other resources may need to be evaluated with respect to temporal and spatial distributions of alternative resources. Kriged maps of various soil properties can be inteipreted within the context of the rela- tionship between the particular soil parameter and A. tridentata. Our data indicate that shape and orientation of resource islands under A. tridentata vaiy with the specific soil property considered, need not be centered on the axis of an A. tridentata plant, and need not be sym- metrical. The maps also provide evidence that suggests a vertical projection of the plant canopy is not well correlated to the distribu- tion of soil variables and thus should not be used as a basis for sampling designs (Fig. 6). For some soil properties (e.g., H2O-C) the dif- ference between values characterizing the resource island and those characterizing the surrounding matrix may be large and the resource island may appear to have sharp boundaries. Conversely, the range of data val- ues for other soil properties (e.g., SIR-C) may be smaller and the transition from resource island to the surrounding soil matrix more complicated. Resource island boundaries may also change with direction, making sampling designs based on only a few transects ques- tionable. Finally, resource islands do not occur under all A. tridentata or for all soil properties. 1994] Geostatistical Analysis of Resource Islands 327 Other plants like annual and perennial grasses can be the focal points for resource islands of some variables (Jackson and Caldwell 1993b). Geostatistics allows estimation and map- ping of resource islands in considerable detail. Such maps can be used to further our under- standing of the ecology of A. trident ata, refine nutrient budgets for shrub-steppe ecosystems, reveal the existence of resource and process- dependent patterns, and help provide a ratio- nale for sampling designs based on natural boundaries. Besides two-dimensional space, geostatistics can be used to consider differ- ences in spatial continuity with soil depth (i.e., a third dimension) or time (via repeated mea- surements). However, even with geostatistics, our definition of a resource island can be improved. Whether a resource island is more properly delineated by some minimum differ- ence in resource concentration or related to the ecological significance of small differences in concentration remains to be answered. Further, the resource island "effect" may be related to more than a single environmental parameter. Consequently, methods must be developed to simultaneously integrate infor- mation for several environmental variables and summarize them spatially. Acknowledgments This research was supported in part by the Ecological Research Division, Office of Health and Environmental Research, U.S. Department of Energy (DOE). Pacific North- west Laboratoiy is operated for the DOE by Battelle Memorial Institute under Contract DE-AC06-76RLO 1830. We thank Joe Aufdennaur, Debbie Bikfasy, Kirk Stallcop, and Jeff Sullivan for analytical assistance. M. W. Palmer and J. Price provided helpful editorial comments. Literature Cited Anderson, J. P E., and K. H. Domscii. 1978. A physio- logical method for the quantitative measurement of microbial biomass in soils. Soil Biologv and Biochemistr>' 10: 215-221. Barnes, R. J. 1991. The variogram sill and the sample variance. Mathematical Geology 23: 673-678. Barth, R. C, and J. O. Klemmedson. 1978. Shrub induced spatial patterns of dry matter, nitrogen and organic carbon. Soil Science Society of America Journal 42: 804-809. Belsky, a. J., R. G. Amundson, J. M. Du.xburv, S. J. Riha, A. R. Ali, and S. M. Mwonga. 1989. The effects of trees on their physical, chemical and biological envi- ronments in a semi-arid savanna in Kenva. Journal of Applied Ecology 26; 1005-1024. Bolton, H., Jr., J. L. Smith, and S.O. Link. 1993. Soil microbial biomass and activity of a disturbed and undisturbed shrub-steppe ecosystem. Soil Biology and Biochemistiy 25: 545-552. Bolton, H., Jr., J. L. Smith, and R. E. Wildung. 1990. Nitrogen mineralization potentials of shrub-steppe soils with different disturbance histories. Soil Science Society of America Journal 54: 887-891. Burke, I. C. 1989. Control of nitrogen mineralization in a sagebrush steppe landscape. Ecology 70; 1115-1126. Burke, L G., W. A. Reiners, and D. S. Schimel. 1989. Organic matter turnover in a sagebrush steppe land- scape. Biogeochemistry 7: 11-31. Gharley, J. L., AND N. E. West 1977. Micro-patterns of nitrogen mineralization activity in soils of some shrub dominated semi-desert ecosystems of Utah. Soil Biology and Biochemistiy 9: 357-365. Coppinger, K. D., R. a. Reiners, I. G. Burke, and R. K. Olson. 1991. Net erosion on a sagebrush steppe landscape as determined by Gesium-137 distribu- tion. Soil Science Society of America Journal 55: 254-258. Daubenmire, R. 1970. Steppe vegetation of Washington. Technical Bulletin 62. Washington Agricultural Experiment Station, Gollege of Agriculture, Washington State University, Pullman. Davidson, D. W., and S. R. Morton. 1984. Dispersal adaptations of some Acacia species in the Australian arid zone. Ecology 65: 1038-1051. DoESCHER, P S., L. E. Eddleman, and M. R. Vaitkus. 1987. Evaluation of soil nutrients, pH and organic matter in rangelands dominated by western juniper. Northwest Science 61: 97-102. Doescher, E S., R. E. Miller, and A. H. Winward. 1984. Soil chemical patterns under eastern Oregon plant communities dominated by big sagebrush. Soil Science Society of America JouiTial 48: 659-663. Everett, R., S. Sharrow, and D. Thr.\n. 1986. Soil nutri- ent distribution under and adjacent to singleleaf pinyon crowns. Soil Science Society of America Journal 50; 788-792. Fireman, M., and H. E. Ha\'\vard. 1952. Indicator signif- icance of some shrubs in the Escalante Desert, Utah. Botanical Gazette 114: 143-155. Garcia-Moya, E., and G. M. McKell. 1970. Gontributions of shnibs to the nitrogen economy of a desert wash plant community. Ecology 51; 81-88. Garner, W., and Y. Steinberger. 1989. A proposed mechanism for the foiTiiation of fertile islands in the desert ecosystem. Journal of Arid Environments 16: 257-262. H.^LVORSON, J. J., H. BoLTON, JR., AND J. L. Smith. 1992. Measurement of resource islands underneath Artemisia tridentata: a geostatistical approach. Pages 117-120 in Pacific Northwest Laboratory annual report. U.S. Department of Energy, Office of Scientific and Technical Information PNL-8000 UG- 407;DE-AG06-76RLO 1830, Oak Ridge, Tennessee. Halvorson, J. J., J. L. Smith, H. Bolton, Jr., and R. E. Rossi. In review. Defining resource islands using multiple variables and geostatistics. Halvorson J. J., J. L. Smith, and E. H. Franz. 1991. Lupine influence on soil carbon, nitrogen and 328 Great Basin Naturalist [Volume 54 microbial activity in developing ecosystems at Moimt St. Helens. Oecologia (Berlin) 87: 162-70. Hook, P B., I. C. Burke, .and VV. K. Lauenrotii. 1991. Heterogeneibi' of soil and plant N and C associated with indi\ idiial plants and openings in North American shortgrass steppe. Plant and Soil 138; 247-256. ISAAKS, E. H., A.ND R. .\1. Ski\a.sta\a. 1988. Spatial conti- nuity measures for probabilistic and deterministic geostatistics. Mathematical Geolog>' 20; 313-341. . 1989. An introduction to applied geostatistics. Oxford University Press, New York. Jackso.n, R. B., A.ND M. M. Caldwell. 1993a. The scale of nutrient heterogeneity around individual plants and its (quantification with geostatistics. Ecology 74: 612-614. . 1993b. Geostatistical patterns of soil heterogene- ity around individual perennial plants. Journal ot Ecolog>' 1993 81: 683-692. Jenny, H. 1980. The soil resource, origin and behavior Springer-Verlag (Berlin). JoURNEL, A. G. 1983. Nonparametric estimation of spatial distributions. Mathematical Geology 15: 445—468. KoTLlAR, N. B., and J. A. WiENS. 1990. Multiple scales of patchiness and patch structure: a hierarchical frame- work for the study of heterogeneit\'. Oikos 59: 253-260. Luster, G. R. 1985. Raw materials for portland cement: applications of conditional simulation of coregional- ization. Unpublished doctoral dissertation, Depart- ment of Earth Sciences, Stanford University, Stanford, California. M.\CK, R. N. 1971. Mineral cycling in Artemisia tridentata. Unpublished doctoral dissertation, Washington State University, Pullman. NiSHlTA, H., AND R.M. Hauc. 1973. Distribution of differ- ent forms of nitrogen in some desert soils. Soil Science 116: 51-58^ PlERSON E B., AND J. R. WiGHT. 1991. Variability of near- surface soil temperature on sagebrush rangeland. Journal of Range Management 44: 491-497. Reynold.s, J. E, R. A. Virginl\, and J. M. Cornelius. 1990. Resource island formation associated with tlie desert shrubs, creosote bush {Larrea tridentata) and mescjuite {Prosopis glandidosa) and its role in the stability of desert ecosystems: a simulation analysis. Supplement to Bulletin of the Ecological Societ>' of America 70: 299-300. Robertson, G. P. 1987. Geostatistics in ecology: interpo- lating with known variance. Ecology 68: 744-748. Robertson, G. P, M. A. Huston, E C. Evans, and J. M. TiEDjE. 1988. Spatial variability in a successional plant community: patterns of nitrogen availability. Ecology 69: 1517-1524. Rogers, L. E., and W. H. Rickard. 1988. Introduction: Shrub-steppe lands. Pages 1-12 in W. H. Rickard, L. E. Rogers, B. E. Vaughan, and S. E Liebetrau, eds.. Shrub-steppe, balance and change in a semi- arid terrestrial ecosystem. Elsevier, New York. Rossi, R. E. 1989. The geostatistical inteqiretation of eco- logical phenomena. Unpublished doctoral disserta- tion, Washington State University, Pullman. Rossi, R. E., P W. BoHTiL A\i) J. J. Tollefson. 1993. Stochastic simulation lor characterizing ecological spatial patterns and appraising risk. Ecological Applications 3: 719-735. Rossi, R. E., D. J. MuLLA, A. G. Journel, and E. H. Franz. 1992. Geostatistical tools for modeling and interpreting ecological spatial dependence. Ecological Monographs 62: 277-314. Sc;iilesinger, W. H., J. E Reynolds, G. L. Cunningham, L. E Huenneke, W. M. Jarrell, R. A. Virginia, and W G. Whitp'ord. 1990. Biological feedbacks in global desertification. Science 247: 1043-1048. Smith J. L., ,\nd E. A. P\ul. 1990. The significance of soil microbial biomass estimations in soil. Pages 3.57-396 in G. Stotzky and J. Bollag, eds.. Soil biochemistrx-. Vol. 6. Marcel Dekker, New York. Smith, J. L., B. L. McNeal, and H. H. Cheng. 1985. Estimation of soil microbial biomass: an analysis of the respiratory response of soils. Soil Biology and Biochemistry 17: 11-16. SxHTii, S. D., W. E. Smith, and D. T. Patten. 1987. Effects of artificially imposed shade on a Sonoran desert ecosystem: arthropod and soil chemistry responses. Journal of Arid Environments 13: 245-257. Srivastava, R. M., and H. M. Parker. 1989. Robust mea- sure of spatial continuity. Pages 295-308 in M. Armstrong, ed., Geostatistics. Vol. 1. Kluwer Academic, Dordrecht, The Netherlands. Virginia, R. A., and W M. Jarrell. 1983. Soil properties in a mesquite-dominated Sonoran desert ecosystem. Soil Science Society of America Journal 47: 138-144. ViRGiNi.A, R. A., W M. Jarrell, and E. Erango-Vizcaino. 1982. Direct measurement of denitrification in a Prosopis (mesquite) dominated Sonoran desert ecosystem. Oecologia (Berlin) .53: 120-122. Webster, R. 1985. Quantitative spatial analysis of soil in the field. Advances in Soil Science 3: 1-70. Webster, R., and M. A. Oliver. 1992. Sample adequate- Iv to estimate variograms of soil properties. Journal of Soil Science 43: 177-192. West, N. E. 1981. Nutrient cycling in desert ecosystems. Pages 301-324 in D. W Goodall and R. A. Perry eds.. Arid land ecosystems: structure, function and management. Cambridge University Press, Cambridge, United Kingdom. . 1983. Temperate deserts and semi-deserts: ecosys- tems of the world. Elsevier Scientific Publishing Co., New York. 522 pp. Whitford, W. G. 1986. Decomposition and nutrient cycling in deserts. Pages 93-117 in W. G. Whitford, ed.. Patterns and processes in desert ecosystems. University of New Me.xico Press, Albuquercjue. ZiNKE, P J. 1962 The pattern of influence of indi\ idual forest trees on soil properties. Ecology' 43: 130-133. Received 5 October 1993 Accepted 23 March 1994 Great Basin Naturalist 54(4), © 1994, pp. 329-334 HABITAT PREFERENCE AND DIURNAL USE AMONG GREATER SANDHILL CRANES Donald E. Mclvorl and Michael R. Conover^ Abstract. — We e.xamined patterns of habitat use by Greater Sandhill Cranes {Grits canadensis tahicla) in the Interniountain West, April-October 1991-92, to determine whether cranes exhibited a specific preference for crops, fields, and areas within a field. This information will help farmers and wildlife managers direct nonlethal control meth- ods to the sites where crane damage is most likely to occur. We conducted surveys along two 37-km transects weekly in Cache Valley, Utah, and biweekly in Bear River Valley, Rich County, Utah, and Lincoln County, Wyoming. We recorded 5814 cranes in 662 separate groups. Most were located in pasture/hay (34%), small grain (39%), alfalfa (9%), plowed (9%), fallow (4%), or com (1%) fields. An index of feeding activity for each field and habitat type suggested cranes fed at approximately the same rate in each field and habitat type. Crane diurnal activity patterns during summer and fall revealed that grainfields were used heaviK- throughout the day. Key words: Greater Sandhill Cranes, Grus canadensis tabida, habitat, depredation, diurnal activity, Utah, Wyoming. The most recent population estimate for the Rockv Mountain Greater Sandhill Crane is 17,000-20,000 (Drewien et al. 1987:27). Records of local summer populations are less complete, but the crane population in Cache Valley, Utah, has increased from 14 individuals in 1970 (Drewien and Bizeau 1974) to approx- imately 200 in 1990 (Bridgerland Audubon Society 1990). Between 1985 and 1987, Rowland et al. (1992) reported 255 cranes sum- mering in Lower Bear River Valley, Wyoming. Crop depredation complaints attributed to cranes are rising concomitantly with population numbers (Lockman et al. 1987). In response to depredation complaints, Wyoming instituted a limited Sandhill Crane hunt in 1982. Utah in- stituted a hunt in 1989, but the decision gen- erated enough public controversy that the hunt was canceled in 1992. Cranes are omnivorous (Mullins and Bizeau 1978) and readily feed in agricultural lands, although habitat use seems to vary widely. Agricultural fields comprised 91% of habitat used by wintering cranes in western Texas (Iverson et al. 1985). During spring staging in Nebraska, Krapu et al. (1984) reported that 70% of habitat use was in agricultural lands. Within agricultiual fields 99% of use was in com stubble. Approximately 80% of spring diurnal habitat use in Alaska was in barley (Iverson et al. 1987). In Wyoming crane use of wet mead- ows and grainfields ranged from 69 to 100% (Rowland et al. 1992). We examined habitat preferences and for- aging habits of summer resident Sandhill Cranes because of increasing depredation complaints from farmers growing corn and small grains (e.g., barley, oats, rye, wheat) in Cache and Rich counties, Utah. As one means of evaluating these problems and potential solutions, we tested the hypothesis that crane use was concentrated in corn and small-grain fields in particular and in agricultural fields in general. High use of a field may alarm a farmer, but little damage may occur if birds are not foraging. Hence, we also tested the hypothesis that cranes forage in habitats in proportion to their availability. In addition, we assessed whether habitat use varied diumally during summer and fall. Additional questions relevant to selecting an appropriate scale for management include (1) whether cranes use all fields available to them or concentrate their activities in a few fields, and (2) how cranes distribute their activities within fields. Methods The study area is in Cache Valley, Utah, and Bear River Valley, Utah and Wyoming, and includes three contiguous counties: Cache and Rich counties in northern Utah and Lincoln ^Department of Fisheries and Wildlife, Utah State University, Logan, Utah 84322. 329 330 Great Basin Naturalist [Volume 54 County in southwestern Wyoming. A compre- hensive description of the area is inchided in Mclvor and Conover (1992). Cranes normally occup\' the region from April until early October To determine patterns of field use, we established a 37-km transect in Cache Valley and another in Bear River Valley. The tran- sects traversed a sample of habitat types avail- able to cranes, including cultivated fields, pas- tures, and natural habitats. Sampling was con- ducted based on a visual sui^vey method simi- lar to that used by Iverson et al. (1985, 1987). Transects were surveyed weekly in Cache Valley and biweekly in Bear River Valley from April through mid-October 1991-92. Surveys began 2 h after sunrise from a vehicle moving at 40 km/h. Habitats on both sides of the transects were scanned systemati- cally. As cranes were located, a variety of para- meters, including type of habitat in use, were recorded. Habitat was categorized by crop (alfalfa, com, small grain, pasture, hay, or mixed use) or groundcover type (riparian, sage [Arte- misia spp.] scrub). To examine the distribution of cranes within fields, we recorded the dis- tance between field edge and the individual crane closest to the edge using a range finder These data produced a distance-to-edge esti- mate, which we used as a general indication of whether cranes preferentially used edges or interiors of fields. Using the range finder, we also recorded minimum distance from the transect to the crane flock. Each sighting of cranes was given equal weighting in constructing contingency tables to maintain statistical independence among field-use obsei"vations. A few observations of cranes were made in mixed-use fields and on rural roads. These sightings were combined under the miscellaneous category. Early sea- son hayfields were difficult to distinguish from pastures, and these obsen'ations were pooled. Habitat availability was quantified along each transect in July 1991 and 1992. A sample of 125 random points on each transect was selected a priori, and each point was located and its habitat type recorded. To be selected as representative of habitat, each point had to meet two criteria. First, any sampling point not visible from the transect was not used. Second, the peri:)endicular distance from tran- sect to sampled habitat locations was bounded by the distance within which 90% of all cranes had been located during weekly siuAcys. An index of feeding activity was developed to allow comparison among habitat types. When > 1 crane was sighted, an individual was chosen at random from the flock and observed for 1 min to determine if the bird was feeding. The result was a logical variable (feed/no feed), and these data were compiled and com- pared across habitat types. Quantitative analyses were based on meth- ods devised b>' Neu et al. (1974). We used a goodness-of-fit test (F < .05) to examine the hypothesis that cranes used habitats in pro- portion to habitat availability, and to deter- mine whether cranes fed preferentially in cer- tain habitat types. We used a Bonferroni Z-sta- tistic to test for habitat and feeding prefer- ence. The Z-statistic and resulting family con- fidence intei^val for testing each contingency table cell were generated using a Monte Carlo sampling simulation from a binomial distribu- tion, on a mainframe computer using Minitab (1989). In 1992 we mapped die distribution of grain- and cornfields along the sui"vey transects. We then compared distribution of available fields with the frequency distribution of cranes obsei"ved to test the Hg of equal use among all grain- and cornfields. These data were ana- lyzed using a goodness-of-fit test. Patterns of diunial habitat use were recorded over a 5-d period in June 1991 using both tran- sects and during September 1992 using only the Cache Valley transect. Data-collection methods were identical to those used in the haliitat-use sui"vey described above, except that transects were sampled 5 times/day: sunrise, 2 h after sunrise, noon, 4 h before sunset, and 2 h before sunset. We used PC-SAS (SAS Institute, Inc. 1988) and the PROC CATMOD routine to examine June 1991 diurnal-use data, and the PROC FREQ routine to examine September 1992 data. Both SAS routines used a goodness-of-fit test (F < .05) to examine the null hypothesis that cranes maintained the same pattern of field use throughout the day. Results Fifty-three sui-veys were conducted in Cache Valley and 29 in Bear River Valley. During two field seasons we recorded 5814 cranes in 662 1994] Sandhill Crane Habitat Preference 331 groups. Most groups were observed in pas- ture/hay (34%), small grains (39%), alfalfa (9%), plowed fields (9%), fallow (4%), or cornfields (1%). Remaining cranes were located in ripari- an (3%), sagebrush (1%), and miscellaneous (2%) habitats. Habitat availability differed between the two survey transects (Table 1). Although the Cache Valle\' transect contained no sagebrush habitat, the Bear River Valley transect con- tained extensive sagebrush (61% in 1991, 58% in 1992). Conversely, the Cache Valley tran- sect contained a small amount of corn (7%) in 1991-92, a crop not cultivated in Bear River Valley. Analysis indicated variation in habitat availability between years along each transect, although the change was not statistically sig- nificant (P = .2230). For these reasons, col- lapsing the contingency tables across sample sites or across years would have made the results ambiguous. Cranes were not distributed randomly among nine available habitats in either 1991 (X2 = 374.0, df = 13, P < .005) or 1992 {X^ = 464.1, df = 14, P < .005). Along the Cache Valley transect, cranes avoided alfalfa and Table L Habitat availabilitv, use, and selection among Sandhill Cranes in Cache Valley (C), Utah, and Bear River Valley (B), Utah and Wyoming.' in 1991 and 1992. Habitat Study # crane Expected Proportion Proportion 95% fami 'ly Use area obser- # crane of study observed con fidence preference" vations obser- area (pjo) in each interval on vations area (Pi) 1991 Alfalfa C 6 2,5 0.130 0.031 .0< P.^ .067 - B 14 12 0.1.58 0.182 .065 < Pi^ .325 0 Corn C 6 13 0.065 0.031 .0< Pi^ .067 0 B 0 0 0.0 0.0 ntb ntb Fallow C 1 4 0.023 0.005 .0< P,^ .021 _ B 0 0 0.0 0.0 ntb ntb Grain C 60 34 0.174 0.308 .205 < Pi^ .405 + B 18 7 0.096 0.234 .104 < Pi^ .364 + Misc. C 4 30 0.152 0.021 .0< P,^ .051 - B 1 3 0.044 0.013 .0< Pi^ .065 0 Pasture C 87 70 0..359 0.446 .346 < Pi^ .544 0 B 3.5 5 0.070 0.455 .299 < Pi^ .610 + Plowed C 22 6 0.033 0.113 .051 < Pi^ .174 + B 1 1 0.009 0.013 .0< Pi^ .065 0 Riparian C 10 13 0.065 0.051 .010 < Pi^ .097 0 B 6 1 0.009 0.078 .013 < Pi^ .169 + Sage C 0 0 0 0 ntb ntb B 2 47 0.614 0.026 .0< P,^ .091 - 1992 Alfalfa C 29 61 0.194 0.092 .051 < P,^ .153 - B 9 9 0.121 0.123 .027 < Pi^ .246 0 Corn C 2 20 0.0065 0.006 .0< Pi^ .022 - B 0 0 0.0 0.0 ntb ntb Fallow C 20 24 0.075 0.064 .025 < Pi^ .102 0 B 3 1 0.010 0.041 .0< Pi^ .123 0 Grain C 144 57 0.183 0.4.59 .382 < Pi^ .535 + B 32 4 0.061 0.438 .274 < Pi^ .616 + Misc. C 7 34 0.108 0.022 .003 < Pi^ .051 - B 0 1 0.020 0.0 ntb ntb Pasture C 78 84 0.269 0.248 .172 < Pi^ .322 0 B 25 11 0.1,52 0.342 .192 < Pi^ .507 + Plowed C 28 14 0.043 0.089 .051 < Pi^ .134 + B 5 1 0.010 0.068 .0< Pi^ .164 0 Riparian C 6 20 0.065 0.019 .003 < Pi^ .048 - B 1 4 0.051 0.014 .0< Pi^ .068 0 Sage C 0 0 0.0 0.0 ntb ntb B 2 42 0..576 0.027 .0< P,^ ,096 - •'Habitat use is expressed as selection for ( + ), use in proportion to a\'ailabilit\' (0), and avoidance t- "Not tested; this habitat t)pe was not recorded in the study area. 332 Great Basin Naturalist [Volume 54 miscellaneous habitats in both years, selected grain and plowed habitats in excess of their availabilit\', and used pasture in proportion to its availability. Along the Bear River transect, cranes avoided sagebrush habitat, selected grain and pasture habitats, and used alfalfa and plowed habitat types in proportion to their availability. Results from other habitat types along the two transects either varied between years or were not tested due to pat- terns of sampling or structural zeros in the contingency tables. We examined distribution of cranes using grain- and cornfields in 1992 and found that certain grainfields received preferential use in Cache Valley {X^ = 272.4, df = 72, F < .001) and in Bear River Viilley {X^ = 42.6, df = 10, P < .001). Insufficient data were available for cornfields in 1992. Cranes tended to exploit field interiors but were broadly distributed within fields. In 1991-92 mean distance-to- field-edge for flocks in corn was 82.2 m (n = 7, SE = 21.2) and 72.1 m {n = 2,50, SE = 7.26) for flocks using grainfields. Cranes were recorded feeding in 75% of our observations. A goodness -of- fit test was used to examine the distribution of cranes feeding in each habitat type in comparison to habitat availability (Table 2). Feeding cranes were not distributed randomly in 1991 (A- = 242.8, df = 13, P < .0005) 'or 1992 {X^ = 332.4, df = 14, P < .0005). Distribution of feeding cranes approximated distribution of all cranes obsei^ved, except in the case of riparian habitat along the Bear River transect. While cranes used this habitat type disproportionate- ly to its availability in 1991, they appeared to feed in this habitat type in proportion to its availability. Data for 1992 were insufficient for analysis. Crane diurnal use of field types varied with time of day (summer diurnal sampling: X^ = 91.04, df = 48, P = .0002; fall diurnal sam- pling: X2 = 72.65, df = 24, P < .01). Crane numbers peaked after sunrise, decreased steadily throughout the day, and then increased again before sunset. Discussion Crop depredation attributed to cranes was reported by farmers in Cache, Rich, and Lincoln counties (Mclvor 1993). Crane dam- age occurred in spring in the Cache Valley transect, primarily with newly planted corn crops. Cranes pulled up corn plants and con- sumed the still-attached seed. Farmers also reported minor damage from cranes trampling emergent alfalfa and small grains (winter wheat, barley, oats). The growing season along the Bear River transect in Rich and Lincoln counties is too short for corn production, and crop damage occurred primarily in the fall, affecting small-grain crops (Lockman et al. 1987, Mclvor and Conover 1994). Some tram- pling damage in spring was also leported in this area. Cranes concentrated activities in small-grain fields during our surveys. Fields planted in corn constituted only 7% of available habitat, and <3% of cranes sighted were in corn. Most activity in cornfields occurred during germi- nation or while plants were young. Thereafter, cranes avoided cornfields until han^est. Large expanses of sagebrush habitat were little used, although they constituted about 60% of available habitat. Sagebrush habitat may have reduced crane foraging efficiency by creating dense cover, limiting movement, and offering few plant foods. Agricultural fields in Bear River Valley were surrounded by vast expanses of sagebrush, a condition that may have concentrated cranes into agricultural fields. Feeding activity closely approximated pat- terns of habitat use, suggesting cranes fed with the same intensity in each habitat t\'pe. Migrat- ing cranes in Nebraska relied on a diversity of habitats to provide various components of their diet (Reinecke and Krapu 1986). Alfalfa fields (Walker and Schemnitz 1987) and grass- lands (Reinecke and Krapu 1986) provided a source of invertebrates for cranes. Although invertebrates may provide certain proteins absent from plant foods (Reinecke and Krapu 1986), they comprise only a small component of the diet, varying from 3% (Reinecke and Krapu 1986) to 27% (Mullins and Bizeau 1978). In this study cranes appeared to avoid feeding in Cache Valley alfalfa fields, possibly obtain- ing invertebrates from pastures or plowed fields. In Bear River Valley cranes fed actively in pasture. Corn (Reinecke and Krapu 1986) and cere- al grains (Krapu and Johnson 1990) provide important nutrient sources for fat synthesis in cranes. Habitat use and feeding activity in grainfields, along both transects and in both 1994] Sandhill Crane Habitat Preference 333 Table 2. Distribution of Sandliill Cranes observed feeding in various habitat types in Cache Valley (C), Utali, and Bear River Valle>' (B), Utah and WVoming, in 1991 and 1992. Habitat Study # crane Expected Proportion Proportion 95% family Use area obser- # crane of study obsei-ved coni Pidence preference" vations obser- area (pjo) in each intei"val on vations area (P,) 1991 Alfalfa C 5 17 0.130 0.037 .0< Pi^ .090 - B 11 9 0.158 0.186 .068 < Pi^ .356 0 Corn C 4 9 0.065 0.030 .0< Pi^ .075 0 B 0 0 0.0 0.0 ntb nt^' Fallow C 0 3 0.023 0 nt"^ nt'' B 0 0 0.0 0.0 nt'' nt'' Grain C 44 23 0.174 0.328 .209 < Pi^ .440 -f- B 14 6 0.096 0.237 .102 < Pi^ .407 -1- Misc. C 2 20 0.152 0.015 .0< P,^ .045 - B 1 3 0.044 0.017 .0< P,^ .068 0 Pas tine C 57 48 0.359 0.425 .313 < P,^ .560 0 B 27 4 0.070 0.458 .271 < Pi^ .644 -1- Plowed C 18 4 0.033 0.134 .052 < Pi^ .239 + B 1 1 0.009 0.017 .0< Pi^ .082 0 Riparian C 4 9 0.065 0.030 .0< Pi^ .075 0 B 4 1 0.009 0.068 .0< Pi^ .169 0 Sage C 0 0 0.0 0.0 ntl' ntl' B 1 36 0.614 0.017 .0< Pi^ .085 - 1992 - - - Alfalfa C 20 43 0.194 0.091 .041 < Pi^ .155 - B 7 6 0.121 0.149 .021 < v< .319 0 Com C 1 14 0.065 0.005 .0< Pi^ .023 - B 0 0 0.0 0.0 nt'^ nt'^ Fallow C 13 17 0.075 0.059 .023 < P,^ .100 0 B 1 1 0.010 0.021 .0< Pi^ .085 0 Grain C 106 40 0.183 0.482 .391 < Pi^ .577 + B 19 3 0.061 0.404 .213 < Pi^ .617 + Misc. C 2 24 0.108 0.009 .0< Pi^ .032 - B 0 1 0.020 0.0 nt^ nf^ Pasture C 54 59 0.269 0.245 .173 < Pi^ .327 0 B 15 7 0.152 0.319 .170 < Pi^ .532 + Plowed C 22 9 0.043 0.100 .050 < Pi^ .164 + B 4 1 0.010 0.085 .0< Pi^ .234 0 Riparian C 2 14 0.065 0.009 .0< Pi^ .032 - B 0 2 0.051 0.0 nt^ nt^ Sage C 0 0 0.0 0.0 ntb nt^ B 1 27 0.576 0.021 .0< P,^ .085 - •'Habitat use is e.xpressed as selection for ( + ), use in proportion to availability (0), and avoidance (-). ''Not tested; this habitat type was not recorded in the study area. •^Not tested; insufficient observed frequencies to test hypothesis. years, were greater than expected. Although midseason grainfields are unhkely to provide dietaiy components other than invertebrates, cranes probably forage for waste grain in spring stubble and for ripening and waste grain before and after fall hai^vest. Certain grainfields, and possibly certain cornfields, are more attractive than others to cranes. Any burden imposed on the agricul- tural community by crane depredation is not shared evenly by producers. Determining why certain fields are more attractive to cranes and lessening these attractants may help reduce crane problems. Iverson et al. (1987:456) reported that "over 90% of the variation in dis- tribution of staging cranes [in Nebraska] could be explained by the composition and juxtapo- sition of essential habitat types." Certain fields in our study area may receive chronic use because of their proximity to other habitat types, such as wetlands and roost sites, or because they possess characteristics that enhance predator detection and escape. It is unlikely that crane presence has a significant negative effect on productivity of pasture, hay, and alfalfa fields. However, the 334 Great Basin Naturalist [Volume 54 concentration of cranes in small-grain fields, particularly in the fall, poses a potential eco- nomic threat to farmers. Delayed harvest of grains in fall due to wet weather is likely to exacerbate the problem because standing grain remains available to an increasing num- ber of prestaging cranes (Lockman et al. 1987). Diurnal changes in habitat use may allow cranes to forage while minimizing heat stress. Cranes using pastine and hayfields in midafter- noon were probably loafing before feeding prior to sunset. For reasons that are unclear, activity patterns observed in Cache Valley were less distinct in Bear River Valley. Cranes may have moved less, visiting fewer habitat types as a result of the pattern of habitat distri- bution in Bear River Valley. Additionally, the Bear River Valley survey may have included a greater proportion of paired individuals, which remained on territories during early summer (Johnsgard 1991) and subsequently visited fewer habitat types. Crane depredation occurs under two dis- parate conditions: in association with spring planting of corn and just before fall hai-vest of cereal grains. Encouraging rapid germination of corn and early harvest of grains would mini- mize availability of these resources to cranes during periods of susceptibility to depreda- tion. Crane damage was concentrated in a few fields, rather than being evenly distributed in all fields, indicating that nonlethal techniques to alleviate these problems need to be focused in these same fields. Farmers who experience chronic depredation problems may wish to consider the economic feasibility of producing crops less prone to crane damage. Acknowledgments Project funding was provided by the Utah State University Wildlife Damage Management Program and the Utah Division of Wildlife Resources. We thank D. A. Brink, K. V Dustin, and H. Low for field assistance. R. S. Krannich, J. A. Bissonette, and P. M. Meyers provided helpful manuscript reviews, and S. L. Durham gave invaluable assistance with matters statis- tical. Literature Cited Bridgerland Audubon SociEri. 1990. Resident Saiulhill Crane popiilatioTis in Cache CJounty, Utah; numbers and distribution. L'Tipuhhshed report, Logan, Utah. 8 pp. Dhkwikn, R. C, and E. G. Bizeau. 1974. Status and dis- tribution of Greater Sandhill Cranes in the Rocky Mountains. Journal of Wildlife Management 38: 720-742. Drewien, R. C, W. M. Brown, and J. D. Varley. 1987. The Greater Sandhill Crane in Yellowstone National Park: a preliminaiy survey. Proceedings of the North American Crane Workshop 4: 27-38. IVERSON, G. C, R A. VOHS, AND T. C. Tacha. 1985. Habitat use by Sandhill Cranes wintering in western Texas. Joumalof Wildlife Management 49: 1074-1083. . 1987. Habitat use by mid-continent Sandhill Cranes during spring migration. Journal of Wildlife Management 51: 448-458. Johnsgard, E A. 1991. Crane music: a natural history of American cranes. Smithsonian Institution Press, Washington, D.C. 136 pp. Krapu, G. L., and D. H. Johnson. 1990. Conditioning of Sandhill Cranes during fall migration. Journal of Wildlife Management 54: 23-^238. Krapu, G. L., D. a. Facey, E. K. Fritzell, and D. H. Johnson. 1984. Habitat use by migrant Sandhill Cranes in Nebraska. Journal of Wildlife Management 48: 407-417. Lockman, D. C, L. Serdiuk, and R. Drewien. 1987. An experimental Greater Sandhill Crane and Canada Goose hunt in Wyoming. Proceedings of the North American Crane Workshop 4: 47-57. McIvoR, D. E. 1993. Incidence and perceptions of Sandhill Crane crop depredations. Unpublished master's thesis, Utah State University, Logan. 75 pp. McIvoR, D. E., and M. R. Conover. 1992. Sandhill Crane habitat use in northeastern Utah and south- western Wyoming. Proceedings of the North American Crane Workshop 6: 81-84. . 1994. Impact of Greater Sandhill Cranes foraging on corn and barley crops. Agriculture, Ecosystems, and Environment 49: 233-237. MiNlTAB. 1989. Minitab statistical software. Release 7.2. Minitab, Inc., State College, Pennsylvania. MuLLiNs, W H., and E. G. Bizeau. 1978. Summer foods of Sandhill Cranes in Idaho. Auk 95: 175-178. Neu, C. W, C. R. Byers, and J. M. Peek. 1974. A tech- nique for analysis of utilization availability data. Journal of Wildlife Management 38: 541-545. Reinecke, K. J., and G. L. Kr.4PU. 1986. Feeding ecolog>' of Sandhill Cranes during spring migration in Nebraska. Journal of Wildlife Management 50: 71-79. Rowland, M. M.. L. Kinter, T. Banks, and D. C. Lockman. 1992. Habitat use by Greater Sandhill Cranes in Wyoming. Proceedings of the North American Crane Workshop 5: 82-85. SAS Institute, Inc. 1988. SAS/STAT user's guide. Release 6.03 edition. SAS Institute, Inc., Caiy North Carolina. 1028 pp. Walker, D. L., and S. D. Schemnitz. 1987. Food habits of Sandhill Cranes in relation to agriculture in cen- tral and southwestern New Mexico. Proceedings of the North American Crane Workshop 4: 201-212. Received 2 September 1993 Accepted 20 April 1994 Great Basin Naturalist 54(4), © 1994, pp. 335-341 SELENIUM GEOCHEMICAL RELATIONSHIPS OF SOME NORTHERN NEVADA SOILS Stephen Pooled Glenn Gross^, and Robert Potts' Abstract. — Soil samples, one fioni each of 10 locations in northern Nevada, were evaluated for redox potential, total and e.xtractable seleniimi, phosphate, free iron oxide, total and ferrous iron. Mole fractions for extractable selenium species were calculated from redox potentials. Data were used to extrapolate general geochemical relationships for soil selenium at the sample sites. Results obtained from one sample per location allowed only the most general conclusions to be drawn. Soil phosphate levels, which affect the adsorption of selenite species on iron oxide by competing for adsoiption sites, were not correlated with levels of extractable selenium in this study. This would suggest that selenium would exist in solution, having been displaced from adsorption sites by phosphorus. Ferrous iron, iron oxides, and redox potential had a combined effect on the level of extractable selenium at all sites. Soils in this study support selenite species that are not readily available to plants and therefore could not support vegetation adequate in Se. Key words: selenium, soil, redox potential, geocheinisfnj, plant bioavailability. Selenium (Se) is a significant micionutrient in production agriculture; because of this, knowledge of the Se status of rangelands is important. Distribution of total and extractable Se can vary widely over short geographic dis- tances (Fisher et al. 1990). Because the geology of Nevada is complex, relationships between critical plant Se levels and geological forma- tions are difficult to define. Recently, a review of the Se status of soils, plants, and animals in Nevada reported deficiency problems in west- em Nevada, variable amounts in northern and central portions of the state, and adequate lev- els in the southern portion of the state. Selenium accumulator plants grow throughout Nevada on limited seleniferous geological for- mations (Poole et al. 1989; Fig. 1). The narrow gap between essential and toxic concentra- tions of Se makes it imperative that processes controlling the distribution of this element be understood (McNeal and Balistrieri 1989). Uptake of Se by plants is governed by many soil and plant factors including type of plant, soil pH, clay content, and mineralogy. Most important factors determining uptake are form and concentration in the soil. Chemical form is controlled by redox potential parameters (pe -f pH; Elrashidi et al. 1989, Mikkelsen et al. 1989). Although Se may exist in four oxidation states, selenate (VI) and selenite (IV) are pre- dominant mobile forms in a soil solution and are available for plant uptake. Redox potentials are important in soils, and theoretical relationships can be used to pre- dict and interpret metal solubilities (Lindsay and Sadiq 1983). Redox potentials have been used in Nevada to interpret observed sequences of minerals in an alteration zone in Ely, Nevada (Raymahashay and Hollard 1969), interpret hydrogeochemistry of the Red Rock, Nevada, area (Fricke 1983), and evaluate trace- element content of sediment and water in west central Nevada (Rowe et al. 1991). Soil redox potential data are lacking for the state. The puipose of this study was to investigate soil Se geochemical relationships for 10 Nevada sites using redox potential (pe -I- pH) and extractable and total Se levels. Phosphate (P), iron (Fe), and iron oxide (Fe203) levels were also investigated to determine their effect on Se bioavailabilit\' for plants growing on the soils. Experimental Procedure A soil sample was taken from each of 10 sites: Battle Mountain and Gund Ranch in central Nevada (Eureka and Lander counties); Minden (Douglas County); Reno, Red Rock area north of Reno, Spanish Springs (Washoe 'Sierra Environmental Monitoring, Inc.. 1135 Financial Blvd., Reno, Nevada 89502. 335 Great Basin Naturalist [Volume 54 Fig. 1. Selenium in Nevada forage. Very low = 81% of samples with Se concentration of < 0.01-0.05 ppm Se; variable — 74% of samples with Se concentration of 0.05-0.5 ppm Se; adequate — 78% of samples with Se concentration of 0.1-1.0 ppm Se. Fig. 2. Soil sample locations in northern Nevada: 1-Battle Mountain, 2-Gund Ranch, 3-Spanish Springs, 4-Reno, 5-Minden, 6-Fallon, 7-Salmon Falls Creek, 8-Huntington Viilley, 9-Clover Valley, 10-Red Rock. County); Fallon (Churchill County); Salmon Falhs Creek near Contact and two locations near the Ruby Mountains (Elko County) in Nevada (Fig. 2). Samples were taken appro.xi- mately 12-15 cm below the surface so as to in- clude the root zone. Air-dried samples <2 mm (No. 10) were used for analysis. Redox potentials were measured according to the procedure of Lindsay and Sadiq (1983). Soil suspensions were prepared in conical flasks to contain 50 g air-dried soil and 100 niL deionized water. Each treatment was pre- pared in duplicate, degassed with argon (Ar), stoppered, and shaken. Millivolt readings were taken on soil suspensions with a plat- inum (Ft) electrode and a glass Ag/AgCl refer- ence electrode using an Altex Selection 5000. The platinum/reference electrode system was standardized using a ferrous/ferric ion refer- ence solution (ZoBells; ASTM 1978). Soil sus- pension pH was determined using a combina- tion electrode that was calibrated with stan- dard buffers (ASTM 1978). Suspension pe was calculated from millivolt readings using the relationship pe = Eh(millivolts)/59.2. To obtain total soil Se levels, we digested samples in aliquots of 1:1 hydrochloric acid and 5% potassium persulfate for 15 min fol- lowed by 3.5% o.xalic acid solution for 15 min. The resulting solution was then treated with concentrated hydrochloric acid for 42 min prior to diluting to 100 niL volume with deionized water. Se concentrations of the digests were determined using hydride gener- ation atomic absorption spectroscopy (AAS; Varian SpectrAA 10 with VGA accessoiy). Soluble Se was measured in a saturation paste extract from each soil (Jump and Sabey 1989) using hydride generation AAS. The extract Se concentration was used for calcula- tion of Se species. The mole fraction of soluble Se species was calculated emploving methods ofElrashidietal. (1987). Bicarbonate extractable P was determined using the method of Olsen (Council on Soil Testing and Plant Analysis 1980). To evaluate 1994] Selenium in Northern Nevada Soils 337 Fe203 levels, we extracted 4 g soil overnight with 4 g sodium dithionite (Na2S204) and 75 mL deionized water. Suspensions were fil- tered, brought to volume (Kilmer 1960), and analyzed for Fe by AAS. Total soil Fe was determined by flame AAS on nitric acid digests of soil samples. Ferrous iron (Fe II) in sample soils was determined colorimetrically. Samples were digested using concentrated sulfuric acid and 30% hydrofluoric acid, neutralized with 4% boric acid, and made to volume with deion- ized water (Walker and Sherman 1962). To an aliquot of the digest we added 0.001 M batho- phenanthroline in 50% ethanol and acetate buffer. Isoamyl alcohol extracted the ferrous- bathophenanthroline complex from the solu- tion. The alcohol layer was drained into a 25- niL volumetric flask, made to volume with 95% ethanol, and the absorption of the solu- tion read on a spectrophotometer (Baush and Lomb Spectronic 710) at a wavelength of 538 nm. Standards and blanks were treated simi- larly. Ferrous iron standards were derived from a stock solution of ferrous ammonium sulfate. For the interpretation of data, we used redox and adsorption relationships developed by Howard (1977), Balistrieri and Chao (1987, 1990), and Schwab and Lindsay (1983) for the behavior of Se, Fe, and FO4 and equilibria described by Elrashidi et al. (1987) for Se in soils. Regression and multiple regression analy- ses were performed following methods of Damon and Hai-vey (1987). Regressions were evaluated for significance at the 95% confi- dence level. Results Total Se, extractable Se, redox parameters, and general site descriptions are presented in Table 1 for sample soils. Mole fi-actions of the Se species for each sample are presented in Table 2. Howard (1977) summarized Se geochem- istiy on an Eh-pH diagram and found that Fe, with which Se is closely associated in both oxi- dizing and reducing environments, controls Se geochemistiy. In aerated soil suspensions the Se (IV) oxyanions HSeO^" and Se03^~ are strongly adsorbed by hydrated surfaces of fer- ric oxides over the pH range 2-8; above pH = Tablk 1. Total selenium, extractable selenium, redox potential (pe + pH), and predominant selenium species in soil samples. Selenium* Sample Total Extractable Predominant location (mg/kg) (M.ii/kg) pe + pH Species Battle Mountain'' 0.62 4 10.0 Selenite Gund Ranchl- 0.74 8 11.7 Selenite Spanish Springs'' 0.10 102 7.5 Selenite Reno' 0.11 <2 11.1 ** Minden'l 0.20 50 10.9 Selenite Fallon** 1.2 3 11.1 Selenite Salmon Creekf 0.37 26 8.5 Selenite Clovers 0.13 <2 7.6 ** Huntington« <0.10 <2 9.6 ** Red Rockl' 0.10 7 11.9 Selenite *Air-dried basis **Could not be calculated due to lack of measurable selenium in extract. ^Stewart and McKee 1977 bUSDA-SCS 1978, USDA-SEA 1980 'USDA-SCS 1983 dUSDA-SCS 1984 '■Willden and Speed 1974 fSchrader 1934, Tueller 1975 STueller 1975 ■lUSDA-SCS 1983, Fricke 1983 8, adsorption decreases to complete desorp- tion at pH = 11. Selenite shows a strong affin- ity for Fe203 surfaces (Balistrieri and Chao 1987), forming stable ferric oxide-selenite (Fe2(OH)4Se03) complexes that cause immo- bilization of Se. Selenate on the other hand shows a weaker affinity for oxide surfaces, forming compounds that are soluble and, therefore, mobile (Howard 1977, Elrashidi et al. 1987, Fresser and Swain 1990) and easily transportable in groundwater and available for plant uptake (Lakin 1961). Levels of extractable Fe203 are presented in Table 3. Because adsorption of selenite increases with increasing concentration of ^'^2^3 '^^^^ ^^ the greater number of available binding sites (Balistrieri and Chao 1987), it follows that soils at Gund Ranch and Spanish Springs have the potential to adsorb the largest amounts of selenite. The Clover Valley sample would be least likely to adsorb selen- ite. pH would not have an effect on the ability of Fe203 to adsorb selenite for all the soils except Clover Valley (pH > 8). Levels of Fe.903 and Fe(II) had a combined effect on the amount of Se extracted from soils (r = .3196). Iron oxide and Fe(II) were not affected by redox potential (r= .0705). 338 Great Basin Naturalist [Volume 54 Table 2. Log mole fraction of selenium species-' Species Sample'^ SfO, HSfO., IloSeO, St'03 HSe03 HoSeO., Se^- liSe li.Se 1 -9.2 -14.9 -29.2 -0.15 -0.55 -5.6 -22.9 -15.6 -19.6 2 -5.9 -12.3 -22.6 -0.02 -1..32 -7.0 -33.3 -26.6 -31.1 3 -12.2 -17.9 -24.4 -0.18 -0.48 -7.3 -13.9 -6.3 -10.4 4 -7.4 -12.3 -21.2 -0..55 -0.15 -4.5 -29.9 -21.9 -24.9 5 -6.1 -10.7 -19.1 -0.86 -0.06 -3.9 -.35.1 -26.6 -29.9 6 -7.5 -12.9 -22.2 -0.30 -0.30 -5.0 -28.5 -20.8 -25.3 7 -19.9 -19.8 -28.8 -0.59 -0.39 -4.9 -8..39 -0.5 -3.79 ■'Base 10 loiiarithiii ''Sample icientification: l-Battle Mountain, 2-Gnnd Ranch, .3-Salmon Creek, -4-Fallon, .5-Recl Rock, 6-Minden, -Spanish Springs Activity of Fe(II) is controlled by FeC03 (siderite) at pe + pH < 8 and by Fe3(OH),c^ (ferrosic hydroxide) at pe + pH > 8. In sys- tems below pH = 6.0 with stable redox, less- soluble iron oxides such as geothite (alpha- FeOOH) can control Fe solubility (Schwab and Lindsay 1983). Levels of Fe(II) are pre- sented in Table 3. Ferrous iron levels had a significant effect on the amoimt of Se extract- ed from soils (r = .5843). Siderite would con- trol Fe(II) activity in the Spanish Springs and Clover Valley samples. Fen'ous iron activity in the remaining samples would be controlled by ferrosic hydroxide. Hematite would control ferric iron activity of sample soils except for Spanish Springs and Clover Valley where fer- roselite would control ferric iron (Fe III). Any remaining Fe(III) could be associated with hydrous selenite complexes. ' Total Fe levels ranged from 8700 to 28000 mg/kg. Total Se and total Fe were not correlat- ed for these soils (r = -.1028). Redox potential and total Fe had an effect on soil Fe(II) and FeaOg content (r = .4565 and r = .3998, respectively). A decrease in selenite adsorbed on iron oxide would depend on the adsoiption density of selenite (moles of ion adsorbed/kg of oxide; Balistrieri and Chao 1987). Other ions in a soil solution, including P, can com- pete with selenite for adsoiption sites on solid surfaces. Anion adsorption relies on several factors including pH, formation of solution complexes, and competing adsorbates (Mikkelsen et al. 1989). Phosphate displaced all the adsorbed selenite on allophane clays (Rajan and Watkinson 1976) and has been shown to desorb selenate (Singh et al. 1981). Most P found in alkaline soil exists as calci- um phosphate (CaHP04; Lindsay and Moreno 1960, Boyle and Lindsay 1986). Phosphate levels of study soils ranged from 8.9 to 147 mg/kg P To evaluate the effect of P on selenite adsorption, we calculated total anion concen- tration ratios {(anion)/(selenite)}. Results are presented in Table 4. A stronger affinity and larger concentration of one anion should result in more sites being occupied by that anion vs. another (Balistrieri and Chao 1990). Levels of P and Fe203 and P did not have an effect on the amount of Se extracted from the study soils (r = .3030 and r = .1019, respectively). Absence of a significant correlation between levels of Fe203 and P and extractable Se sug- gests that Se would exist in solution for these soils. Phosphorus would have displaced Se from available binding sites. Selenium-phos- phate interactions are generally not of conse- quence for plant uptake of Se except for plants growing where levels of Se are inadequate to meet animal nutritional needs (Mikkelsen et al. 1989). Mean concentration of total Se in soils and surficial materials for the western United States is 0.23 mg/kg, with an obsei-ved range of < 0.10-4.3 mg/kg Se. Most soils from low- Se areas in the United States contain <0.5 mg/kg Se (National Research Council 1983, Boon 1989). A limited sui-vey of Nevada soils, as part of a trace-element survey of soils throughout the United States, revealed a variety of Se levels (Shacklette et al. 1974). Observa- tions in the Fallon area demonstrate that seleniferous spots may be found in alluvial Pliocene deposits occurring over a large part of Nevada, particularly in the Carson and Humboldt sinks (Lakin and Byers 1948, Rowe et al. 1991). Total soil Se levels in this study ranged from <0.10 to 0.74 mg/kg. As with many other elements, total concentration of Se in soils shows little relationship to Se concen- tration in plants grown in those soils. 1994] Selenium in Northern Nevada Soils 339 Table 3. Total iron, extractable iron oxides, and ferrous ron in soil samples. Table 4. Extractable phosphate phosphorus, selenite selenium, and phosphate/selenite molar ratios. Mill igrams/kilogr am'' Sample location Total Fe Oxide Fe Ferrous Fe Battle Mountain 17000 4.575 3290 Gund Ranch 28000 12175 6090 Spanish Springs 19000 8900 7920 Reno 27000 6550 5800 Minden 19000 3950 10940 Fallon 12000 3700 3010 Salmon Creek 12000 2525 8580 Clover Vallev 8700 875 7060 Huntington Vallev 19000 4350 5300 Red Rock 17000 6350 3220 ■'Air-drit'd basis Workmann and Soltanpour (1980) have reported that water-soluble Se is usually <50 /ig/kg in normal cultivated soils. Soils in this study had soluble Se levels of < 1-102 [xg/kg, existing primarily as selenite. The significant correlation between pe + pH and levels of extractable Se (r = -.4475) suggests a relation- ship between the amount of Se available for plant uptake and soil redox potential at the study sites. Certain native plants of the Great Basin have tendencies to aggregate in relation to temperature gradients, precipitation patterns, physiography, and soils (Tueller 1975, Reveal 1979). Approximately 80% of all forage and grain sampled in western Nevada has been shown to contain <0.10 ppm Se, less than the dietary requirement of 0.10 ppm for grazing animals (Kubota et al. 1967, McDowell et al. 1983, National Research Council 1983). Soil Se concentration can vaiy widely over a veiy short geographic distance (Fisher and Munshower 1991). Upper rangeland forage of extreme northeastern Nevada growing on Idavada volcanics and silicic rocks of volcanic origin was found to contain low levels of Se (Carter et al. 1969). In contrast, lower range- lands surrounding a portion of these areas pro- duce forage adequate in Se (Carter et al. 1968). Alfalfa samples taken from the Carson Valley area were found to be below (<0.05 ppm) the dietary requirement of 0.1 ppm Se (Allaway and Hodgson 1964). Forage at Gund Ranch has been shown to contain 0.13-0.17 ppm Se (Poole et al. 1989). Selenium indicator plants are limited to localized areas on selenif- erous geological formations in Nevada (Poole et al. 1989) and are not reported to occur with- in sample site areas. Sample Phosphate Selenite {(Phosphate)/ location (lOE-5 M) (lOE-7 M) (selenite)} Battle Mountain 3.4 1.3 271 Gund Ranch 3.5 2.5 136 Spanish Springs 15.2 28.9 52 Reno 23.9 0 0 Minden 1.4 29.1 5 Fallon 3.4 1.3 269 Salmon Creek 4.3 5.0 85 Clover 5.8 0 0 Huntington 4.6 0 0 Red Rock 4.1 2.5 160 Forage at the Fallon site would not be expected to contain appreciable amounts of Se. Soil redox potential does not allow for for- mation of plant-available selenate. In areas adjoining Carson Valley, including Fallon, white muscle disease in sheep has been a rec- ognized problem (Vawter and Records 1947, Kuttler and Marble 1958) for animals raised on native forage. Soil at the Gund Ranch site supports a small fraction of selenate, allowing for growth of forage marginally deficient in Se. Grazing cattle have been found to be border- line deficient in plasma Se at the Gund Ranch site (Poole et al. 1986). Upper rangeland for- age in the Salmon Creek area would be defi- cient in Se because of lack of available soil selenate. Samples lacking measurable amounts of extractable Se would not support growth of Se-bearing forage. Conclusion Total and extractable Se, redox potential, pH, and P Fe(II), and Fe203 levels were dif- ferent for each of the sample sites. Redox potential and Fe(II) and free Fe203 levels would affect the c^uantity of Se available for plant uptake in study soils. Anion concentra- tion ratios indicate that P would influence adsorption of selenite on iron oxide. Soils in this study support selenite species that are not readily available to plants and therefore could not support vegetation adequate in Se. Soil Se concentration can vary widely over a very short geographic distance. Nevada's complex geology therefore requires evaluation of the Se status of soils and vegetation on a site basis. Further studies are needed to devel- op a better understanding of the Se status of the state. 340 Great Basin Naturalist [Volume 54 Literature Cited AlL/Way, W. H., and J. E Hodgson. 1964. Sympo.siuni on nutrition, forage and pastures: selenium in forages as related to the geological distribution of nuiscular dystrophy in livestock. Journal of Animal Science 23: 271-277.' ASTM. 1978. Annual book of ASTM standards. Part .31. Water. Philadelphia, Pennsylvania. BaLISTRIKRI, L. S., and T. T. ChaO. 1987. Selenium adsorption by geothite. Soil Science Society of America Journal 51: 1145-1151. . 1990. Adsorption of selenium b\' amorjohous iron ().\\hydro.\ide and manganese dio.xide. Geochimica et Cosmochimica Acta 54: 739-751. Boon, D. Y. 1989. Potential selenium problems in Great Plains soils. Pages 107-121 in L. W. Jacobs, ed.. Selenium in agriculture and the environment. Soil Science Society of America Special Publication No. 23. Madison, Wisconsin. Boyle, E W, and W. L. Linds.w. 1986. Manganese phos- phate equilibrium relationships in soils. Soil Science Society of America Journal 50: 588-593. Carter, D. L., M. J. Brown, W H. All^way, and E. E. Gary. 1968. Selenium content of forage and hay crops in the Pacific Northwest. Agronomy Journal 60: 532-534. Carter, D. L., C. W Robins, and M. J. Brown. 1969. Selenium concentrations in forage on some high northwestern ranges. Journal of Range Management 23: 234-238. Council on Soil Testing and Plant Analysis. 1980. Pages 47-51 in Handbook on reference methods for soil testing. Athens, Georgia. Damon, R. A., Jr., and W. R. Harvey. 1987. Experimental design, ANOVA, and regression. Harper and Row Publishers, New York. Elrashidi, M. a., D. G. Adriano, and W. L. Lindsay. 1989. Solubility, speciation and transformation of selenium in soils. Pages 51-63 in L. W. Jacobs, ed.. Selenium in agriculture and the environment. Soil Science Society of Aiuerica Special Publication No. 23. Madison, Wisconsin. Elrashidi, M. A., D. C. Adriano, S. M. Workman, and W. L. Lindsay. 1987. Chemical equilibria of seleni- um in soils: a theoretical development. Soil Science 144: 141-152. Fisher, S. E., and E E Munsiiower. 1991. Selenium issues in drastically disturbed land reclamation plan- ning in arid and semiarid environments. In: R. C. Severson, S. E. Fisher, Jr, and L. R Cough, eds.. Proceedings of the 1990 Billings land reclamation symposium on selenium in arid and semiarid envi- ronments, western United States. U.S. Geological Survey Circular 1064. FrickE, R. a. 1983. The hydrogeochemistr> and acjueous uranium distribution of Petersen Mountain and Red Rock Valley, Washoe County, Nevada. Unpublished master's thesis. University' of Nevada, Reno. Howard, J. H. 1977. Geochemistiy of selenium: formation of ferroselite and selenium behavior in the vicinity of oxidizing sulfide and uranium deposits. Geochimica et Cosmochimica Acta 41: 166.5-1678. Jump, R. K., and B. R. Sabey. 1989. Soil test extractants for predicting selenium in soil. Pages 95-105 in L. W Jacobs, ed.. Selenium in agriculture and the environ- ment. Soil Science Society of America Special Publication No. 23. Madison, Wisconsin. Kilmer, V. J. 1960. The estimation of free iron oxides in soils. Soil Science Socieb,- Proceedings 24: 420-421. Kubota, J., W. H. Allaway, D. L. Carter, E. E. Gary, AND V. A. Lazar. 1967. Selenium in crops in the United States in relation to selenium-responsive dis- eases of animals. Journal of Agricidtural Food Chemistry 15: 448-453. KUTTLER, K. L., AND D. W Marble. 1958. Relationships of serum transaminase to naturally occurring and artificial!} induced white muscle disease in calves and lambs. American Journal of Veterinary Research 19: 632. Lakin, H. W. 1961. Geochemistiy of selenium in relation to agriculture. Selenium in agriculture. Agricultural Handbook 200, US DA, 3-12. Lakln, H. W, and H. G. Byers. 1948. Selenium occur- rence in certain soils in the United States, with a discussion of related topics: seventh report. USDA Technical Bulletin 950. Washington, DC. Lindsay, W L., and E. C. Moreno. 1960. Phosphate phase equilibria in soils. Soil Science Society of America Proceedings 24: 177-182. Lindsay, W L., and M. Sadiq. 1983. Use of pe -I- pH to predict and interpret metal solubility relationships in soils. Science of the Total Environment 28: 169-178. McDowell, L. R., J. H. Conrad, G. L. Ellis, .\nd J. K. LoosLl. 1983. Minerals for grazing ruminants in tropical regions. Bulletin of the University of Florida, Gainesville. McNeal, J. M., and L. S. Balistrieri. 1989. Geochemistr>' and occurrence of selenium: an overview. Pages 1-13 in L. W. Jacobs, ed.. Selenium in agriculture and the environment. Soil Science Society of America Special Publication 23. Madison, Wisconsin. Mikkelsen, R. L., a. L. Page, and E T Bingham. 1989. Factors affecting selenium accumulation by agricul- tural crops. Pages 65-94 in L. W. Jacobs, ed.. Selenium in agriculture and the environment. Soil Science Societ>- of America Special Publication No. 23. Madison, Wisconsin. National Research Council. 1983. Selenium in nutri- tion. Revised edition. National Academy Press, Washington, D.C. Poole, S. C., V. R. Bohman, L. A. Rhodes, and R. TORELL. 1986. The selenium status of range cattle in northeastern and central Nevada. Proceedings of the Western Section of the Aiuerican Society' of Animal Science 37: 220-223. Poole, S. C, V. R. Bohman, and J. A. Young. 1989. Review of selenium in soils, plants and animals in Nevada. Great Basin Naturalist 49: 201-213. Presser, T. S., and W C. Swain. 1990. Geochemical evi- dence for Se mobilization by the weathering of pyritic shale, San Joaquin Valley, California, USA. Applications of Geochemisti-y 5: 703-717. R\JAN, S. S. S., AND J. H. W.\TKlNSON. 1976. Adsorption of selenite and phosphate on an allophane clay. Soil Science Society of America Journal 40: 51-54. Raymahashay, B., and H. D. Holu^rd. 1969. Redox reac- tions accompanying hydrothenual wall rock alter- ation. Economic Geolog\' 64: 291-305. Reveal, J. L. 1979. Biogeography of the Interniountain region. A speculative appraisal. Mentzelia 4: 1-87. 1994] Selenium in Northern Nevada Soils 341 RowE, T. G., M. S. Lico, R. J. Hallock, A. S. Maest, and R. J. Hoffman. 1991. Physical, chemical, and biolog- ical data for detailed study of irrigation drainage in and near Stillwater, Femley, and Humboldt Wildlife Management areas and Carson Lake, west-central Nevada, 1987-89. U.S. Geological Survey Open File Report 91-185. SCHRADER, E C. 1934. The Contact Mining District. United States Department of the Interior Bulletin 847-A. Schwab, A. P, and W. L. Lind.s.ay. 1983. Effect of redo.x on the solubility and availability of iron. Soil Science Society of America Journal 47: 201-205. Shacklette, H. T, J. G. Boerngen, and J. R. Keith. 1974. Selenium, fluorine, and arsenic in surficial materials of the conterminous United States. U.S. Geological Survey Circular 692. Singh, M., N. Singh, and P S. Relan. 1981. Adsoiption and desoiption of selenite and selenate selenium on different soils. Soil Science 132: 134-141. Stewart, J. H., and E. H. McKee. 1977. Geology and mineral deposits of Lander Count\', Nevada. Nevada Bureau of Mines and Geology Bulletin 88. Tueller, R T. 1975. The natural vegetation of Nevada. Mentzelia 1: 3-6, 2.3-28. USDA-SCS. 1978. Advance soil sui^vey data, Gund Ranch, Eureka and Lander counties, Nevada. . 1983. Soil sun'ey of Washoe County, Nevada, south part. . 1984. Soil suwey of Douglas County area, Nevada. USDA-SEA. 1980. Physical, biological', and cultural resources of the Gimd Research and Demonstration Ranch, Nevada. J. A. Young and R. A. Evans, eds. Vawter, L. R., and E. Records. 1947. Muscular dystro- phy (white muscle disease) in young calves. Journal of the American Veterinary Association 110: 152-157. Walker, J. L., and G. D. Sherman. 1962. Determination of total ferrous iron in soils. Soil Science 5: 325-328. Willden, C, and R. C. Speed. 1974. Geologv' and mineral deposits of Churchill County, Nevada. Nevada Bureau of Mines and Geology Bulletin 83. University of Nevada, Reno. Workman, S. M., and P. N. Soltanpour. 1980. Importance of prereducing selenium (VI) and decomposing organic matter in soil extracts prior to determination of seleniimi using hydride generation. Soil Science Society of America Journal 44: 1331-1333. Received 6 August 1993 Accepted 19 April 1994 Great Basin Naturalist 54(4), © 1994, pp. 342-350 STATUS AND DISTRIBUTION OF THE LARIDAE IN WYOMING THROUGH 1986 Scott L. Findliolt' Abstmct. — To date, 17 species of Laridae have been reported in Wyoming. Six of these species have known breed- ing populations in the state: the Ring-billed Gull {Lanis delawarensis), California Gull [Larus californicus), Herring Gull {Lanis argentatus), Caspian Tern {Sterna caspia), Forster's Tern (Stenia forsteri), and Black Tern iChlidonias niger). Of these species, the California Gull is the most abundant and widespread. In 1984 approximately 7300 nests existed in Wyoming at six breeding locations consisting of 10 different colonies. In contrast, only small breeding populations have been discovered for the remaining five species. The Herring Gull is the most recent addition among Laridae known to nest in Wyoming. Likewise, two Ring-billed Gull colonies were recenth' found after not having been documented as lireeding in the state for over 50 years. Although some nesting colonies are threatened b\- habitat loss and human disturbance, most seem secure at present. Limited nesting and foraging habitat precludes establishment of large breeding populations of most Laridae in the state. Key words: Laridae, historical records, inventory, population status, distribution, breeding, Wyoming. Considerable interest and concern exist regarding conservation and management of colonially nesting waterbirds in the United States and elsewhere. These species occupy high trophic levels on aquatic food chains and are sensitive to disturbance of aquatic ecosys- tems, especially loss of wetland habitat and con- tamination by chemical pollutants. In addition, because most of these species nest in colonies, they are vulnerable to human intervention. Findholt (1984) and Findholt and Berner (1988) reported on the status and distribution of the Ciconiiforms in Wyoming. The puipose of this paper is to provide information on the historical and present status and distribution of the Laridae in the state. Methods Data-collection methods utilized were pre- viously reported (Findholt 1984, 1986a, Find- holt and Berner 1988). From 1981 through 1986, but more intensively during the 1984-86 period, I conducted a comprehensive statev^dde inventory for colonially nesting waterbirds in Wyoming. From 4 April to 31 May 1984 and from 28 March to 5 June 1986, I made 15 aerial surveys in fixed-wing aircraft totaling 67.1 h of flight time to locate new nesting areas. Reservoirs, lakes, marshes, and other potential breeding locations not observed during aerial searches were checked from the ground with binoculars or a 20-45X spotting scope. Breeding colonies were usually censused by making total ground counts of nests. Where ground counts were not feasible, I estimated the number of nests (ground estimates). Colonies were censused when most birds were in late incubation or early hatching stages, and cen- suses were based on a single visit. As discussed by Buckley and Buckley (1979), a waterbird colony is difficult to define. There- fore, I used Kushlans (1986) definition, which is an assemblage of nesting birds. Nests were con- sidered active if adult birds were sitting or standing on nests, incubation was observed, or eggs or young were present (McCrimmon 1982). Additional sources of information included a literature review, an examination of the files of the Wyoming Game and Fish Department, and con-espondence with biologists, naturalists, birdwatchers, and others considered knowl- edgeable of the Laridae in Wyoming. This paper includes records through 31 December 1986. 'Wyoming Game and Fish Department, 260 Buena Vista Drive, Lander, Wyoming 82.520. Present address: Oregon Department of Fish and Wildhfe, Forestry and Range Sciences Laboratory, 1401 Gekeler Lane, La Grande, Oregon 978.50. 342 1994] Laridae of Wyoming 343 Results and Discussion Pomarine Jaeger There is one record of the Pomarine Jaeger {Stercorarius pomarinus) from Wyoming. J. and V Herold obsei-ved an adult individual at Burlington Lake (Goldeneye Reservoir), 24 km northwest of Casper, Natrona County, on 15 May 1980. On the following day, O. K. Scott and B. Stratton saw the jaeger at the same location and confirmed its identity. According to the AOU Check-list of North American Birds (1983), the Pomarine Jaeger breeding range occurs along northern coastal areas in North America. Thus, only accidental occurrence is expected in Wyoming. Parasitic Jaeger O. K. Scott discovered the first Parasitic Jaeger {Stercorarius parasiticiis)in Wyoming at Soda Lake, Casper, Natrona County, on 2 September 1962. Two more Parasitic Jaegers were observed at Jackson Lake, Grand Teton National Park, Teton Count\', on 22 June 1975 by M. and B. Raynes. On 24 October 1977, H. Downing and M. Collins reported one imma- ture individual at Lake DeSmet, near Buffalo, Johnson County. One year later another immature bird was observed at Lake DeSmet by M. Collins on 28 August. An immatine Para- sitic Jaeger was seen at Soda Lake, Natrona County, by J. and V. Herold on 14 November 1981. On 4 September 1983, G. Scott found one individual at Bates Creek Reservoir, south of Casper, Natrona County. The most recent record of this species is from Healy Reservoir, east of Buffalo, Johnson County, when H. Downing and M. Collins observed an adult bird on 21 June 1985. The Parasitic Jaeger is mostly pelagic, breeds north of the conterminous United States, and generally winters offshore along ocean coasts (AOU 1983). Therefore, only acci- dental occurrence is anticipated in Wyoming. Franklin's Gull The first record of the Franklin's Gull {Lams pipixcan) in Wyoming is a specimen collected near Wheatland, Platte County, on 6 May 1912 (Grave and Walker 1913). McCreary (1939) stated that this species occasionally occurred in tlie state, and on 5 May 1933, 37 individuals were seen near Torrington, Goshen County; some birds remained in the area until 12 Mav. Also, A. B. Mickey observed a Franklin's Gull near Lake Hattie, soutliwest of Laramie, Albany County, on 7 May 1933 (McCreary 1939). Oakleaf et al. (1982) considered this species a common summer resident and recorded it in 20 (71%) of 28 degree blocks and breeding in one block. The only nesting record is from Beck Lake near Cody, Park County, where U. Kepler found 10-20 nesting Franklin's Gulls in 1977 (Kingeiy 1977). In my intensive statewide survey for Franklin's Gull nesting sites, I found none at Beck Lake or elsewhere in Wyoming. Based on the lack of suitable nest- ing habitat at Beck Lake, the validity of this breeding record is questionable. Breeding records exist for this species in adjacent states of Idaho (Larrison et al. 1967, C. H. Trost personal communication), Montana (Skaar 1980), South Dakota (Johnsgard 1979), and Utah (Behle and Periy 1975). Bonaparte's Gull Knight (1902) considered the Bonaparte's Gull {Lams Philadelphia) a rather rare migrant in Wyoming and provided details of several records from the state. In addition to Knight's records. Grave and Walker (1913) reported one individual taken from near Sheridan, Sheridan County, by Metz. McCreaiy (1939) indicated that the Bonaparte's Gull was a frequent migrant in eastern Wyoming and sometimes common. This species has been listed as occur- ring in Yellowstone National Park (Skinner 1925). More recently, Oakleaf et al. (1982) con- sidered the Bonaparte's Gull an uncommon migrant and reported it from 10 (36%) of 28 latilong blocks. Because this species breeds north of the conterminous United States (AOU 1983) and does not appear to be expanding its range southward, it is highly unlikely that Bonaparte's Gulls will be discovered nesting in Wyoming. Heermann's Gull On 26 September 1984, O. K. Scott et al. dis- covered a Heermann's Gull {Lams heermanni) at Soda Lake, approximately 3 km north of Casper, Natrona County (Kingery 1985). This is the first record of this species in Wyoming. The Heermann's Gull breeds in the vicinity of Baja California and is a coastal species rang- ing from southern British Columbia south to Guatemala (AOU 1983). Thus, only accidental occurrence is expected in Wyoming. 344 CiREAT Basin Naturalist [Volume 54 Mew Gull Two historical records exist in Wyoming for the Mew Gull {Larus canus). One juvenile bird was collected by V Bailey on Lake Fork, a tributar\' of the Careen River in the Wind River Mountains, Sublette County, on 28 August 1893 (Oberholser 1919). This speci- men is located in the U.S. National Museum. Another Mew Gull was taken near Laramie, Albany County, by A. E. Lockwood prior to 1913 (Grave and Walker 1913). No recent records exist for this species in the state. According to the AOU Check-list of North American Birds (1983), the Mew Gull breeds north of the contiguous United States. Based on the paucity of reports for this species in states that adjoin Wyoming, only accidental occurrence is anticipated in the state. Ring-billed Gull In the 1920s the Ring-billed Gull {Lams delawarensis) nested on the Laramie plains, Albany County, and on Yellowstone Lake, Yellowstone National Park (Knight 1902, Skinner 1917, Kemsies 1930). It is difficult to assess when the Ring-billed Gull disappeared as a breeding species in these two areas of the state. This species no longer breeds on the Laramie plains (Raper 1975, Findholt personal observation). Also, the Ring-billed Gull no longer nests in Yellowstone National Park (Schaller 1964, Diem and Condon 1967, K. L. Diem personal communication). Two active Ring-billed Gull colonies were present at two locations in Wyoming during the 1984-86 period. On 21 May 1984, I count- ed 102 adults of this species and 70 nests with eggs at Soda Lake (42°54'N, 106° 18' W), about 3 km north of Casper, Natrona County (Findholt 1986b). Although Ring-billed Gulls continued to nest at Soda Lake in 1985 and 1986, the colony was not censused. One addi- tional Ring-billed Gull nesting colony was found in Wyoming at Ocean Lake (43°07'N, 108°35'W), approximately 24 km northwest of Riverton, Fremont County. On 22 May 1985, I counted 10 adults and 6 nests containing 2-3 eggs each on Peninsula Island. Twenty-three active nests were present on 31 May 1986. Breeding records exist in adjoining states of Idaho (Larrison et al. 1967, C. H. Trost per- sonal communication), Montana (Skaar 1980), and South Dakota (Johnsgard 1979). California Gull Two California Gull {Larus californicus) nesting colonies existed historically in Wyoming. One colony was discovered on the Molly Islands, Yellowstone Lake, Yellowstone National Park, in 1898 when Skinner (1917) estimated about 1000 gulls were present. The other colony, which contained an unknown number of California Gulls, was located on an island in Bamforth Lake, about 15 km north- west of Laramie, Albany County, since 1934 (McCreary 1939). In 1984 there were six breeding locations consisting of 10 different colonies that included approximately 7300 nests (Findholt 1986a). The six sites included both Yellowstone Lake and Bamforth Lake in addition to four recently occupied nesting areas. The new California Gull colonies are located at Pathfinder Reservoir, Carbon County (42°23'N, 106°56'W); Ocean Lake, Fremont County (43°07'N, 108°35'W); Sand Mesa, Fremont County (43°19'N, 108°20'W); and Soda Lake, Natrona Countv (42°54'N, 106°18'W). Although California Gulls continued to nest at all six locations dur- ing the 1985-86 period, none of the colonies were censused. Also, only 5-10 pairs appeared to be present at Sand Mesa in 1985 and none in 1986. The decline in the Sand Mesa nesting population is a result of intentional destruc- tion of nests by the Wyoming Game and Fish Department to supposedly enhance Canada Goose {Branta canadensis) production. The overall increase in the California Gull nesting population in Wyoming since historical times is most likely a result of human-induced environmental changes. These changes have created additional breeding habitat and new food sources (Findholt 1986a). This species breeds in adjacent states of Colorado (Ryder 1978), Idaho (Larrison et al. 1967, C. H. Trost personal communication), Montana (Skaar 1980), and Utah (Behle and Periy 1975). Herring Gull Knight (1902) considered the Herring Gull {Larus argentatus) very rare in Wyoming and noted that there was only one record from the state. This species apparently increased in numbers during the early 1900s and was reported as being a common summer resident at Yellowstone Lake and in the Big Horn Basin (Grave and Walker 1913). Later, McCreary 1994] Laridae of Wyoming 345 (1939) considered the Herring Gull a moder- ately common migrant seen around the lakes of the eastern part of the state and along the North Platte River. Recently, Oakleaf et al. (1982) reported diis species to be an uncommon migrant that had been observed in 12 (43%) of 28 degree blocks. In 1984 three Herring Gull nests were located at Bamforth Lake, Albany Count\' (B. H. Pugesek personal communication). This is the first record of this species breeding in Wyoming. One to three pairs of Herring Gulls continued to nest at Bamforth Lake in 1985 and 1986. Although the Herring Gull has been re- ported from adjoining states of Colorado (Bailey and Neidrach 1965, Ryder 1978), Idaho (Larrison et al. 1967), Montana (Skaar 1980), Nebraska and South Dakota (Johnsgard 1979), and Utah (Behle and Peny 1975), I am unaware of breeding records from these states except for recent evidence of nesting at Antero Reservoir, Park County, Colorado (Chase 1987). Glaucous Gull The first record of the Glaucous Gull {Lams hijperboreus) in Wyoming is of a bird collected by E. Isberg at Lake Hattie, Albany County, on 23 November 1933 (McCreaiy and Mickey 1935, McCreaiy 1939). Another report of this species by A. B. Klots in McCreaiy (1930) was not mentioned later (McCreaiy 1939), possibly because the validity of the report was ques- tionable. There are three recent obsei-vations of the Glaucous Gull in Wyoming. On 23 September 1969, K. L. Diem obsei-ved one individual near Laramie, Albany County. A second Glaucous Gull was seen south of Laramie by W. Hep- worth on 20 June 1979. The most recent re- port of this species from Wyoming is of a bird seen by O. K. Scott at Soda Lake, Natrona County, on 1 May 1982. Few observations of Glaucous Gulls are expected in Wyoming because this species prefers coastal areas and large inland bodies of water and its breeding range is north of the contiguous United States (AOU 1983). Black-legged Kittiwake The Black-legged Kittiwake {Rissa tridactijla) was first reported in Wyoming by Knight (1902). One bird was collected by M. Jeserum near Douglas, Converse County, on 18 November 1898. Two more birds were observed at Dubois, Fremont County, on 22 October 1974 by M. Back (Kingery 1975). This is the only recent record of this species in Wyoming. Because the Black-legged Kittiwake is pri- marily a pelagic species and breeds north of the contiguous United States (AOU 1983), only accidental occurrence is anticipated in Wyoming. Sabine's Gull McCreaiy (1939) indicated that the Sabine's Gull {Xema sabini) is rare in Wyoming. Two specimens were taken by A. E. Lockwood in the fall near lakes on the Laramie plains, Albany County (Grave and Walker 1913). Another Sabine's Gull was found dead near Douglas, Converse County, by K. Cook and A. Hay on 24 October 1937 (McCreary 1939). Since 1954 there have been approximately 24 reports of Sabine's Gulls in Wyoming con- sisting of 28 individual birds. All sightings were made in September and October e.xcept for a subadult observed at Lake DeSmet, Johnson County, by J. Daly on 7 June 1981. The Sabine's Gull has been located in 7 (25%) of 28 latilong blocks and is considered a rare migrant in the state (Oakleaf et al. 1982). According to the AOU Check-list of North American Birds (1983), the Sabine's Gull is primarily pelagic and breeds north of the con- tiguous United States. Thus, this species is ex- pected to be seen rarely in Wyoming and then mostly during migration. Caspian Tern Skinner (1917) observed Caspian Terns {Sterna caspia) on the Molly Islands, Yellow- stone Lake, Yellowstone National Park, but was unable to determine whether they were nesting. On 4 June 1932, Wright (1934) also saw this species on the Molly Islands and pre- sumed it to be a breeder but failed to locate a nest. Kemsies (1930) first documented breed- ing Caspian Terns on the Molly Islands when he found eggs and downy young on 29 June 1929. Between 1932 and 1966 the number of Caspian Tern nests varied from a low of 4 nests on 5 July 1959 to a high of 18 on 24 June 1966 (Diem and Condon 1967). In 1955 Warkley found evidence of Caspian Terns nesting at Ocean Lake WHMA (Scott 1955). Also, McCreary (1939) indicated that A. B. Mickey located a pair on an island at Bamforth Lake, 346 Great Basin Naturalist [Volume 54 Albany County, in the summer of 1936. One nest of this species was discovered at Bamforth Lake in 1974, and two pairs appeared to nest there in 1975 (E. Raper personal communica- tion). The most recent evidence of Caspian Terns nesting at Bamforth Lake is from 1983 when I counted four nests with eggs on 10 June. All nests were later destroyed by high water. In recent years Caspian Terns have nested at five locations in Wyoming (Table 1). For unknown reasons, there has been a precipi- tous decline in the state's breeding popvilation during recent surveys. The only active colony in 1985 and 1986 was at Pathfinder Resei'voir. Breeding records exist in Idaho (Larrison et al. 1967, C. H. Trost personal communica- tion) and Utah (Behle and Perry 1975). I am unaware of nesting records in other states that adjoin Wyoming. Common Tern Bond (1885) was the first to list the Common Teni {Sterna hinindo) as occurring in Wyoming. This species was considered rare by both Knight (1902) and McCreaiy (1939). Two speci- mens were collected by McCarthy along the Sweetwater River, Natrona County, in 1859, and another bird was taken at Cheyenne, Laramie County, by F. Bond prior to 1902 (Knight 1902). Black'welder may have seen a Common Tern in the Teton region (Grave and Walker 1913). Apparently, Woodbury (1937) collected a specimen at Yellowstone Lake, probably in 1931. Oakleaf et al. (1982) report- ed that the Common Tern was an uncommon summer resident in Wyoming, occurring in 9 (32%) of 28 latilong blocks. This species may occur in the state more frequently than reports indicate because of its similarity in appearance to the more common Forster's Tern. Breeding records exist for the Common Tern in adjacent states of Idaho (C. H. Trost person- al communication). South Dakota (Johnsgard 1979), and Montana (Skaar 1980). Forster's Tern Both Knight (1902) and Grave and Walker (1913) considered the Forster's Tern {Sterna forsteri) a rare migrant to be found only in the southeastern part of the state. McCreary (1939) indicated that this species was a com- mon migrant in eastern Wyoming and summer resident in the southeastern portion of the state. One nest with two eggs was found on 31 May 1936 by A. B. IVIickey at Bamforth Lake, Albany County, and a colony containing 12 nests was found at the same location on 2 July 1933 (McCreary 1939). Another nest of this species was discovered in Albany County as late as 21 July (McCreary 1939). Although Kemsies (1935) speculated that the Forster's Tern occurred fairly frequently in Yellowstone National Park and indicated that it may possi- bly breed in the marshes bordering Yellowstone Lake, thus far there has been only one record for the park. Oakleaf et al. (1982) considered the Forster's Tern a common summer resident and reported it as a breeding species from one (3.6%) lati- long and occurring in 20 (71%) of 28 latilongs. During the 1982-86 period, Forster's Terns Table 1. Location, number of nests, and habitat of Caspian Tern colonies in Wyoming, 1983- Location Number of nests Name 1983 1984 1985 1986 Habitat Albany County Bamfortli Lake Lake Bamforth Island 4r24'N,105°44'W 4 0 0 0 Carbon County Pathfinder Resei'voir Reserv'ior Bird Island 4.3°2.3'N,106°56'W NC^' 15-20 23 29 Natrona County Soda Lake Resei'voir West Island 42°54'N,1()6°19'VV 13 0 0 0 Rattlesnake Island 42°54'N,106°18'W 0 1 0 0 Yellowstone National Park'' Yellowstone Lake Lake Molly Islands 44°19'N,110°16'W 12 3 0 0 ''NC = not censvist'd. ■'Data Irom K. L. Diem (personal coniniunication). 1994] Laridae of Wyoming 347 Table 2. Location, number of nests, and habitat of Fprster's Tern colonies in Wyoming, 1982, 1984-86. Location Number of nests Name 1982 1984 1985 1986 Habitat Albany County Caldwell Lake 4r09'N,105°48'W NCa NC NC 19 Lake Carroll Lake 4r25'N,105°44'W NC NC 15-20 3 Lake Hutton Lake NWR 4ril'N,105°44'W 8-15 0 3 2-3 Marsh Kay Ranch 4ri5'N,105°42'W 2-3 0 0 0 Lake Pilger Lake 4r23'N,105°50'W NC NC NC 9 Lake iTemont Counb.' Ocean Lake 43°07'N,108°35'W NC 10 36 12 Reservoir Lincoln Count}' Bear Rivei' 42°01'N,110°58'W 0 2-3 0 0 Marsh "NC = not censused. nested at seven locations in Wyoming (Table 2). However, not all of these sites were active each year. Based on the 1986 colony censuses, approximately 45—46 nests were present. This compares to 10-18 active nests in two colonies during 1982. The increase in the breeding population is primarily a result of locating four new nesting areas during recent surveys. I am imcertain why Forster s Terns failed to nest on the Kay Ranch and Bear River in 1985 and 1986. Significant declines in nesting Forster's Terns were also noted at Ocean Lake and Carroll Lake in 1986. Fewer terns probably nested at Carroll Lake because of very low water levels that reduced nesting habitat. At Ocean Lake the decline may have been caused by the addition of more cobble to the man-made nesting islands, which made them more dome-shaped and less suitable as nest- ing substrate. Flooding of nests may also be a serious problem at Ocean Lake. Breeding records exist for this species from adjoining states of Colorado (Bailey and Neidrach 1965), Idaho (Larrison et al.'l967, C. H. Trost personal communication), Montana (Skaar 1980), Nebraska and South Dakota (Johnsgard 1979), and Utah (Behle and Perry 1975). Least Teni McCreary (1939) indicated that the Least Tern {Sterna albifrons) was a summer resident along the North Platte River. The first sighting of this species was at Torrington, Goshen County, on 11 June 1929 (McCreary 1934, McCreaiy and Mickey 1935). J. W Scott noted 8 or 10 individuals near Fort Laramie, Goshen County, on 25 June 1932 (McCreary 1934). One year later on 27 May the Least Tern was again reported fi^om Torrington, Goshen County (McCreary 1939). No recent records exist in the state for this species. The Least Tern breeds locally and irregularly in South Dakota and Nebraska (Johnsgard 1979). I am imaware of nesting records from other states that are adjacent to Wyoming. Black Tern Bond (1885) was the first to list the Black Tern {Chlidonias niger) as occurring in Wyoming. This species was considered a rare migrant in the state by Knight (1902). Grave and Walker (1913) indicated that there were records of Black Tenis fiom Cody, Park County; Sheridan, Sheridan County; Lake Como, Albany County; Cheyenne, Laramie County; and Douglas, Converse County. McCreary (1939) noted this species as being a common migrant in eastern Wyoming and a summer resident in the southeastern portion of the state. Henninger (1915) found a nest contain- ing one egg near Bamforth Lake on 12 June 1914. This was the first documentation of nesting by Black Terns in Wyoming. In Yellow- stone National Park, Kemsies (1930) reported that the Black Tern was a frequent migrant and probable summer resident. This species is considered a common summer resident by Oakleaf et al. (1982) and has been reported from 20 (71%) of 28 latilong blocks with strong evidence of breeding from 2 latilongs. On 3 June 1982, I discovered 2-4 nesting pairs of Black Terns with eggs on the Kay Ranch, about 10 km southwest of Laramie, Albany County. From 1984 through 1986 this species nested at three locations in Albany County and in the marshes associated with the Bear River, south of Cokeville, Lincoln County (Table 3). Since 1984, new Black Tern colonies have been discovered at Carroll and Caldwell 348 GiiEAT Basin Naturalist [Volume 54 Table 3. Location, number of nests, and habitat of Black Tern colonies in Wyoming, 1984—86. Location Number of nests Name 1984 1985 1986 Habitat Albany Connt\ Cakluell Lake 4r09'N,105°48'W NC NC 2-3 Lake (Carroll Lake 4r25'N,105°44'W NC 10-15 2-3 Lake Hntton LakeNWR 4ril'N,I05°44'VV 7-10 8-10 1-2 Marsh Ka\ Ranch 41°15'N,1()5°42'W 0 0 0 Lake Lincoln County Bear River 42''01'N.11()°58'W 100-150 NC NC Marsh *NC = not censuscd. lakes. For unknown reasons, this species failed to nest at the Kay Ranch during the 1984-86 period and has not been documented as breeding there since 1982. Population trends of Black Terns are unknown in Wyoming because most colonies have been monitored an insufficient number of years. Also, numbers of Black Terns nesting in the marshes adjoin- ing the Bear River have not been censused since 1984. This species has been found nesting in the following states that adjoin Wyoming: Colorado (Bailey and Neidrach 1965), Idaho (Larrison et al. 1967, C. H. Trost personal com- munication), Montana (Skaar 1980), Nebraska and South Dakota (Johnsgard 1979), and Utah (Behle and Peny 1975). Conclusions In recent years observations of nonbreeding species of gulls, terns, and jaegers have in- creased in Wyoming. I believe these increases are primarily a result of more surveys being conducted by professional biologists and more time spent in the field by greater numbers of amateur birdwatchers. Of the 11 nonbreeding Laridae docimiented in the state, the Pomarine Jaeger and Heermann's Gull were reported for the first time since 1980. Also, the majority of sightings of other nonbreeding species have occurred during the last 10-15 years. Observations of these species will most likely continue to increase as more individuals take up birdwatching as a hobby in Wyoming. Alternative explanations for increased reports of nonbreeding species of Laridae are range expansions or changes in migration routes. I am unaware of evidence from Wyoming or else- where for either explanation. It is unknown whether breeding populations of some Laridae in Wyoming have recently increased or whether new colonies are the result of intensive surveys. I believe that evi- dence exists for recent population increases of the Ring-billed Gull, California Gull, and Her- ring Gull in the state. Reasons for proliferation of California Gull, and possibly Ring-billed Gull and Hening Gull, populations in Wyoming include construction of large reservoirs with isolated islands for nesting as well as creation of new food sources such as garbage dumps, other human refuse, and agricultural land (Findholt 1986a). Breeding populations of these species apparently are expanding throughout the western United States (Conover 1983, Chase 1987). In contrast, I believe most new colonies of Caspian Tems, Forster's Terns, and Black Terns are a result of current sur- veys, and not the result of recent breeding range expansions into Wyoming. However, the addition of at least a few new colonies of Caspian Tems and Forester's Tems in Wyoming since historical times appears to be the result of human-caused environmental changes, especially the construction of reservoirs, which have created nesting and foraging habitat. With the exception of the California Gull, which is a relatively abundant and widespread nesting species in Wyoming, breeding popula- tions of the other five species of Laridae that nest in the state are small. It appears that lim- ited nesting and foraging habitat restricts pop- ulation sizes of most gulls and terns. Also, because Wyoming is at the edge of the breed- ing range of most species currently nesting in the state, populations may remain small. It seems unlikely that nesting populations of the 11 nonbreeding Laridae will be docu- mented in Wyoming, except for the Franklin's Gull and Common Tern, because the breeding range of most species occurs along coastal areas or north of the contiguous United States. The only other species, in addition to the 1994] Laridae of Wyoming 349 Franklin's Gull and Common Tern, that nests in states that adjoin Wyoming is the Least Tern. Since the Least Tern nests locally and irregularly in South Dakota and Nebraska and does not appear to be expanding its range, it seems unlikely that it will be found nesting in Wyoming (Johnsgard 1979). It is difficult to assess long-term population trends of most Laridae that currently breed in Wyoming because of the limited number of years that population data are available. However, results presented in this paper will sei^ve as baseline data that can he used to eval- uate future population changes in the state. Because most breeding colonies are cur- rently protected in Wyoming, prospects for maintaining viable nesting populations appear good. It is my hope that natural resource man- agement agencies will continue efforts to monitor long-term population changes and will implement appropriate management strategies to ensure that currently unprotected breeding populations of Laridae are main- tained in the state. Acknowledgments I thank K. L. Diem, B. H. Pugesek, B. and M. Raynes, O. K. Scott, H. Downing, and oth- ers for allowing me to use their obsei^vations of rarely sighted Laridae in this paper. Much information comes from Sheridan Area Birds, an unpublished collection of sightings edited by H. Downing. K. L. Berner provided invalu- able field assistance in 1986. I thank N. E. Findholt, S. H. Anderson, J. Burger, L. Spear, W. H. Behle, and an anonymous reviewer for helpful comments on the manuscript. This study was funded by the Wyoming Game and Fish Department, Nongame Program. Literature Gited American Ornithologists' Union. 1983. Check-list of North American birds. 6th edition. Allen Press, Inc., Lawrence, Kansas. Bailey, A. M., and R. J. Neidrach. 1965. Birds of Colorado. Denver Museum of Natural History, Denver, Colorado. 2 volumes. Behle, VV. H., and M. L. Perry. 1975. Utah birds: check- list, seasonal and ecological occurrence charts and guides to bird finding. Utah Museum of Natural History, University of Utah, Salt Lake City. Bond, F. 1885. A list of the birds of Wyoming. Pages 1138-1140 in Annual report of the Governor of Wyoming to the Secretary of the Interior. In Annual report of the Secretary of the Interior. 49th Congress, 1st Session, House Executive Document No. 1, Part .5, Vol. 2. Buckley, R A., and F G. Buckley. 1979. What constitutes a waterbird colony? Reflections from the northeast- ern U.S. Proceedings of the 1979 Colonial Waterbird Group 3: 1-15. Chase, C. A., III. 1987. Hybridization of Herring and CalifoiTiia gulls. Pages 16-17 in Proceedings of the 14th annual meeting. Pacific Seabird Group, Pacific Grove, California. Conover, M. R. 1983. Recent changes in Ring-billed and California Gull populations in the western United States. Wilson Bulletin 95: 362-383. Diem, K. L., and D. D. Condon. 1967. Banding studies of water birds on the Molly Islands, Yellowstone Lake, Wyoming. Yellowstone Library and Museum Association, Yellowstone National Park. Findholt, S. L. 1984. Status and distribution of herons, egrets, ibises, and related species in Wyoming. Colonial Waterbirds 7: 55-62. . 1986a. Status and distribution of California Gull nesting colonies in Wyoming. Great Basin Naturalist 46: 128-1.33. . 1986b. The Ring-billed Gull: a rediscovered nest- ing species in Wyoming. Western Birds 17; 189-190. Findholt, S. L., and K. L. Berner. 1988. Current status and distribution of the Ciconiiforms nesting in Wyoming. Great Basin Naturalist 48: 290-297. Grave,' B. H., and E. R Walker. 1913. The birds of Wyoming. University of Wyoming, Laramie. HenNINGER, W F 1915. June birds of Laramie, Wyoming. Wilson Bulletin 27: 221-242. Johnsgard, R A. 1979. Birds of the Great Plains: breeding species and their distribution. University of Nebraska Press, Lincoln. Kemsies, E. 1930. Birds of the Yellowstone National Park, with some recent additions. Wilson Bulletin 42: 198-210, . 19.35. Changes in the list of birds of Yellowstone Naional Park. Wilson Bulletin 47: 68-70. KiNGERY, H. E. 1975. Mountain West. American Birds 29: 93-98. . 1977. Mountain West. American Birds 31: 1168. . 1985. Mountain West. American Birds 39: 82-85. Knight, W C. 1902. The birds of Wyoming. University of Wyoming Agricultural Experiment Station Bulletin 5.5. Kushlan, J. A. 1986. Commentary: colonies, sites, and surveys: the terminology of colonial waterbird stud- ies. Colonial WaterbirdV9: 119-120. Larrison, E. J., J. L. Tucker, and M. T. Joblie. 1967. Guide to Idaho birds. Journal of the Idaho Academy of Science 5: 1-220. McCreary, O. 1930. Some recent obser\ations on Wyoming birds. Journal of the Colorado-Wyoming Academy of Science 1(2): 26. . 1934. The Platte River as a migration route for birds. Nebraska Bird Review 2(2): 38-39. . 19.39. Wyoming bird life. Burgess Publishing Co., Minneapolis, Minnesota. McCreary, O., and A. B. Mickey. 1935. Bird migration records from southeastern Wyoming. Wilson Bulletin 47: 129-157. McCrimmon, D. a., Jr. 1982. Populations of the Great Blue Heron {Ardea herodias) in New York State from 1964 to 1981. Colonial Waterbirds 5: 87-95. 350 Great Basin Naturalist [Volume 54 Oakleaf, B., H. Downing, B. Raynes, M. Raynes, and O. K. Scott. 1982. Wyoming avian atlas. Wyoming Game and Fish Department, Cheyenne. OberHOLSEH, H. C. 1919. Lams canus brachtjrlujnchus in Wyoming. Auk 36: 276-277. Raper, E. L. 1975. Influence of the nesting halntat on the breeding success of California Gulls {Lams califonii- cus), Bamforth Lake, Albany County, Wyoming. Un- published master's thesis. University of Wyoming, Laramie. Ryder, R. A. 1978. Gulls in Colorado: their distribution, status, and movements. Proceedings of the 1977 Conference of the Colonial Waterbird Group 1: 3-9. SCHALLER, G. B. 1964. Breeding behavior of the White Pelican at Yellowstone Lake, Wyoming. Condor 66: 3-23. Scott, O. K. 1955. Great Basin, central Rock>' Mountain region. Audubon Field Notes 9; .392-393. Skaar, R D. 1980. Montana bird distribution. Published by P D. Skaar, Bozeman, Montana. Skinner, M. R 1917. The birds of Molly Islands, Yellow- stone National Park. Condor 19: 177-182. . 1925. The birds of the Yellowstone Park. Roosevelt Wild Life Bulletin 3; 1-192. Woodbury, L. A. 1937. Parasites of fish-eating birds at Yellowstone Lake, Wyoming. Condor 39: 12.5-126. Wricht, G. M. 1934. The primitive persists in bird life of Yellowstone Park. Condor 36: 14.5-153. Received 2 September 1993 Accepted 7 February 1994 Great Basin Naturalist 54(4), © 1994, pp. 351-358 SEED PRODUCTION IN GENTIANA NEWBERRYI (GENTIANACEAE) Myra E. Barnes^ and Richard W. Rusti Abstract. — Experimental manipulations and observations in one population ofGentiana newberryi Gray flowers over 2 years showed significant variation in seed production relative to pollinator and soil water availability. When pollinators were rare, there was a significant relationship between number of bees present and number of mature seeds produced, and supplemental hand cross-pollination (xenogamy) did improve seed set in Gentiana neivberryi Gray. When pollina- tors were abundant, supplemental hand cross-pollination did not increase seed set. Self-fertilized seeds (autogamy) ger- minated at the same rate as cross-pollinated seeds. Seed production in unvisited flowers is probably limited anatomical- ly and is not influenced by the type of fertilization. There was a significant relationship between soil moisture and flower size in G. newbemji, with larger flowers found in wetter areas. Key words: Gentiana, seed production, pollination, bumblebees, soil water potential. Seed set can be limited by insufficient pol- linator visits (Levin and Anderson 1970, Thomson 1980, 1981, Bierzychudek 1981, Gross and Werner 1983, Pleasants 1983, Waser 1983a, Motten 1986, Galen and Newport 1988, Calvo and Horvitz 1990, Harder 1990, Ashman and Stanton 1991) or by other resources, such as water or nutrients, in populations with suf- ficient pollinators (Stephenson 1981, Evenson 1983, McDade and Davidar 1984, Primack and Kang 1989). Resource limitation may result in aborting the whole fruit or only some seeds in a fruit (Lee 1988). Multiple reproduc- tive strategies in perennials, including cross- pollination, self-compatibility (Levin 1971, Jain 1976, Barrett 1988, Karoly 1992), and vegeta- tive reproduction (Evenson 1983, Waller 1988), are advantageous in populations where pollinators and other resources are un-pre- dictable (Motten 1982, Sutherland 1986, Ehrlen 1992). While self-fertilized and vegeta- tively produced plants increase the risk of inbreeding depression, those that are success- ful may have coadapted genes that are advan- tageous for current environmental stresses (Lloyd 1979, Waser and Price 1982, Barrett 1988). Optimal flowering time is when a plant can attract the most visitors and still be able to set seed during the growing season (Pleasants 1983, Waser 1983b, Primack 1987). When bees are abundant, pollination does not limit seed set. When bees are infrequent, there is often a correlation between seed set and polli- nator visitation rate (Zimmerman 1980, McDade and Davidar 1984, Zimmerman and Pyke 1988). Supplemental hand pollination can be used to determine whether pollinators or climatic factors are limiting seed set (Motten 1983). Here we present both observational and experimental data on seed production in Gentiana newberryi Gray. These perennial plants are restricted to high-elevation wet meadows in eastern California, western Nevada, and southern Oregon (Munz 1973). Gentiana newbernji has protandrous, funnel- shaped flowers that are usually white with greenish spots (Munz 1973). Each ramet has one or two flowering shoots with one or two terminal flowers (personal observation). They can reproduce sexually and vegetatively (Spira 1983, Spira and Pollak 1986). Initial observational data included pollina- tion mode, pollinator activity, and soil mois- ture effects on seed production. Based on observational information, we then measured soil water potential and pollinator visitation across the study area and throughout the sea- son to determine relationships between polli- nator availability or soil moisture and seed production. Methods and Materials Study Site In August 1991 we selected a 2700-m2 study site at Little Valley, 27.3 km southeast of 'Biolog\' Department, University of Nevada, Reno, Nevada S9.5.57 351 352 Great Basin Naturalist [Volume 54 Reno, Nevada (119°52'W, 39°15'N). The site, located at an elevation of 2000 m along the eastern edge of the Sierra Nevada escaipnient, is part of the 120()-ha Whittell Forest and Wildlife Area owned In the University of Nevada, Reno (Rust 1987). The area is cov- ered by snow each winter but has an average of 120 days with minimum temperatures above 0°C (Houghton et al. 1975). Gentiana newberryi populations are found in meadows that collect and retain snowmelt water longer than surrounding meadow areas. A small creek running through the southern portion of the site keeps the areas near the creek wet throughout the growing season. Central and northern portions of the site dry out toward the end of the growing season. The study was conducted during the fifth and sixth years of a drought when snowfall was only 50% of nor- mal (James 1992). Most flower species began flowering 4 weeks earlier than usual in 1992 (personal observation) after a warm spring and early snowmelt (James 1992). Population Characteristics In mid-September 1991 and 1992, num- bers of plants (ramets) and flowers on the study site were estimated using 11 transects (38-54 m long) placed 5 m apart and extend- ing across the population to include all G. newberryi plants. At each meter along each transect all plants and flowers were counted and percentage of G. newberryi coverage was estimated within 0.5-m2 circular quadrats. Floral Characteristics In 1992, after observing the differential drying of the study site in 1991, we divided the site into five areas. Area 1 was adjacent to the southern creek (always wet during 1991). Areas 2 to 5 were equally spaced away from the creek, with area 2 the closest to area 1 and area 5 the farthest removed. Soil water poten- tial was measured in each area weekly with a Quickdraw Series 2900 Soil Moisture Probe. Moisture was measured at approximately 30 cm. Three measurements were taken per area per week. Petal length and maximum corolla tube width were measured for 10 random G. newberryi flowers (each from a different plant, and with dehiscing anthers) in each area in 1992 (n = 50). An index was developed to compare flower size by multiplying petal length by corolla tube width. Ultra\iolet re- flectiveness was determined b\' photographing numerous buds and open flowers on live plants in the field with a Wratten ISA UV filter. On 21 August 1991, 10 mature bud flowers (ready to open) were marked and covered with waxed paper bags (1 in the wet area and 9 in the dry). Another 13 buds were marked and left uncovered (7 in the wet and 6 in the dry area). At 0900 the following morning, nectar volume in each flower was measured using a 1-Atl capillaiy tube. On 13 September 1991, 10 mature bud flowers in the dry area and 13 in the wet area were bagged. The following after- noon (1300) nectar was measured from each flower. Each week during 1992 nectar was measured at approximately 0900 from one ran- domly selected, uncovered, dehiscing flower in each of the five soil moisture areas (n = 60). Pollinators Bumblebees were common throughout August and September 1991, but the number of bees present was not regularly recorded. Individual foraging bees were followed; and the flowers visited, distance between flowers, and times were recorded. In 1992 a 100-m transect was established across the drier part of the site and a 50-m (shorter due to limits of the wet area) transect placed across the wet area. These transects were walked hourly at least 2 days each week, and any bees obsei^ved within 5 m of the transect were recorded. A sample of all insect visitors to G. newberryi was collected for identification by R. Rust, University of Nevada, Reno, and R. Brooks, University of Kansas, Lawrence. Seed Production In 1991 flower buds of G. newberryi were randomly selected with not more than one flower per plant and marked with numbered paper tags (/i = 113). Three times each week the phenology (bud, flower opening, dehiscing anthers, receptive stigma, and seed capsule formation) of each marked flower and soil moisture conditions adjacent to the plant (visually) were recorded. During 1992 two newly opened flowers on different plants with dehisced anthers in each area were randomly selected and marked each week fiom early July through early October {n = 84). Mature seed capsules for all marked flowers were collected 1994] Gentiana newberryi Seed Production 353 and placed in individual waxed paper bags. Mature seeds and undeveloped ovules were counted using a dissecting microscope. Seeds for each flower with any mature seeds were placed in individual waxed paper bags and stored outdoors in Reno, Nevada. During August 1991, 53 G. newberryi buds were randomly selected and marked, and the plants covered with white nylon organdy (100- niesh) bags over wire fi-ames. When the flow- ers opened, 9 were hand-pollinated using a paint brush bearing pollen from a flower on the same plant (geitonogamy), 12 with pollen from a different plant (xenogamy), and 34 were left to self-pollinate (autogamy). During the week of 18 August 1992, 10 plants (2 in each soil moisture area) with bud flowers were selected at random, marked, covered with nylon organdy bags, and allowed to self-polli- nate. Ten other newly opened flowers in the different areas were marked and cross-polli- nated by hand after the stigma became recep- tive. Another 10 flowers in the different areas were marked and left alone for natural pollina- tion. Mature seed capsules for treatments used in 1991 and 1992 were collected and seeds counted and stored using the same method. Each year five randomly chosen seed cap- sules from each treatment and the open flow- ers were germinated. Seeds for each flower were placed in a petri dish on brown paper over kimpack moistened with a 400 ppm giberellic acid solution. Seeds were kept at 15 °C for 7 d and then alternated between 15 °C for 12 h and 25 °C for 12 h for 7 d in the dark. Statistical Analyses Analysis of variance (GLM in SAS 1990) was used for all comparative analyses between years and between areas or weeks within each year. Bonferroni t tests were used for multiple comparisons when analysis of variance indi- cated a significant difference. Arcsine transfor- mations were used for analysis of percentage data (Zar 1974). Linear regression (SAS 1990) was used to determine if there was a relation- ship between soil moisture and flower size, soil moisture and seed production, or seed production and bee visits each week. Plant distribution was determined using the stan- dardized Morisita index (I • Krebs 1989). Results Population Characteristics There was no significant difference between 1991 and 1992 in number of plants (F = 0.43, df = 1,984, P = .51) or number of flowers (F = 1.16, df = 1,984, P = .28) per quadrat (Table 1). Percentage cover was significantly different between years (F = 3.97, df = 1,984, P = .04) (Table 1). Distribution of plants is clumped throughout the study site as indicat- ed by the Morisita index (Ip = 0.51). Floral Characteristics A significant difference in G. newberryi flower size (petal length X maximum corolla tube width) was found among the five soil moisture areas of the study site in 1992 (F = 37.04, df = 4,45, P < .0001; Table 2). There was a significant regression (y = 693.4 + 12. 5x) between soil water potential and flower size (F = 117.79, df = 1,48, P < .0001; R^ = .71), with larger flowers found in wetter areas. Ultraviolet images of G. newberryi flowers show a dark, central, UV-absorbing bullseye pattern in the corolla tube and a dark longitu- dinal stripe on the outside of each petal from the base to the apex. Outer petal stripes are also visible on flower buds. In 1991, in a sample of newly opened flow- ers at 0900 h, there was no difference in the amount of nectar available between flowers covered with a bag overnight (0.1 ±0.1 [mean and standard deviation], range 0-0.3 jjlI) and those left open (0.1 ± 0.1, range 0-0.3 /xl) (F = 0.03, df = 1,22, P = .97) or in the amount of nectar in flowers between wet area (0.4 ± 0.4, range 0-1.4 /xl) and diy area (0.2 ± 0.2, range 0-0.7 /xl) (F = 2.02, df = 1,21, P = .17). In 1992 there was no difference in the amount of nectar available in open flowers between Table 1. Number of G. newberryi plants, flowers, and percentage of cover per 0.5-m- quadrat at Little Valley, Nevada. Values are means ± standard de\iation; n = 49.3. Total numbers of plants and flowers in 2700-m- study site are in parentheses. Plants (no.) Flowers (no.) Percent cover 1991 - L2 ±3.7 (3115) 2.3 ±3.1 (773) 1.7 ±5.5 1992 - LI ±4.1 (2710) 0.2 ±1.1 (476) 1.1 ±5.4 354 Great Basin Naturalist [Volume 54 Table 2. Relationship !>etwecn soil water potential and flower size in G. newhemji at Little Valley, Nevada, in the second week of Jul\ 1992. Soil water potential, petal length, corolla width, and a flower size index (petal length X corol- la width) are indicated for five areas of decreasing soil water potential. Values are means ± standard deviation; /i = 10. Area Soil water Petal Corolla Flower size potential length width index (MPa) (mm) (mm) 1 0.0 ±0.0 46.5 ±2.4 16.4 ± 1.0 764.1 ±79.0 A^' 2 -3.5 ±0.5 44.0 ±3.5 14.0 ±1.8 617.6 ± 106.8 B 3 -6.0 ± L6 4L7 ±3.6 13.3 ± 1.9 557.7 ± 105.5 B 4 -12.0 ±4.9 38.3 ±3.1 11.0 ±1.9 425.1 ± 102.3 C 5 -35.0 ±0.0 33.1 ±2.9 8.6 ±1.8 287.5 ± 74.4 D wer size index aiiione areas. Me; ins with Ihe same letter are not sienificantlv different (P < ,0.5i weeks (week 1 — 0.1 ± 0.1, range 0-0.3 /xl; week 2—0.6 ± 0.9, range 0-2.1 fxl; week 3— 0.4 ± 0.3, range 0-0.7 /xl; week 4—0; week 5—0.7 ± 0.4, range 0.4-1.3/11; week 6—0.1 ± 0.2, range 0-0.4 {A; and week 7—0.2 ± 0.2, range 0-0.4^11) (F = 1.99, df = 6,28, P = .10). No difference in amount of nectar was found between areas (area 1 [driest] — 0.1 ± 0.2, range 0-0.5 /xl; area 2—0.3 ± 0.2, range 0-0.6 ^tl; area 3—0.2 ± 0.3, range 0-0.7/^1; area 4— 0.4 ± 0.4, range 0-1.3 /Ltl; and area 5 [wettest]— 0.6 ± 0.8, range 0-2.2 /xl) (F = 1.01, df = 4,30, P = .41). Most nectar was usually found in one or two of the five nectar tubes. Pollinators In 1991 four species of bumblebees were observed visiting G. newhemji flowers (Table 3). Most visits were for nectar Nectar foragers would pick up pollen on their ventral surface from the centrally located anthers. Bonibus appositus Cresson and Boinbus edwardsii Cresson were frequent visitors to G. newhemji from early August until the end of September Bomhus vosnesenskii Radoszkowski was fre- quently observed visiting Lupinus seUuhis Kell. adjacent to G. newhemji and occasional- ly visited a few G. newberryi flowers. Bomhus fervidus (Fabricius) was seen visiting G. new- hemji flowers during only one week in August. Usually between one and six bumblebees could be found visiting G. newhemji flowers on the study site anytime during August and September when the weather was warm and calm. During 1992, bumblebee visits to G. new- herryi were rarely observed. A few visits by Bomhus edwardsii were seen. Bomhus vosne- senskii were observed visiting other flower species and occasionally robbed nectar from G. newberryi from outside the flower. Anthophora homhoides Kirby, A. urhana Cresson, and A. terminalis Cresson were occa- sionally observed visiting G. newhemji. Apis mellifera L. were common visitors to adjacent Lupinus seUuhis, and one was seen visiting a G. newhemji flower Anthophora species and A. meUifera were not seen in 1991. In 1992 there were always many flowers open with pollen available. Less than one bee per 750 m^ was observed when walking tran- sects. There was a significant positive correla- tion between number of bees observed each week along the combined transects and num- ber of seeds produced per flower marked that week (Fig. 1). Seed Production There was no significant difference in the mean number of matine G. newherryi seeds produced per marked flower between 1991 (116.2 ± 143.6, n = 58) and 1992 (135.4 ± 114.4, „ = 76) (F = 0.75, df = 1,132, P = .38). When we eliminated flowers with aborted seed capsules from the analysis (50% aborted 1991, 23% 1992), the number of mature seeds per capsule was higher in 1991 (232.4 ± 118.3, n = 29) than 1992 (174.6 ± 100.0, n = 59) (F = 5.76, df= 1,86, F= .01). In 1991 significantly more seeds were pro- duced by G. newherryi flowers in the areas with a wet soil surface (210.3 ± 175.1, n = 20) than in the dry areas (66.7 ± 93.4, n = 38) (F = 16.70, df = 1,56, P < .0001). More seed capsules aborted in the diy than the wet area 1994] Gentiana newberryi Seed Production 355 Table 3. Total number of visits, mean length of visits, and mean distance traveled between flowers for vai-ying numbers of indi\ iduals of four foraging Boinbus species at Little Valle\', Ne\ada, in August 1991. Values for time and distance are mean ± standard de\ iation. Species I ndi\ iduals Visits Time Distance (sec) (cm) B. apposittis 1 6 10.3 ±0.7 13.5 ±8.3 B. cdwardsii 1 20 12.8 ±4.5 9.1 ±24.7 B.jcrvidus 5 45 14.7 ±9.2 64.4 ±98.2 B. vosnesenskii 3 26 8.2 ±5.6 23.5 ±23.2 (58% vs. 35%). A higher percentage of ovules per capsule matured to seed in the wet area (84.7 ± 10.8%, n = 13) than in the dry area (60.2 ± 24.0%, n = 16) (F = 5.88, df = 1,27, P = .01). In 1992 there was no significant difference in the number of mature seeds produced per flower between the areas with varying soil water (area 1 [driest]— 133.7 ± 112.5, area 2— 153.1 ± 98.4, area 3—57.0 ± 79.3, area 4— 161.9 ± 135.8, and area 5 [wettest]— 142.1 ± 122.8) (F = 2.26, df = 4,79, P = .07) or in the percentage of ovules that matured to seed per capsule (F = 0.56, df = 4,60, P =.69). There was, however, a significant positive relation- ship between soil water at the time a flower opened and number of mature seeds pro- duced (y = 158.9 - 2.3x and F = 5.96, df = 1,69, P = .02). There was a significant difference in the average number of seeds produced (F = 8.44, df = 3,83, P < .0001) and the percentage of ovules that matured to seeds (F = 9.06, df = 3,42, P < .0001) between the open, xenoga- mous, and geitonogamous versus autogamous pollination treatments in 1991 (Table 4). In 1992 there was also a significant difference between pollination treatments of open and xenogamous versus autogamous in the average number of seeds produced (F = 11.37, df = 2,20, P = .0005) and the percentage of mature seeds per capsule (F = 8.15, df = 2,25, P = .002) (Table 4). Hand cross-pollinated (xenogamy) seed production was highest but was not statistically different from open bee pollination in both years. Autogamous seed production was lowest but not statistically dif- ferent from open pollination. The percentage of germinated G. newber- ryi seeds was not significantly different 400 300 200 I 00 J: r = 0.79 p = 0.01 0 0.0 0.2 0.4 0.6 0.8 Mean Bees/Transect/Week Fig. 1. Correlation between mean number of bees obsewed along combined transects and mean number of seeds produced by G. newberryi flowers that were open the week bees were obsei^ved; n — 7. between treatments in 1991 (F = 2.00, df = 3,15, P = .15) or 1992 (F = 2.93, df = 2,9, P = .10). Seed germination ranged from 91.5 to 98% in 1991 and 84.8 to 98.8% in 1992 for all treatments. Discussion Seed production per marked Gentiana newberryi flower was not significantly differ- ent between 1991 and 1992 in Litde Valley even though pollinator numbers and soil mois- ture were different between the 2 years. The large variation observed in all measurements of G. newberryi population characters and flower/seed characters in the Little Valley population both among and between years suggests much individual variability. Within the G. newberryi habitat area, individuals respond to a variety of localized microenviron- mental parameters with the resulting variation in sexual reproductive output. Bumblebees visiting G. newberryi were abundant during 1991. It is unlikely that polli- nators limited seed set in 1991 since flowers with hand cross-pollination did not set signifi- cantly more seeds. There was a higher num- ber of aborted fruits in the dry area in 1991. During 1992, when soil was wetter, G. newber- ryi plants appeared more vigorous throughout the study area. Low pollinator availability in 1992 did appear to limit seed set, as the num- ber of flowers that matured seeds each week was coiTclated to the number of bees observed. Zimmerman (1980) and McDade and Davidar (1984) both found that seed set was correlated with visitation rates when pollinator numbers 356 Great Basin Natufl^list [Volume 54 Table 4. Percentage of mature seeds per capsule and number of mature seeds per flower in lour pollination treat- ments. Treatment 1 = open, 2 = hand cross-pollinated, 3 — self-pollinated, 4 = hand-pollinated from flower on same plant. Aborted seed capsules were not included in the percentage of mature seeds per capsule analysis. Values are mean ± standard deviation; n is in parentheses. Mature seeds Mature seeds Treatments per capsule ■{%) per flower ( no. ) 1991 -- 1 72.5 ±25.2 (16) A^' 124.3 ± 155.8 (32) A B 2 69.9 ±27.8 (12) A 189.6 ±91.6 (12) A 3 22.5 ±23.6 (11) B 20.4 ±49.5 (34) B 4 49.3 ±34.0 (7) A B 105.6 ± 130.0 (9) A B 1992 1 58.5 ±27.5 (6) A B 93.8 ±94.7 (8) A B 2 90.8 ±8.0 (10) A 156.5 ±35.3 (10) A 3 45.9 ±24.8 (7) B 46.0 ±42.5 (10) B ^BonfeiToni ( tests comparing treatments within each \ear. Means with same letter are not sisniflcantK diflercnt (P < .05). were low. When aborted fruits were eliminat- ed from the analysis in 1991 (50%), more mature seeds were produced per flower in 1991 than in 1992. This also suggests that the larger number of bees present in 1991 increased the numbers of seeds produced in plants with sufficient moisture to produce mature fruits. Fruit initiation may be pollina- tor limited, but mature fruit and seed produc- tion are primarily limited by resources such as water (Galen and Newport 1988, Horvitz and Schemske 1988, Ashman and Stanton 1991). It is not known whether the rarity of bees obsened in 1992 could be related to drought conditions or snow and cold weather in late June. Many species of bees were seen foraging on other flower species in Little Valley during May 1992 (personal observation). Numerous male Bombus vosnesenskii were observed for- aging on several flower species on the study site during July, but few females of any Bombus species were seen during the G. new- bernji flowering period. In contrast, an open flower in 1991 could rarely be found with pollen remaining on the anthers. Throughout 1992, most open G. newberryi flowers had nectar and abundant pollen available as poten- tial pollinator rewards. There were large populations of Lupinus selluhis adjacent to patches of G. newberryi and Aster species and Perideridia bolanderi (Gray) Nels. & Macbr. were common from July through September In 1991 Bombus edward- sii, B. fervidus, and B. appositus were seen vis- iting only G. newberryi. Bombus vosnesenskii was usually obsei-ved visiting L. sellulus with occasional visits to several G. newberryi. In 1992 the few bees observed visited a variety of flowers, and none showed a preference for G. newberryi. We observed a few nectar-robbing visits from outside the flower by male B. vos- nesenski in 1992. Ultraviolet markings in the center of the corolla and on the outside of petals attract and guide bees to nectar sources (Silberglied 1979, Kevan 1983, Waddington 1983, Waser 1983b). Bumblebees were often observed flying quick- ly and directly in a straight line from G. new- berryi flower to flower. When entering a flower, a nectar-foraging bee positions its ven- tral surface over the anthers in the staminate phase and over the stigma in the pistillate phase. Bees were rarely observed collecting G. newberryi pollen. There was no difference in the rate of seed germination between self- and cross-pollinat- ed flowers. Seed production is not affected by type of pollination, but it may be limited ana- tomically in unvisited self-pollinated flowers. As the flower closes, the stigma bends over and touches only one or two of the anthers. Caged plants that were hand-pollinated with pollen from a flower on the same plant pro- duced as many seeds as open plants or plants with supplemental hand cross-pollination. Plants in wetter soil had larger, showier flowers than those in diy soil. Pollinators are usually attracted to larger flowers (Galen and NewiDort 1988, Ashman and Stanton 1991). In 1991 the surface soil was very diy in a large part (1800 m-) of the study site. Fewer seeds matured per flower and more seed capsules 1994] Gentiana newberryi Seed Production 357 aborted in the diy area, suggesting that inade- quate soil water can hmit the number of seeds produced per flower. Surface soil throughout the study site appeared wetter in 1992. The positive relationship between soil water at flower opening and number of mature seeds produced in 1992 indicates the importance of sufficient water resources in determining seed set. Facultative self-compatibility allows G. newbernji to produce seeds even when polli- nators are rare. Vegetative reproduction re- quires less energy (Waller 1988) and may be used in addition to or in place of flower pro- duction or sexual reproduction. Of 10 plants that were dug up, only one did not have an attached lateral rhizome. Since data will con- tinue to be collected from this site, we did not dig up a sufficient number of plants to be able to determine size of clones or average number of rhizomes per ramet. There was no significant difference in pop- ulation size, number of flowers, or mean num- ber of seeds produced per flower between 1991 and 1992. Some plants marked in 1991 that remained covered with 1 cm of water or more from the natural creek diversion did not sui'vive. Larger G. newberryi flowers are found in wetter areas and produce more mature seeds than flowers in drier soil areas. There was a significant relationship between number of pollinators present and number of seeds pro- duced only when pollinators are rare. There was no difference in seed production between flowers with xenogamous and geitonogamous pollination. Facultative self-compatibility and vegetative reproduction allow plants to pro- duce seeds or ramets when pollinators are limiting. Gentiana newberryi appears well adapted to survive during unpredictable peri- ods of pollinator availability and soil moisture. Acknowledgments We thank James Cane (Auburn University) and Gary Vinyard and Robert Nowak (Uni- versity of Nevada, Reno) for valuable com- ments on the original manuscript. Comments from journal reviewers and associate editor Jeanne Chambers were also constructive and helpful. Funding assistance was provided by the George Whittell Forest Board. Literature Cited Ashman, T. L., and M. Stanton. 1991. Seasonal variation in pollination dynamics of sexually dimorphic Sidalcea oregaiui ssp. spicata (Malvaceae). Ecology 72: 993-1003. Barrett, S. C. H. 1988. The evolution, maintainance, and loss of self incompatibility systems. Pages 98-124 in L. Lovett-Doust and J. Lovett-Doust, eds.. Plant reproductive ecology: patterns and strategies. Oxford University Press, New York. BlERZYcnUDEK, P. 1981. Pollinator limitation of plant reproductive effort. American Naturalist 117: 838-840. Calvo, R. N., and C. C. Horvitz. 1990. Pollinator limita- tion, cost of reproduction, and fitness in plants; a transition-matrix demographic approach. American Naturalist 136; 499-516. Ehrlen, J. 1992. Proximate limits to seed production in a herbaceous perennial legume, Lathyrus vernns. Ecology 73: 1820-1831. EvENSON, W. E. 1983. Experimental studies of the repro- ductive energy allocation in plants. Pages 249-274 in C. E. Jones and R. J. Little, eds., Handbook of experimental pollination biology. Reinhold, New York. Galen, C, and M. A. Newport. 1988. Pollination quality, seed set, and flower traits in Polemoniwn viscoswn: complementary effects of variation in flower scent and size. American Journal of Botany 75: 900-905. Gross, R. S., and F A. Werner. 1983. Relationships among flowering phenology, insect visitors, and seed set of individuals: experimental studies on four co- occuring species of goldenrod (Solidago; Compositae). Ecological Monographs 53; 9.5-117. Harder, L. D. 1990. Behavioral responses by bumble bees to variation in pollen availabilib.'. Oecologia 85; 41-47. Horvitz, C. C., and D. W. Schemske. 1988. A test for the pollinator limitation hypothesis for a Neotropical herb. Ecology 69:200-206. Houghton, J. G., C. M. Sakamoto, and R. O. Gifford. 1975. Nevada's weather and climate. Nevada Bureau of Mines and Geology Special Publication 2: 1-78. Jain, S. K. 1976. The evolution of inbreeding in plants. Annual Review of Ecology and Systematics 7; 469-495. James, J. W. 1992. Nevada climate summary'. Vol. 10, No. 5. Universit\' of Nevada, Reno. Karoly, K. 1992. Pollinator limitation in the facultatively autogamous annual, Lupinus nanus (Leguminosae). American Journal of Botany 79: 49-56. Kevan, P G. 1972. Floral colors in the high arctic with ref- erence to insect-flower relations and pollination. Canadian Journal of Botany .50; 2289-2316. . 1983 Floral colors through the insect eye: what they are and what they mean. Pages 3-.30 in C. E. Jones and R. J. Litde, eds.. Handbook of experimen- tal pollination biology. Reinhold, New York. Krebs, C. J. 1989. Ecological methodology. Harper and Row, New York. Lee, T. L. 1988. Patterns of fruit and seed production. Pages 179-202 in J. Lovett-Doust and L. Lovett- Doust, eds.. Plant reproductive ecology: patterns and strategies. Oxford University Press, New York. 358 Great Basin Naturalist [Volume 54 Levin, D. A. 197L Competition for pollinator .ser\ice; a stimulus for the evolution of autogamy. Evolution 26: 668-669. Levin, D. A., and W. VV. Anderson. 1970. (Competition for pollinators between simultaneousK' flowering speeies. Ameriean Naturalist 104: 4.55—467. Lloyd, D. G. 1979. Some reproduetive factors affecting the selection of self-fertilization in plants. American Naturalist 113: 67-79. McD.^DE, L. A., ANi:) P. Dwidak. 1984. Determinants of fruit and seed set in Pavonia (hisijixiala (Malvaceae). Occologia 64: 61-67. MoTTEN, A. F. 1982. Autogamy and competition for polli- nators in He})atica mnericana (Ranunculaceae). American Journal of Botany 69: 1296-1.305. . 1983. Reproduction of Erythronhim umbilicatwn (Liliaceae): pollination success and pollinator effec- tiveness. Oecologia 59: 351-359. 1986. Pollination ecology of the spring wildflower community of a temperate deciduous forest. Ecological Monographs 56: 21-42. MuNZ, P. A. 1973. A California flora and supplement. University' of California Press, Berkeley. Pleasants, J. M. 1983. Structure of plant and pollinator communities. Pages 37.5-393 in C. E. Jones and R. J. Little, eds.. Handbook of experimental pollination biolog}'. Reinhold, New York. Primack, R. B. 1987. Relationships among flowers, fiiiits, and seeds. Annual Review of Ecology and Systematics 18: 409-4.30. Primack, R. B., and H. Kang. 1989. Measuring fitness and natural selection in wild plant populations. Annual Review of Ecology and Systematics 20: 367-396. Rust, R. VV. 1987. Collecting oiPinm (Pinaceae) pollen by Osmia bees (Hymenoptera: Megachilidae). Environ- mental Entomology 16; 668-671. SAS. 1990. SAS/stat users guide. Vols. 1 and 2. SAS Institute, Caiy, North Carolina. SiLBERGLlED, R. E. 1979. Communication in the ultravio- let. Annual Review of Ecolog\ and Systematics 10: 373-398. Spira, T. R 1983. Reproductive and demographic charac- teristics of alpine biennial and perennial gentians {Gentiana spp.) in the White Mountains. Unpublished doctoral dissertation. University of California, Berkelev. Spira, T. R, and O. D. Poll.ak. 1986. Comparative repro- ductive biology of alpine biennial and perennial gen- tians (Gentiana: Centianaceae) in California. American Journal of Botany 73: 39^7. Stephenson, A. G. 1981. Flower and fruit abortion; pro.\i- mate causes and ultimate functions. Annual Review of Ecology and Systematics 12; 253-279. Sutherland, S. 1986. Patterns of fruit-set: what controls fruit-flower ratios in plants? Evolution 40; 117-128. Thomson, J. D. 1980. Skewed flowering distributions and pollinator attraction. Ecology 61: 572-579. . 1981. Spatial and temporal components of resource assessment by flower-feeding insects. Journal of Animal Ecology 50: 49-59. Waddington, K. D. 1983. Foraging behavior of pollina- tors. In: L. Real, ed., Pollination biology. Academic Press, Orlando, Florida. Waller, D. M. 1988. Plant moiphologx' and reproduction. Pages 203-227 in J. Lovett-Doust and L. Lovett- Doust, eds.. Plant reproductive ecology: patterns and strategies. Oxford University Press, New York. Waser, N. M. 1983a. Competition for pollination and floral character differences among sympatric plant species: a review of evidence. Pages 277-293 in C. E. Jones and R. J. Little, eds.. Handbook of experimental pol- lination biology-. Reinhold, New York. . 1983b. The adaptive nature of floral traits; ideas and evidence. Pages 277-293 in L. Real, ed.. Pollination biology. Academic Press, Orlando, Florida. Waser, N. M., and M. V. Price. 1982. Optimal and actual outcrossing in plants, and the nature of plant-polli- natorinteraction. Pages 341-359 in C. E. Jones and R. J. Little, eds.. Handbook of experimental pollina- tion biologv'. Reinhold, New York. Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall, New Jersey. Zimmerman, M. 1980. Reproduction in Polemoniwn: com- petition for pollinators. Ecology 61:497-501. ZiMMER.MAN, M., AND G. PiKE. 1988. Reproduction in Polemonium: assessing the factors limiting seed set. American Naturalist 131: 723-738. Received 2 July 1993 Accepted 25 January 1994 Great Basin Naturalist 54(4), © 1994, pp. 359-365 USE OF A SECONDARY NEST IN GREAT BASIN DESERT THATCH ANTS {FORMICA OBSCURIPES FOREL) James D. Mclver^ and Trygve Steen^ Abstiuct. — Workers of Great Basin Desert thatch ants {Formica ubscuripes Forel) dig simple secondary nests at the base of plants upon which they tend aphids and scales. These secondar\' nests house only foragers, with the number of foragers occupying each nest positively correlated with the number of worker-tended Homoptera feeding on plant foliage above. Thatch ant secondan' nests are cooler than 25 cm below the dome top of the primary nest and maintain a significantly more constant temperature than is obsei-ved on the ground sui-fiice or in the plant canopy Thatch ant foragers use secondary nests for at least two purposes: as a cool refuge for Homoptera tenders when midday plant canopy tem- peratures rise during the summer months, and as the primary place within which Homoptera tenders transfer honey- dew to larger "honeydew transporters" for ultimate transport back to the primaiy nest. Key words: honeydew harvest, thermal refiigia, behavioral thermoregulation, red wood ants, desert adaptation, satel- lite nests. Although most ant species use a nest struc- ture consisting of a single central location (a primary nest), many species also employ "sec- ondary" nests in which a portion of the colony population is dispersed among several alter- nate sites (Wheeler 1910). Several species of Camponotus, for example, use a secondary nest to which workers tiansport late-instar larvae and pupae fiom a central loca- tion occupied by the queen and brood (Hansen and Akre 1987). Similarly, the dolichoderine Iridomyrmex sanguineus maintains secondaiy nests containing older larvae and pupae, but workers bring young from several locations within oligogynous colonies (Mclver 1991). Many other ant species {Polyrachis simplex, Lasius niger, L. emarginatus, Formica pratensis, F exsectoides, Crematogaster pilosiis) are known to use secondaiy nests in which only foraging workers reside (Forel 1921, Andrews 1929, Ofer 1970). These secondary nests are thought to serve as refuges for the workers from the physical environment, as a defense against enemies, or as a protected site within which to tend Homoptera for honeydew (Wheeler 1910). This paper characterizes the secondary nest used by the thatch ant Formica obscuripes Forel living in the Great Basin Desert and dis- cusses its possible fimction within the context of the desert environment. Study Area and Species Thatch ants were studied between June 1987 and September 1991 at Pike Creek, 160 km soudieast of Bums, Oregon. The Pike Creek study site is at 1300 m elevation at the base of Steen's Mountain in the northern Great Basin Desert. Sagebmsh {Artemisia tridentata), rabbit- brush [Cknjsothamnus nauseosus), horsebrush {Tetradymia sp.), lupine {Lupinus caudatus Kellogg), and cheatgrass {Bromus spp.) are dominant plants at the site, which was grazed moderately by cattle throughout the study period. A total of four colonies o{ Formica obscuripes Forel were observed for various parts of the study. F. obscuripes is a widespread and abun- dant North American ra/fl-group species (Wheeler and Wheeler 1983). Like F. rufa- group species elsewhere, F. obscuripes builds symmetrical, dome-shaped primary nests of thatch, from which radiate trunk trails that access foraging territory. In all four study colonies workers foraged for honeydew on sagebrush, rabbitbrush, horsebrush, and/or lupine, and scavenged for arthropods in the area surrounding each nest. Although broodless satellite nests were occasionally observed, there was no evidence of primary nest poly- do my in any study colonies. ^ Departments of Biology and Education, Eastern Oregon State College, La Grande, Oregon 97850, and Blue Mountains Natural Resources Institute, 1401 Gekeler Lane, La Grande, Oregon 978.50. ^Department of Biology, Portland State University, Box 75L Portland, Oregon 97207. 359 360 Great Basin Naturalist [Volume 54 Methods and Materials Secondary Nest Characteristics The aliovegroiind structure of the sec- ondan' nest is portra\ ed by a photograpli taken from colony 5 at Pike Creek, August 1988. The belowground structure was investigated by pouring a measured quantity of dental labstone down 10 different secondary nest entrances of two colonies (colonies 4 and 26) during August 1991. Quantit)' of labstone required to fill each secondaiy nest was then correlated with basal plant diameter and number of workeis tending Homoptera in the plant canopy. Actual struc- ture of the secondaiy nest interior was deter- mined by excavating two nests, photographing the labstone "plug" in place, and drawing one of these to scale using the photograph as refer- ence. Temperature at 6 cm depth in a typical sec- ondary nest (plant A, colony 2) was measured during summer 1987 and compared to mea- surements for tending localities in the plant canopy, ground surface, and 25 cm below the top of the primaiy nest dome. Secondaiy Nest Use Use of the secondary nest by thatch ant workers was explored by conducting intensive observations on a selected sagebrush plant (plant 13) at the Pike Creek study site during July 1987. Beginning 1 July 1987, thatch ants working in the vicinity of plant 13 were indi- vidually marked with "beenumbers" (Charles Graz Co., Frankfurt, Germany) so that the activity pattern of each could be determined. By 23 July, a total of 66 workers had visited plant 13 and been marked, 30 of which were still using the plant daily. At noon on 23 July, we began a 24-h continuous period of obsei-va- tion of worker behavior on plant 13. We record- ed the location and task of each worker at 15- min intei^vals throughout the 24-h period and noted its interaction with other workers. The result was a time budget for 30 different work- ers that frequented plant 13 during the 24-h period, from which we could infer how work- ers of various task specializations used the sec- ondary nest at the plant base. Results Secondary Nest Characteristics Thatch ant secondaiy nests were found at the base of each plant upon which workers tended Homoptera at Pike Creek during the study period. Viewed from above, secon^iary nests were simple openings in the ground adjacent to plant trunks (Fig. 1). Ground around an opening was typically littered with thatch material, fallen from the plant canopy, blown in, or excavated from the gallery beneath. Volume of 10 secondary nests beneath active tending groups of workers ranged from 35 to 125 cc. Secondaiy nest volume was not signifi- cantly correlated with basal plant diameter (K- = .02, P > .05, N = 10) but was significantly correlated with number of tenders (K- = .33, P < .05, Y = .54X + 43.3, N = 10). Excavations of secondary nests into which labstone had been poured revealed that cavities essentially conformed to morphology of the plant trunk itself (Fig. 2). Thatch ant workers typically removed dirt, small stones, and other debris from within 5-20 mm of the plant trunk, leaving a cavity punctuated with large stones and roots. The nest represented in Figure 2 was 10.8 cm deep and consisted of three separate chambers totaling 175 cc in vol- ume. Temperature within the secondaiy nest dif- fered considerably from temperatures record- ed simultaneously on the ground surface, in the plant canopy, or deep within the primaiy nest (Fig. 3). Over the 1-wk period 13-19 June 1987, for example, the secondaiy nest we measured was an average of about 7°C cooler than 25 cm fi-om the dome top of the primaiy nest (18.8° vs. 26.1°), with a little over twice the variance over time (12.6 vs. 5.7). Compared to ground surface, the secondary nest was slightly cooler (18.8° vs. 19.2°) but much less variable, exhibiting a variance of about one-ninth the ground sur- face (12.6 vs. 112.4). Compared to tlie canopy of the same plant, the secondaiy nest was slightK' warmer on average (18.8° vs. 18.0°) but about one-fifth as variable (12.6 vs. 67.1). Temperature trends over the entire summer were similar to those measured in this 1-wk sample period in mid-June. Secondaiy Nest Use Obsei-vations of individually marked work- ers on plant 13 of colony 2 clearly show that the secondaiy nest is used throughout the day (Fig. 4). The greatest percentage of workers was found in the secondaiy nest during mid- afternoon, corresponding to highest daily tem- 1994] V Secondary Nests in Thatch Ants 361 ^^^y:^*¥ ^^ J^'"- iiUmS^ ^\ '* ?£^'^'''V ^ ./iVil Thatch ant worker ^VV^KtfHP^^'V' *^' ''■"'* 1 ^4"r* %Ji ^^ Secondary nest entrances ^^^^ ^^" ^ V > |. '* ^^ ^v J'T ^^ ■•;^«4^»';v: -i.s^^ls^ Fig. 1. Aboveground appearance of secondar\- nest at base of sagebrnsli plant. Pike Creek, southeastern Oregon, June 1994 (photograph by Tiygve Steen). peratures. Secondaiy nest population was low- est between 1700 and 2000, and between 0600 and 0900, during principal times when work- ers deliver honeydew to the primary nest. Two typical patterns of activity were obsei-ved for plant-associated workers (Fig. 5). Tenders spent the majority,' of tlieir time tending Homoptera for honeydew in the plant canopy. Worker 84, for example, spent 54% of her time tending aphids, with each visit to the plant canopy lasting between 2 min and about 3 h. Her visits to the plant canopy were inter- spersed with frequent visits to the secondaiy nest at the plant base, where it is likely she transferred honeydew to larger nontending individuals like worker 13 (chain transport). Twice per day she returned to the primary nest: once in the early evening and once in the morning. Honeydew transporters spend the majority of their time in the secondary nest itself Worker 13, for example, spent 66% of her time in the secondaiy nest, 23% scavenging on the ground surface, and 9% on twice-daily returns to the primaiy nest. On her returns to the pri- mary nest, worker 13 often had a distended gaster, indicating a crop swollen with honey- dew. Typically, workers like #13 were scaven- gers, secondaiy nest excavators, and/or honey- dew transporters, receiving the majority of their honeydew from workers that concentrat- ed on tending Homoptera in the plant canopy. Of the 30 workers associated with plant 13 during the intensive observation period, 19 were classified as tenders, 6 as honeydew transporters/scavengers, 2 had behavior inter- mediate between tender and transporter/scav- enger, and 3 were not obsei-ved often enough to classify. Discussion Great Basin Desert thatch ants use sec- ondary nests as a refuge from high midday temperatures and as a site within which honey- dew is transferred from workers who collect it in the plant canopy to those who help trans- port it back to the primaiy nest. Ground tem- peratures above 50 °C have been reported as lethal to F. obscuripes (O'Neill and Kemp 1990), and Mackay and Mackay (1984) 362 GiiKAT Basin Naturalist [Volume 54 observed that F. haeniorrhoidalis workers hide under pine cones or retreat to shady places during midday heat. Chain transport appears to be an effective way to increase deliveiy of honeydew to the primary nest (Mclver and Yandell 1994); thus, it is not surprising that honeydew transfer occurs at a site offering refuge from midday heat. The use of cool midday refugia by workers may also reduce metabolic costs and increase worker longevit)-. In a study on fire ant thermal preferences, Porter and Tschinkel (1993) report- ed that fire ant workers consistently choose cooler temperatures than those selected for the brood. They postulate that this tendency increases longevity of workers not directly associated with brood care. This idea is sup- ported by Calabi and Porter (1989), who demonstrated that because temperature and metabolic rate are highly corrrelated, fire ants reared and maintained under high tempera- ture regimes have lower longevity. nsfi%*m- ENLARGED If AREA Fig. 2. Scale drawing of secondaiy nest, taken from photograph of labstone plug, Pike Creek, southeast- ern Oregon, August 1991. Thatch ants living at other sites in the Great Basin also use secondary nests of this kind (Mclver personal observation); Weber (1935) described secondary nests in his study of South Dakota thatch ants. However, Weber reported that the function of these nests was to sei-ve as (1) an arborescent chamber within which to tend Homoptera and (2) a potential site for development into primary nests. Certainly, colonies of Formica rufa-group species often reproduce b>' budding (Mabelis 1979; F polyctena), and the site of a new pri- mary nest is very often a secondary nest (Scherba 1959, Mclver personal observation). It is not known whether F rufa-group species living in other habitats employ secondary nests for these or other reasons. Other Formica species are also known to employ secondaiy nests. The moimd-building ant F exsectoides {exsectoides-group) uses sec- ondary nests as shelters for treehoppers and as sites for food exchange (Andrews 1929). 1994] Secondary Nests in Thatch Ants 363 O O LU GC LU LU 60 50 40 30 20 10 GROUND Pike Creek Colony 2 -- 1987 SECONDARY 0 8 16 0 8 16 0 8 16 0 8 16 0 8 16 0 8 16 0 8 16 13 June 14 June 15 June 16 June 17 June 18 June 19 June TIME Fig. 3. Temperature (°C) during week of 13-19 June 1987, on ground surface, in sagebmsh canopy, 25 cm below dome top of primaiy nest, and in secondaiy nest of colony 2, Pike Creek, Oregon. Formica Integra of North America and F. pratensis of Europe construct secondary nests along covered paths (Wheeler 1910, Forel 1921). Other Homoptera-tending ants, including the formicines Lasius niger (Forel 1921), L. emarginatus (Forel 1921), L. flavus (Soulie 1961), and Polyrachis simplex (Ofer 1970), and the myrmicines Crematogaster pilosus (Forel 1921) and C. auherti (Soulie 1961), use sec- ondary nests as shelters for their homopteran symbiotes. Acknowledgments Bryce Kimberling drew the secondaiy nest from photographs. We thank Courtney Loomis, Deborah Coffey, Joseph Furnish, and Bill Clark for assistance in the field. Jeffrey C. Miller provided the datapod for temperature recordings. Andre Francour kindly identified Formica ohscuripes Forel. Research was sup- ported by the National Geographic Society and the Systematic Entomology Laboratory of Oregon State University (Dr. John Lattin), where voucher specimens are held. Literature Cited Andrews, E. A. 1929. The mound-building ant, Formica exsectoides E, associated with treehoppers. Annals of the Entomological Society of America 22: 369-391. Calabi, P, and S. D. Porter. 1989. Worker longevity in the fire ant Solenopsis invicta: ergonomic considera- tions of conelations between body size and metabol- ic rates. Journal of Insect Physiolog\- 35; 643-649. Forel, A. 1921. Le monde social des fourmis du globe. Vols. 1-3. Geneve, Libraire Kundig. 1921-1923. Hansen, L. D., and R. D. Akre. 1987. Biolog>' of carpen- ter ants. Pages 274-280 in R. K. Vander Meer, K. Jaffe, and A. Cedeno, eds.. Applied myrmecology — a world perspective. Westview Press, Boulder, Colorado. Mabelis, a. a. 1979. Nest splitting by the red wood ant {F. pohjctena Foerster). Netherlands Journal of Zool- ogy' 29: 109-125. Mackay, E. E., and W R Mackay. 1984. Biology of the thatching ant Formica haemorrhoidalis Emery 364 Great Basin Natufulist [Volume 54 SECONDARY NEST □ PLANT CANOPY O GROUND ^ PRIMARY NEST 1200 1400 1600 1800 2000 2200 2400 200 400 600 600 1000 1200 TIME OF DAY Fig. 4. Activity of marked workers of plant 13, Pike Creek colony 2, 23-24 July 1987. Number of workers obser\ed in secondaiA' nest, in plant canopy, on ground, at primaiy nest, and temperature in degrees Celsius over 24-h period. % TIME 54% 2% 36% 8% LOCATION PLANT CANOPY GROUND SECONDARY NEST Worker 84: Tender, Honeydew Transporter I I I I 1 L 12 13 14 15 16 17 18 PRIMARY NEST I I I I I I \ 1 1 U 20 21 22 £3 24 1 mklnighl 2 3 4 5 2% 23% 66% 9% Worker 13: Scavenger, Honeydew Transporter PLANT CANOPY SECONDARY NEST J I I I L 12 13 14 15 16 17 PRIfAARYNEST I I I I I 1 I L I I I L 920 212223 24 123 456 7 mklnlahl 10 11 12 noon Fig. 5. Activity over 24-h period of workers 84 and 13 on plant 13, colony 2, Pike Creek, Oregon, 23-24 July 1987. 1994] Secondary Nests in Thatch Ants 365 (Hvmenoptera: Fomiicidae). Pan-Pacific Entomology 60: 79-83. Mch'EK, J. 1991. Dispersed central-place foraging in Australian meat ants. Insectes Sociaux 38: 129-137. McTver, J., AND K. Yandell. 1995. Honeydevv hai-vest in Great Basin Desert thatch ants. National Geographic Research 11: In press. Ofer, J. 1970. Pohjrachis simplex, the weaver ant of Israel. Insectes Sociaux 17: 49-82. O'Neill, K. M., and W. P Kemp. 1990. Worker response to thermal contraints in the ant Formica ohscuripes (H:F). Journal of Thermal Biology 15: 133-140. Porter, S. D., and W. R. Tschlnkel. 1993. Fire ant ther- mal preferences: behavioral control of growth and metabolism. Behavioral Ecology and Sociobiology 32: 321-329. Scherba, G. 1965. Analysis of inter-nest movements by workers of the ant Formica opaciventris Emery. Animal Behavior 12: 508-512. Soulie, J. 1961. Les nids et le comportement nidificateur des Fourmis du genre Crematogaster d Europe, d'Afrique du nord et d Asie du sud-est. Insectes Sociaux 8: 213-297. Weber, N. A. 1935. The biology of the thatching ant, Formica rufa ohscuripes Forel, in North Dakota. Ecology 5: 166-206. Wheeler, W M. 1910. Ants. Their structure, develop- ment and behavior Columbia University Press, New Y'ork. 663 pp. Wheeler, G. C., and J. N. Wheeler. 1983. The ants of Nevada. Natural History Museum of Los Angeles County, Los Angeles, California. 138 pp. Received 24 August 1993 Accepted 26 January 1994 Great Basin Naturalist 54(4), © 1994, pp. 36(>-.370 SPAWNING CHRONOLOGY AND LARVAL EMERGENCE OF JUNE SUCKER {CHASMISTES LIORUS) Timothy Modde ^-^ and Neal Muirheadi Abstract. — June sucker {Cluisinistes lioriis) spawned in the Provo River, Utah, over a 2-wk period in earl\- June dur- ing both 1987 and 1988. Emergent larvae emigrated from the river to Utah Lake over a 2- to 3-wk period. Drift into the lake peaked between 1200 and 0400. Dining da\'light hours, emergent laiAae tended to occur in pools. Peak emergence of lanal drift was appro.ximateh' 1.2 larvae/ni^ during late June in 1987 and 1988. Recruitment failure of June sucker is not dui' to reproductive failure. Key words: June sucker, Chasniistes liorus, spaicnin<^, larvae, liabitat, drift, emergence, river. The June sucker {Chasinistes Horns) is one of three contemporary species of the genus Chasmistes (Miller and Smith 1981) and is endemic to Utah Lake, a 38,000-ha remnant of prehistoric Lake Bonneville. Once June sucker numbered in the millions (Jordan 1891) and were one of the most abundant fishes in Utah Lake. During the last century population size of June sucker declined drastically. In a survey of Utah Lake fishes, less than 0.4% of fish col- lected were June sucker (Radant and Sakaguchi 1981). The population of June sucker has been estimated to be <1000 adults and is listed on the federal register as an endangered species (US. Fish and Wildlife Service 1986). Suspected factors contributing to the decline of this species include water loss to irrigation and drought, degradation of water quality, and negative interactions with nonnative fishes (Radant and Hickman 1984). Reduction of water quantity and quality impacted both the lake and spawning tributaries. The direct cause of decline in the June sucker population has been lack of recruit- ment (Sigler et al. 1985). In a sui-vey of Utah Lake, Radant and Sakaguchi (1981) did not capture any June sucker <400 mm total length. Scoppettone (1988) reported that June sucker may live to be 42 years of age; thus, in the absence of recruitment, senescent individ- uals would dominate the population. None of the 18 fish he examined was younger than 20 years of age. June sucker have been described as spawn- ing on gravel cobble substrate in relatively high-velocity habitats (Radant and Hickman 1984). Sex products are broadcast over the substrate, and eggs are adhesive to the sub- strate (Shirley 1983, Radant and Hickman 1984). Although information on spawning behavior and larval morjohology (Shirley 1983, Snyder and Muth 1988) exists, no information is available on spawning success of the June sucker. Because natality is a vital element of recruitment, information on spawning success is important in understanding declining abun- dance of this species. The objectives of our study were to (1) estimate timing and magni- tude of downstream drift of emergent June sucker lai^vae and (2) describe habitats occu- pied by larval June sucker in the Provo River. Methods Drift Sampling Drift netting was conducted in the lower Provo River to capture emergent lai'vae during the 1987 and 1988 spawning periods. Netting began 1 June 1987 and terminated when lar- vae ceased to appear in collections. Five drift nets, each with a mouth size of 30 X 45 cm and a mesh size of 560 microns, were placed at a single site about 3 km upstream of Utah Lake, immediately downstream of the lowermost observed June sucker spawning activity. Nets 'National Biological Survey, Ufali CooperaKve Fish and Wildlife Research Unit, Department of Fisheries and Wildlife, Utah State University, Logan, Utali 84322-5210. The Utah Cooperative Fish and Wildlife Research Unit is a program jointly sponsored by tlie U.S. Department of Interior, Utah Division of Wildlife Resource, and Utah State University. ^Present address: Colorado River Fish Project, U.S. Fish and Wildlife Ser\ice, 26(i West 100 North, Suite 2. Vernal, Utah 84078. 366 1994] June Sucker Spawning 367 were anchored with 0.64-cm-diameter rebar along a single transect perpendicular to the channel. When depth permitted, nets were placed alternately at the surface and bottom. In 1988, drift netting using the same sampling scheme began 6 June. The netting site was moved about 50 m downstream of the 1987 site because of physical changes in the chan- nel. Only four nets were used during 1988. Nets were set on alternate days (MWF) each week. Each 24-h day was divided into six 4-h periods, and drift was sampled continu- ously during the middle 1.5 h of each. Starting times were 1315, 1715, 2115, 0115, 0515, and 0915 h. Drift from each net was rinsed, placed in watertight plastic bags, and preserved in 5% buffered formalin; 420 samples were taken. Velocity (10-sec average) through each net and water depth were measured before and after each set. Volume sampled was estimated by multiplying the average of the two velocity measurements by time sampled and area of the net opening. Water temperature was re- corded during each 4-h interval. All samples were sorted for eggs and larvae, which were identified to species (Snyder and Muth 1988), counted, and measured to the nearest 0.1 mm (total length). Habitat Sampling Fishes in a 2.25-km section of the lower Frovo River were sampled during the 1988 spawning season to determine larval habitat use. Eighty-four transects, about 27 m apart and peipendicular to the thalweg, were estab- lished from aerial photographs of the river. Three samples were taken along each transect, one near each shore and one in the middle of the river Samples were collected with a 1-m^ bag seine with a 560-micron mesh. Substrate in a l-m^ area immediately in front of the seine was mechanically stirred at each sam- pling site, and the seine was quickly pulled through. Samples were taken only during day- light hours. Habitat types were described and widths measured along each transect using a modifi- cation of Bisson et al. (1982). All fish collected were placed in plastic containers and pre- served in 5% buffered formalin. Lai-vae were identified to species, measured to the nearest 0.1 mm (total length), and counted. Analysis Means for egg and larval density in the drift were determined for daily and 4-h peri- ods. Standard deviations were calculated from daily means among periods and for periods with days as replicates. Drift densities were estimated by dividing eggs and larvae collected during each sampling period by water volume passing through drift nets. Daily estimates were detemiined by computing the means of all six time periods. Estimates of total larvae on the peak drift date were determined by averaging discharge recorded at the Provo City gauge station (USGS) on both days sam- ples were made and multiplying the volume estimate by daily mean lai^val density. Because of the few sites in which June sucker larvae were present, habitats were grouped into pool and nonpool categories. Chi-square analysis was used to test the signif- icance of differences in the incidence of larval June sucker in pool and nonpool habitats, and odds ratio analysis (Fienberg 1980) was used to quantify the magnitude of differences obsei'ved. Resuuts Drift Spawning, as defined by egg drift, was high- er on 3^ June 1987 and peaked on 6-7 June 1988 (Figs. 1, 2). A malfunction of the velocity meter on 6 and 8 June 1988 prevented accu- rate estimation of egg and larval concentra- tions. However, absolute numbers of eggs cap- tured on 7 June (0.007 eggs/sec) and 8 June (0.005 eggs/sec) exceeded those caught on 11 June (0.0007 eggs/sec). Average river tempera- tures during the spawning period were 13-14 °C in 1987 and 12-17° C in 1988. Spawn- ing occurred over a relatively short time; eggs were collected for 1 wk in 1987 and 11 d in 1988. Spawning duration was probably longer in both years than shown in Figures 1 and 2 because eggs were already present in the river when sampling began. However, collections from both years suggested that June sucker spawning activity does not last more than 2 wk, with the greatest number of eggs spawned witliin a 3- (1987) to 5-d (1988) period. Density' of egg drift was variable and showed no diel pattern (Fig. 3). Thus, either fish were spawning in both light and dark hours or eggs were being randomly dislodged from the 368 Great Basin Naturalist [Volume 54 - 1988 larval density r 3 ' 5' 7 ' 9 ' 1 1'13'15' 17'19'2 1'23'25'27'29' 1 June 13 5 7 9 1113 15 17 19 2123 25 27 29 1 June Fig. 1. Drift rates of June sucker larvae and eggs, and Fig. 2. Drift rates of June sucker Iai-vae and eggs, and daily average temperature collected from the Provo River, daily average temperature collected from the Provo River, Utah, in June 1987. Utah, in June 1988. substrate throughout the 24-h period. Duriug drift uetting operations June sucker were obsei'ved spawning during both day and night. Larval June sucker first appeared in the drift on 3 June 1987 and 6 June 1988 (Fig. 1). Although velocity error precluded absolute measurement until after 10 June, few larvae were collected until 20-21 June. Peak densi- ties of larvae in the drift occurred on 22-23 June 1987 and 22-23 June 1988. Minimum estimates of the time between egg deposition and swim-up, measured as the period between peak egg drift and peak lai-val drift, were 19 d in 1987 and 16 d in 1988. The difference in incubation time between years is probably due to warmer river temperature in 1988 (15-19°C) than in 1987 (12-16°C). Drift of June sucker larvae continued for about 3 wk during both study years. All June sucker lar- vae collected were identified as either proto- or mesolai^vae. A distinct daily pattern of lai"val drift densi- ty was observed, with most larvae captured between 2000 and 0400 h (Fig. 4). Few larvae were collected in drift nets during daylight. Peak daily estimates of drifting June sucker lanae in the Provo River were approximately 60,200 in 1987 and 73,000 in 1988. Habitat Use A total of 57 June sucker lai-vae were col- lected in 7 of 115 collections. Incidence of lar- vae in pool-type habitats was different from nonpool halMtats (X^ = 7.04, .05 = 5.99). June sucker lanae were 7.5 times more likely to be found in pool than nonpool habitats during daylight hours. Discussion Shirley (1983) reported June sucker spawn- ing in mid- June when mean water tempera- ture was between 11 and 13 °C. Similar obser- vations were made by Radant and Hickman (1984) and Radant and Sakaguchi (1981). Radant and Hickman (1984) also observed a short spawming period that lasted only 5-8 d. The cui-ui {Chasinistes cujus) also spawns dur- ing a brief period: males occupying the Tmckee River, Nevada, 6.5-16.5 d and females 4.0-10.5 d (Scoppettone et al. 1986). Temperatures of the Truckee River during cui-ui spawning 1994] June Sucker Spawning 369 0.06 - 0.05 - 0.04 0.03 1987 ii^^il 0200 0600 1000 1800 2200 0.06 0.05 0.04 0.03 I 1988 ■-■ ■ 0200 0600 1000 1400 Hour Fig. 3. Diel drift rate of June sucker eggs collected in the Provo River, Utah, in June 1987 and 1988. Vertical bars represent one standard deviation. ranged from 12 to 15 °C. Cui-ui spawned between 2000 and 0600 h during a 3-d period (Scoppettone et al. 1981), whereas egg drift densities of June sucker and observation dur- ing both 1987 and 1988 indicated spawning occurred during all hours of the day and night. Scoppettone et al. (1983) reported peak emer- gence of cui-ui lai^vae occurred 14 d after peak spawning. Differences between peak June sucker spawning and peak emergence varied between years, from 19 d in 1987 (tempera- ture range 13-15 °C) to 16 d in 1988 (tempera- ture range 17-19 °C). Like cui-ui (Scoppettone et al. 1986) and other catostomids (Geen et al. 1966), June sucker larvae emigrate from spawning tribu- tary into receiving lake shortly after emer- gence. Drift activity of larval June suckers was nearly identical to that of cui-ui, most drift occuring just prior to 0000 and declining to negligible numbers by 0600. In spite of large numbers of larvae captured in the drift, rela- tively few were captured by seining. Those larvae seined during daylight hours were mostly in pool-type habitats, as reported by both Radant and Hickman (1984) and Shirley (1983). Few, if any, larvae remained in the 1987 1988 3 - 2.5- 2 - 1.5 - 0.5 - 0 - 1 1 1 1 0200 0600 1000 1400 1800 2200 Hour Fig. 4. Diel drift rate of June sucker lan'ae collected in the Provo River, Utah, in June 1987 and 1988. Vertical bars represent one standard deviation. Provo River for an extended time. Most lai-vae drifted out of the Provo River during a 2- to 3- wk period, whereas cui-ui were reported to drift through the Truckee River for nearly 30 d. Differences between the two species in dura- tion of lai-val emergence and drift may result from a larger cui-ui spawning population. Although the abundant numbers of June sucker lai"vae produced in 1987 and 1988 are surely less than historic numbers, substantial numbers of larvae drifted into Utah Lake. Sigler et al. (1985) suggested the decline in the Pyramid Lake cui-ui population was due to failure of natural reproduction. Based on the large numbers of larvae captured in the drift, despite the relatively small population of adult June sucker, insufficient spawning or emergent success seemingly did not limit recruitment to Utah Lake. Instead, factors affecting sumval after lai-val emergence, such as nonnative predators, seem likely. Acknowledgments This study was funded by the Utah Division of Wildlife Resources and the U.S. Fish and Wildlife Service. Randy Radant 370 Great Basin Naturalist [Volume 54 assisted with all phases of the stud\- and pro- vided infonnation on the histoiy of Jmie suek- er spawning in the Provo River. Dennis Shirley was instrumental in providing field coordination and logistical support at the study site. Roger Mellinthin, Brad Schmitz, and Don Archer provided field assistance. Literature Cited BissoN, E A., J. L. Nielsen, R. A. Palmason, and L. E. Grove. 1982. A .system for naming habitat types in small streams, with examples of habitat utilization by salmonids dnring low stream flow. Pages 62-73 ;')! N. B. Armantrout, ed., Accjuisition and ntilization of aqnatie habitat inventoiy information. Western Divi- sion of the American Fisheries Society, Bethesda, Maiyland. FlENBERG, S. E. 1980. The analysis of cro.ss-classified cate- gorical data. MIT Press, Cambridge, Massachusetts. Geen, G. H., T. G. Northcote, G. E H.'vrt.man, and C. C. LiNDSEY. 1966. Life histories of two species of catostomid fishes in Si.xteenmile Lake, British Columbia, with particular reference to inlet stream spawning. Journal of the Fisheries Research Board of Canada 23: 1761-1788. Jordan, D. S. 1891. Report of exploration in Colorado and Utah during the siunmer of 1889, with an account of the fishes found in each of the river basins exam- ined. United States Fisheries Commission Bulletin 9: 1-40. Miller, R. R., and G. R. Smith. 1981. Distribution and evolution of Chasmintes (Pisces: Catostomidae) in western North America. University of Michigan, Museum of Zoolog\', Occasional Papers 696: 1-46. Radant, R. D., and T. J. Hickman. 1984. Status of the June sucker {Chasmistes liorus). Proceedings of the Desert Fishes Council 15(1983); 277-282. Radant, R. D., and D. K. Sak-AGUCHI. 1981. Utah Lake fisheries inventory. United States Bureau of Reclamation Contract 8-07-40-.50634 Modification 04. Division of Wildlife Resources, Salt Lake Cit\, Utah. SCOPPEITONE, G. G. 1988. Grov\ th and longevit\' of the cui-ui and longevity of other catostomids and cyprinids in Western North America. Transactions of the American Fisheries Society 117: •301-307. ScoppETTONE, G. G., M. Coleman, H. Burce, and G. Wedemeyer. 1981. Cui-ui life history-: river pha.se. Annual report. United States Fish and Wildlife Service, Fisheries Assistance Office, Reno, Nevada. 63 pp. ScoppETTONE, G G., M. Coleman, and G A. Wedemeyer. 1986. Life history and status of the endangered cui- ui of Pyramid Lake, Nevada. Fish and Wildlife Research 1: 1-23. SCOPPETTONE, G. G., G. A. Wedemeyer, .M. Coleman, AND H. Burge. 1983. Reproduction by the endan- gered cui-ui in the lower Truckee River. Transactions of the American Fisheries Society' 112: 788-793. Shirley, D. L. 1983. Spawning ecology' and lai-val devel- opment of the June sucker. Proceedings of the Bonneville Chapter, American Fisheries Societv (198.3): 18-36. SiGLER, W. F, S. ViGG, AND M. Bres. 1985. Life histoiy of the cui-ui, Chasmistes ciijiis Cope, in P\ ramid Lake, Nevada: a review. Great Basin Naturalist 45: 571-603. Snyder, D. E., and R. T. Muth. 1988. Description and identification of June, Utah and mountain sucker lar- vae and early juveniles. Contract 87-2891, Division of Wildlife Resources, Salt Lake City, Utah. U.S. Fish and Wildlife Service. 1986. Endangered and threatened species listing and recovery priority guidelines. Federal Register 48: 16756-167759. Received 5 January 1994 Accepted 11 April 1994 Great Basin Naturalist 54(4), © 1994, pp. 371-375 COMPARISON OF REPRODUCTIVE TIMING TO SNOW CONDITIONS IN WILD ONIONS AND WHITE-CROWNED SPARROWS AT HIGH ALTITUDE Martin L. Morton^ Abstr\CT. — Timing of reproduction was assessed for wild onions and White-cro^vned Sparrows in relation to snow conditions on the same subalpine meadow in the Sierra Nevada for 21 years. Flowering date and clutch initiation date were both highK- correlated with snow conditions, being later as snowpack was deeper Interannual \'ariation in sched- ule was 46 davs for onions and 33 days for spaiTOws. There was nearly a fixefold difference in snowpack depth, and date of snow disappearance \'aried interannually b>' 72 days. Compensation for late-lying snows occurred in both species but was greater in sparrows than in onions because the nest-building behavior of sparrows was flexible. In years of deeper snow, sparrows were able to lay eggs earlier because they built more nests than usual in trees and shrubs rather than waiting for groundcover to develop. Key icords: Allium, Zonotrichia, snmvpack, high altitude, proximate factors, reproduction. Montane settings are useful for the study of environmental adaptation in organisms because their brief, sharply delimited growing seasons and variable climates can be potent agents of natural selection. Diurnal and sea- sonal cycles of abiotic factors, principally air temperature (Ta), moisture, and wind speed, shift in level and amplitude as elevation increases (Rosenberg 1974). The resulting decrease in mean Ta, high winds, decreased availability of soil moisture due to freezing, and variable snowpack can greatly influence the phenology, distribution, and productivity of plants (Billings and Bliss 1959, Scott and Billings 1964, Weaver and Collins 1977, Osder et al. 1982). Annual schedules of reproduction and survival of hibernating mammals (Morton and Sherman 1978) as well as reproductive success of both sedentary (Clarke and Johnson 1992) and migratory birds (Morton 1978, Smith and Andersen 1985) also are known to be affected, especially by spring storms and snowpack depth. It follows that long-term studies of the annual rhythm of reproduction of organisms at high altitude should provide valuable information for understanding the pathways and scope of adaptations to climatic conditions and for determining the efficacy of environmental variations to act as cues or proximate factors in the control of reproduc- tion schedules. Such studies are few, however. and, to my knowledge, there are none that have compared plants and animals on the same study area. Herein I present 21 seasons of data that index reproductive schedules of the wild onion {Allium validum) and the Mountain White-crowned Sparrow {Zonotrichia leitcophrys oriantha) at the same location in the Sierra Nevada in relation to interannual variations in snow conditions. Materials and Methods The study site, Tioga Pass Meadow (TPM), is a subalpine meadow with an area of about 50 ha (0.5 X 1.0 km) and elevation of 3000 m located in the upper end of Lee Vining Canyon, Mono County, California. It is bound- ed by Tioga Lake on the northern edge and the boundary of Yosemite National Park at Tioga Pass on the southern edge. Allium validum, which grows in large clumps in wet meadows at elevations of 1200-3350 m in the Sierra Nevada, is usually 0.5-1.0 m in height and has numerous small (6-10 mm) flowers organized into terminal umbels (Munz 1970). For 21 summers, 1968-70, 1973, 1976, and 1978-93, when I was on TPM daily, I kept notes on the flowering schedule of one partic- ular, fairly compact patch of A. validum that covered an area of about 0.1 ha (25 X 40 m) near the center of TPM and that usually 'Biolo};\ Department, Occidental College, Los Angeles, California 90041-3392. 371 372 Great Basin Naturalist [Volume 54 contained about 1200 umbels. Technically, flowering includes the period from floral bud initiation through floral persistence (Rathcke and Lacey 1985), but I use the tenii to describe the date on which buds opened to reveal the mass of flowers within. In any given year this date varied by 2 wk or more among the vari- ous patches of A. validnvi scattered across TPM, but flowering within a particular patch, including the study patch, was highly synchro- nous, occurring within a 3-d period for most individuals. The last day of this opening peri- od was noted every year and is the datum used in this analysis. The primar)^ focus of my field studies on TPM was the reproductive biology of Z. /. ori- antha, a migratory finch that winters in Mexico and breeds in montane meadows of the western United States. Individuals arrive at breeding areas in May and June and depart for wintering grounds in September and October. Wet subalpine meadows like TPM are a preferred breeding ground habitat, but more xeric locations at lower elevations are sometime utilized (Morton and Allan 1990). Nests are built only by females and are placed on the ground or in shrubs, such as willows (Salix sp.), or in small trees. Data gathered on banded females included the date they laid their first egg of the season, i.e., the clutch ini- tiation date. Herein I use the mean date of the first 10 clutch initiations on TPM each season to indicate the onset of reproduction in Z. /. oriantha. Most nesting data were obtained from females that were fre(|uently observed and trapped (each had a unique combination of color bands), allowing me to follow changes in their behaviors, body weights, and brood patches. This was important to data quality because females quickly renested if a nest was lost from storms or predation. Information on snow conditions was avail- able because, first, I estimated from direct observations the snowcover on TPM as the season progressed up to the day when all patches of snow had disappeared — the date of 0% snowcover. Second, information on snow- depth in TPM could be obtained because it is a site traditionally used by the State of California Department of Water Resources to measure snow depth in order to predict water storage and runoff. Maximum snowpack occurs about 1 April, and this measurement is published in their bulletin 120. Results During this study snow depths ranged from a low of 79.0 cm inl976 to a high of 375.7 cm in 1983. The earliest 0% snowcover date was 1 June 1992, and the latest 11 August 1983, a range of 72 d. The earliest date for onion flow- ering was 6 July 1968 and 1976, and the latest was 21 August 1983, a range of 46 d. The earli- est mean date for clutch initiations, based on the first 10 nests of the season, was 27 May 1992, and the latest mean date was 29 June 1983, a range of 33 d (Table 1). Thus, repro- ductive schedule in relation to snowpack was affected more in onions than in sparrows. In both species, however, timing of reproduction was highly correlated (P < .001) with maxi- mum snowqDack (Fig. 1). The final disappear- ance of snow (0% snowcover) was tightly cou- pled to snow depth, and flowering and clutch initiation schedules were related accordingly to time of snow disappearance (Table 2). Slope values for the relationship of flowering to both measures of snow conditions (snow depth and 0% snowcover) were about twice those obsei^ved for clutch starts in relation to these same two measures (Table 2). Slopes for both comparisons were significantly different {t tests, P < .001). Discussion Reproduction was delayed by deep, late- lying snow more so in A. validuin than Z. /. oriantha, but there was a compensaton' mech- anism operating to lessen the temporal impact of heavy snows even in the onion. When flow- ering date was regressed on snow disappear- ance date, the slope was 0.53, far less than 1.0 (Table 2). How then does the onion adapt? Only three major physical environmental fac- tors have been identified as cues that initiate flowering: temperature, moisture, and photo- period (Rathcke and Lacey 1985). Photoperiod- responsive or long-day plants are relatively unaffected by snowpack because they flower and are pollinated late in the summer. In con- trast, nonphotoperiod-responsive plants tend to bloom in spring or early summer, and phenophases may be affected by as much as 6 wk by temperature and moisture conditions (Owen 1976). Some plants can "catch up, ' at least somewhat, by condensing or telescoping phenophases when delayed by overbing snow (Billings and Bliss 1959, Scott and Billings 1994] Plant-Animal Reproductive Schedules 373 Table L Twenty-one years of data on snow conditions, time of flowering in A. valkhim, and mean of clntcli initiations and nest locations in Z. /. oriantha. The first 10 nests of the season were used to calculate mean date of clutch initiation, and all nests found in a given season were used to calculate percentage of those built aboveground. Number of nests found per season: mean = 59.6, S.D. = 21.9, range = 18-100, total of all years = 1252. Snow depth 0% snowcover Onion flowering Clutch initiation Aboveground Year (cm) (Julian da)) (Julian day) (mean Julian day) nests (%) 1968 113.5 175 188 160.6 40.0 1969 342.1 215 233 170.1 72.3 1970 176.3 183 198 158.3 34.9 1973 204.2 181 207 163.1 59.3 1976 79.0 158 188 153.8 10.9 1978 263.4 215 222 177.2 50.8 1979 227.1 191 207 160.7 43.6 1980 262.6 214 217 177.0 52.4 1981 173.0 167 197 156.4 25.3 1982 294.4 216 225 169.0 43.8 1983 375.7 224 234 181.3 32.1 1984 205.0 197 208 158.3 44.6 1985 145.8 172 197 156.8 25.5 1986 243.3 213 213 160.9 56.9 1987 113.3 163 201 152.5 18.6 1988 121.2 170 198 161.3 11.1 1989 158.0 176 201 156.2 39.4 1990 90.9 172 196 157.4 30.0 1991 167.4 181 214 164.7 46.7 1992 108.2 152 201 146.8 37.0 1993 227.1 199 216 169.5 49.2 240 230 220 (D 210 200 - 190 - 180 - 1 70 - 1 60 150 o c o 3 140 Onion Flowering R2 = 0.87 n = 21 years Sparrow Clutch Initiation R2 = 0.67 CO D cn < =3 CD C 13 o 50 100 150 200 250 300 Snow Depth (cm) 350 400 Fig. 1. Date of flowering in A. validwn and mean clutch initiation in Z. /. oriantha as a function of snow depth on 1 April at Tioga Pass Meadow. 374 Great Basin Naturalist [Volume 54 Table 2. Slopes and coefficients of determination (R~) for linear regressions involving date of flowering in A. vcilidiitn and mean date of clutch initiation and percentage ot nests placed in sites off the ground in Z. /. oriantha in relation to snow conditions (snow depth on 1 April and date of 0"% snowco\ei) at Tioga Pass Meadow. N = 21 years. Y X Slope K2 (%) 0% snowcover Snow depth 0.25 Flowering date Snow depth 0.15 Flowering date 0% snowcover 0.53 Clutch initiation date Snow depth 0.09 Clutch initiation date 0% snowcover 0.34 Aboveground nests Snow depth 0.12 Aboveground nests 0% snowcover 0.47 86.1 <.001 86.9 <.001 76.1 <.001 68.9 <.001 75.7 <.001 40.2 .002 42.4 .001 1964, Weaver 1974, Weaver and Collins 1977). Alpine plants often store large quantities of carbohydrate in underground organs; mobi- lization of these reserves to shoots can occur even before snowcover is gone. Relatively high carbohydrate levels are then maintained in the shoot portion until after fruiting, whereupon a return of peak reserve levels to underground parts occurs at the beginning of fall dormancy (Mooney and Billings 1960). The mechanism whereby events in the growing season can be accelerated has not been studied to my knowl- edge, but this seasonal cycle of transport and utilization of stored energy must be a vital constituent. A puzzling aspect of A. validums flowering response is that it was strongly affected by snow conditions, typical of nonphotoperiod- responsive plants that usually flower in May or June. A. validum flowers in July or August, a time that is more typical of plants cued by long days (Owen 1976). Perhaps A. validum has seasonal changes in its photoresponsivity and is following a mixed strategy energetically, using stored reserves early in the season and then later switching to a greater reliance on photosynthate of the current year, the latter being a trait common to photoperiodic species (Mooney and Billings 1960). The lessened impact of snow conditions on reproductive schedule in Z. /. oriantha, as compared to A. validum, appears to occur because the bird has flexible nest-building habits. Only about 11% of all nests construct- ed at TPM in dry years, such as 1976, were placed in aboveground sites (Table 1). In wet years, such as 1969, when snowpack was unusually heavy, this increased to 72% (Table 1). The 1983 data seem anomalous, but hot spring weather induced rapid snowmelting, and more nests were placed on the ground than might have been otherwise expected. The main point here is that when groundcover was adequate for hiding nests, females seemed to prefer nesting on the ground. When plant growth and development were impeded by late-lying snow, they did not wait a long peri- od for this cover to develop, but instead built a greater proportion of their nests aboveground, usually in pines {Pimis sp.) and willows. Thus, behavioral plasticity in selection of nesting sites allowed Z. /. oriantha to proceed with rearing young with less delay than might be predicted from snow conditions or even from plant phenophases. In summary, this correlative study presents temporal indices of reproductive schedules during 21 years in a plant and an animal occu- pying the same high-altitude environment, thus permitting a comparison of their respons- es to a proximate or environmental factor experienced in common, namely interannual variation in snowpack. Both organisms were affected by this factor and both exhibited com- pensatoiy adjustments of their schedules. The adjustment was greater in the animal because it possesses a basic trait, not present in the plant, that can be acted on by natural selec- tion, its behavior. Literature Cited Billings, W. D., and L. C. Bliss. 1959. An alpine snow- bank environment and its effects on vegetation, plant development, and productivity. Ecology 40: 388-397. CUARKE, J. A., AND R. E. JOHNSON. 1992. The influence of spring snow depth on White-tailed Ptarmigan breeding success in the Sierra Nevada. Condor 94: 622-627. Mooney, H. A., and W. D. Billincs. 1960. The annua! carbohydrate cycle of alpine plants as related to growth. American Joimial of Botany 47: 594-598. Morton, M. L. 1978. Snow conditions and the onset of breeding in the Mountain White-crowned Sparrow. Condor 80: 285-289. 1994] Plant-Animal Reproductive Schedules 375 Morton, M. L., and N. Allan. 1990. Effects of snowpack and age on reproductive schedules and testosterone levels in male White-crowned Sparrows in a mon- tane environment. Pages 235-249 in M. Wada, S. Ishii, and C. G. Scanes, eds.. Endocrinology of birds: molecular to behavioral. Japan Scientific Society Press, Springer Verlag, Tokyo. Morton, M. L., and P W. Sherman. 1978. Effects of a spring snowstorm on behavior, reproduction and survival of Belding's ground squirrels. Canadian Journal of Zoolog>' 56: 2578-2590. MUNZ, P A. 1970. A California flora. University of Cali- fornia Press, Berkeley. 1681 pp. Ostler, W. K., K. T Harper, K. B. McKnight, and D. C. Anderson. 1982. The effects of increasing snowjiack on a subalpine meadow in the Uinta Mountains, Utah, U.S.A. Arctic and Alpine Research 14: 203-214. Owen, H. E. 1976. Phenological development of herba- ceous plants in relation to snowmelt date. In: H. W. Steinhoff and J. D. Ives, eds.. Ecological impacts of snowpack augmentation in the San Juan Mountains, Colorado. San Juan Ecology Project, Final Report, Colorado State University Publications, Fort Collins. Rathcke, B., and E. P L.acey. 1985. Phenological pat- terns of terrestrial plants. Annual Review of Ecology Systems 16: 179-214. Rosenberg, N. J. 1974. Microclimate. The biological environment. John Wiley and Sons, New York. 315 pp. Scott, D., and W D. Billings. 1964. Effects of environ- mental factors on standing crop and production of alpine tundra. Ecological Monographs 34: 243-270. Smith, K. C, and D. C. Andersen. 1985. Snowjiack and variation in reproductive ecology of a montane ground-nesting passerine, Jiinco hyemalis. Ornis Scandinavica 16: 8-13. Weaver, T. 1974. Ecological effects of weather modifica- tion: effects of late snowmelt on Festiica klalioensis Elmer meadows. American Midland Naturalist 92: 346-356. Weaver, T, and D. Collins. 1977. Possible effects of weather modification (increased snowpack) on Festuca idahoensis meadows. Journal of Range Management 30: 451-456. Received 5 April 1993 Accepted 9 March 1994 Great Basin Naturalist 54(4), © 1994, pp. 376-379 OPPORTUNISTIC BREEDING AFTER SUMMER RAINS BY ARIZONA TIGER SALAMANDERS Linda J. Allison', PanI E. Brnnkow', and fames R Collins' Key words: Ambystoina, Arizona, ainphihidns, opportioiistic breed uifj,. Identifying factors influencing the number of times organisms breed during a lifetime and the seasonal timing of reproductive episodes is central to understanding the evolution of life histoiy traits (Steams 1992). In this regard, am- phibian reproductive cycles are often consid- ered adaptations to the seasonalit)' of oviposi- tion opportunities (Joly 1971, Lofts 1974, Saltlie and Mecham 1974). This is seen in most north temperate zone amphibians that breed in tem- porary' aquatic habitats and have annual repro- ductive cycles (Bishop 1947, Wright and Wright 1949). These frogs and salamanders typically come into reproductive condition once a year at the same time of year, often in spring as ephemeral habitats predictably fill from snowmelt or winter rains. Where aquatic habitats fill unpredictably or irregularly, such as in dry temperate or tropical regions, amphibians may have acyclical reproductive periods allowing them to breed opportunisti- cally (Salthe and Mecham 1974, van Beurden 1979). Opportunistic breeding in ephemeral habitats is commonly understood as an adapta- tion for avoiding predaceous fish (e.g., Webb 1969, Heyer et al. 1975, Wilbur 1977, CoUins and Wilbur 1979). Tiger salamanders {Amhijstoma tigriniiin Green) range across North America and have an obligatory aquatic larval stage (Stebbins 1985). Like most ambystomatids (Bishop 1947), the six subspecies of tiger salamanders in western USA {calijorniense, niavortium, nehulosiim, diaboli, melanostictwn, stebbinsi [Collins et al. 1980, Jones et al. 1988]) all breed in late winter, spring, or even early summer at high elevations. Differences between populations in the timing of the pri- mary breeding period correspond to differ- ences in availability of water in their breeding habitats (Houghton 1976). Some A. t. mavor- tiiiin populations breed both in spring and in summer (Webb and Roueche 1971), and Tanner et al. (1971) documented A. t. nebulo- sum breeding in spring and summer in a con- tinuously filled lake. In this study we report that A. t. nebulosion populations can also breed twice per year in ponds that fill with water during the winter, dry during the late spring and early summer, and refill during summer rains. Arizona tiger salamanders (A. t. nebulosiiin) are found commonly at high elevations in montane Colorado and Utah (Stebbins 1985) and in Arizona between 1500 and 2900 m (Collins 1981). Aquatic habitats discussed in this study are in Rocky Mountain montane conifer forest (Pase and Brown 1982). Arizona tiger salamanders breed regularly in late winter and spring following snowmelt (Sexton and Bizer 1978, Collins and Cheek 1983, Holomuzki 1986, Jones and Collins 1992). During the course of other fieldwork, we realized that a second breeding pattern also occurs. Here, we present obsenaitions that led us to conclude that these tiger salamanders can breed "opportunistically," defined as any breeding outside the usual late winter and early spring breeding period. On 30 March 1990 we noted typical spring oviposition activity when we observed thou- sands of A. /. nebidosiim eggs in Horseshoe Lake (34°22'53"N, liri4'38"W) on the Mogollon Rim in central Arizona. We visited the lake again on 6 August 1990, before simi- mer rains began. It was completely diy at this time, and we photographed thousands of des- iccated salamander lai'vae on the lake bed. We presumed this represented elimination of the spring 1990 cohort. We sampled this lake again 'Depaiinu-iit of Zoolof^; Arizona State UiiivcTsily. Tciiipe, Arizona 85287-1501. 376 1994] Notes 377 on 23 September 1990 after it had refilled fol- lowing summer rains and noted several hun- dred small A. t. nehulosum lai-vae. Live mea- surements were taken from a sample of lai-vae on 29 September and compared to measure- ments taken from projections of two close-up slides of some of the dead spring cohort. We also visited other ponds in this region that had dried and refilled to gauge the extent and suc- cess of this breeding tactic. Larvae caught in Horseshoe Lake in Sep- tember were either a second cohort that hatched after summer rains, or animals from the spring cohort that survived the lake drying, presumably by burrowing. Metamorphosed tiger salamanders burrow in soft lake mud (Webb 1969), but this has not been reported for larvae or branchiate adults. Dead larvae photographed in August aver- aged 51.4 mm total length (SE = 1.20 mm, N = 35), and larvae collected in September averaged 35.8 mm total length (SE = 0.78 mm, N = 96). If animals collected in Septem- ber survived diying by burrowing, they should have been at least as large as the dead animals observed in August. Animals in September were, however, significantly smaller than those photographed in August [i = 10.53, P <.0001). This is a conservative test since ani- mals photographed in August were dried and therefore smaller than at death. We do not consider summer breeding to be the primary breeding event for this popula- tion. Because we observed several thousand eggs in March, we conclude that there was a normal spring breeding. Typical lai-val densi- ties (estimated by drop-box and seining through a measured volume of water) in this part of Arizona in June are 2-75 salamanders jj^-3 (Pfennig et al. 1991). The low density of larvae we estimated in September, 0.3 sala- manders m~'^, supports the conclusion that there was a second reproductive episode in which a small number of animals bred oppor- tunistically in Horseshoe Lake during 1990. Adult salamanders breeding in Horseshoe Lake in late summer 1990 were taking advan- tage of a newly filled habitat. At least some lar- vae in this second 1990 cohort ovei-wintered successfrilly (M. Loeb personal obsei-vation, 4 May 1991). In 1990, Charco Tmk (34°07'50"N, 110°07'32"W) in the White Mountains dried following drought conditions that also dried Horseshoe Lake. In spring 1991 we collected larvae several centimeters larger than recently hatched salamanders in this tank, suggesting salamanders produced a second cohort in this tank following rains in summer 1990. Cottonwood Tank (34°08'43"N, 110°09'06"W) dried in June 1992. Although we did not visit this habitat later in 1992, presence of a large larva in March 1993 (before the spring cohort hatched) suggests opportunistic breeding in summer 1992. Late-season breeding may, however, fail. Johnnie Tank (34°10'06^'N, 110°04'02"W) in the White Mountains is at the same elevation as 13 ponds in the surrounding 180 km-. Breeding at all other ponds in this area was completed by late March in 1992 and 1993. Johnnie Tank dried in early spring 1992 and then refilled in late May after early monsoon rains. Hatchlings produced by opportunistic breeding following this refilling were all killed when the tank dried again in late June. Because the May oviposition does not overlap the usual breeding season in this area, we con- sider this to be opportunistic breeding. Odier evidence suggests tliat summer l:)reed- ing is exhibited regularly in this subspecies. Metamorphosed females with yolked follicles were recorded in Arizona by Durham (1956) on 18 July on the Kaibab Plateau and by J. Collins (unpublished observation) on 11 July 1980 in the White Mountains. Tanner et al. (1971) reported sizes of A. t. nehulosum lai^vae in Salamander Lake, a per- manent lake in Utah. While following growth of larvae throughout July and August, they recorded a small size class beginning in late July, which they interpreted as evidence of a second breeding. A pattern of spring and fall breeding in permanent ponds with unpre- dictability in seasonal rainfall is also reported for Triturus alpestrls apuanus in Italy (Andreone and Dore 1992). Despite the fact that we reg- ularly visit a few dozen continuously filled ponds in Arizona, we have never obsen^ed a second breeding in one. Our observations, consistent with the hypothesis outlined below, emphasize that the natural history we are describing for A. t. nehulosum in Arizona dif- fers importantly from that reported in Utah. Webb (1969) argued that an irregular breeding pattern is among the traits that adapt A. t. mavortium for life in the Chihuahuan Desert in southern New Mexico; i.e., this sub- species reproduces whenever water fills the 378 Great Basin Naturalist [Volume 54 ephemeral ponds in which it commonly breeds. A. t. mavortiwn breeds every year in ponds that fill in winter or spring, but may also breed after summer rains. Oui- data sug- gest A. /. nebiiloswn in Arizona has evolved a similar life histoiy tactic. In general, breeding occurs following snowmelt at high elevations, but there are some conditions under which individuals will breed opportunistically fol- lowing sunnner rains. An amphibian larva from a spring cohort is not guaranteed sufficient time to complete development to metamoiphosis in any aquatic habitat that can diy unpredictably. This is an explanation for iteroparity in most ambystom- atids using temporary or "most nearly perma- nent" ponds (Wilbur 1977). It might be adap- tive, however, for adults to take advantage of ponds whenever they refill. In contrast, late- season breeding in permanent aquatic habitats is generally not advantageous since these habitats may harbor older larvae or fish that can prey on embryos and hatchlings (Burger 1950, Reese 1968, Webb and Roueche 1971; but see Dodson and Dodson 1971 and Collins and Holomuzki 1984). Breeding after a habitat refills would be advantageous as the drying would eliminate fish or older larvae. A second advantage of breeding opportunistically arises since metamorphosis is only possible after a minimum size is achieved (Wilbur and Collins 1973), and in temporaiy ponds like those we report here, this size may not be attained before the pond dries; however, larvae from summer cohorts that successfully overwinter will have a growth advantage over larvae from spring clutches. A closer examination of the life histories of other subspecies of tiger salamanders found in the arid and semiarid western USA might reveal that regular breeding in late winter and spring, with opportunistic breeding in sum- mer, also occurs in these subspecies. Acknowledgments We thank the White Mountain Apache Indian tribe for their cooperation and permis- sion to work on their reservation (permits #90-04, #91-01, and #92-03). The Arizona Game and Fish Department provided collect- ing permit #CLNS0000118. This work was supported by NSF grant #BSR-8919901 to JPC. Literature Cited Andreone, E, and B. Dore. 1992. Adaptation of the reproductive cycle in Tritiirtis alpcstris apiianits to an unpredictable habitat. Aniphibia-Reptilia 13: 251-261. Bishop, S. C. 1947. Handbook of salamanders. Comstock Publishing Company, Ithaca, New York. 555 pp. BuRCER, W. L. 1950. Novel aspects of the life histories of two ambystomas. Journal of the Tennessee Academy ofScience 25: 252-257. Collins, J. E 1981. Distribution, habitats, and life histor\' variation in the tiger salamander, Ambijstoina tigrinmn, in east-central and southeast Arizona. Copeia 1981: 666-675. Collins, J. P, and J. Cheek. 1983. Effect of food and density on development of typical and cannibalistic salamander larvae in Ambtjstoma tigrinum nebiilo- siim. American Zoologist 23: 77-84. Collins, J. P, and J. R. Holomuzki. 1984. Intraspecific variation in diet within and between trophic moiphs in larval tiger salamanders {Ainhystoma tigrinum nehulosum). Canadian Journal of Zoology 62: 168-174. Collins, J. P, and H. M. Wilbur. 1979. Breeding habits and habitats of the amphibians of the Edwin S. George Resei-ve, Michigan, with notes on the local distribution of fishes. Occasional Papers of the Museum of Zoology, University of Michigan 686: 1-34. Collins, J. P, J. B. Mitton, and B. A. Pierce. 1980. Amhij- stonui tigrinum: a multi-species conglomerate? Copeia 1980; 938-941. Dodson, S. I., and V. E. Dodso.n. 1971. The diet of Amlnjstoma tigrinum lai-vae from western Colorado. Copeia 1971: 614-624. Durham, F. E. 1956. Amphibians and reptiles of the North Rim, Grand Canyon, Arizona. Herpetologica 12: 220-224. Heyer, \V. R., R. W. Mc.Diarmid, and D. L. Weigmann. 1975. liidpoles, predation and pond habitats in the tropics. Biotropica 7(2): 100-111. Holomuzki, J. R. 1986. Effect of microhabitat on fitness components of larval tiger salamanders. Oecologia 71: 142-148. Houghton, F E. 1976. Climates of die States. New Me.\ico. Climatography of the United States, 60-15: 677-681. Gale Research Company, Detroit, Michigan. JOLY, J. M. J. 1971. Les cycles sexuels de Salamandra (L.). I. Cycle sexuel des males. Annales des sciences naturelles. Zoologie et Biologic Animale 13: 451-504. Jones, T. R., and J. P Collins. 1992. Analysis of a hybrid zone between subspecies of the tiger salamander {Ambystoma tigrinum) in central New Mexico. Journal of E\()luti()nar\' Biology' 5: 375-402. Jones, T. R., J. E Collins, T. D Kocher, .\nd J. B. Mitton. 1988. Systematic status and distribution oi Ambij- stonui tigrinum slebbimi Lowe. Copeia 1988: 621-635. Lofts, B. 1974. Reproduction. Pages 107-218 in B. Lofts, ed.. Physiology of the amphibia. Academic Press, New York. Ease, C. E, and D. E. Brown. 1982. Rocky Mountain (Eetran) and Madrean montane conifer forest. Desert Elants 4(1-4): 43-48. 1994] Notes 379 Pfennig, D. W, M. L. G. Loeb, and J. P Collins. 1991. Pathogens as a factor limiting the spread of cannibal- ism in tiger salamanders. Oecologia 88: 161-166. Reese, R. VV. 1968. The taxonomy and ecology of the tiger salamander {Ambysfoma ti^rinuin) of Colorado. Unpublished doctoral dissertation, University of Colorado, Boulder. Salthe, S. N., and J. S. Mecham. 1974. Reproductive and courtship patterns. Pages 310-521 in B. Lofts, ed., Phvsiology of the amphibia. Academic Press, New York. Sexton, O. J., and J. R. Bizer. 1978. Life histoiy patterns oi Ainhystoma tigrinum in montane Colorado. American Midland Naturalist 99: 101-118. Stearns, S. C. 1992. The evolution of life histories. Oxford University Press, Oxford. 249 pp. Stebbins, R. C. 1985. Western reptiles and amphibians. Houghton Mifflin Company, Boston. 336 pp. Tanner, W. W., D. L. Fisher,' and T. J. Willis. 1971. Notes on the life histoiy of Ambystoma tigrinum neb- ulosiim Hallowell in Utah. Great Basin Naturalist 31: 213-222. van Beurden, E. K. 1979. Gamete development in rela- tion to season, moisture, energy reserve, and size in the Australian water-holding frog, Cyclorana platy- cephahts. Heipetologica 35: 370-374. Webb, R. G. 1969. Survival adaptations of tiger salaman- ders {Ambystoma tigrinum) in the Chihuahuan Desert. Pages 143-147 in C. C. Hoff and M. L. Riedesel, eds.. Physiological systems in semiarid environments. University of New Mexico Press, Albuquerque. Webb R. G., and W Roueche. 1971. Life history aspects of the tiger salamander {Ambystoma tigrinum mavor- tium) in the Chihuahuan Desert. Great Basin Naturalist 31: 193-212. Wilbur, H. M. 1977. Propagule size, number, and disper- sion pattern in Ambystoma and Asclepias. American Naturalist 111:43-68. Wilbur, H. M., and J. P Collins. 1973. Ecological aspects of amphibian metamorphosis. Science 182: 1305-1314. Wright, A. H., and A. A. Wright. 1949. Handbook of frogs and toads of the United States and Canada. Cornell University' Press, New York. 640 pp. Received 19 January 1993 Accepted 31 March 1994 Great Basin Naturalist 54(4), © 1994, pp. 380-383 VEGETATION RECOVERY FOLLOWING FIRE IN AN OAKBRUSH VEGETATION MOSAIC Stephen E Poreda^-^ and Leroy H. WuUstein^''^ Key iv(>r(ls:fire. oak, secondarij succession, soil erosion, shrub, firass, ecosystem. Fire plays a role in maintaining eeosystem diversity, improving forage production, enhancing wildlife habitat, and recycling nutrients in the soil (Wright and Bailey 1982). Postfire succession in the Gambel oak {Qiiercus gamhelii) type, however, has received less attention in the literature than most major vegetation types. The most extensive work on secondaiy succession in the Gambel oak t\'pe was done by McKell (1950). In August of 1990 an intense wildfire burned nearly 3000 ac of oak-dominated vege- tation in the \'icinit>' of Wasatch Mountain State Park near Midway, Utah. This note reports on the first-year vegetation recoveiy in a vegeta- tion mosaic of oakbrush and sagebrush-grass communities during the first year following that fire. Individual study sites are located in Heber Valley near tiie town of Midway, Utali. All are in the lower foothill zone of the central Wasatch Mountains. Gambel oak dominates the hillside vegetation of Heber Valley and grows in a dis- continuous belt extending from approximately 1500 to 2600 m elevation. Study sites lie with- in the ecotone near the lower margin of the scrub-oak belt and the upper margin of the foothill zone. The ecotone comprises a vegeta- tion mosaic of open spaces and oak-clone thickets. Major shrub or tree species associat- ed with Gambel oak include Prunus virgini- ana, Acer grandidentatiiin, Sy)ni)horic(irpos oreophihis, and Ainelancliier alnifolia. Interspersed among oak-clone thickets are open spaces containing vegetation characteris- tic of both the mountain shrub community and the foothill zone. The interspaces characteris- ticalK' support populations of Artetnisia triden- tata, Purshia tridentata, Chnj.sothcnnnus vis- cid ifloriis, Broiniis tcctonun, and Agropyron spicatiiin. Climate of the Heber Valley area is charac- teristically continental. Annual precipitation at the Heber City weather station averages 39 cm. Mean annual snowfall is 175 cm. Annual average daily maximum and minimum tem- peratures are 16.1 °C and -2.6 °C, respectively. The frost-free period is typically 70-80 days (USDA 1976). One study site was selected within the bum, and a similar nearby unbumed area was chosen to provide comparison. Sites were selected for similarity of elevation, aspect, slope, and soils. The burned site (T3S R4E S33 SEl/4) within an elevation range of 1771-1832 m consists of two opposing slopes, one with a generally east-facing aspect, the other generally west- facing. Variability in topography allowed for sampling across a full range (0-360°) of aspect. The unburned site consists of a roughly cir- cular sampling transect centered on Memorial Hill (T3S R4E S35 NEl/4). The sampling ele- vation ranges from 1740 to 1771 m. Slopes of both sites range from 20 to 60%. Soils consist of a complex of Hennefer silt loam and Hennefer cobbly silt loam (Pachic argixe rolls; USDA 1976). Runoff is rapid on Hennefer soils and erosion hazard is considered high. Burned and unburned sites were identified for intensive sampling and quadrat analysis. Two supplemental sites of similar slope, eleva- tion, and topography were selected for recon- naissance sui"vey and identified as area C (T3S R4E S21 NEl/4) and area D (T3S R4E S23 NEl/4). Comprehensive species checklists were compiled for all sites to establish whether 'Department of Geography, University of Utah, Salt Liike City Utah 84n2. ^Present address: 839 E. Garfield Ave., Salt Lake City, Utah 84105. ■^Address correspondence to this author 380 1994] Notes 381 the intensive study sites were representative of vegetation of the vicinit\' as a whole (Poreda 1992). Sampling was done in mid-June, early August, and late September during the 1991 growing season. For the reconnaissance sur- veys, species were recorded as encountered while the surveyor walked arbitrarily selected transects within each of the four study areas. This permitted observation of additional species not found in the quadrats. Quadrat sampling was conducted along transect lines established on both the burned and unburned sites. Quadrats (1.0 m-) were marked at 30-m intei"vals along each transect line. The burned site contained 171 quadrats, the unburned site 39. Cover for each species, total vegetative cover, bare soil, rock, and litter were estimat- ed within each quadrat using a procedure slightly modified from Daubenmire (1959). The modification consisted of adding one extra cover class with limits of 0-1%. This modification provided a more accurate esti- mate of cover for small or subordinate species (Davis and Harper 1989). Plant densities were based on counts of individuals (by species) rooted within the l.O-m^ quadrats. Species frequencies were the percentage of quadrats in which a species occurred. Values for cover-class and density were recorded for each species for each quadrat along with sampling date and community type. Species mean percent cover (%C) was computed separately for oak and shrub-grass communities (including aggregated communi- ties) on both the burned and unburned sites. The Mann-Whitney test was used for signifi- cance testing of mean differences. Percent of total cover (%TC) of each species was expressed as a percentage of the summed maximum cover of all species. Species identification fol- lows Amow et al. (1980). Total vegetative cover was substantially reduced for both communities on the burn even after one full year compared to that of the unburned site (Table 1). Changes in bare soil June through September on both unburned sites are not statistically significant {P >.28), nor is tlie change between June and September on burned shrub-grass sites (F >.4). On burned oak sites, however, the trend of signifi- cantly (P =.0004) decreasing bare soil can be attributed to the dramatic increase in vegeta- tive cover. Species count (Table 2) on burned shrub- grass quadrats (73) was 1.87 times the number on unburned quadrats (39). Total number of species on burned oak sites (61) was 1.33 times greater than the number on unburned sites (46). The data strongly indicate that shiTjb-grass sites are richer in species than oak sites, and that this relationship holds even after fire. Moreover, species diversity is higher on both communities following fire. Many species declined in both frequency and cover following fire (Tables 3, 4). Others increased in both frequency and cover follow- ing fire. A few showed little change. On shrub-grass sites species showing great- est decrease include Agoseris glauca, Agro- pyron spicatwn, Artemisia tridentata, Crepis acuminata, Lathynis pauciflorus, and Loma- tiiim triternatum. Species increasing after Table 1. Seasonal areal coverage (%C)'' of vegetation, litter, bare soil, and exposed rock for burned and unburned sites'\ Shrub-grass Oak June Aug Sept June Aug Sept Burned Vegetation 16.79 12.87 28.14 16.30 41.59 49.50 Litter 6.76 21.67 22.86 7.22 9.21 11.57 Bare soil 35.72 37..38 37.98 42.88 33.63 32.44 Rock 18.09 18.74 19.41 10.28 10.25 10.15 Unburned Vegetation 46.09 31.59 49.50 73.94 65.06 73.18 Litter 37.30 53.95 60.27 88.52 90.59 92.65 Bare soil 7.95 9.05 8.93 6.53 0.79 0.65 Rock 18.80 18.80 19.34 3.09 3.09 2.94 ^%C expressed as mean co\er based on m- ciiiadrats "n = 171 (l)unied); n= 39 (unburned). 382 Great Basin Naturalist [Volume 54 Table 2. Cumulative number of species^ observed in sampled quadrats. Burned Total Annual Forb Grass Shrub UNBURNEI) Total Annual Forb Crass Shruli Shrub- grass Oak Aggregated # 7c # "-/c # % 73 100 61 100 80 100 24 33 20 33 26 32 33 45 28 46 36 45 9 12 4 6 9 11 7 10 9 15 9 11 39 100 29 100 46 100 10 26 9 31 14 30 16 41 9 31 17 37 8 20 6 21 8 17 5 13 5 17 ~ 15 **Jiine throu^li SepteinI>t.T 1991 Table 3. Frequency and cover for major species on shrub-grass sites. Unburned Burned Species Freq. %C %TC Freq. 9^C %TC Agoseris glauca 9.10 .05 .07 .00 .00 .00 Agropyron spicatwn 63.60 10.31 16.04 9.80 .80 2.11 Amelanchier ainifolia .00 .00 .00 1.60 .05 .13 Artemisia tridentata 27.30 4.59 7.15 .00 .00 .00 Aster chilensis 9.10 .16 .25 3.30 .50 1.30 Bronms fectorum 68.20 6.84 10.64 70.50 6.51 17.14 Chenopocliiiin allntin .00 .00 .00 24.60 1.41 3.72 C. leptophijUwn .00 .00 .00 19.70 .50 1.32 Collinsia parviflora 18.20 .21 ..32 26.20 .34 .88 Crepis acuminatum 54.50 2.16 3.36 4.90 .07 .17 Galium aparine 4.50 .02 .04 24.60 .52 1.38 Lathyru.s pauciflorus 13.60 .84 1.31 3.30 .05 .13 Lomatium triternatum 27.30 .91 1.41 4.90 .02 .06 Machaeranthera canescens .00 .00 .00 9.80 .40 1.06 Poa pratcnsis 27.30 6.14 9.56 18.00 2.39 6.28 Pohjgomim ramosissimum 4.50 .02 .04 19.70 .26 .69 Primus virginiana .00 .00 .00 3.30 .30 .78 Qtiercus gamhelii 22.70 5.49 8.55 14.80 1.59 4.17 Solidago sparsiflora .00 .00 .00 4.90 .54 1.42 Symphoricarpos oreaphilus .00 .00 .00 4.90 .11 .28 Verbascum thapsus .00 .00 .00 19.70 .90 2.37 Viguiera multiflora .00 .00 .00 36.10 2.53 6.67 buiTiing include Chenopodiuin album, C. lepto- phyllmn, Collinsia parviflora, Galium aparine, Machaeranthera canescens, Polijf^omnn ramosis- simum, Solidago sparsiflora, Verbascum thap- sus, and Viguiera multiflora. Some species, such as Aster chilensis, may be somewhat less typical in that frequency was lower on burned shrub-grass sites, yet cover was actually greater than on unburned sites (Table 4), suggesting a response of increased size and vigor of the surviving individuals. Relative to the unburned, Bromus tectorum showed little difference in frequency or cover. suggesting only a minor effect in the first year following fire; relative importance of this species was, however, enhanced due to the decline of most other species. Studies by Young and Evans (1978) suggest a potential explosive increase of B. tectorum in the sec- ond and third years after fire as the species rapidly colonizes space made available by fire. On oak-dominated sites there was a similar or greater reduction of those same species exhibiting a lowered frequency on shrub-grass sites. In addition. Aster chilensis and Bromus tectorum (species showing substantial sui-vival 1994] Notes 383 Table 4. Frequency and cover for major species on oak-dominated sites. Unburned Burned Species Free J. %C %TC Freq. %C %TC Agoseris ghiiica 11,80 1.06 1.10 .90 <.01 .01 Agropijron spicatum 29.40 .59 .61 .00 .00 .00 Aiuchinchicr ainifolia 23.50 1.09 1.13 8.10 .95 1.66 Artemisia tndentata 17.60 1.09 1.13 .00 .00 .00 Aster chilensis 5.90 .03 .03 .00 .00 .00 Broimis teetoruin 23.50 4.17 4.33 18.90 .16 .28 Chenopodiwn album .00 .00 .00 24.30 1.24 2.15 C. leptophyUum 5.90 .03 .03 19.80 .39 .67 CoIIinsia parviflora 29.40 .15 .15 27.90 .32 .56 Crepis acuminata 47.10 1.38 1.44 3.60 .02 .03 Galium aparine 11.80 .21 .21 24.30 .69 1.20 Lathijrus pauciflorus 58.80 2.59 2.68 7.20 .30 ..52 Lomatium triternatum 35.30 .77 .79 1.80 .01 .02 Machaeranthera canescens .00 .00 .00 .90 .03 .05 Poa pratensis 52.90 16.81 17.44 17.10 1.23 2.14 Polygonum ramosissimum .00 .00 .00 6.30 .03 .05 Prunus virginiana 29.40 8.38 8.69 10.80 .84 1.46 Qiiercus gambelii 100.00 51.12 53.03 92.80 38.36 66.68 Solidago sparsiflora .00 .00 .00 2.70 .19 .33 Verhascum thapsus .00 .00 .00 19.80 1.12 1.95 Viguiera multiflora .00 .00 .00 24.30 .86 1.49 on shrub-grass sites) were virtually eliminated from oak-dominated sites. Higher burn tem- peratures associated with oak-dominated veg- etation were likely more damaging to these species. Shrubs most common to oak sites, Prunus virginiana and Amelanchier ainifolia, both had decreased frequencies following fire. Amelanchier ainifolia, however, exhibited more vigorous resprouting. Although frequen- cy was lower on burned oak sites, cover of A. ainifolia one year after the burn was only slightly less on the burn site than on the unburned sites. Literature Cited Arnow, L., B. Albee, and A. Wyckoff. 1980. Flora of the central Wasatch Front, Utah. University of Utah, Salt Lake City. Daubenmire, R. 1959. A canopy-coverage method of veg- etational analvsis. NoilJiwest Science 33: 43-64. Davis, J. N., and K. T. Harper. 1989. Weedy annuals and establishment of seeded species on a chained juniper-pinyon woodland in central Utah. Unpublished manuscript. McKell, C. M. 1950. A study of plant succession in the oak brush {Quercus gambelii) zone after fire. Unpublished master's thesis. University of Utah, Salt Lake City. POREDA, S. F 1992. Vegetation recovery and dynamics fol- lowing the Wasatch Mountain fire (1992), Midway, Utah. Unpublished master's thesis. University of Utah, Salt Lake City 162 pp. USDA. 1976. Soil survey of Heber Valley area, Utah. National Cooperative Soil Survey. Wright, H. A., and A. W. Bailey. 1982. Fire ecolog>'. John Wiley and Sons, New York. Young, J. A., and R. A. Evans. 1978. Population dynamics after wildfires in sagebrush grasslands. Journal of Range Management 31: 283-289. Received 2 July 1993 Accepted 19 April 1994 H E GREAT BASIN NATURALIST N D E X VOLUME 54 - 1994 BRIGHAM YOUNG UNIVERSITY Great Basin Naturalist 54(4), © 1994, pp. 386-392 INDEX Volume 54—1994 Author Index Allen, Brant C, 130 Allison, Linda J., 376 Allphin, Loreen, 193 Anderson, Jay E., 204 Anderson, R. Scott, 142 Austin, George T, 97 Bacon, Kerniit L., 106 Barnes, Myra E., 351 Bates, J. William, 248 Beauchamp, David A., 130 Berg, Louis N., 272 Berkliousen, Amy, 71 Blank, Roliert R., 291 Bolton, Harvey, Jr., 313 Bozek, Michael A., 91 Britten, Hugh B., 97 Brunkow, Paul E., 376 Brussard, Peter E, 97 BiTisven, Merlyn A., 64 Budy, Phaedra E., 130 Bursey, Charles R., 189 Busacca, Alan J., 234 Buss, Warren, R., 182 Coates, Kevin P, 86 Collins, James P, 376 Connelly John W, 228 Conover, Michael R., 329 Cramer, Kenneth L., 287 Crawford, John A., 170 Crowe, Elizabeth A., 234 Dalton, Jack, 79 Danvir, Rick, 122 De Rocher, T. R., 177 Drut, Martin S., 170 Elser, James J., 162 Findholt, Scott L., 342 Godfrey Jeffrey M., 130 Goldberg, Stephen R., 189 Goldman, Charles R., 162 Gregg, Michael A., 170 Gross, Glenn, 335 Halvorson, Jonathan J., 313 Haro, Roger J., 64 Haqier, Kimball T, 193 Hepworth, Dale K., 272 Huntly, Nancy, 204 Junge, Christopher, 162 Klemmedson, James O., 301 Koehler, Peter A., 142 LaRue, Charles T, 1 Laundre, John W, 114 Levesque, Steve, 150 Lindauer, Ivo E., 182 Mclver, James D., 359 Mclvor, Donald E., 329 Modde, Timothy, 366 Moretti, Miles 6., 248 Morton, Martin L., 371 Muirhead, Neal, 377 Murphy, Dennis D., 97 Musil, David D., 228 Ottenbacher, Michael J., 272 Paton, Peter W. C, 79 Poole, Stephen, 335 Poreda, Stephen E, 380 Potts, Robert, 335 Rasmuson, Kaylie E., 204 Reese, Kern' P, 228 Reganold, John P, 235 Reichenbacher, Frank W, 256 386 1994] Index 387 Richards, Carl, 106 Ritchie, Mark E., 122 Rossi, Richard E., 313 Rust, Richard W, 351 Scheiiinitz, Sanford D., 86 Shapiro, Arthur M., 71 Smith, Jeffrey L., 313 Smith, Jocelyn, 191 Steeu, Tiygve, 359 Stone, Eric, 191 SutheHand, Steven D., 212 Svejcar, Tony J., 291 Tausch, R. J., 177 Thornton, Polly, 191 Tiedemann, Arthur R., 301 Trent, James D., 291 Ueng, Rengen, 182 Uresk, Daniel W, 156 Vickeiy, Robert K., Jn, 212 Vinson, Mark, 150 Wolfe, Michael L., 122 WuUstein, Leroy H., 380 Yamamoto, Teruo, 156 Young, Michael K., 91 Zamora, Benjamin A., 234 Key Word Index abundance, 130, 272 Agropyron smithii, 182 Allium, 371 Amhijstoma, 376 amphibians, 376 ants red wood, 359 aquatic surface respiration, 150 Aquihi chnjsaetos, 248 Arizona, 1, 376 cliffrose, 256 Artemisia tridentata, 291, 313 Asia flammeus, 191 avifauna sui^vey, 1 jjehavior, 86 behavioral thermoregulation, 359 bighorn sheep, 114 Bighorn Canyon National Recreation Area, biomass prediction, 177 vield, 301 bird(s) breeding, 1 density, 1 4iabitat relationships, 1 Black Mesa, [Arizona], 1 body size, 71 branch orientation, 204 breeding, 342 birds, 1 bulk density 301 bumblebees, 351 calcium, 182 California, 142 carbon isotope, 204 Centrarcus urophasianiis, 122 Centrocercus urophasianiis, 170, 228 Chasmistes liorus, 366 chigger, 189 chloride, 182 chlorophyll, 182 chronology migration, 79 Clirysothammis uauseosus, 71 cliffrose, 256 Arizona, 256 coal mining, 1 cobble embeddedness, 64 Colorado Plateau, 193 cormorants, 272 Cottus beldingi, 64 crayfish, 162 critical habitat, 193 debris torrent, 91 density bird, 1 bulk, 301 depredation, 329 desert adaptation, 359 fishes, 150 diet selection, 191 discrimination, 204 dispersal, 228 distribution, 342 diurnal activity, 329 drift, 366 ecosystem, 380 emergence, 366 endangered species, 256 Erigeron kachinensis, 193 Eutrombicula lipovskyami, 189 fire, 91, 380 fish kill, 91 388 Great Basin Naturalist [Volume 54 foliage biomass, 177 food habits, 272 forbs, 156 Fremont Island, [Utah], 287 Geckobiella texana, 189 gene flow, 97 Gentiami, 351 geoehemistry, 335 geostatistics, 313 Golden Eagle, 248 grass, 380 Great Basin, 97 Great Salt Lake, 287 Greater Yellowstone Ecosystem, 91 Greater Sandhill Cranes, 329 Grwi canadcmis tcibida, 329 habitat, 122, 170, 329, 366 critical, 193 use, 86, 228, 248 herbivor>', 162 Hesperia, jiiba, 71 heterozygosity, 97 hibernation, 71 high altitude, 371 historical records, 342 home range, 228 honeydew harvest, 359 Hopi resei-vation, 1 hyphal length, 291 hyporheos, 106 Idaho, 228 intensity, 189 intermittent streams, 150 introgression, 256 inventory, 342 island biogeography, 287 Joshua tree, 204 June sucker, 366 Juniperus scopulururn, 142 kokanee, 130 kriging, 313 Laridae, 342 larvae, 366 Lepidoptera, 97 life span, 130 litter, 301 Long-billed Curlew, 79 macroinvertebrate, 106 macrophytes, 162 magnesium, 182 mammals, 287 management, 272 migration chronology, 79 Mimidus, 212 mining, 156 coal, 1 mite, 189 Montana, 86 morjihological adaptations, 204 moiphometrics, 256 mountain goats, 114 sheep, 86 mycorrhizae, 291 Navajo resei"vation, 1 nectar, 212 replenishment, 212 variance, 212 volume, 212 neonate, 189 nest(s), 122 satellite, 359 site characteristics, 79 nonlethal effects, 64 nonpoint source sedimentation, 64 Numenhis omericanus, 79 nutrient availability, 313 NymphaUs antiopa, 71 oak, 380 Oncorhijnchus mykiss gairdneri, 150 opportunistic breeding, 376 Oreamnos americanus, 114 Oregon, 170 organic matter, 106 Ovis c. canadensis, 86, 114 oxygen tolerance, 150 packrat middens, 142 paleoecology, 142 Pasifastacus leniusculus, 162 Peromyscus maniculatus, 287 Phalacrocorax auritus, 272 phenology, 71 photosynthetic capacity, 204 Phrynosomatidae, 189 physiology, 182 plant bioavailability, 335 composition, 301 cover, 301 pluvial Owens Lake, 142 pollen cariyover, 71 pollination, 351 pollinator reward, 212 population, 248 status, 342 structure, 97 potassium, 182 predation, 122, 272 predator(s), 191 -prey, 64 1994] Index 389 prevalence, 189 prey relationships, 248 protein electrophoresis, 97 proximate factors, 371 Purshia stansburiana, 256 subintegra, 256 radio telemetry, 228 reclamation, 156 red wood ants, 359 redband trout, 150 redox potential, 335 Reithrodontomys megalotis, 287 reproduction, 371 reservation, Hopi, 1 Navajo, 1 reservoirs, 272 resource islands, 313 overlap, 114 partitioning, 114 Wiinichthijs oscitlus, 150 river, 366 Rock>' Mountain juniper, 142 rosette azimuth, 204 Sage Grouse, 122, 170, 228 sagebrush, 122 salinity, 182 sapwood area, 177 satellite nests, 359 Sceloporus jarrovii, 189 sculpins, 64 secondary succession, 380 sediment, 106 seed production, 352 selenium, 335 Short-eared Owls, 191 shrub(s), 156, 380 singleleaf pinyon, 177 snowpack, 371 sodium, 182 soil, 335 bulk density, 301 characteristics, 234 erosion, 380 pH, 301 moisture, 291 temperature, 291 water potential, 351 C„,g,301 CO3-C, 301 N, 301 1^301 S, 301 spatial correlation, 313 spawner, 130 spawning, 366 species listing [birds], 1 sport fishing, 272 stoneflies, 64 stomi, 91 stream(s) ecology, 106 intermittent, 150 summer range, 301 suspended sediment, 91 taxonomy, 256 themial refugia, 359 Tioga glacial stage, 142 translocation, 228 trees, 156 trout, 272 turnover rate, 130 Utah, 79, 272, 287, 329 vegetation zones, 234 vernal pool, 234 vesicular-arbuscular mycorrhizae, 291 Washington, eastern, 234 water potential, 182 wild horses, Wyoming, 86, 156, 191, 239, 342 Yucca brevifolila, 204 Zonotrichia, 371 390 Great Basin Naturalist [Volume 54 TABLE OF CONTENTS Voluiiie 54 No. 1 — ^January 1994 Articles Birds of northern Black Mesa, Navajo (bounty, Arizona Charles T. LaRue 1 Effects of cobble embeddedness on the microdistribntion of the sculpin Cottus heldingi and its stonefly prey Roger J. Haro and Merlyn A. Brusven 64 Persistent pollen as a tracer for hibernating butterflies: the case of Hesperia jiiba (Lepidoptera: Hesperiidae) Amy Berkhousen and Arthur M. Shapiro 71 Breeding ecology of L()ng-];)illed Curlews at Great Salt Lake, Utah Peter W. C. Paton and Jack Dalton 79 Notes Habitat use and behavior of male mountain sheep in foraging associations with wild horses Kevin P Coates and Sanford D. Scheninitz 86 Fish mortality resulting from delayed effects of fire in the Greater Yellowstone Ecosystem Michael A. Bozek and Michael K. Young 91 No. 2— April 1994 Articles Colony isolation and isozyme variability of the western seep fritillary, Speyeria ixohomis apacheana (Nymphalidae), in the western Great Basin Hugh B. Britten, Peter F Brussard, Dennis D. Muiphy, and George T. Austin 97 Influence of fine sediment on macroinvertebrate colonization of surface and hyporheic stream substrates Carl Richards and Kermit L. Bacon 106 Resource overlap between mountain goats and bighorn sheep John W. Laundre 114 Predation of artificial Sage Grouse nests in treated and untreated sagebrush Mark E. Ritchie, Michael L. Wolfe, and Rick Danvir 1 22 Timing, distribution, and abundance of kokanees spawning in a Lake Tahoe tributaiy David A. Beauchamp, Phaedra E. Budy, Brant C. Allen, and Jeffrey M. Godfrey 1 30 Full-glacial shoreline vegetation during the maximum highstand at Owens Lake, California Peter A. Koehler and R. Scott Anderson 142 Redband trout response to hypoxia in a natural environment Mark Vinson and Steve Levesque 150 Field study of plant sui-vival as affected by amendments to bentonite spoil Daniel W. Uresk and Teruo Yamamoto 1 56 Population structure and ecological effects of the crayfish Pacifastaciis loiiiisciiltis in Castle Lake, California James J. Elser, Christopher Junge, amd Charles R. Goldman 162 Brood habitat use by Sage Grouse in Oregon Martin S. Drut, John A. Crawford, and Michael A. Gregg 1 70 Needle biomass equations for singleleaf pinxon on the Virginia Range, Nevada T. R. De Rocher and R. J. Tiusch 177 Some physiological variations ol' A