HARVARD UNIVERSITY Library of the Museum of Comparative Zoology , MCZ LIBRARY H E ^^^ 0 7 15 1995 , . HARVARD UNIVZ:RSITy GREAT BASIN NMRAUST VOLUME 55 N2 1 _ JANUARY 1995 BRIGHAM YOUNG UNIVERSITY GREAT BASIN NATURALIST Editor Assistant Editor Richard VV. 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. Bowkrs 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. Johanskn Paul T. Tueller Department of Biology, John Carroll University Department of Environmental Resource Sciences University Heights, OH 441 18 University of Nevada-Reno, 1000 Vallev 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, Zoolog>'; 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 Biolog>- 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 Unpublished manuscripts that further our biological understanding of the Great Basin and sunounding areas in western North America are accepted for publication. Subscriptions. Annual subscriptions to the Great Basin Naturalist for 1995 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 © 1995 by Brigham Young University ISSN 0017-3614 Ofticial publication date: 16 January 1995 1-95 750 12935 The Great Basin Naturalist Published at Provo, Utah, by Brigham Young University' ISSN 0017-3614 Volume 55 31 January 1995 No. 1 Great Basin Naturalist 55(1). © 1995, pp. 1-18 LIFE HISTORIES OF STONEFLIES (PLECOPTERA) IN THE RIO CONEJOS OF SOUTHERN COLORADO R. Edward DeWaltl'2 and Kenneth W. Stewartl Abstract. — Thirty-one stonefly species representing eight famihes were collected during the March 1987 to May 1990 study period. Genera represented by more than one species included Capnia. Utacapnia, Taenionema. Siiuallia, Triznaka, Isogenoides, and Isoperla. Peak species richness was recorded on or near the summer solstice in 1988 and 1989. Climatic differences between years were reflected in nymphal development and emergence phenology of most species. New or important corroborative life histoiy data are presented for 1 1 stonefly species of this assemblage. The hyporheic nymphal development of most chloroperlid species limited the number of early instars sampled and our capacity' to inteipret voltinism. Limited nymphal data suggested a univoltine-slow cycle for Phimiperia diversa (Prison). Adults of Suwallia pallidula (Banks) and S\ uardi (Banks) were present for an extended summer period, but the bulk of their respective emergence times was temporally separated. Isogenoides zionensis Hanson, Pteronarcella hadia (Hagen), and Pteronarcijs califomica Newport were all showTi for the first time to have a 9-10-mo egg diapause, and all three species have a semivoltine life cycle. Skwcda americana (Klapalek) and Isoperla ftdva Claassen were further confimied to have univoltine-slow cycles. Univoltine-fast and univoltine-slow life cycles are reported for the first time in /. phalerata and /. qiiinquepunctata, respectively. Regression analysis revealed that si.x of the eight abimdant species had extended emergence patterns (slopes of <5%/d), while only two had synchronous patterns. Warmer spring and summer tempera- tures in 1989 increased the slopes for five of the eight species studied, but did not change their synchrony designation. Nine of 11 abundant species advanced their median emergence date in 1989 over 1988. This and the higher slope values are consistent with a hurried nymphal development and narrower emergence period due to the warmer thermal regime of 1989. Key words: Plecoptera, life history, biodiversity, life cycle. Rocky Mountains. Stoneflies (Plecoptera) are one of the integral understood (Sheldon and Jewett 1967, Stewart and often dominant insect orders in stream and Stark 1988). Precise life histories are ecosystems; therefore, they are important as known for <5% of the more than 575 North biological indicators, as fish food, and as part American species, and knowledge of stonefly of the energy and nutrient economy of streams life histories and ecology in southern Rocky (Stewart and Stark 1988). Taxonomy of the Mountain streams is sparse. This has limited North American fauna is now well known; our ability to increase understanding of eco- however, information on their life histories, logical relationships between cohabiting stone- local species richness, and ecology is still poorly fly species in this region. 'Department of Biological Sciences. University' of North Te.\as, Denton, TX 76203. ^Present address; Department of Zoology and Physiology, Louisiana State University, Baton Rouge, LA 70S0.3. Great Basin Naturalist [Volume 55 One objective of this study was to deter- mine richness of the stonefly assemblage of the Rio Conejos of southern Colorado, a large drainage that has not been previously studied. Second, we documented the important life history events of its dominant species for which sufficient individuals and observations could be gathered by intensive monthly sam- pling and by living streamside during spring and summer. Research was patterned after the classic studies of Harper (1973a, 1973b) and Harper and Hynes (1972), who studied a substantial portion of the eastern Canadian fauna and addressed critical aspects of life histories such as egg development, diapause, and adult behaviors that are often overlooked. H. B. N. Hynes, in an address to the International Ple- coptera Symposium (1992), emphasized the need for more attention to these aspects to support the eventual development of a para- digm of life history evolution within the Ple- coptera. We have also adopted the approaches of Knight and Gaufin (1966), Harper and Magnin (1969), Sheldon (1972), Barton (1980), Ernst and Stewart (1985a, 1985b), and Hassage and Stewart (1990) in comparatively studying an assemblage of species. This report is the first to address, on a large scale, such an assemblage in a western North American stream since the works of Knight and Gaufin (1966), Sheldon (1972), and Stanford (1975). Methods Study Stream The Rio Conejos is located in the southern Rocky Mountains of south central Colorado. The river flows east to west for 145 km from its headwaters in the Rio Grande National Forest of the San Juan range to the Rio Grande 32 km northeast of Antonito, CO. Three sam- pling sites were established along the Rio Conejos to ensure access to at least one of them during the winter and to enhance collec- tion of stonefly species that were not abundant at all sites. These were located at elevations be- tween 2400 and 2600 m above sea level. The primary site (106° 15 'W longitude, 37°03'N latitude) consisted of a 1-km stretch located 24 km west of Antonito, Conejos County, CO, off Colorado highway 17. Sites two and three were located 22.5 km west of Antonito, also on highway 17, and 4 km north of Antonito at the Colorado highway 285 bridge, respectively. Stream temperatures varied from below freezing during the winter months to near 20 °C in August. Ice cover was common from December through March. Snowmelt began in April, usually leading to peak flows in June. Base flows were attained by late August and continued through the winter. Water released from Platoro Reservoir, 48 km upstream, aug- mented river flow during summer low-flow periods. Bottom substrates were characterized by large boulders, cobble, gravel, and sand. These were covered by a thin layer of silt in quiet water. Important organic substrates included the flooded coppices of willows and cotton woods and their entrained leaf packs. Willow {Salix spp.), cottonwoods and aspens {Popidus spp.), and alder {Alnus sp.) con- tributed to the riparian corridor. Physical Conditions Stream temperature was monitored at site one from June through August 1988 using a Ryan™ continuous recording thermograph. High, low, and mean daily stream tempera- tures were calculated from temperatures recorded at 0400, 0800, 1200, 1600, 2000, and 2400 h. Water temperatures were not record- ed during 1989 due to equipment failure. However, summer air temperature highs and lows and rainfall were recorded (1300 h daily, mountain time) for both 1988 and 1989 at the Conejos Peak U.S. Weather Service reporting station at site one. Flow data for site two were gathered from Petsch (1987-90). Nymphal Growth Nymphs were collected monthly (except December due to poor weather conditions) at all sites from March 1987 to May 1988. Additional collections were made at irregular intei-vals until March 1990. Samples were col- lected by disturbing the substrate (mineral and organic) upstream of a BioQuip rectangu- lar dipnet until debris clogged the net. The net was composed of a coarse, 1-mm mesh first stage, modified by the addition of a coni- cal second stage of 153-/xm mesh size. The lat- ter collected even the smallest instars. A plankton bucket was attached to the second stage to facilitate sample removal. Contents of the plankton bucket and the coarse stage con- stituted a sampling unit and were stored in 1995] Stonefly Life Histories 70% isopropyl alcohol. The number of sam- pling units per month varied with the effort necessaiy to secure approximately 50 nymphs of all abundant species. Nymphs were separated from sample debris with the aid of 4-lOX magnification on a stereo-dissecting microscope, sorted to species when possible, and stored in 80% ethanol until measurement. Head capsule width (HCW, greatest distance across the eyes) was measured with a calibrated ocular micrometer fitted to a stereo-dissection micro- scope. Nymphs from all sites for the 3-yr sam- pling period were pooled by species and month of collection to increase the number of nymphs per month and to allow construction of more robust growth histograms. Gender of nymphs was assessed by a gap in the posterior setal margin of the eighth sternum of females (Stewart and Stark 1988) and by developing external genitalia of females. Sex-specific kite diagrams were constructed by placing male and female nymphs into 0.1- or 0.2-mm size classes. The frequency of these classes was converted to a percentage of the total number of nymphs (males + females + unsexed nymphs) collected for that month. Polygons were constructed for each month depicting the relative proportion of all nymphs at that size class. Adult Emergence Adults of winter- and early spring-emerg- ing stoneflies were collected from bridge abut- ments, from shoreline debris, and under the cobble at streamside to provide a general emergence period for each species. Adults were also reared from preemergent nymphs. A combination of sampling methods and observational procedures was used during the summers of 1988 and 1989 to evaluate emer- gence, duration of adult presence, and behav- ior of these species. Adult traps and methods included a 2.25-m^ basal area BioQuip malaise trap, two 0.25-m2 basal area floating emer- gence traps, pitfall traps, sweepnetting of streamside vegetation, exuviae collection, and day and night transect walks. Pitfall traps were emptied on alternate days, and the others were emptied daily between 0900 and 1100 h. All of these methods were used at site one; sweepnetting was employed at site three on several occasions. The malaise trap was deployed among wil- low and Cottonwood coppices, where its olive- drab coloration mimicked the surrounding vegetation. Flying, or crawling, adults inter- cepted by the trap ascended the screening into a dry apical collection chamber Addition- ally, all adults on the trap mesh were collected using an aspirator Emergence traps were anchored over shal- low riffles during the 1988 field season. Natural diurnal changes in water level and erratic discharges due to water release from Platoro Reservoir rendered these ineffective at times; therefore, their use was discontinued in 1989. Pitfall traps consisted of 28.3-cm2 modified aluminum soda cans that were buried flush in streamside substrates. A mixture of 70% ethanol and ethylene glycol (the latter to retard evaporation) was used as a preservative. In 1988, 12 traps were installed 1 m from the stream at 1-m intervals on an open beach with nearby vegetation. This was expanded in 1989 to three transects, each consisting of 30 cans set 1 m apart in transects 1 m, 5 m, and 8 m from the initial shoreline. These traps moni- tored not only adult presence of ground-tra- versing, brachypterous stoneflies, but also their potential to move laterally from the stream. Sweepnetting was conducted over a 15 x 2- m willow and cotton wood riparian zone. The entire area was methodically swept, working from the base of each clump of vegetation up- ward. Exuviae removal was the only method used to assess emergence of Claassenia sabu- losa (Banks) and was used for no other species. In 1988 exuviae were removed daily from the same 15 X 1-m area of cobble shoreline, and the frequency of each sex was noted. In 1989 the removal area was expanded to 30 x 1 m of shoreline area and up to 5 m into the water for collecting exuviae from emergent substrates. Year and sex-specific kite diagrams of adult presence were produced for all abundant sum- mer stoneflies by pooling all methods and expressing daily catches as a percentage of the total catch. Duration of emergence of Ptero- narcijs californica Newport would be greatly overestimated by including pitfall trap collec- tions due to its synchronous emergence and since pitfall traps were emptied on alternate davs. Great Basin Natur.\list [Volume 55 Dates of first capture, 50% cumulative catch, and last collection, plus total duration of adult presence, were determined for the 11 most abundant species collected in the sum- mers of 1988 and 1989. Emergence synchrony was estimated using linear regression of the cumulative percentage catch (all methods pooled) versus days since first capture. Slopes generated for each species were used as an index of synchrony. Steeper slopes indicated a more synchronous emergence. Slopes > 5%/d were chosen to be indicative of synchronous emergence since species with these slopes emerged their entire population within a few days and had steep, j-shaped, cumulative emer- gence curves. Differences between slopes for 1988 and 1989 were tested using a modified t test (Zar 1984). Common slopes were calculated if no differences between years were noted. This was a purely descriptive approach de- signed to detect and compare patterns; there- fore, it is not our aim to model emergence for the purpose of prediction, but only to describe patterns of emergence. Since most adult collection methods em- ployed in this stvidy collected adults of unknown age, results reflected adult presence rather than, in the strictest sense, emergence. No attempt was made to discard old males and females using any index of age. However, pat- terns of adult presence should follow that of a true emergence pattern, and since longevit\' of most adults approached only 1 wk in the labo- ratory, we believe these results to be useful. Behavioral observations were made from 0800 to 1300 h and from 2000 to 2300 h for sev- eral days during emergence of each species. Observations made during intervening hours produced little adult behavior. Timing of adult activities, their relative distance from the stream, and substrates on which activities took place were monitored by walking the stream margin, turning logs and rocks, and exposing leaf-entrained bases of marginal vegetation. Details of these observations have been narra- tively described for each species in this paper. Fecundity and Egg Incubation Eggs of several species were incubated in the laboratory to confirm proposed voltinism based on growth histograms. Eggs were placed into 1-cm-diameter dialysis tubing bags and reared in a Frigid Units Living Stream"', or they were stored in 100 x 15- mm plastic petri dishes in an environmental chamber. In both instances these were incu- bated at approximate stream temperature and light regime. Fecundity was estimated from number of egg batches deposited, number of eggs per batch, and, for Skwala americana (Klapalek) only, total number of eggs remaining in the ovarioles. Females were housed at streamside in screened, glass containers and provided with moist cotton balls as a source of water. Alternatively, some species were reared in Denton and held under simulated streamside conditions in large cotton-stoppered shell vials. Results Physical Conditions Mean daily stream temperatures in 1988 increased from near 10 °C in early June to 15 °C in mid- July (Fig. 1). The stream cooled dramatically between 8 and 12 July. This coin- cided with cool, damp weather conditions (Fig. 2). Summer air temperature highs rarely exceeded 30 °C in 1988, and rainfall occurred at regular intervals throughout the summer (Fig. 2). However, 1989 was marked by many days above 30 °C with rainfall relegated to late July and August (Fig. 2). The mean montlily dis- charge of the Rio Conejos during 1987-1989 fluctuated predictably. Peak discharge occurred typically in June but occurred in Mav during the warm, windy spring of 1989 (Fig. 3). Species Richness More than 13,000 nymphs and adults were studied over the 3-yr period. Among these ^0 mean low high 18 f\ £ 16 hi ^ 1 ;•.;■ \ r. A s f '/l i. \N/^ a 12 CO !\M/^-/i fv^-A/y 10 it^f'-- •• ,; V ', 8 ■--' ••■•/ 7/14 Dates Fig. 1. Daily mean, high, and low stream temperatures in the Rio Conejos, summer 1988. 1995] Stonefly Life Histories 5/27 6/6 6/16 6/26 7/6 7/16 7/26 8/5 8/15 8/25 Dates C . 5/1 5/11 5/21 5/31 6/10 6/20 6/30 7/10 7/20 7/30 8/9 Dates Fig. 2. Daih' high and low air temperatures and rainfall for summer 1988 and 1989. Fig. 3. Mean, minimum, and maximum monthly stream discharge of the Rio Conejos during the study period. were 31 species (Table 1) in eight families. The Chloroperlidae, Perlodidae, and Capniidae were the most speciose families with six, seven, and seven species, respectively. Seven genera were represented by more than one species: Capnia, Utacapnia, Taenioneina, Suwal- lia, Triznaka, Isogenoides, and Isoperla (Table 1). Peak species richness occurred on or near the summer solstice in both years (Fig. 4). Pattern differences existed between years. including an early waning and a more peaked distribution of species richness in 1989. Leuctridae Paraleuctra vershina Gaufin and Ricker. This was the only leuctrid found at our sites. No nymphs were recovered from the stream, indicating a probable hyporheic existence. Adults were abundant in riparian vegetation during June and July (Fig. 5). No variation in adult presence parameters was noted for P. vershina (Table 2). Emergence was classified as extended in both years, although slopes of these cumulative emergence curves were sig- nificantly different over the 2 yr (Table 3). Table 1. Stoneflies collected from the Rio Conejos, Colorado, March 1987 through March 1990. Euholognatha Capniidae Capnia coloradensis Claassen^ Capnia confitsa Claassen Capnia vernalis (Newport) Isocapnia crinita (Needham & Claassen)^ Utacapnia logana (Nebeker & Gaufin)* Utacapnia poda (Nebeker & Caufin)! Leuctridae Paraleuctra vershina Gaufin & Ricker* Nemouridae Amphinemura banksi Baumann & Gaufin* Prostoia hesemetsa (Ricker)* Zapada frigida (Claassen)* Taeniopterygidae Taenionema pallidum (Banks)* Taenionema pacificum (Banks)* Doddsia occidentalis (Banks)* Systellognatha Chloroperlidae Paraperia frontalis (Banks)* Phimiperla diversa (Prison)* SuualJia lincosa (Banks)* Suuallia pallidula (Banks)* Suuallia icardi Kondratieff & Kirchner* Triznaka pintada (Ricker)* Triznaka signata (Banks)* Perlidae Claassenia sabulosa (Banks)* Hesperoperla pacifica (Banks)* Perlodidae Isogenoides zionensis Hanson* Isogenoides prob. colubrimis (Hagen)* Isoperla fill va Claassen Isoperla monnona Banks* Isoperla phalerata (Smith)* Isoperla quinquepunctata (Banks) Skwala americana (Klapalek) Pteronarcyidae Pteronarcella badia (Hagen) Pteronarcijs californica Newport* ^New drainage and county records. Great Basin Naturalist [Volume 55 5/19 5/29 6/8 6/18 6/28 7/8 7/18 7/28 8/7 8/17 Dates Fig. 4. Temporal species richness pattern of adult stoneflies collected daily fiom the Rio Conejos during the summers of 1988 and 1989. Chloroperlidae Representatives from two subfamilies in- habited the stream. The early- and mid-instar nymphs of the Chloroperlinae genera could not be reliably identified to genus. This neces- sitated the illustration of a portion of the nymphal growth of Plumiperla cliversa (Prison) and Triznaka signata (Banks) as Chloroperlinae spp. (Fig. 6). Growth of reliably identified mid- to late-instar nymphs was illustrated separately. Paraperla frontalis (Banks) (Paraperlinae). Nymphs were collected infrequently among marginal substrates during the colder months of the year. All were pale, very thin, and had eyes set far forward as described for mature nymphs (Stewart and Stark 1988). These limit- ed data are presented for the first year of the presumed semivoltine growth pattern of this large chloroperlid (Fig. 6). Less than 10 adults were collected in early June during the 3-yr study. Plumiperla diversa (Chloroperlinae). No adults were collected on which to base specif- ic identity; however, nymphs of this genus are distinctive, and only P. diversa has been col- lected in this region (Baumann et al. 1977). Nymphs were identifiable to genus by March. Females were readily distinguished from males at this time. Growth continued through May when females attained a median HCW 9.6% larger than males. The limited nymphal data suggested a univoltine-slow life cycle for this species. Suwallia pallidula (Banks) (Chloroper- linae). Only 59 nymphs of Suwallia spp. were collected from the Rio Conejos, even though adults were abundant. Nymphs were hyporhe- ic until immediately prior to emergence. This habitat preference and our present inability to distinguish congeners of Suwallia nymphs precluded generation of meaningful his- tograms and designation of voltinism for either species. Adults of Suwallia wardi Kondratieff & Kirchner were consistently larger than S. pallidula. This trend followed in nymphs, too, with proposed female nymphs of S. wardi in June (peak emergence) being 22.0% larger Paraleuctra vershina M : F 25 % of total catch 36:58 9 M. kk 1988 20 30 9 19 29 9 19 MAY JUNE JULY 29 Fig. 5. Emergence of Paraleuctra vershina from the Rio Conejos, 1988 and 1989. Polygons indicate daily relative proportion of total catch. 1995] Stonefly Life Histories Table 2. Range of dates for adult presence parameters (appears first) and 1989 from the Rio Conejos. Duration is able for C. sabulosa, which emerged past our study period. for 11 summer-emerging stonefly species collected in 1988 mean ± SD of the number of days. All parameters not avail- Species n Date 1^' captiue Date 50% catch Last date capture Duration (d) P. vershina 94 58 2 June 1 June 12 June 17 June 5 Julv 7 July 35.0 ± 0.0 S. wardi 467 352 6 June 19 May 10 July 25 June 15 August 18 July 66.0 ±7.1 S. pallidida 276 162 30 June IJuly 28 July 14 July 23 August 4 August 44.5 ± 4.8 T. signata 662 2697 9 June 2 June 28 June 19 June 23 August 12 July 59.0 ± 24.0 C. sabulosa 356 1195 19 July 16 July — — — I. fulva 19 61 9 June 9 June 22 June 18 June 7 July 28 June 24.5 + 7.8 I. phalerata 12 20 20 June 18 June 28 June 25 June 24 July 8 July 22.5 ± 3.5 I. quinqiiepunctata 9 12 24 June 19 June 14 Julv 5 July 27 Julv 15 July 30.0 ± 4.2 I. zionensis 200 75 8 June 10 June 19 June 17 June 28 June 24 June 15.5 ±4.9 P. badia 215 480 10 June 7 June 22 June 20 June 17 July 7 July 34.5 ± 3.5 P. califomica 55 21 6 June 4 June 8 June 5 June 12 June 13 June 6.0 ±2.8 than the July (peak emergence) females of S. pallidula. Only two proposed male nymphs of the latter were collected over the 3-yr period. Adults of S. pallidula were collected in July and August in both years (Fig. 7, Table 2). Slopes from regression models were different between years (t = -64.7, p < .0001), but below the 5%/d criterion. We categorized this species as an extended emerger (Table 3). The median emergence date was advanced by 2 wk in 1989 over that of 1988 (Table 2). The adult sex ratio over the two seasons was 13 6 : 415 9. Six field-collected and laboratory- maintained females produced only one egg batch (Table 4). Suwallia wardi (Chloroperlinae). This was the most abundant of the three Suwallia species collected from the Rio Conejos. Adults were first collected in late May or early June, reached 50% cumulative catch by mid- July, and disappeared from streamside by early August (Table 2). It had the longest mean duration of presence (66 d) for any stonefly studied on the Rio Conejos (Table 2). Like its congener, S. wardi' s 1989 date of median catch was advanced by 2 wk over that of 1988 (Table 2, Fig. 7). Emergence of S. wardi was extend- ed, and no significant slope differences were noted between years (Table 3). No egg data were collected for this species. Riparian vegetation was used by this large, yellow-green chloroperlid as a staging ground for adult behaviors. Suwallia wardi was active throughout the morning on sunny days and again for 2-3 h before sunset if conditions were warm and dry. During cool, rainy days the low vegetation was devoid of S. wardi or any other stonefly species. Triznaka signata (Banks) (Chloroperlinae). Identifiable, late-instar nymphs were collect- ed during a 5-mo period in the spring and summer. Nymphs of this univoltine-slow Great Basin Naturalist [Volume 55 Table 3. Synchrony and linear regression statistics lor the years 1988 (appears first) and 1989. Slopes between years were tested: * = significance .()5-.0I, ** = <.0()1 level or lower probability, and NT = not tested. = 2 individuals Species Slope fi2 Svnchronv P. vershina T. signata S. piiUidttla S'. wardi C. sahiilosa I. zionensis P. badia 3.0 3.5* 1.5 3.0 2.5 3.5* 2.2 2.2 2.8 4.1 7.5 7.9 3.6 4.4* P. califomica 13.3 18.9NT .85 .90 .87 .91 .96 .90 .84 .85 .94 .99 .92 .95 .81 .92 .84 ,97 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .0001 .004 .103 extended extended extended extended extended extended extended extended extended extended synchronous synchronous extended extended synchronous synchronous species were largely full grown by April (Fig. 6) with some degree of sexual dimorphism present at this time. Adults first appeared in early June, reached 50% cumulative catch 2 wk later, and could no longer be collected by late August (Fig. 7, Table 2). Emergence was protandrous, but slightly female-skewed sex ratios dominated in both 1988 and 1989 (Fig. 7). Triznaka signata displayed the greatest variation in last date of capture and duration of presence of all stone- flies in the river (Table 2). It advanced its 1989 median emergence date by 9 d over that of 1988. Regression slopes indicated an extended emergence in both years (Table 3). Differences between slopes for 1988 and 1989 were signif- icant (t = -11.35, p < .0001). Attempts during the entire study to obtain eggs from laboratory-reared and -mated females were unsuccessful. The mean number of eggs from six females caught during ovipo- sition flights was lower than any first batches for other stoneflies studied (Table 4). Although these females were held for a prolonged peri- od of time, no additional egg batches were laid. Adults were never seen emerging in the field, despite many hours of oliservation along the shoreline, day and night, in habitats where I — I = 50 % of monthly catch n = 83 ♦ * -- oviposition \ emergence Paraperia frontalis > 1.2 . "4f4 ^^ ^ n = 453 Plumiperia diversa n = 653 Triznaka signata ;^^ n = 694 Chloroperlinae spp. Ill IV V VI VII VIII IX X XI Months Fig. 6. Growth of Chloroperlidae nymphs collected from the Rio Conejos, 1987-1990. they were collected in abundance during the day. Adults inhabited marginal vegetation, where males were observed actively searching willow stems and leaves for females. No drum- ming was observed during the two summers of intensive fieldwork. Large flights of adults of both sexes took place just before dark, at which time females were observed oviposit- ing. Egg masses were dropped from up to 4-5 m above the stream. Perlidae Claassenia sabulosa. Although two perlids were present in the Rio Conejos (Table 1), only C. sabulosa was sufficiently abundant for growth and emergence interpretation. Nymphs of this species were found among larger rub- ble of midstream. The life cycle was semivol- tine and appeared to require 3 yr of nymphal growth (Fig. 8). Recruitment occurred throughout the fall with possibly some addi- tional recruitment in March from overwinter- ing eggs. Eggs containing eyespots were recovered from the stream in October and November. Sexual dimorphism in both size and external genitalia occurred when nymphal size reached 2.2 mm HCW. The size disparity increased until the third year of growth when little overlap between the sexes remained. A protandrous emergence began in mid- July in both years (Fig. 9, Table 2). Exuviae of this species were abimdant throughout August, possibly into September. Emergence ol C. sabulosa was extended and slopes were signif- icantly different between years (t = -10.7, /; < .0001,' Table 3). 1995] Stonefly Life Histories = 2 ind. I = 5% of total catch -^g^"- Sex ratio M:F Triznaka signata ^^^ 263:446 1989 1181:1515 Suwallia wardi ,w^ ^|-g^ 1988 119:348 1989 150:191 Suwallia pallidula A 1988 13:253 1989 0:162 20 30 9 19 29 9 19 29 8 18 28 May June July August Fig. 7. Emergence of Chloroperlidae from the Rio Conejos, 1988 and 1989. Polygons indicate daily relative proportion ot total catch. Claassenia sahulosa produced the greatest mean number of eggs of any stonefly species studied (Table 4), with females producing up to four batches. Longevity of seven females was 3.9 ± 1.9 d. Egg production lasted through 80% of the adult life. Several egg batches were incubated, but none hatched within 6 mo of obsei'vation. Emergence occurred between 2000 and 2200 h. Nymphs crawled out of the water onto emergent cobble and boulders to transform, the entire molting process taking less than 5 min. Hardened and newly transformed males ran over all emergent substrates, searched for females in a circular pattern, and drummed mostly on large mineral substrates. Pitfall trap collections of 1989 caught a total of 115 male adults in transect 1 and only 12 in transects 2 and 3. Only two females were collected in the pitfall traps, presumably because of their less- intensive and unidirectional movement pat- tern. Therefore, excursions of great distance away from the water's edge for either sex were infrequent. Females were often found in the morning under dry cobble with abdomens de- void of eggs or with large egg masses suspend- ed between the cerci. Several females were observed at night running over the surface of the water, but the cause of this behavior could not be determined. No females were actually observed ovipositing. Males were distinctly cursorial, which fits with their brachypterous morphology; however, females were never observed flying, nor did they inhabit tall sub- strates, even though they had full wings. Perlodidae Isogenoides zionensis Hanson (Perlodinae: Perlodini). The large range in size of nymphs from July samples (Fig. 10) could not be accounted for by nymphs hatching from eggs laid by June-mated females. June eggs reared at simulated stream conditions hatched in March and April, 9-10 mo after oviposition. Therefore, at least some individuals of this species have a semivoltine life cycle with eggs diapausing over their first summer and winter. Early-instar nymphs were missed in benthic samples during their second spring, possibly due to high water or their occurrence deep in the substratum. Sexual dimorphism in size and morphology was apparent by July of the second year when nymphs approached 1.8 mm HCW (Fig. 10). This disparity increased steadily throughout the rest of their growth. Little overlap in size of the sexes existed by May prior to emergence. The adult presence parameters of /. zionen- sis showed little variation over the 2 yr studied (Table 2). Emergence was not protandrous, but the sex ratio was heavily skewed towards males (Fig. 11). This species was one of two that emerged synchronously (Table 3). No dif- ference in slope was found between years (t = 0.82, p > .2); therefore, a common slope of 7.6%/d was calculated. Laborator\'-reared females put nearly 75% of their total egg complement into a first batch (Table 4). Only one of four females produced additional batches. Transformation of /. zionensis took place from 2030 to about 2200 h. Nymphs crawled away from the stream until they reached wil- lows or other vegetation, then ascended < 1 m vertically where they molted. Daylight activity- began by 0700-0800 h at the base of small wil- low coppices, where adults were often found in emergent leafpacks. Adults ascended stream- side willows as the sun rose. Drumming, mat- ing, and egg batch formation took place from these perches. Females crawled to the tops of these willows and flew to the stream where they fluttered on the water to release their black egg masses. Most activity ceased by 1300-1400 h on days when air temperature reached near 25 °C. On cloudy, cool days this ascendance did not occur. Most adults could then be found in the leaf-entrained bases of riparian vegetation. Drumming on willow stems 10 Great Basin Naturalist [Volume 55 Table 4. Mean eggs per hatch, nuinher of hatches, and mean total egg complement for nine species of stoneflies occurring in the Rio Conejos, Colorado. Eggs / hatch n Species 1 2 3 4 Total S. pallidula 54.7 ± 26.6 6 — — — 54.7 ± 26.6 6 T. signata 42.2 ±17.4 6 — — — 42.2 ± 17.4 6 C. salmlosa 2166.0 ±774.0 7 902.0 ± 246.2 5 158.0 ±91.0 5 40.0 1 3188.0 ±613.0 7 I. zionemis 588.0 ± 86.0 4 327.0 1 185.0 1 — 843.2 + 141.4 4 I. fulva 231.5 ±7.8 2 — — — 231.5 ± 7.8 2 I. phalerata 703.0 1 — — — 703.0 1 S. amcricana 884.7 ± 267.3 6 — — — 884.7 ± 267.3^ 6 P. badia 339.0 ± 86.0 30 58.4 ± 37.2 5 56.8 ± 39.7 4 — 351.0 ±101.0 30 P. calif ornica 393.0 ± 125.6 4 191.3 ±130.2 4 94.3 ± 49.5 4 69.8 ± 24.7 4 5 6 7 51.3 ±29.3 4 58.5 ± 23.3 2 57.0 1 845.3 ± 90.5 4 ^Total fecundity includes those eggs remaining in o\aii(ilts was observed at night, even when tempera- tures approached 10 °C. Isoperla fulva Claassen (Isoperlinae). We collected this species in benthic samples only occasionally, but enough individuals were obtained to allow a tentative interpretation of voltinism. Recruitment of nymphs was first detected in August (Fig. 12). These measured 0.4-0.8 mm HCW and grew at a slow rate throughout the fall until a winter decrease in growth rate. Their size increased dramatically after February, until emergence in June and July. This species conformed to a univoltine- slow growth pattern. Adults were collected for the first time on 9 June in both years (Fig. 11, Table 2). Sex ratios for the small number of 1988 adults were approximately equal, but heavily skewed towards males in 1989. Numbers of adults col- lected in both years were too small to warrant an analysis of synchrony. Fecundity was difficult to assess since few mature nymphs were available for rearing. One egg batch from each of two field-oviposit- ing females was collected (Table 4). Longevity of three field-collected adult females was 5.7 ± 0.58 d. Isoperla phalerata (Smith) (Isoperlinae). Although the number of nymphs collected was small, no month supported more than one size class (Fig. 12). Therefore, we have tentatively proposed a univoltine-slow growth pattern for this species. Adults were taken from mid- June through mid- July (Table 2, Fig. 11). No assess- ment of synchrony was made for 7. phalerata due to low numbers of adults captured. Females did not produce eggs in captivity. A single egg batch from a field-collected individ- ual contained 703 eggs. Four field-caught females lived 11.3 ± 3.6 d past date of capture. Isoperla quinquepunctata (Banks) (Isoper- linae). This species was more common at site 1995] Stonefly Life Histories 11 Claassenia sabulosa 4 % of total catch I II III IV V VI VII Vlli IX X XI Months Fig. 8. Growth of Claassenia sahtilo.sa n\'mphs collected from the Rio Conejos, 1987-1990. three. The data suggested that /. qiiinquepunc- tata had a univoltine-fast growth pattern. Recruitment occuiTed in Januaiy and Februaiy (Fig. 12), and growth was rapid from March through May. Sexual dimoiphism in nymphal size was not as evident in this species as in its congeners. Emergence began in mid-June and lasted through much of July (Table 2, Fig. 11). No eggs were collected. Skwala americana (Klapalek) (Perlodinae). This species displayed a univoltine-slow growth pattern and grew faster during sum- mer and fall months than all other perlodids in the Rio Conejos (Fig. 13). Nymphs were re- cruited in June and increased their median HCW from 0.4 mm to about 2.8 mm by January. Growth was nearly completed by this time. Sexual dimorphism was apparent as early as August, and female nymphs reached a medi- an HCW before emergence that was 21.4% greater than males. Female nymphs in April were found to contain fully sclerotized eggs in their oviducts; hence, this species is fully capable of mating and egg-laying immediately upon emergence. Emergence was in April and early May when our sampling was still on a monthly basis; therefore, no detailed analysis of emergence phenology and synchrony can be offered. Adults were collected mainly from emergent logjam debris or under cobble at the stream margin. Egg batches collected in mid-April from four laboratory-reared females hatched syn- chronously after a mean of 61.0 ± 7.3 d. This corroborates field collections of early-instar nymphs in June. Only a single egg batch was collected from each of six laboratory-reared females (Table 1). 19 July 18 August Fig. 9. Emergence of Claassenia sabulosa from the Rio Conejos, 1988 and 1989. Polygons indicate daily relative proportion of total catch. Pteronarcyidae Pteronarcella hadia (Hagen). This species was found to have a semivoltine growth pat- tern. Recruitment of nymphs began in March and April from eggs laid the previous June (Fig. 14). Many small nymphs were available in benthic samples by mid-April when they were at 0.2-0.4 mm HCW. This scenario was coiToborated by laboratoiy incubation of several egg batches that hatched in March and April after a 9-10-mo diapause. Growth of nymphs was rapid throughout their first spring. Size differentiation among sexes was not apparent until August, a full 14 mo after oviposition. Median size of females just before emergence the following May was 21% greater than that of males. Emergence began by early June, with slight protandry and a preponderance of males being collected (Fig. 15). Median emergence occurred in the third week of June in both years (Table 2). Emergence was extended (Table 3) and slopes were significantly different between years (t = -2.2, p < .05). Females generally laid only single egg batches, but a small number produced up to three egg batches (Table 4). Most females laid their first egg batch within 24 h of mating and often waited 2-d intei-vals before laying others. Longevity of seven females under simulated field conditions was 7.7 ± 4.2 d. Pteronarcella hadia emerged just after dusk and typically used willows, cottonwoods, and stream margin sedges as transformation sites. Males were observed actively searching the willows and drumming for females at night. 12 Great Basin Naturalist [Volume 55 E E S r Isogenoides zionensis if emergence — ' ~ oviposiiion (-(•

T Isoperia phalerata ■^ n = 25 4> f Y Isoperia qumquepunctata n = 229 III IV V VI VII VIII IX Months Fig. 12. Growth of Isoperia spp. n>iTiphs collected from the Rio Conejos, 1987-1990. and distribution of this order of aquatic insects in at least some portions of the southern Rocky Mountains. Responses to Altered Thermal Regime We became aware of substantial climatic differences (Fig. 2) between the two summers when adults were intensively studied. Though no water temperatures were available for 1989, air temperatures (Fig. 2) and hydrologic data (Fig. 3) suggested that the stream warmed more quickly and attained peak summer highs much earlier than in 1988. Consequently, development of several species was hurried, which narrowed the window of time adults were present streamside. At the assemblage level of organization, this trend is demonstrat- ed by the species richness pattern of Figure 4. The 1989 pattern was more peaked and great- ly truncated over that of 1988. Species-level responses can be demonstrated by inspection of the flight diagrams for each species. Nine of the 11 species presented in Table 2 show in- creased median emergence dates. Additionally, slopes produced by linear regression that were different between years (Table 3) were always higher in 1989. This result was consis- tent with a hurried nymphal development and shorter emergence period for each species. Life History Parameters Leuctridae Paraleiictra vershina. Harper (1973b) reports that most Leuctra ferruginea in an Ontario stream are semivoltine, but that some univol- tine individuals exist. Huryn and Wallace (1987) propose a 2-yr life cycle for a composite of Leuctra spp., most of which were probably c?9 Skwala americana ^ = 30 % of monthly catch ■ =3 individuals n = 227 Fig. 13. Growth of Skwala americana nymphs collected from the Rio Conejos, 1987-1990. L. ferruginea (Walker). Snellen and Stewart (1979) record univoltine fast cycles for Zealeuctra claasseni and Z. hitei in streams of north Texas. Additionally, Ernst and Stewart (1985a) report Leuctra tenuis as univoltine-fast in an Ouachita Mountain stream. Chloroperlidae Most Chloroperlidae exhibit a univoltine- slow or -fast growth pattern. Haploperla brevis (Banks) is widespread from Oklahoma to Quebec and west to Alberta, Canada. Ontario (Harper and Magnin 1969), Quebec (Harper et al. 1994), and Oklahoma (Ernst and Stewart 1985a) populations exhibited univoltine-fast growth with a 2-5-mo diapause, while Alberta populations were univoltine-slow (Barton 1980). European populations of Chloroperla tripunctata (Scopoli) (Elliott 1988), Siphono- perki torrentium (Pictet) (Elliott 1967), and S. hunneisteri (Pictet) (Benedetto 1973) also exhibited univoltine-slow growth. Species with semivoltine growth include Sweltsa onkos (Ricker) and possibly Utaperla gaspe- siana Harper and Roy (Haiper 1973a, Harper et al. 1994), S. mediana (Banks) (Cushman et al. 1977), and S. lateralis (Banks) (Huiyn and Wallace 1987). Paraperla frontalis. Stanford and Gaufin (1974) presented some evidence for semivol- tine growth of this species. Emergence for this species and for P. wilsoni Ricker occurs from May through July (Stewart and Stark 1988). Paraperlinae are radier robust chloroperlids that tend to be hx^Dorheic for most of their nymphal development. Their larger size, the more sta- ble stream temperatures in the hyporheic environment (Hendricks 1993), and the possi- blv low availabilitv of some nutrients in the 14 Great Basin Naturalist [Volume 55 Pteronarcella badia oviposition V VI VII VIII Months Fig. 14. Growth of Pteronarcella haclia inmphs collect- ed from the Rio Conejos, 1987-1990. hyporheic habitat (Stanford and Ward 1993) may have contributed to a preponderance of semivoltinism in this subfamily. Plumiperla diversa. Stewart et al. (1990) re- ported a univoltine-slow cycle for this species on the North Slope of Alaska. Emergence occurred from May through September, with recruitment of nymphs from a direct hatch in July. Growth occurred through the summer months with most nymphs attaining maximimi size before a winter quiescence. This assess- ment compared well with our limited data. Failure to collect adults was probably due to our infrequent sampling during their pre- sumed early May emergence. Suwallia pallidula and Suwallia wardi. No aspects of the life histories of either S. paUidiiki or S. wardi have been reported. The latter was recently described from a Colorado Front Range springbrook (Kondratieff and Kirchner 1990). It was one of the most abundant chloro- perlids in the Rio Conejos. This suggests that its ecological tolerance is wide and that it may soon be found in a variety of streams in the southern Rocky Mountains. Several explanations are possible for the heavily female-skewed sex ratio (13 cj:425 ?) of S. pallidula adults. The most probable is a combination of limited use of emergence traps coupled with an inaccessible microhabitat of adult males, probably high in the vegetation. Parthenogenesis may also be possible, but it is exceedingly rare in stoneflies. HaqDcr (1973a) reported that a few eggs of a perlid, Parag- netina media (Walker), hatched without fertil- ization. We did not attempt rearing of eggs from virgin females to check for parthenogen- esis in either Suwallia spp. These sex ratios are a perplexing problem, compounded by the Reronarcyidae = 20% of total catch M:F 9 /^f(s. Pteronarcys calif ornica 1988 23:32 <3 W"" M; - 1989 11:10 ^ Pteronarcella badia _ /v_:^»'^r>w„ 1988 134:79 ■•nr^r^ . nr^-j^^T^^ «. 1989 308:172 1 1 1 1 1 30 9 19 29 9 1 19 1 29 June July Fig. 15. Emergence of Pteronarcella badia and Pteronarcys californica from the Rio Conejos, 1988 and 1989. Polygons indicate daily relative proportion of total catch. fact that 0 6 : 657 ? of the closely related S. lineosa were caught during concurrent sam- pling on Massey Creek, a tributary of the Rio Conejos. Triznaka signata. Hassage and Stewart (1990) studied the widely distributed T. signata in the Rio Vallecitos of northern New Mexico. They reported a univoltine-slow growth pat- tern, with which we concur. No study of the emergence of this species has previously been published. Perlidae Claassenia sabulosa. Hassage and Stewart (1990) and Barton (1980) report a merovoltine (>2 yr) growth pattern for New Mexico and Alberta populations of this species. No egg batches from the Rio Conejos hatched in our laboratory, but this Colorado population showed some evidence of an extended hatch leading to cohort splitting (Stewart and Stark 1988). Eggs may undergo a temperature-dependent quies- cence as occurs in Dinocras cephalotes (Curtis) when fall temperatures decline to 8°C (Lille- hammer et al. 1989). Presence of first-instar nymphs in the fall, eyed eggs in October and November, and more first-instar nymphs in March supported this contention. Life histories have been reported for at least one species in every genus in the tribe Perlini, to which C. sabulosa belongs. All growth pat- terns involve 2-3 yr of development. Agnetina flavescens (Walsh), from an Ozark stream, exhibits a 2-yr life cycle, a short egg incuba- tion period, and an extended emergence period 1995] Stonefly Life Histories 15 §3 I — I = 20 % ol monthly catch emergence ■ = 1 individual i* Pteronarcys califomica ^^ oviposition VI VII Vlll IX X XI Fig. 16. Growth of Pteronarcys califomica nymphs col- lected from the Rio Conejos, 1987-1990. (Ernst and Stewart 1985b). Agnetina capitata (Pictet) was shown to have a 3-yr cycle, ex- tended emergence, and a 40-80-d egg incuba- tion period in Ontario (Harper 1973a). This range of incubation coupled with a long emer- gence promotes great differences in size of nymphs that ultimately prevents the separa- tion of cohorts and determination of voltinism. This was also a problem for C. sabulosa in the Rio Conejos. Perlodidae This family contains over 115 species (Stark et al. 1986, Stewart and Stark 1988) in the Nearctic. Although life histories of only 26 species are known, a clear trend toward uni- voltine-slow cycles occurs among the subfami- lies Isoperlinae and Perlodinae (Stewart and Stark 1988). Growth and emergence had not previously been studied for three of the seven perlodids in the Rio Conejos. These include /. zionensis, 1. quinquepunctata, and /. phalerata. Isogenoides zionensis. Few detailed life his- tory studies of the genus have been reported (Stewart and Stark 1988). Barton (1980) supect- ed semivoltinism for an Alberta population of I. cohibrinus, since two size classes of nymphs were collected in early May. Flannagan (1977) reported great body length variation in May for this species in another Alberta watershed but concluded a univoltine-slow cycle. Hilsenhoff and Billmeyer (1973) and Dosdall and Lehm- kuhl (1979) proposed univoltine growth pat- terns for the May-June-emerging /. frontalis in Wisconsin and Saskatchewan streams, respectively, based on samples taken a few months of the year. Semivoltinism, as reported for /. zionensis in the Rio Conejos, may also occur in its congeners, but this will be con- firmed only when detailed studies using small mesh nets, frequent sampling, and egg rearing have been conducted. Isoperla spp. Of the three Isoperla whose partial growth patterns are presented here, only 7. fulva has been previously reported. Hassage and Stewart (1990) reported a univol- tine-slow cycle, with a June emergence in the Rio Vallecitos of New Mexico. We concur with the New Mexico study. Our results agree well with reviews of Isoperla biology, summarized for 12 Nearctic species through 1987 (Stewart and Stark 1988). Ten species were univoltine- slow, while only two were univoltine-fast. In more recent literature Stewart et al. (1990) reported univoltine-slow growth for /. petersoni Needham & Christenson of Alaska. Additionally, Harper et al. (1994) added as univoltine-slow /. francesca Harper and 7. montana (Banks) from Quebec populations. These and our Rio Conejos work bring to 17 the Nearctic Isoperla species known to exhibit univoltine-slow cycles, while only three species appear to be univoltine-fast. Isoperla grammatica (Poda) and 7. difformis (Klapalek) (Malmqvist and Sjostrom 1989) and 7. obscura (Zetterstedt) studied by Ulfstrand (1968) are univoltine- slow in the Palearctic. Up to seven species of Isoperla commonly occur in streams in North America (Stewart and Stark 1988); conversely, in Scandinavia rarely more than two species occur simultane- ously (Malmqvist and Sjostrom 1989). Congen- erics of aquatic insects often partition resources along one or more resource gradients (Grant and Mackay 1969). Though only small numbers of adults were collected, a pattern of succes- sive emergence of 7. fulva, I. quinquepunctata, and 7. phalerata was clear in the Rio Conejos. Fifty percent cumulative catch dates for 7. fulva, I. phalerata, and 7. quinquepunctata were 22 June, 28 June, and 14 July, respectively, for 1988. These dates for 1989 were 18 June, 25 June, and 5 July. Temporal segregation brought about by a gradual change in domi- nance (lilies 1952) of these species may have accounted for the pres-ent coexistence of these stoneflies. Skwala americana. Two other studies re- ported univoltine-slow cycles with emergence from February through April for this species in northern New Mexico and central Colorado (Short and Ward 1980, Hassage and Stewart 16 Great Basin Naturalist [Volume 55 1990). Skwala curvata (Hanson) of" California also exhibited a univoltine-slow cycle, with emergence in April and May (Sheldon 1972). Other Arcynopterygini with univoltine-slow growth include Frisonia picticeps (Hanson) in California (Sheldon 1972), Megarcys signata (Hagen) in Utah (Cather and Gaufin 1975), and Perhnodes aurea (Smith) in California and Alberta (Radford and Hartland-Howe 1971, Sheldon 1972). Sheldon (1972) estimated average total fecundity of S. curvata to be near 1780 eggs for preemergent nymphs. This is much greater than that proposed for S. americana from the Rio Conejos. He used interocular width as an index to predict fecundity. Conversion of inter- ocular width to HCW likely involves a factor of 2X, which would make S. curvata the larger of the two stoneflies. This largely accounts for differences in fecundity. Mutch and Pritchard (1986) reported that S. americana (as S. paralle- la) had a warm, stenothermal egg development. Most species in this family have conserved the life history traits that Lillehammer et al. (1989) proposed as ancestral. These traits include uni\'oltine-slow cycles, temperature- dependent growth, and direct egg develop- ment. Isoperla quinquepunctata and /. zionen- sis have likely abandoned all of these except temperature-dependent growth. Pteronarcyidae Pteronarcella badia. Gaufin et al. (1972) reported that a 2-yr life cycle was possible for this species in Utah; however, S. Perry et al. (1987) and Stanford (1975) reported a univol- tine life history in Montana. No eggs were reared in either Montana study, and it is apparent from their growth histograms that early instars were missed entirely. Therefore, semivoltine life history is most probable throughout its range. Nymphs of this species are more likely to be found aggregated on filter paper leaf mod- els than alone (Hassage et al. 1988). We have also observed nymphs aggregating under mar- gin cobble immediately before emergence. Adults aggregate in leaf debris at the base of willow and cottonwood coppices at the Rio Conejos. This behavior may be attributable to the transformation and nighttime refuge sites being contagiously distributed. Hassage et al. (1988) also postulated that aggregation in P. badia lowers individual risk to predation. Pteronarcys californica. The egg diapause plus 38-mo nymphal life span places total life span of this population at 4 yr. This is one of the longest-lived aquatic insects known to occur in the Nearctic. Additionally, this species is perhaps our most synchronously emerging stonefly. Two- to 3-yr life cycles with a 9-10-mo egg diapause occur in other Pteronarcys such as P. dorsata (Barton 1980), P. proteus (Holdsworth 1941a, 1941b, W. Perry et al. 1987), and P scotti in the southern Appalachian Mountains (Folsom and Manuel 1983). However, Lechleit- ner and Kondratieff (1983) detailed a 1-yr life histoiy for P. dorsata in Virginia. Multiple-year life histories are common among larger species of the Pteronarcyidae (Stewart and Stark 1988). Accompanying this long nymphal growth, and perhaps contribut- ing to it, is another life history trait, long egg diapause. Univoltine growth patterns and direct egg development are ancestral patterns, while the semivoltine growth and diapause of P. badia and P. californica are derived traits (Lillehammer et al. 1989). Future studies of egg incubation in lower latitudes of North America will enable us to outline the range of responses of which Pteronarcys and Pteronar- cella are capable. Unanswered Questions Several largely unanswered questions per- sist about the life histories of stoneflies in and along the Rio Conejos. We have found that nymphs of many chloroperlids are not avail- able in surface sediments until just prior to emergence. They must be hyporheic in their habitat choice. Second, chloroperlids of the present study did not readily produce eggs in captivity, and those incubated never hatched. We can still ask many questions about their life histories. The answers would require a detailed study of the hyporheic habitat of an open-sediment stream like the Rio Conejos. This study should concentrate only on the chloroperlids, since they are generally abun- dant and diverse. Such a study would still fit within the comparative study approach of Sheldon (1972), but the guild would involve hyporheic chloroperlids. To settle the dilemma of aberrant sex ratios in this family, studies must concentrate on the presence of male nymphs in the stream. In this way the search for adult males whose 1995] Stonefly Life Histories 17 whereabouts are unknown need not take place, since both sexes of nymphs presumably enjoy a similar microhabitat. If no male nymphs are located, then incubation of eggs from virgin females should be conducted to confirm the possibility of parthenogenesis. An exciting observation we made during the study was that of basking in the sun of nearly all adults of summer-emerging species. Most displayed a remarkably consistent pat- tern of ascendence of riparian vegetation be- ginning at about 0800 h. Activity usually ceased by 1300 h when air temperatures were hottest. This ascendence culminated for females in egg batching and oviposition flights, while males used these riparian staging grounds for mate searching, drumming, and mating. Stoneflies should be investigated for potential to benefit from basking, an unreported phe- nomenon for Plecoptera. Acknowledgments We thank the Conejos Peak District of the U.S. Forest Sei"vice for providing lodging and laboratory space during the study. Special thanks go to J. B. Moring for his help with sample collection and D. Ziegler for providing some fecundity data for P. badia. This project was partially funded by faculty research funds of UNT and a National Science Foundation grant, BSR 8308422, to BCWS. Literature Cited Barton, D. R. 1980. Obsei-vations on the life histories and biology of Ephemeroptera and Plecoptera in north- eastern Alberta. Aquatic Insects 2: 97-111. Baumann, R. W, a. R. G.wfin, a.nd R. E Surdick. 1977. The stoneflies (Plecoptera) of the Rocky Mountains. Memoirs of the American Entomological Society 31. Benedetto, L. A. 1973. Growth of stonefly nymphs in Swedish Lapland. Entomologisk Tidskrift 94: 15-19. Gather, M. R., and A. R. Gaufin. 1975. Life history and ecology of Megarctjs signata (Plecoptera: Perlodidae), Mill Greek, Wasatch Mountains, Utah. Great Basin Naturalist 35: 39^8. GusHMAN, R. M., J. W. Elwood, and S. G. Hildebrand. 1977. Life history and production dynamics of A//o- perla mediana and Diplectrona modesta in Walker Branch, Tennessee. American Midland Naturalist 98: 354-364. Dosdall, L., and D. M. Lehmkuhl. 1979. Stoneflies (Plecoptera) of Saskatchewan. Quaestiones Entomo- logicae 15: 3-116. Elliott, J. M. 1967. The life histories and drifting of Plecoptera and Ephemeroptera in a Dartmoor stream. Journal of Animal Ecolog\' 36: 343-362. . 1988. Interspecific and intraspecific variation in egg hatching for British populations of the stoneflies Siphonoperla torrenfium and Chloroperla tripunciata (Plecoptera: Chloroperlidae). Freshwater Biology' 20: 11-18. Ern.st M. R., and K. W Stewart 1985a. Growth and drift of nine stonefly species (Plecoptera) in an Oklahoma Ozark footliills stream, and conformation to regression models. Annals of the Entomological Society of America 78: 635-646. . 1985b. Emergence patterns and an assessment of collecting methods for adult stoneflies (Plecoptera) in an Ozark foothills stream. Ganadian Journal of Zoology 63: 2962-2968. Flannagan, J. F 1977. Life cycles of some common aqua- tic insects of the Athabasca River, Alberta. Alberta Oil Sands Environmental Research Program, Report 11.20 pp. FoLSOM, T. G., and K. L. Manuel. 1983. The life cycle of Pteronarctjs scotti (Plecoptera: Pteronarcyidae) from the southern Appalachians, U.S.A. Atjuatic Insects 5: 227-232. Gaufin, A. R., W E. Ricker, M. Miner, R Milam, and R. A. Hays. 1972. The stoneflies (Plecoptera) of Montana. Transactions of the American Entomological Society 98: 1-161. Grant, P R., ,\nd R. J. Mackay. 1969. Ecological separa- tion of systematically related stream insects. Gana- dian Journal of Zoology 47: 691-694. Harper, P. P. 1973a. Emergence, reproduction, and growth of setipalpian Plecoptera in southern Ontario. Oikos 24: 94-107. . 1973b. Life histories of Nemouridae and Leuctri- dae in southeiTi Ontario (Plecoptera). Hydrobiologia 41:309-356. Harper, R P, and H. B. N. Hynes. 1972. Life histories of Capniidae and Taeniopteiygidae in southeiTi Ontario (Plecoptera). Archiv fiir Hydrobiologie, Supplement 40: 274-314. Harper, R R, and E. Magnin. 1969. Gycles vitaux de quelques Plecopteres des Laurentides (Insecta). Ganadian Journal of Zoology 47: 483-494. Harper, P R, M. Lauzon, and F. Harper. 1994. Life cycles of sundry stoneflies (Plecoptera) from Quebec. Review d'Entomologie Quebec 36: 28-41. Hassage, R. L., and K. W. Stewart. 1990. Growth and voltinism of five stonefly species in a New Me.xico mountain stream. Southwestern Naturalist 35: 130-134. Hassage, R. L., R. E. DeWalt, and K. W. Stewart 1988. Aggregation of Pteronarcella badia nymphs and effects of interaction with Claassenia sabulosa (Plecoptera). Oikos ,53: 37-40. Hendricks, S. R 1993. Microbial ecology of the hyporhe- ic zone: a perspective integrating hydrology and biology. Journal of the North American Benthologi- cal Society 12: 70-78. Hilsenhofe W. L., and S. J. Billmeyer. 1973. Perlo- didae (Plecoptera) of Wisconsin. Great Lakes Ento- mologist 6: 1-14. HoLDSWORTH, R. P 1941a. The life histor>' and growtli of Pteronarcys proteiis Newman (Pteronarcyidae: Plecoptera). Annals of the Entomological Society of America 34: 495-502. . 1941b. Additional information and a correction con- cerning the growth of Pteronarcys protean Newman (Pteronarcyidae: Plecoptera). Annals of the Entomo- logical Society of America 34: 714—715. 18 Great Basin Naturalist [Volume 55 HURYN, A. D., AND J. B. Wallace. 1987. The exopterygote insect community of a mountain stream in North Carohna, USA: life histories, production, and func- tional structure. Aquatic Insects 9: 229-251. Hynes, H. B. N. 1992. Some thoughts on unanswered questions about stoneflies. XI International Sympos- ium on Plecoptera, Tomahawk, Wl. IllIES, J. 1952. Die Plecoptcren luid das Monardsche Prinzip. Berlin Limnologische Fluhstation Freuden- thid 3: 53-69. Knight, A. W, and A. R. Gaufin. 1966. Altitudinal distribu- tion of stoneflies (Plecoptera) in a Rocky Mountain drainage system. Journal of the Kansas Entomological Society 39: 668-675. Kondr.at[eff, B. C, and R. K Kirchner. 1991. New Nearctic Chloroperlidae (Plecoptera). Journal of the New York Entomological Society 99: 199-203. Lechleitner, R. a., and B. C. Kondr.\tiefe 1983. The life history of Pteronarcijs dorsata (Say) (Plecoptera: Pteronarcvidae) in southwestern Virginia. Canadian Journal Zoology 61: 1981-1985. LiLLEHAMMER, A., J. E. BRITTAIN, S. J. SaLTVEIT, AND R S. Nielsen. 1989. Egg development, nymphal growth and life cycle strategies in Plecoptera. Holarctic Ecology 12: 173-186. Malmqvist, B., and R Sjostrom. 1989. The life cycle and growth of Isoperla grammatica and /. diffonnis (Plecoptera) in southernmost Sweden: intra- and interspecific considerations. Hydrobiologia 175: 97-108. Mutch, R. A., and G. Pritchard. 1986. Developmental rates of eggs of some Canadian stoneflies (Plecoptera) in relation to temperature. Journal of the North American Benthological Society 5: 272-277. Nelson, C. R., and R. W. Baumann. 1989. Systematics and distribution of the winter stonefly genus Capnia (Plecoptera: Capniidae) in North America. Great Basin Naturalist 49: 289-366. Perry, S. A., W. B. Perry, and J. A. Stanford. 1987. Effects of thermal regime on size, growth rates and emergence of two species of stoneflies (Plecoptera: Taeniopterygidae, Pteronarcyidae) in the Flathead River, Montana. American Midland Naturalist 117: 83-93. Perry, W. B., E. F. Benfield, S. A. Perry, and J. R. Webster. 1987. Energetics, growth, and production of a leaf-shredding stonefly in an Appalachian Mountain stream. Journal of the North American Benthological Society 6: 12-25. Petsch, H. E., Jr. 1987. Water resources data — Colorado water year 1986. Vol. 1. Missouri River Basin, Arkansas River Basin, and Rio Grande River Basin water data report CO-86-1. United States Geological Survey, Department of the Interior. . 1988. Water resources data — Colorado water year 1987. Vol. 1. Missouri River Basin, Arkansas River Basin, and Rio Grande River Basin water data report CO-87-1. United States Geological Survey, Department of the Interior. . 1989. Water resources data — Colorado water year 1988. Vol. 1. Missouri River Basin, Arkansas River Basin, and Rio Grande River Basin water data report CO-88-1. United States Geological Survey, Department of the Interior. . 1990. Water resources data — Colorado water year 1989. Vol. 1. Missouri River Basin, Arkansas River Basin, and Rio Grande River Basin water data report CO-S9-1. United States Geological Survey, Department of the Interior. Radford, D. S., and R. Hartland-Rowe. 1971. The life cycles of some stream insects (Ephemeroptera, Plecoptera) in Alberta. (Canadian Entomologist 103: 609-617. Sheldon, A. L. 1972. Comparative ecology o{ Arcijnoptenjx and Diura (Plecoptera) in a California stream. Archiv fiir Hydrobiologie 69: 521-546. Sheldon, A. L., and S. Jewett. 1967. Stonefly emergence in a Sierra Nevada stream. Pan-Pacific Entomologist 43: 1-8. Short, R. A., and J. V. Ward. 1980. Life cycle and produc- tion of Skwahi parallela (Frison) (Plecoptera: Perlo- didae) in a Colorado montane stream. Hydrobiologia 69: 273-275. Snellen, R. K., and K. W Stewart. 1979. The life cycle and drumming behavior of Zealeiictra claasseni (Frison) and Zealeuctra hitei Ricker and Ross (Plecoptera: Leuctridae) in Te.xas, U.S.A. Aquatic Insects 1: 65-89. Stanford, J. A. 1975. Ecological studies of Plecoptera in upper Flathead and Tobacco rivers, Montana. Un- published doctoral dissertation. University of Utah, Salt Lake City. Stanford, J. A., and A. R. Gaufin. 1974. Hyporheic com- munities of two Montana rivers. Science 185: 700-702. Stanford, J. A., and J. V. Ward. 1993. An ecosystem per- spective of alluvial rivers: connectivity and the hyporheic corridor. Journal of the North American Benthological Society 12: 49-60. Stark, B. P, S. W. Szczytko, and R. W Bau.mann. 1986. North American stoneflies (Plecoptera): systematics, distribution, and taxonomic references. Great Basin Naturalist 46: 383-397. Stewart, K. W, and B. P Stark. 1988. Nymphs of North American stonefly genera (Plecoptera). Entomological Society of America, Thomas Say Foundation 12: 1-460.' Stewart, K. W, R. L. Hassage, S. J. Holder, and M. W OswooD. 1990. Life cycles of six stonefly species (Plecoptera) in subarctic and arctic Alaska streams. Annals of the Entomological Society of America 83: 207-214. Szczytko, S. W, and K. W Stewart. 1979. The genus Isoperla (Plecoptera) of western North America: holomorpholog\' and systematics, and a new stonefly genus Cascadoperla. Memoirs of the American Entomological Society 31. Ulfstrand, S. 1968. Life cycles of benthic insects in Lapland streams (Ephemeroptera: Plecoptera: Trichoptera: Diptera-Simuliidae). Oikos 19: 167-190. Zar, J. H. 1984. Biostatistical analysis. 2nd edition. Prentice- Hall, Inc., Englewood Cliffs, NJ. 718 pp. Received 1 October 1993 Accepted 11 April 1994 Great Basin Naturalist 55(1), © 1995, pp. 19-28 POLLINATOR SHARING BY THREE SYMPATRIC MILKVETCHES, INCLUDING THE ENDANGERED SPECIES ASTRAGALUS MONTH S. M. Geerl-3, V J. Tepedino^-l, T. L. Griswold^, and W. R. Bowlinl Abstfl\ct. — Insects visiting flowers of the endangered Heliotrope milkvetch. Astragalus montii, were compared with those visiting two common sympatric congeners, A. kentrophyta and A. miser, on three sites on the Wasatch Plateau of central Utah for 2 yr. We recorded 27+ species of bees, most of which were uncommon, visiting the three species. All three species were primarily visited by native bees of the genera Osmia (15 species) and/or Bombus (4 species). Most Osmia species visited the three species oi Astragalus indiscriminantly; binnblebees preferred A. miser and avoided A. montii. Our hypodiesis that A. montii flowers would receive fewer total bee visits and be visited by fewer bee species than their common congeners was rejected: A. montii was intermediate to the two common species in its attractiveness to bees. Also rejected was our hypothesis that the greater similarity between A. montii and A. kentrophyta in flower size, flower moiphology, and microhabitat would be associated with greater similarity of flower visitors than either had with A. miser The data suggest that, rather than competing with each other for pollinators, the three species of Astragalus facilitate each other's visitation rates. Key words: Astragalus, milkvetch, endangered plant, reproduction, pollination, facilitation, bee diversity, conservation, Fabaceae, Osmia. Many insects such as dipterans and lepidop- terans use flowers only as fuel stations (Elton 1966); they collect nectar and burn it as they search for suitable spots to lay eggs. Such insects may merely pass through areas where flowers are sparse. Bees, in contrast, are central- place foragers (Orians and Pearson 1979) that must consistently reap profits in both nectar and pollen, for they forage not simply to under- wiite their own movements, but to provide food to rear their progeny as well (Stephen et al. 1969). Because bees are under strong, selective pressure to be profitable foragers, they are attracted to dense patches of flowers (Heinrich 1976, 1979, Thomson 1982). Bumblebees, for example, quickly recognize and exploit partic- ularly rewarding flower patches (Heinrich 1976, 1979); other bees probably do so also. Density-dependent foraging behavior by bees has important implications for certain rare plants. Rabinowitz (1981) distinguished seven types of rarity in plants using the following three criteria: (1) local abundance, (2) habitat specificity (narrow or wide), and (3) geographic range (large or small). Those species with both narrow habitat specificity and small local pop- ulations (regardless of geographic range) are sparse and likely to attract foraging bees only incidentally. We expect such species to be pol- linator-vulneiable and, therefore, to be highly self-compatible and perhaps primarily self- pollinating (Karron 1987). It is less clear whether plants in other categories of rarity, especially endemics (Rabinowitz 1981, Kruck- berg and Rabinowitz 1985), are also pollinator- vulnerable. Endemics have narrow habitat specificit)' but may be locally abundant. One such endemic, the rare Heliotrope milkvetch, Astragalus montii Welsh, is limited to a few isolated populations in limestone gravel outcrops on the Wasatch Plateau of central Utah at about 3350 m. There it grows with two common congeners, A. kentrophyta var. tegetarius (S. Wats.) Dorn, hereafter A. kentrophyta, and A. miser var. oblongifoliiis (Rydb.) Cron., hereafter A. miser In all three species, seed production requires, or is in- creased by, pollinator visits to flowers (Geer and Tepedino 1993). Information on the identity and biology of these pollinators is important, for A. jnontii occurs on rangelands that are grazed by domestic livestock and sprayed with insecticides to control grasshoppers. Successful management of this rare species requires ^Department of Biolog\. Utah State University, Logan, UT 84322-5305. 2USDA, ARS Bee Biology and Systematics Laboratory, Utah State University, Logan, UT 84322-5310. ^Present address: Wallowa Whitman National Forest, Highway 82, Box 88401, Enterprise, OR 97828. ^Author to whom correspondence should be addressed. 19 20 Great Basin Naturalist [Volume 55 knowledge of how such spraying may affect its pollinators. In this report we compared composition and abundance of pollinator fauna of A. montii with those of its two sympatric congeners. Because there may be wide variation in a species' pollinators between years and sites (Tepedino and Stanton 1981, Herrera 1990, Eckhart 1992), we censused pollinators of A. montii and its congeners for 2 yr at three sites. We hypothesized that A. montii would (1) attract fewer individual pollinators, (2) have lower pollinator species diversity than its two common congeners, and (3) share more species of flower visitors with A. kentrophyta than with A. miser because similarity in plant and flower size, flowering time, and microhabitat is greater with the former than with the latter. Species and Study Areas All three species of Astragalus are small perennial herbaceous legumes. A. montii is re- stricted to three mountaintops on the Wasatch Plateau in central Utah. Although Isely (1983) proposed that A. montii be reduced in status to a variety of A. limnocharis Barneby, it was listed as endangered under the Endangered Species Act in 1987 as A. montii and remains so (Anonymous 1991). Therefore, we refer to this taxon as A. montii. A. kentrophijta and A. miser are widespread species that occur with A. montii at three sites on two of the mountains; the third mountain is less accessible and was not included in the study. A. kentrophyta is widespread and abun- dant in the Rocky Mountains, mostly between 2280 and 3650 m. A. miser, one of the most common species of Astragalus in the Rocky Mountains, is locally abundant from sagebrush foothills to the spruce-fir belt (Barneby 1989). The three species co-occur at 3250 to 3350 m in an Engelmann spruce {Picea engelmannii Parry)/subalpine fir {Ahies lasiocarpa [Hook.] Nutt.) community. A. montii and A. kentrophyta are intenningled in limestone gravel outcrop- pings where A. miser is found only occasional- ly. A. miser is most abundant nearby where soil is deeper and less rocky. A. montii and A. miser occur at similar local densities on Heliotrope Mountain (9.3 ± O.l/m^ and 12.6 ± 8.3/m2; Geer unpublished data). There are fewer A. kentrophyta (2.6 ± O.S/m^; Geer un- published), but individuals cover more ground than do those of its congeners. The three species overlap in bloom time for about 3 wk (Fig. 1). Heliotrope milkvetch is a subacaulescent plant 1-5 cm tall that arises from a branched caudex. Flowers are deep purple with white wingtips. There may be a dozen to a hundred or more flowers (7.8 ± 1.5 mm long, N = 10; Geer unpublished) per plant, two to eight per raceme (Barneby 1989). It does not appear to reproduce vegetatively (personal observation). In 1989 and 1990 A. montii commenced flow- ering with final snowmelt beginning as early as June and continuing for about 4 wk until mid-July (Fig. 1). The common species A. kentrophyta started to flower approximately 1 wk before A. montii and continued to flower through early August. It is prostrate, with stems that fork repeatedly and closely to form low convex cushions cov- ered with small blue-white to puiplish flowers (6.6 ±1.2 mm long, N = 10; Geer unpublished), only two per raceme (Barneby 1989). The other common congener, A. miser, com- menced flowering 1-2 wk after A. montii and continued flowering until September. It is taller (2-20 cm) than A. montii or A. kentrophyta. Flowers are larger (11.4 ± 1.4 mm long, N = 11; Geer unpublished) and vaiy in number per raceme (3-15; Barneby 1989) and in color; flowers may be white, pink, or lavender. All Astragalus species have papilionaceous blossoms composed of a showy standard or banner petal, a keel that protects the joined stamens and pistil, and two wings that, along with the keel, typically serve as a landing plat- form (Kalin Arroyo 1981). To trip A. miser flowers, bees land on the keel and force their way under the banner (personal obsei-vation) as they do for other species of Astragalus (Green and Bohart 1975, Fliegri and van der Fiji 1979). Visitors to A. montii or A. kentrophyta spread the wing petals with their midlegs and take nectar, or comb pollen from the anthers to their abdominal pollen baskets with their forelegs (personal observation). Stylar hairs (termed a brush mechanism) aid in the collec- tion of pollen by transporting it from the keel outward (Kalin Arroyo 1981). Sexual reproduction by A. miser and A. kentrophyta requires insects to transfer pollen; A. montii is capable of unassisted self-pollina- tion (autogamy). However, fruits produced autogamously by A. montii may be inferior in 1995] Pollinators of Sympatric Milkvetches 21 A. kentrophyta wind, and no precipitation). Initially, sight identification of some taxa was attempted so as to reduce impact on the poHinator community. It soon became obvious that it was impossible to identify Osmia and other individuals with- out laboratory examination. Subsequently, all flower visitors were collected whenever possi- ble. Few insects other than bees visited the flowers. Diversity of bee visitors to each Astragalus 1 7(5 ^ 30 ^ ^ '^ To 20 species was calculated using Simpson's diver- J""^ J"iy A"g"=t sity index, D = 1 - Z^j^j (Pi)2, where Pj = the „. 1 Di J i f u • r proportion of individuals that belong to each I'lg. 1. Blooming dates tor tliree co-occumng species ot f" '^ ^ , Astragalus at the SSH site. Solid line = 1989; dashed line Dee species (Southwood 1978). Simpson S = 1990. index gives little weight to rare species and more weight to common ones. Similarity of ,.,,., 1 , , . the bee fauna visiting Astragalus species was qualitv to tliose produced by geitonogamous .. . i . /-. i ^ •' ■ i -^ ^ ' 1 1 11- .• 1 estimated using Lzekanowski s similarity or xenogamous hand pollinations, or open-pol- . j ^ ^,,7/ ii i \ i m • .^i ,. ^ J* ^ 1 ^ f 1. /M r index: C<. = NJ/(a+b + ...n), where N is the Imated control treatments (there are fewer , r i 1 • i • i t • J r -1. J J u r^ number oi plant species being compared, I is seeds per truit and seeds are smaller; Geer ^i i ri • i i i i J T- J- inno\ T-u 11 ..L • the number ot bee species shared by those and lepedino 1993). Ihus all three species , , . , , , \ 1111 r.. r . ,.■■...■ plant species, and a, b, etc., are the total num- probably beneht from insect visitation. f pV . ... , , ber or bee species visiting each plant species Methods (Southwood 1978). C^ is based on species presence alone. We also calculated Cj, which Insect visitors were collected for about 3 wk adjusts for the number of individuals per in 1989 and for 2 wk in 1990 at the following species (Southwood 1978). The indices range three sites, starting when A. montii was in peak ^om 0 (no similarity) to 1.0 (complete similari- bloom: the head of Mill Stream on Ferron ^v)- They were calculated between pairs of Mountain (HMS), south side of Heliotrope species and among all three species. Mountain (SSH), and east end of Heliotrope Probable pollinators of the three Astragalus Mountain (EEH). In 1990 collections from all species were ascertained by examining flower three Astragalus species were made only at visitors and recording areas of their bodies on the SSH site because only two insect collec- which pollen was found. Specimens were then tors were available instead of four, as in 1989. relaxed and pollen was removed using an We concentrated on the SSH site in 1990 to insect pin or by dabbing it with acid-fuchsin make the number of collector hours there gel (Beattie 1971). The pollen was placed on a equivalent to the 1989 effort. In 1990 visitors glass slide with acid-fuchsin gel, warmed until to A. kentrophyta were collected at the SSH liquid, and a cover slip applied (modified from and HMS sites, and visitors to A. miser were Faegri and Iverson 1964). One slide per leg or collected at the SSH and EEH sites. Following two slides per abdomen were made for each are approximate direct distances between sites: insect. All slides were viewed at lOOX magni- HMS to SSH = 3.6 km, HMS to EEH = 2.4 fication and the pollen compared to a pollen km, and EEH to SSH = 1.2 km. reference collection of species in bloom at the Pollinators were collected with a standard study sites, butterfly net and killed in cyanide jars. Cold temperatures, strong winds, and frequent pre- Results cipitation (snow and rain) prohibited pollina- tors from flying during all but brief windows Bees were scarce at the study sites in both of calm, sunny weather, so opportunistic collec- years (Table 1, Appendices I, II). Bee visitors tion was necessary to ensure an adequate sam- per plant species ranged from about 0.5 to just pie size. Collections were made from all three over 3 per hour, a small number considering species contemporaneously, whenever weath- that many flowers of each species were being er permitted (i.e., temperatures >13°C, little monitored. Bee numbers were higher in 1990; 22 Gkeat Basin Naturalist [Volume 55 Tablf. 1. Number of person hours spent collecting and number ot bee individuals collected or observed visiting flow- ers of Astragalus montii (Asmo), A. kentrophyta (Aske), and A. miser (Asmi) at three sites on the Wasatch Plateau in 1989 and 1990. SSH, EEH = south and east side Heliotrope Mountain, respectively; HMS = head of Mill Stream, Perron Mountain. SSH EEH HMS Asmo Aske Asmi Asmo Aske Asmi Asmo Aske Asmi 1989 Hours 24 8 10 30 24 22 8 16 16 Individuals 28 9 10 30 19 10 5 11 18 Individuals/hour 1.2 1.1 1.0 1.0 0.8 0.5 0.6 0.7 1.1 Species 7 3 5 7 7 3 3 4 / 1990 Hours 30 15 15 12 12 — 12 — 12 Individuals 57 7 35 40 24 — 16 — 24 Individuals/liour 1.9 0.5 2.3 3.3 2.0 — 1.3 — 2.0 Species 10 5 11 5 5 — 6 — 3 when categorized by site and Astragalus species visited, six of seven categories had more indi- viduals per hour in 1990 than in 1989. The initial hypothesis, that A. montii would have fewer individual flower visitors than would its common congeners, received little support (Table 1, Appendices I, II). In 1989 there was little difference among species in visitors per person hour at SSH. At EEH A. montii flowers were visited more often than the other species. Conversely, at HMS A. montii flowers received the fewest visits. In 1990 comparisons of number of visitors among all three Astragalus species could be made only at the SSH site where A. montii had an intermediate number of visitors per hour. At EEH, A. montii again had more visits per hour than A. kentrophyta, and at HMS it had fewer visits per hour than A. miser The prediction that species richness and species diversity of bees visiting the three Astragalus species would be lowest for A. montii was also provisionally rejected. The number of species captured on A. montii commonly ex- ceeded those captured on the other species, both when more hours were spent collecting from A. montii than the other species (1989 SSH) and when collecting hours were equal (1990 HMS; Table 1). Only once, when fewer hours were spent collecting on A. tnontii than on the other Astragalus species (1989 HMS), was A. montii visited by the fewest species of bees. When all sites were considered, total number of species collected on A. montii in 1989 exceeded those captured on A. kentro- phyta and equaled those captured on A. miser (Table 2). In 1990 more species were caught visiting A. montii than the other two species, but this difference is probably because we col- lected at three sites for A. montii but at only two for each of the other two species. Calculations using species diversity, D', also failed to yield expected trends (Table 2). In 1989 diversity of visitors to flowers of A. montii was very similar to diversity recorded for A. kentrophyta and A. miser Comparisons for 1990 are more tenuous because of the differ- ences among species in number of sites sam- pled. However, diversity of flower visitors was highest for A. miser and similar for A. montii and A. kentrophyta. Diversity in 1990 was generally lower than in 1989, although num- ber of individuals captured was greater. The most frequent visitors to these Astra- galus species in both 1989 and 1990 were Osmia bees (Table 3). For the small-flowered A. montii and A. kentrophyta, in both years >70.0% of all visitors were Osmia bees. Only for A. miser in 1990 did the percent Osmia visitors drop below 50%. A. miser was more frequently visited by bumblebees, especially at SSH. The abundance of bumblebees caused SSH to have the lowest percentage of Osmia individuals recorded at any site in both years. Even so, Osmia bees were always more than 60% of the total flower visitor fauna recorded in any site -year. Because of greater similarities in flower size, color, and microclimate, we expected A. montii and A. kentrophyta to have more visitors in common than either did with A. miser This was not true in either year. The three pairings of Astragalus did not differ much in the num- ber of bee species they shared, though results 1995] Pollinators of Sympatric Milkvetches 23 Table 2. Number of individuals, number of species, and species diversity (D) of bees found visiting three species of Astragalus at three sites on the Wasatch Plateau. In 1989 collections were made for each species at all three sites; in 1990 collections were made at all sites for A. montii, but at only two sites for the other two species. For comparative pur- poses, collection data for the latter two species are shown in 1989 for all three sites and for only the two sites collected at in 1990. D = Simpson's diversity index. Astragalus Individuals .Species D' species 3 sites 2 sites 3 sites 2 sites 3 sites 2 sites 1989 montii 63 — 13 — 0.87 kentrophyta 39 28 9 8 0.79 0.81 miser 38 28 13 11 0.88 0.87 1990 montii 113 — 13 — 0.62 kentroplujta — 31 — 7 — 0.60 miser — 59 — 12 — 0.79 'In 1989 onK- indiviuals that were collected were used in calculations, because uncaptured Otmia indi\iduals were not identifiable to species. Table 3. Percent visitors that were Osmia bees to the flowers of three Astragalus species (abbreviations as in Table 1). Data showai grouped by species across sites, and by site across species, for 2 yr. For comparative purposes, 1989 data are shown in entirety' (3 sites or 3 species) or only for the 2 sites or 2 species sampled in 1990. Asmo Aske Asmi SSH EEH HMS — Across sites - - - - - Across species - - 1989 3 (sites/species) 88.9 71.8 73.7 62.3 88.1 76.5 2 (sites/species) — 78.6 64.3 — 85.7 87.0 1990 93.8 74.2 47.5 62.6 87.5 95.0 varied somewhat with year and with index used (Table 4). In 1989 the three pairings of Astraga- lus species had about the same number of bee species in common. In 1990 A. miser and A. montii had about twice the number of species in common as did the other pairings. Neither coefficient of similarity, C^ or Cj, consistently supported the hypothesis; in 1989, but not 1990, C^ and Cj were highest for the A. mon- tii-A. kentrophyta comparison. Many bees visiting Astragalus flowers car- ried pollen on their bodies: 43% of the bees captured, primarily females of the genus Osmia, had been collecting pollen. Pollen loads com- prised primarily Astragalus pollen (all means >80%; Table 5). It is unknown whether loads commonly contained more than one species of Astragalus because pollen grains could not be distinguished to species with the light micro- scope. Our observations of foraging bees suggest some interspecific movement. In 1989 few Osmia individuals flew between A. montii and A. miser or A. kentrophyta; of 74 interplant movements only two were interspecific. In 1990, 4 of 21 observed interplant movements were between species. Interspecific visits occurred most commonly where species grew intermingled. Discussion Two hypotheses make predictions about the abundance and diversity of visitors to the flow- ers of rare plants. For entomophilous plants, Levin and Anderson (1970), Straw (1972), and Karron (1987) proposed that pollinators should be more flower constant to abundant plant species than to rare ones, that this differential flower constancy would result in more suc- cessful reproduction by "majority" species than by "minority" species, and that over time minority species would become extinct because of dwindling recruitment or would evolve some method of self-reproduction (Levin 1972). A corollar)' of this hypothesis is that both the number and diversity of visitors to the flowers of rare plants should be lower than they are to abundant ones. 24 Great Basin Naturalist [Volume 5." Table 4. Number of Ix-e species (S) collected on each A.str(i^alii.s species, and number of species shared (C) and simi- larity' indices for each j^airin^ for each year. C^ = Czekanowski s similarity index for bee species presence-absence; Cj = index weighted by individuals captured. Astragalus species pair 1989 1990 montii 13 miser 13 iiioiitii 13 kcntrophijta 9 kentrophyta 13 miser 9 All three species 20 0.50 0.55 0.45 0.35 0.34 0.43 0.43 0.27 13 12 13 7 12 21 0.56 0.37 0.40 0.35 0.32 0.53 0.28 0.30 In conticLst, the facilitation hypothesis (re- viewed by Rathcke 1983) predicts that rare species growing with attractive, more abun- dant species may actually reproduce more successfully because the latter draw many more pollinating insects into the area than would otherwise be present. If so, rare and abiuidant sympatric species should have simi- lar visitor diversity, and visitor abundances should reflect respective frequencies of the plants. This study indirectly assessed the importance of facilitation and competition. A direct assessment is difficult because (1) the experiments necessar>' to distinguish between alternatives cannot be conducted when the "plant protagonist" is protected by the Endan- gered Species Act; and (2) A. montii did not occur in the absence of its congeners on our study sites, so visitation rates of "facilitated" and "unfacilitated" populations could not be compared. Our results supply consistent, though indi- rect, support for the facilitation hypothesis. Except for bumblebees, which foraged almost exclusively from large-flowered A. miser, bees did not discriminate against A. montii but rather seemed to treat all three Astragalus species as one taxa. First, A. montii did not consistently attract fewer visitors per hour than did the other species. Indeed, visitation rates to A. montii were higher than to the other species in three of six site-years (Table 1). Second, neither species richness nor species diversity of pollinators was consistently lower for A. montii than for the other species (Table 2). In fact, an ecjual or greater nimiber of species visited A. montii than visited the others in both years. And finally, bees were observed moving between species on individual foraging trips. Gross (1992) also reported that bees for- aging on closely related legumes commonly moved between species. Thus, there was no detectable rare species disadvantage and no evidence that endemics, at least those growing in close proximity to abundant congeners, are pollinator- vulnerable. The shared microhabitat and similarities in flower size and morphology of A. montii and A. kentrophyta led us to expect that facilitation would be more likely between these two species and, therefore, that they would have more visi- tors in common than either would with A. miser For example, Thomson (1978, 1981, 1982) found that, in two-species mixtures, the degree of intermingling and the similarity in structure and appearance of congeners' flowers deter- mined the importance of competition and mutualism. The more similar the flowers, the more likely that visitation rates to rare species would be bolstered by the presence of abun- dant species and the more likely that visitors would be shared. Our data supported this expectation for 1989 but not for 1990 (Table 4). In 1990 C^ for the A. montii-A. kentrophyta comparison was intermediate to the other comparisons; for C; it was lower than the other comparisons. Thus, results for the similarity analyses also tend to support the hypothesis that most bees do not distinguish among these Astragahis species when foraging, and that the Astragalus species tend to facilitate each other's visitation rates. Only bumblebees seem uninfluenced by Astragalus flowers in the aggregate. They clearly preferred flowers of A. miser and avoided those of the other Astragalus species. Flowers of A. miser are large, probably more rewarding, and provide a landing platfonii from 1995] Pollinators of Sympatric Milkvetches 25 Table 5. Percent Astragalus pollen grains in pollen loads, and location of pollen loads carried by bees collected on three Astragalus species at three sites on the Wasatch Plateau in 1989 and 1990. Astragalus species montii kentrophyta Number po'len lo •of ads Mean % Astragalus pollen ( + SE) Location of pollen Abdomen Legs 45 82 ±4 42 3 19 90 ± 1 19 — 5 95 ± 1 5 — which large, energy-demanding bumblebees can readily forage. Other large-flowered Astra- galus species also attract numerous large bees such as bumblebees {Bombiis spp.) and antho- phorids (Green and Bohart 1975, Sugden 1985, Karron 1987). In comparison, bumblebees seemed unable to land on the small, weakly supported A. montii flowers which are borne above the foliage; they did occasionally exploit the tiny A. kentrophyta blossoms while perched on the foliage of that cushion plant. Factors other than flower abundance can influence the flight path of foraging bees. Because bees are central-place foragers (Orians and Pearson 1979), travel time and energy expended between flower patches and nest are also important. Thus, bees may patronize a flower patch because of its proximity to their nest, even though flowers are more abundant elsewhere. For example, Osniia bees mated and nested at the sheltered EEH site where relatively few A. kentrophyta or A. miser plants grew; the population of A. montii was small but dense. Nevertheless, bees visited flowers at least as frequently at EEH as at the other, more flower-rich, sites (Tables 1, 2). Thus, suit- ability of nesting habitat at EEH, rather than Astragalus flower abundance, may best account for the abundance of bees there. The effect of wild bee nesting sites on seed production of surrounding vegetation is poorly studied and warrants additional attention. Rigorous subalpine communities of the Wasatch Plateau, with frequent high winds, thunderstorms, and below-freezing tempera- tures during the blooming season, support a surprisingly rich bee fauna. In 2 yr we collected 27+ bee species foraging on Astragalus flow- ers during 2-3 wk (Appendices I, II). These bees are invaluable pollinators of native plants both rare and common. Their welfare must also be considered in management plans for rare plants. Land managers must eliminate losses of bees to insecticide applications made for rangeland grasshoppers and minimize physical damage to nest sites. The present insecticide- free buffer zone (currently 4.8 km) around rare plant populations should continue to be main- tained. Areas where bees nest in soil should also be protected from livestock trampling, off-road vehicle use, and foot traffic (Sugden 1985). Such diversity, comparable to or greater than that of other subalpine areas in North America (Moldenke and Lincoln 1979), is to be marveled at and preserved. Acknowledgments We are grateful to the many people who assisted in this study. Etta Sechrest and Mike Cram were reliable field and laboratory assis- tants. John Healey, Don Riddle, and Bob Thompson of the U.S. Forest Service and Lairy England, U.S. Fish and Wildlife Service, helped in a variety of ways, fi^om locating plant popu- lations to putting a roof over our heads. The manuscript was constructively reviewed by M. Barkworth, K. Harper, and E. Sugden. This study was funded as part of the APHIS Grass- hopper IPM Project. It is Journal Paper #4436 from the Utah Agricultural Experiment Station. Literature Cited Anonymous. 1991. Endangered and threatened wildlife and plants; .50 CFR 17.11 and 17.12, July 15, 1991. Publication unit, U.S. Fish and Wildlife Service, Washington, DC. Barneby, R. C. 1989. Fabales. Pages 12-167 in A. Cronquist, A. H. Holmgren, N. H. Holmgren, J. L. Reveal, and P K. Holmgren, eds., Intennountain flora 3, Part B. New York Botanic Garden, Bronx, NY. Beattie, a. J. 1971. A technique for the study of insect- borne pollen. Pan-Pacific Entomologist 47: 82. EcKHART, V M. 1992. Spatio-temporal variation in abun- dance and variation in foraging behavior of the polli- nators of gynodioecious Phacelia linearis (Hydro- phyllaceae).' Oikos 64: 573-586. 26 Great Basin Naturalist [Volume 55 Elton, C. S. 1966. The pattern of animal eoiiiniunities. Methuen, London. 432 pp. F.\EGRI, K., AND J. IvERSON. 1964. Textbook of pollen analysis. Hafrier Co., NY. 237 pp. Faegri, K., and L. VAN DER FiJL. 1979. The principles of pollination ecology. 3rd revised edition. Perganion Press, Oxford. 244 pp. Geer, S. M., and V. J. Tepedino. 1993. Breeding .systems of the rare Heliotrope milkvetch {Astragalus montii Welsh: Fahaceae) and two conmion congeners. Pages 334-344 ill R. Sivinski and K. Lightfoot, eds., Proceed- ings of the Southwestern Rare and Endangered Plant Conference. New Mexico Forestry and Resources Conservation Division, Santa Fe. Green, T. W., and G. E. Bohart. 1975. The pollination ecology of Astragalus cibarius and Astragalus uta- hensis (Leguminosae). American Journal of Botany 62: 379-386. Gross, C. L. 1992. Floral traits and pollinator constancy: foraging by native bees among three sympatric legumes. Australian Journal of Ecology 17: 67-74. Heinrich, B. 1976. Foraging specializations of individual bumblebees. Ecological Monographs 46: 105-128. . 1979. "Majoring" and "minoring" by foraging bumblebees, Bombus vagans: an experimental analy- sis. Ecology 60: 245-255. Herrera, C. 1990. Daily patterns of pollinator activity, differential pollinating effectiveness, and floral re- source availability, in a summer-flowering Mediter- ranean shrub. Oikos 58: 277-288. Isely, D. 1983. New combinations and two new varieties in Astragalus, Orophaca, and Oxytropis (Legmnino- sae). Systematic Botany 8: 422. Kalin Arroyo, M. T. 1981. Breeding systems and pollina- tion biology in Leguminosae. Pages 723-769 //; R. M. Polhill and P H. Raven, eds.. Advances in legume systematics. Part 2. Royal Botanic Gardens, Kew, UK. Karron, J. D. 1987. The pollination ecology of co-occur- ring geographically restricted and widespread species of Astragalus (Fabaceae). Biological Consei-vation 39: 179-193. Kruckeberg, A. R., and D. Rabinowitz. 1985. Biological aspects of endemism in higher plants. Annual Review of Ecolog\' and Systematics. 16: 447-479. Levin, D. A. 1972. Competition for pollinator sei-vice: a stimulus for the evolution of autogamy. Evolution 26: 668-674. Levin, D. A., and W. VV. Anderson. 1970. Competition for pollinators between simultaneously flowering species. American Naturalist 104; 455-467. Moldenke, a. R., and P Lingoln. 1979. Pollination ecol- ogy in montane Colorado: a community analysis. Phytologia 42: 349-379. Orians, G. H., and N. E. Pearson. 1979. On the theory of central place foraging. Pages 155-177 in D. J. Horn, G. R. Stairs, and R. D. Mitchell, eds., Analysis of ecological systems. Ohio State University Press, Columbus. Rabinowitz, D. 1981. Seven fomis of rarity. Pages 20.5-218 in H. Synge, ed.. The biological aspects of rare plant conservation. Wiley, New York, NY. R.ATHGKE, B. 1983. Competition and facilitation among plants for pollination. Pages 305-329 in L. A. Real, ed., Pollination biology. Academic Press, New York, NY Southwood, T. R. E. 1978. Ecological methods. 2nd edi- tion. Chapman and Hall, London. 524 pp. Stephen, W. P, G. E. Bohart, and P F Torchio. 1969. The biology and external moiphology of bees. Agricultural Experiment Station, Oregon State University, Coi"vallis. 140 pp. Str.\w, R. M. 1972. A Markov model for pollinator con- stancy and competition. American Naturalist 106: 597-620. Sugden, E. A. 1985. Pollinators of Astragalus monoensis Barneby (Fabaceae): new host records; potential impact of sheep grazing. Great Basin Naturalist 45: 299-312. Tepedino, V. J., and N. L. Stanton. 1981. Diversity and competition in bee-plant communities on short-grass prairie. Oikos 36: 35-44. Thomson, J. D. 1978. Effect of stand composition of insect visitation on two-species mixtures of Hieracium. American Midland Naturalist 100: 431-440. . 1981. Field measures of constancy in bumble- bees. American Midland Naturalist 105: 377-380. . 1982. Patterns of visitation by animal pollinators. Oikos 39: 241-250. Received 29 April 1993 Accepted 2 June 1994 Appendi.x I. Species of bees collected and obsei-ved visiting flowers of A. montii (Asmo), A. miser (Asmi), or A. kentro- phyta (Aske) at three sites in 1989. Entries represent number of males/females collected. Obsei-vations are in parenthe- ses. Site abbreviations as in Table 1. SSH 1-21 June EEH 14-25 June HMS 1-J^22 June Bee species Asmo Aske Asmi Asmo Aske Asmi Asmo Aske Asmi Andrenidae Andrena transnigra Vier. Andrena spp. 0/1 (1) 1995] Pollinators of Sympatric Milkvetches 27 Appendix I. Continued. SSH 1-21 June EEH 14-25 June HMS 14-22 June Bee species Asino Aske Asmi Asnio Aske Asmi Asnio Aske Asmi API DAE Boinbiis hifarius Cr. Boiiibiis Jlavifrons Cr. Botnbtts hiintii Greene Boinbiis nevadensis Cr 0/1 (0/2) (0/2) 0/1 0/2 Halictidae Evtjlaeus niger (Viereck) Megachilidae Anthidium temiiflorae Ckll. Megachile spp. Osmia cijanopoda Ckll. Osinia htirdii White Osinia longula Cr. Osmia nigrifrons Cr. Osinia atf. nigrifrons Osinia paradisica Sanh. Osinia penstcinonis Ckll. Osinia pikei Ckll. Osmia piisilla Cr Osmia sladeni Sanh. Osmia sladeni &/or alpestris Osmia taniwri Sanh. Osmia spp. 0/1 (2)1/0 (1/0) 0/1 0/3 0/1 (1/0) (1) (1)1/2(1) 1/2(2) 1/0 0/1 0/4 1/3 (5/9) 2/0 (1/4) (0/3) 0/2 0/1 0/3 0/1 0/1 1/0 2/2 0/1 0/1 0/1 1/0 1/0 4/0 3/0 2/0 0/2 0/3 0/2 0/1 0/2 0/5 1/2 0/1 0/1 (8/9) (1/1) (1/3) (1/1) (0/1) (0/1) Appendix H. Species of bees collected and observed visiting flowers of A. montii (Asmo) at three sites and A. miser (Asmi) and A. kentrophijta (Aske) at two sites each in 1990. Entries represent number of males/females collected. Observations are in parentheses. Site abbreviations as in Table 1. SSH 19 June-4 July EEH 19-29 June HMS 21-29 June Visitor Asmo Aske Asmi Asmo Aske Asmo Asmi Andrenidae Andrena nigrihirta (Ashm) Andrena transnigra Vier. 0/1 0/1 Apidae Apis mellifera L. Bombus hifarius Cr Bombus flavifrons Cr Bombus hunt a Greene Bombus nevadensis Cr. 0/1 0/1 0/1(5) 0/1 0/3(6) 0/3(5) Megachilidae Anthidium temiiflorae Ckll. 1/0(2) Hoplitis fidgida Cr. Megacile melanophaea Smith Megachile perihirta Ckll. 1/0 2/0 2/0 1/0 0/1(3) 3/0 1/0 1/0 28 Great Basin Naturalist [Volume 55 Appf.ndix II. ("ontinued. SSH EEH HMS 19 J une-4 July 19- -29 June 21- -29 June Visitor Asmo Askc Asmi Asmo Aske Asmo Asmi Megachilidae (continued) Osmia lon^ida Cr. 2/0 Osmia montana Cr. 1/0 Osmia afF. nig.n,frons 0/1 0/1 Osmia paradisica Sanh. 1/0 0/2 3/0 1/2 1/0 Osmia penstemonis Ckll. 0/1 Osmia pusilla Cr. 0/1 Osmia sculleni Ckll. 2/0 1/0 Osmia sladeni Sanh 19/13 4/0 1/0 8/16 7/8 3/6 1/21 Osmia suhawstralis Ckll. 4/0 1/0 Osmia tanneri Sanh. 9/2 1/0 1/0 9/1 1/0 0/1 Great Basin Naturalist 55(1), © 1995, pp. 29-36 FACTORS AFFECTING SELECTION OF WINTER FOOD AND ROOSTING RESOURCES BY PORCUPINES IN UTAH Dave Strickland -2, Jerran T. Flinders 1'3^ and Rex G. Gates ^ Abstract. — Ecological and phytochemical factors potentially affecting winter dietar)' discrimination by porcupines {Erethizon dorsatiim) in tlie mountain brush zone of Utah were studied. Porcupines utilized gaml^el Oiik {Quercits gamhelii) as their primary winter food and roosting resource. Big-tooth maple [Acer grandidentatum) was the most conmion tree species in the study area but was rarely utilized by porcupines. Conifer species were used as a food and roosting resource significantly less often than they occurred in the study area, despite themial advantages provided by their rela- tively dense canopies. Oak feed trees were successfully separated from conifer feed trees by discriminant analysis 100% of the time. Oak trees were correctly classified as feed and nonfeed trees 71% of the time. Gambel oak contained higher amounts of crude protein, fiber, and tannins, but was lower in ether extract fractions and fatty acid content than conifers. A layer of adipose tissue used as an energy reserve by porcupines may have relaxed energy intake demands sufficiendy to permit them to concentrate on a diet of oak tissue, which is high in protein, rather than a high-fat conifer diet. A diet relatively high in protein may have facilitated digestion of food material high in fiber. Temperature did not affect selec- tion of tree species used for roosting. Rock and snow caves were utilized infrequently and the study population ranged widely. Three of 15 study animals were eaten by predators. Keij words: porcupine, Erethizon dorsatum, gambel oak, Quercus gambelii, dietary selection, mountain brush predation. '.one. Porcupines {Erethizon dorsatum) roost and feed in canopies of deciduous trees and shrubs for extended periods during winter in much of western North America (Oveson 1983, Craig and Keller 1986, Sweitzer and Berger 1992). Apparent localized interspecific and intra- specific preferences for food and shelter resources by porcupines imply that chemical and/or physical advantages are available to them. Further, since snow caves, rock dens, and cover in canopies of coniferous tree species likely offer increased thermal advan- tages in the form of energy savings to porcu- pines (Clarke and Brander 1973, Roze 1987, 1989), their dependence on a deciduous food and roosting resource (which does not offer those advantages) further strengthens the implication that chemical and/or physical selective advantages are realized by dietary selection. Predator avoidance may also be an important force in food and roost tree selec- tion. The objective of this research was to investigate physical, phytochemical, and eco- logical agents involved in selection of gambel oak by porcupines in south central Utah. Study Area The study was conducted in the mountain brush zone near the mouth of Spanish Fork Canyon in north central Utah. Elevations at the study site range from 1650 to 2075 m. The general exposure is northern, and terrain is steep. Overstory woody vegetation is dominated by gambel oak {Quercus gambelii) and big- toodi maple {Acer grandidentatum). Aspen {Pop- ulus tremidoides), chokecherry {Primus virgini- ana), Douglas fir {Pseudostuga menziesii), white fir {Abies concolor), and mountain maple {Acer glabrum) are also represented in the woody flora. The climate in Spanish Fork Canyon during the winter of 1984-85 was not atypical. Data from the Spanish Fork U.S. Climatological Station, located approximately 5.5 km from the study site, indicate that temperatures were slightly colder and precipitation was slightly higher than average (U.S. Climatological Data for Utah 1984-85). Coyote {Canis latrans) and mountain lion {Felis concolor) tracks were fre- quently encountered in the study area. Private access into the study area allowed observation ^Department of Botany and Range Science, Brigham Young University, Provo, UT 84602. ^Present address: USDA Forest Service, Pleasant Grove, UT 84062. ■'Address correspondence and reprint requests to this author 29 30 Great Basin Naturalist [Volume 55 of a porcupine population relatively free from human disturbance. Methods Fieldwork We conducted fieldwork from late Decem- ber 1984 through April 1985, at which time the study population had shifted from a diet of inner bark (phloem and cambium) of woody vegetation to herbaceous vegetation. The study area was systematically searched by researchers on snowshoes. Study animals were captured by hand, usually while they were still in tree canopies. This was accomplished by grasping distal guard hairs at the posterior end of the tail between thumb and forefinger and pulling the tail taut. The captured animal was then secured by grasping the tail with the free hand using a backward stroking motion to flatten the quills. Fifteen porcupines, 10 females and 5 males, were instrumented with radio transmit- ter collars (Telonics, Inc.). Animals were located daily by triangulation, and visual sightings were made on each animal approximately weekly. Percent occurrence of woody species was calculated from point-quarter measurements using the feed/roost tree as the center point (Cottam and Curtis 1956). Percent occurrence of woody species vs. percent utilization of each feed tree species was compared using chi- square analysis to test whether feed tree selec- tion was random. Diameter at breast height (dbh), species, and distance from the feed tree center point were recorded for the nearest woody stem in each quadrant. Point-quarter measurements were repeated using the near- est neighbor nonfeed tree of the same species as the center point. Tissues from feed and nonfeed trees were collected to investigate possible differences in chemical makeup. Tissue samples from feed trees were collected where fresh bark removal indicated the roost- ing animal had foraged. Samples from nearest neighbor nonfeed trees were taken from branches at the same height and with a diame- ter similar to those from corresponding feed trees. Bark samples were frozen and analyzed for dietary components. Results from those analyses reasonably approximated values reported for gambel oak (Smith 1957, Kufeld et al. 1981, Welch 1989). Location, slope, aspect, snow depth, and climatic conditions were recorded at each feed tree site. High and low temperature readings were taken dail>' at an elevation of 1597 m, as well as from the Spanish Fork climatological station. Laboratory and Statistical Methods Tissues from feed and nonfeed trees were analyzed for protein and phosphorus using the auto analyzer semiautomated method #12 for feeds (Horwitz 1980). Calcium, magnesium, potassium, and sodium content were deter- mined by the atomic absorption method #2 for plants (Horwitz 1980). Sulphur content was determined by a wet-ash process using nitric and perchloric acid. Crude fiber was determined by the acid detergent fiber and lignin #21 method (Hoi-witz 1980). An evalua- tion of crude fat was made using the direct method (Hoi-witz 1980) on a Lab Con soxlet extractor. A limited number of tissue samples were analyzed on a Hewlett Packard model 5995 gas chromatograph/mass spectrometer (GCMS) for fatty acids and terpenes. Tannin content was measured by the radial diffusion method (Hagerman 1987) with quebracho tan- nin being the standard, and by astringency (Gambliel et al. 1985). Soluble carbohydrates were determined according to daSilveira (1978). Urine samples of captive porcupines on a strict diet of gambel oak were analyzed for calcium and phosphorus content when lab- oratory results indicated the Ca/P ratio in the tissue of food materials was greater than ex- pected. Eight oak tissue samples were chosen at random and retested for calcium and phos- phorus content according to Horwitz (1980) on a Beckman DU-30 spectrophotometer Differences between oak, white fir, and Douglas fir feed and nonfeed trees were statis- tically analyzed to help discern foraging pat- terns used by instrumented porcupines. Chemical and ecological factors were evaluat- ed for between-species differences using two- sample t tests, and for within-species differ- ences with paired t tests (Minitab 1982). Statistical results are reported at the p < .05 and p < .1 levels. Chi-square analysis was used to determine if utilization of feed tree species by porcupines differed from the ex- pected. Discriminant analysis using backward elimination and forward selection (SAS 1985) was used to determine chemical and ecologi- cal factors that best discriminate between tree species, and between feed and nonfeed trees of the same species. 1995] Wintering Porcupines in Gambel Oak 31 Table 1. Mean values for factors tested for possible effects on porcupine herbivory. Oak(l) White fii :• (2) Douglas fir (3) N on feed Feed Nonfeed Feed Nonfeed Feed tree tree tree tree tree tree **„ = 46 n =46 n =3 « =3 n = 7 n =7 Distance from conifer (m) — 207 — 0 0 Distance to feed tree, same sp. (m 0 - 3582 — 5431 — .377 Wind speed (mph) — 5.53 — 3.7 — 9.71 Slope (%) — 33.53 — 36.5 — 42.41 Elevation (m) — 17792 — 19371-3 — 16802 Dbh (cm) *13.2 16.52-3 *25.4 40.41 33.3 34.51 Crude fiber (%) 43.3 44.23 43.6 48.03 42.4 40.11-2 Protein {%) 4.9 5.02-3 4.0 4.21 4.0 3.91 Phosphorus {%•) 0.038 0.039 0.087 0.064 0.038 0.042 Ether extract fractions (%) 9.0 9.12-3 15.7 12.71-3 16.5 I8.9I-2 Water 41.0 39.52.3 46.1 49.71 53.4 50.71 Potassium (%) 0.39 0.393 0.36 0.31 0.16 0.191 Calcium (%) 2.7 2.73 2.8 2.7 1.7 1.71 Magnesium (%) 0.137 0.1422-3 0.083 0.0921-3 0.068 0.0651-2 Sodium (ppm) 51.1 54.0 53.7 60.0 71.6 58.0 Sulfur (%) 0.20 0.19 0.70 0.48 0.14 0..35 pH 4.7 4.73 4.7 4.73 4.3 4.41-2 Tannins (radius in cm) 30.5 29.62.3 17.5 17.91-3 26.2 25.02-3 ***Astringency (mg/g fw) 85.8 83.0 *48.7 66.3 104.3 95.8 Sodium salts (%) 2.9 3.0 3.0 3.3 2.9 3.4 Soluble carbohydrates {%) 16.02 16.20 17..33 16.,58 FA (GCMS count units) 827,905 399,239 — 2,609,969 — 1,2.59,531 Superscript values indicate differences between species at thep < .1 level or less. 1 = oak, 2 = white fir, 3 = Douglas fir. *V'alues different between feed and nonfeed trees of the same species at or below p < I. ** Multiple locations in the same tree responsible for different n values used in calculations of cbemistn, and climatic data. Climatic data n vah as reported in Table 3. *** Not comparable across species boundaries. n \alues for factors below dashed line not as reported for rest of column. Not statistically comparable due to smaller sample size. i are the same Results Oak and white fir feed trees were larger than nonfeed trees of the same species {p < .05, Table 1). Herbivory by porcupines in decidu- ous species occurred in the canopies of large trees or in shrubs where branch diameters were relatively small. In coniferous species herbivory was also concentrated in the canopy rather than on the tree bole. Only two instances of chipping bark off the bole to expose the inner bark were noted in our study, both on deciduous tree species. There were no trends correlating calendar date or temperature to selection of feed tree species. Douglas fir feed trees contained greater amounts of crude pro- tein than Douglas fir nonfeed trees (p < .05). Crude protein content of both conifer species was less than that of oak trees (Douglas firp < .05, white fir p < .1). Total tannins (as mea- sured by radial diffusion) were higher in oak than in conifers (Douglas fir p < .1, white fir p < .05). Astringency (protein binding capacity) was not comparable among species but was greater for white fir feed trees than nonfeed trees (p < .1). Ether extract fractions were lower in oak than in conifers (p < .05) and lower in white fir than Douglas fir (p < .1). Tissue from Douglas fir contained less crude fiber than tissue from oak and white fir (p < .05), and Douglas fir feed trees contained still less than nonfeed trees (p < .1). Water con- tent was lower in oak tissue than in conifer tis- sue (p < .05). Oak contained higher levels of potassium and calcium than Douglas fir (p < .05). White fir was also higher than Douglas fir in calcium (p < .05). Magnesium levels for oak were greater than for either conifer species (p < .05). White fir and oak tissue had higher pH values than tissues from Douglas fir (p < .05). Oak feed trees were higher in sodium salts than Douglas fir feed trees (p < .1). Calcium- phosphorus ratios for feed trees were higher in oak than in Douglas fir (p < .05). The calcium- phosphorus ratio for oak is well above accept- able limits for mineral absoiption by mammals 32 Great Basin Naturalist [Volume 55 (Underwood 1966). High calcium-phosphorus ratios have also been reported by Masslich (1985) for aspen {Popidus tremuloides) tissue utilized by beaver. After an independent test of feed tree tissue confirmed the high ratio, we tested the mineral content of feces and urine from captive porcupines on an oak diet. Calcium-phosphorus ratios from fecal material were 10:1, while ratios from urine were approximately 221:1. Tissue samples from feed trees were ana- lyzed by GCMS primarily as a check on ether extract fractions. The small sample size did not permit statistical analysis, but trends showing lower fatty acid content in oak than in conifers concurred with our observation of lower ether extract fractions in oak. The amount of fatty acids was lower in oak than in either conifer species. Discriminant analysis correctly classified feed trees as either conifer or oak 100% of the time (Table 2). Six factors were important con- tributors to the model. Conifer feed trees had higher amounts of phosphorus and a greater ether extract fraction than oak feed trees. Alternatively, oak feed trees were higher in protein, calcium, tannins, and magnesium. Although tannins entered into the model, they were not a significant contributor These dif- ferences between oak and conifer feed trees generally are in agreement with differences in Table 1. The classification of oak feed and non- feed trees was less successful (71%, Table 2). Oak feed trees were significantly higher in sodium and fiber than nonfeed trees, while nonfeed trees were higher in water content. Porcupines used gambel oak as a food source more often than it occurred in the study site {p values listed in Table 3). Six of 15 animals were found roosting and feeding exclusively in oak, while 9 roosted and fed in conifer species at least once. Snow depths and temperatures were analyzed for the winter period before the main snowmelt (judged to be 18 March). Average snow depths at porcupine location sites for that time period were 0.60 m. Maximum snow depth was 1.20 m (median 0.65 m). Mean minimum temperature for the night previous to locating study animals was -10°C; the extreme low was -27°C. Mean temperature for the night previous to locating animals in rock or snow caves was -12 °C. There was no statistical difference between the minimum nightly temperature previous to locating porcupines in station trees compared to locating porcupines in rock or snow dens. There were approximately 7.0 porcupines/ km^ in the study area. Radio-collared animals were far ranging and did not utilize a single den or station tree as a base from which to launch foraging expeditions. Rather, they roosted and fed in a single tree for one to sev- eral days and then moved to another roost and feeding tree. Death loss due to predation and other causes left only 3 of 5 male and 6 of 10 female porcupines instrumented with radio transmitting devices for the entire winter This sample size made statistical analysis of home ranges unreliable. Several animals spent the winter in relatively small areas, but most had relatively large, overlapping home ranges. Male home range extremes were 6.8 and 47.5 ha. Extremes for females were 9.2 and 61.8 ha. One female's home range overlapped those of three males and at least four other females. Movements of up to 400-500 m between relo- cations of some of the larger, mature animals were not uncommon. Some juvenile animals had reduced home ranges and movements, which generally agrees with observations by Roze (1989). Mean distance from oak feed trees to a potential conifer feed tree was sig- nificantly less {p < .05) than the distance of an average move by a porcupine from an oak feed tree to any other feed tree (Table 1). Three of 15 porcupines (20%) were eaten by predators in a 4-mo period. Tracks in the snow indicated that one porcupine was pur- sued, worried, and killed by two coyotes. The other two porcupines eaten by predators died late in the season on south-facing slopes bare of snow; neither the cause of death nor carni- vore species could be positively determined. Carcasses of two other porcupines that died presumably of starvation and/or exposure dur- ing the course of the study were not scav- enged by coyotes. Discussion Chemical Factors Dietary alternatives in the form of different feed tree species, with significantly different chemical makeup, were available to the study population. In winter, vegetative oils have the potential to be the most important source of energy for porcupines. Data from ether extract fi-actions derived fi-oni feed tree tissues indicate 1995] Wintering Porcupines in Gambel Oak 33 Table 2. Standardized canonical discriminant function coefficients for factors that discriminated between oak and conifer feed trees (100% correct classification), and between oak feed trees and oak nonfeed trees (71% correct classifi- cation). Oak ( + ) vs. conifer (-) feed trees n = 56 Oak feed ( + ) vs. nonfeed (-) trees n = 46 Coefficient Prob > b Coefficient Prob > b Phosphorous -1.24 .00001 Water content -0.62 .006 Ether extract fractions -0.60 .0001 Sodium +0.61 .02 Protein +1.18 .0005 Fiber +0.59 .001 Calcium +0.39 .019 Tlmnins +0.29 .175 Magnesium +0.24 .006 that gambel oak, the major food source of our animals, had lower values of ether extract fiac- tions than tissues from conifers. Evaluation of fatty acids by GCMS confirmed that fatty acid content was higher in conifer tissue. Additional research on known digestible fractions is needed, but until data indicating otherwise are presented, we will operate under the premise that for porcupines conifers provide a greater source of useable fats than do oaks. Discriminant analysis was used to determine if, when all variables were taken together, there would be general support from this analysis with the t test. Significant differences found by these analyses comparing oak and conifer feed trees were in agreement (Tables 1, 2). Phosphorus and the ether extract frac- tion were higher in conifer feed trees com- pared to oak feed trees, and protein, calcium, tannins, and magnesium were higher in oak feed trees. Discriminant analysis was less suc- cessful in classifying feed and nonfeed trees within oak (Table 2). An important reason for this less successful classification was that the cloning nature of oak was emphasized by the point-quarter method. This method may have resulted in selecting nonfeed trees from the same clone as the feed tree. Future research should involve delineating the boundary of the clone and selecting a nonfeed tree from a clone different from the feed-tree clone. Conifer roost sites also offer greater thermal advantages than deciduous roost sites (Clarke and Brander 1973, Roze 1989). Despite multi- ple options, porcupines depended heavily on an oak diet low in fats and associated themial advantages but higher in tannins. The advan- tage of the oak diet may well be that it is high- er in protein. High levels of crude fiber (e.g., cellulose) reduce the digestibility of crude protein in monogastrics (Glover and Duthie 1958a, 1958b). Therefore, herbivores on a high-fiber diet would be expected to maximize the intake of crude protein to compensate for a low digestibility rate. Implications of a diet high in calcium and tannins are less clear, but it is possible that porcupines may deal with high levels of calcium in their food material by concentrating calcium in the urine. Tannins function as protein binding agents (Rhoades and Gates 1976). It is now evident that some insects can circumvent tannins through a higher gut pH and the presence of surfactants (Bernays 1981, Martin and Martin 1984, Martin et al. 1985). However, pH values for the mid-caecum (6.6), and the pyloric (1.8) and esophageal (3.2) regions of the stomach of a laboratory porcupine on a diet of oak were consistent with gut pH for monogastrics of comparable size (Hume 1982). Oveson (1983) measured subcutaneous adi- pose concentrations on the rump of porcu- pines and reported a thickness of 15.1 mm (± 2.6 mm) in early winter. By late February and early March fat reserves were virtually non- existent. A similar phenomenon was observed by Sweitzer and Berger (1993) in Nevada, where porcupine body condition decreased significantly throughout the winter season. Those authors suggested the change in body mass was an indication that porcupines deplet- ed energy reserves early in the winter and were stressed nutritionally during late winter. The heavy accumulation of fat serves as an energy reserve for porcupines to draw upon through- out the winter, allowing them to concentrate on a food source relatively high in crude pro- tein. The reduced capabilities of protein digestibility associated with a high-fiber diet may have encouraged our study animals to maximize dietary protein by selecting oak. Porcupine herbivory was generally noted on small branches. In large trees porcupines fed high in the canopy where limbs are smaller. 34 Great Basin Natufuijst [Volume 55 Table 3. Chi i-square analysis of per cent occunenct ■ and uti lization of trees by porcnpines.* % occurrence % nsecl Chi -square value /; \ alue Oak Conifer 43.5 2.7 82.1 16.4 3.23 0.10 Maple Conifer .52.1 2.7 1.5 16.4 52.41 0.01 Oak Maple 43.5 ,52.1 82.1 1.5 59.14 0.01 *n values differ from those reported in Table 1 due to the extended use of some feed trees by porcupines. Occupancy of the same feed tree during more than one sampling event counted as multiple utilization of oak but not double sampled for chemistry data. Df = 1. We obsei-ved only two instances in which por- cupines chipped bark of large tree boles and fed on tissue from large dbh limbs or trunks. Selection of larger feed trees by porcupines may be related to the texture of bark and ease of climbing (Roze 1989) rather than chemistry. Deciduous Food and Roosting Resource Roze (1989) discussed the thermal advan- tages of dens and/or conifer roost trees in rela- tion to maintenance of a core body temperature. Citing Ii-ving et al. (1955) and Clarke (1969), he indicated that the critical external tempera- ture below which porcupines must increase their metabolic rates to maintain a core body temperature is a range between -12 and -4°C. He suggested dens are temperature-averaging devices that protect porcupines against convec- tional and radiational heat loss. Station trees provide thermal advantages to porcupines (Clarke and Brander 1973) and may serve as a substitute for rock caves and snow dens. However, none of these are requisite to porcu- pine survival. Roze (1989) noted that porcu- pines may spend winters in trees away from dens and that in every report the tree species have been evergreens. Our data conflict with this observation. Porcupines throughout western North America are able to survive using a variety of deciduous species as food and roost tree resources. Despite the prominence of literature concern- ing dens and conifer station trees, use of a deciduous food and roosting resource without dependence on caves or snow dens is not an anomaly for porcupines. Craig and Keller s (1986) study site in southern Idaho was at an elevation of 1525-2089 m in desert shrub habitat. Animals in this study were not observed using dens during the winter or fol- lowing runways in feeding areas. They re- mained in the tops of hawthorne {Crataegus douglasii) thickets or utilized other deciduous food sources throughout the winter. Sweitzer and Berger (1993) identified buffalo-berry {Shepherdia argentea), willow {Salix spp.), bit- terbrush {Purshia tridentata), and juniper {Jiiniperus osteosperma) as primary winter food sources of porcupines in Nevada. We have also observed the extensive use of hack- beny {Celtis occidentalis) and green ash {Fraxi- nus pennsylvanica) by porcupines as a food and roosting resource in the Sand Hills of Nebraska and the Missouri River Breaks of South Dakota. Caves and conifers (except plan- tation forests and eastern red cedar [Juniperus virginiana]) are not available in the Sand Hills (Swinehart 1989). Oveson (1983) reported that a porcupine remained virtually motionless while perched in a gambel oak tree for a 24-h period when the ambient temperature was as low as -37°C. During a 13-d period from 30 January through 11 Februaiy, when the mean low temperature was -17°C, 3 of 25 (12%) loca- tions of our study animals were in conifers, 4 (16%) were in rock or snow caves, and 18 (72%) were in oak. Although porcupines did select trees with a larger dlih as roosting/feed- ing sites, they were also often found in smallish shrubs even though large trees were readily available. It is therefore difficult to link possi- ble benefits presumed to be available to porcu- pines that roost in larger trees, such as protec- tion from the elements or from predators, to the selection shown by animals in this study. Despite the availabilit\' of snow caves, dens, and conifer species that could provide thermal advantages, the study population was heavily 1995] Wintering Porcupines in Gambel Oak 35 dependent on gambel oak for a roosting and feeding resource. Considering that this re- hance was during a season of energetic stress, it is hkely that remaining motionless in the canopy of oak trees to consei-ve energy while exploiting a high-protein food source is an adaptive strategy. Movements and Predation The availability of conifer feed trees was not limiting since the average distance between locations of study animals was significantly greater than the mean distance of a move from any roost tree to a conifer roost tree (Table 1). It does not appear that spatial relationships of the various feed tree species played a role in feed tree selection by our study population. The relatively large overlapping winter home ranges of animals in this study differ from reports of other researchers. Home ranges for porcupines in northwestern Minnesota were small enough to be reported in square meters (Tenneson and Oring 1985). Curtis (1941), Dodge (1967), Brander (1973), Roze (1987, 1989), and others have documented that por- cupines move short distances from dens to feed trees, sometimes along permanent trails in the snow. Craig and Keller (1986) and Smith (1979) also reported reduced ranges in the winter. However, Dodge and Barnes (1975) did not indicate a similar restriction in winter movements. Roze (1987) suggested the reason may be crusted snows that bear the weight of the animals. Porcupines in our study did adeptly toboggan on crusted snows down extreme slopes in an attempt to avoid capture. However, one female moved over 450 m in fresh snow. Trails in powdery snow were often direct and suggested that a destination may have been predetermined. Common use of oak and conifer feed trees by different porcupines occurred several times during the study, sometimes concurrently. Hedging in the canopies of gambel oak trees in- dicated that some trees were used consistently over time by porcupines while others were not. Consistent foraging in common trees over time may indicate a learned behavior such as that described by Glander (1981) for howler monkeys, but we hesitate to attribute it to such because porcupine young-of-the-year were usually separated from their mothers during the winter. It is possible that some young accompanied their mothers for limited periods in the winter or that more subtle cues were used to transfer the information. Long movements between feed trees in dense oak cover by some study animals sug- gest that predator-prey relationships may have influenced movements. Sweitzer and Berger (1992) found that habitat use was related to the age or size class of porcupines, presum- ably in response to increased risk of predation to smaller porcupines. Our observations gen- erally agree with their findings. Mountain lion and coyote tracks were seen regularly in the study area. Both species are known to prey on porcupines (Keller 1935, Robinette et al. 1959, Toweill and Meslow 1977, Maser and Rohweder 1983). The strong urine scent at station trees or dens makes porcupines readily detectable. Mountain lions are capable of knocking porcu- pines from the canopies of trees (Taylor 1935). If long moves decreased the predictability of mountain lions locating porcupines in station trees, it would be an adaptive strategy. How- ever, long moves expose porcupines to terres- trial predation by mountain lions, coyotes, and wolves {Canis lupis, which are now extirpated fi-om the study area) and would presumably be nonadaptive. Since ample forage exists throughout the study site and long moves to locate food resources do not appear to be a dietary necessity, long movements may be an adaptive strategy to avoid arboreal predation by mountain lions. This hypothesis deserves further examination. Acknowledgments We thank S. H. Jenkins and two anony- mous reviewers for helpful suggestions to this manuscript. Literature Cited Bernays, E. a. 1981. Plant tannins and insect herbivores: an appraisal. Ecological Entomology 6: 353-360. Brander, R. B. 1973. Life historv' notes on the porcupine in a hardwood-hemlock forest in upper Michigan. Michigan Academician 5: 425—433. Clarke, S. H. 1969. Thermoregulatory' response of the porcupine, Erethizon dorsatiim, at low temperatures. Special report. Department of Forestry' and Wildlife Management, University of Massachusetts, Amherst. Clarke, S. H., and R. B. Brander. 1973. Radiometric determination of porcupine surface temperature under two conditions of overhead cover. Physiological Zoology 46: 230-237. Cottam, C, and J. T. Curtis. 1956. The use of distance measures in phytosociological sampling. Ecology 37: 451-460. 36 Great Basin Naturalist [Volume 55 Craic;, E. H., and B. L. Kkller. 1986. Movements and home range of porcupines {Erethizoii dorsutwn) in Idaho shnib desert. Canadian Field-Naturahst 100; 167-173. Curtis, J. D. 1941. The silvicuhural sigiiilicanee of the porcupine. Journal of Forestry 39: 583-594. D.^SiLVEiiu, A. J., K K Fkitosa Teles, and J. W. Stull. 1978. A rapid teclini(jue lor total nonstructural car- bohydrate determination of plant tissue. Journal of Agricultural and Food (]hemistr>' 26: 771-772. Dodge, \V. E. 1967. Life histor\ and biology of the porcu- pine {Erethizon dorsatum) in western Massachusetts. Unpublished doctoral dissertation. University of Massachusetts, Amherst. 167 pp. Dodge, W. E., and V. G. Barnes. 1975. Movements, home range, and control of porcupines in western Washington. U.S. Department of Interior, Fish and Wildlife Service Leaflet 507. Gambliel, H. a., R. G. Gates, M. K. Gaffey-Moquin, and T. D. Paine. 1985. Variation in the chemistry of loblolly pine in relation to infection by the blue-stain limgus. Pages 177-185 ;/i S. Branhani and R. Thatcher, eds.. Proceedings of the Integrated Pest Mangement Symposium, Asheville, NG. USDA Forest Service, Southern Experiment Station, New Orleans, LA. Glander, K. E. 1981. Feeding patterns in mantled howl- ing monkeys. Pages 231-257 in A. G. Kami!, and T D. Sargent, eds.. Foraging behavior: ecological, etho- logical, and psychological approaches. Garland STPM Press, New York, NY. Glover, J., and D. W Duthie. 1958a. The nutritive ratio/ crude protein relationship in nmiinant and nonnmii- nant digestion. Joinnal of Agricultural Science 50: 227-229. . 1958b. The apparent digestibility of crude protein by nonruminants and nmiinants. Journal of Agricul- tural Science 51: 289-293. Hagerman, a. E. 1987. A radial diffusion method for determining tannin in plant e.xtiacts. Journal of Chem- ical Ecologv' 13: 437-449. HORWITZ, W, ED. 1980. official methods of analysis of the Association of Official AnaKtical Chemists. Association of Official Analytical Chemists, Washington, DC. 1018 pp. Hume, I. D. 1982. Digestion physiology and nutrition of marsupials. Cambridge University Press, Cambridge, MA. 256 pp. Irving, L. H., H. Krog, and M. Monson. 1955. The metabolism of some Alaskan animals in winter and summer. Physiological Zoology 28: 173-185. Keller, E L. 1935. Porcupines killed and eaten by a coy- ote. Journal of Mammalogy 16: 232. Kufeld, R. C, M. Stevens, and D. G. Bowden. 1981. Winter variation in nutrient and fiber content and in vitro digestibility of gambel oak {Qucrcus gamhelii) and big sagebrush {Artemisia tridentata) from diver- sified sites in Colorado. Journal of Range Manage- ment 34: 149-151. Martin, J. S., and M. M. Martin. 1984. Surfactants: their role in preventing the precipitation of proteins by tannins in insect guts. Oecologia 61: 342-345. Martin, M. M., D. G. Rockholm, and J. C. Martin. 1985. Effects of surfactants on precipitation of pro- teins by tannins. Journal of Chemical Ecology 11: 485-494. Maser, C, and R. S. Rohvveder. 1983. Winter food habits of cougars from northeastern Oregon. Great Basin Naturalist 43: 42.5-428. Masslich, W J. 1985. Aspen-beaver relationships in the Strawberry Valley of central Utah. Unpublished master's thesis, Brigham Young University', Provo, UT 34 pp. MiNITAB. 1982. Release 82.1. Copyright, Penn State Uni- versity, State College, PA. Oveson, M. C. 1983. Behavioral and metabolic adapta- tions of porcupines (Erethizon dorsatum) to winter stress. Unpublished master's thesis, Brigham Young University, Provo, UT. 20 pp. Rhoades, D., and R. G. Gates. 1976. Toward a general theory of plant antiherbivore chemistry. Pages 168-213 in Biochemical interaction between plants and insects. J. W Wallace and R. W. Marshall, eds.. Recent Advances in Phytochemistry. Volume 10. Plenum Press, New York-London. ROBINETTE, W L., J. S. GaSHWILER, AND O. W. MORRIS. 1959. Food habits of the cougar in Utah and Nevada. Journal of Wildlife Management 23: 261-273. ROZE, U. 1987. Denning and winter range of the porcu- pine. Canadian Journal of Zoology 65: 981-986. . 1989. The North American porcupine. Smithsonian Institution Press, Washington, DC. 261 pp. SAS. 1985. Copyright, SAS Inc., Gary, NG. Smith, A. D. 1957. Nutritive value of some browse plants in winter. Journal of Range Management 10: 162-164. Smith, G. W 1979. Movements and home range of the porcupine in northeastern Oregon. Noithwest Science 53: 277-282. SWEITZER, R. A., AND J. Berger. 1992. Size-related effects of predation on habitat use and behavior of porcu- pines [Erethizon dorsatum). Ecology 73: 867-875. . 1993. Seasonal dxnamics of mass and body condi- tion in Great Basin porcupines {Erethizon dorsatuni). Journal of Mammolog>' 74: 198-203. SwiNEHART, J. B. 1989. Wind-blown deposits. Pages 43-56 in A. Bleed and G. Flowerday, eds.. An atlas of the Sand Hills. Resource Atlas No. 5, Conservation and Survey Division, Institute of Agriculture and Natural Resources, Lhiiversitv' of Nebraska-Lincoln. Taylor, W R 1935. Ecology and life history- of the porcu- pine {Erethizon epixanthum) as related to the forests of Arizona and the southwestern United States. University' of Arizona Bulletin 6: 1-177. Tenneson, C., and L. W Oring. 1985. Winter food pref- erences of porcupines. Journal of Wildlife Manage- ment 49: 28-33. TowEiLL, D. E., and C. E. Meslow. 1977. Food habits of cougars in Oregon. Journal of \Vildlife Management 41: 576-578. Underwood, E. J. 1966. The mineral nutrition of live- stock. Central Press Ltd., Aberdeen, Great Britain. 237 pp. U.S. Glimatological Data (Utah). 1984 and 1985. National Climactic Center, Asheville, NG. 86: 12 and 87: 13. Welgh, B. L. 1989. Nutritive value of shrubs. Pages 405-^24 in G. M. McKell, ed., Biology and utilization of shrubs. Academic Press, Inc., New York, NY. Received 18 August 1993 Accepted 30 September 1994 Great Basin Naturalist 55(1), © 1995, pp. 37-45 HISTORIC EXPANSION OF JUNIPERUS OCCIDENTALIS (WESTERN JUNIPER) IN SOUTHEASTERN OREGON Richard E Miller^ and Jeffeiy A. Rosel Abstract. — The chronology of Junipents occidentalis (western juniper) expansion in eastern Oregon, the effect of plant canopy and interspace on / occidentalis seedling establishment and growth rates, and the age of/, occidentalis maximum reproductive potential were determined. Measurements were recorded in twenty-two 0.4-ha plots estab- lished in sagebrush-grassland communities and six 0.1-ha plots in Populus tremuloides (quaking aspen) communities. /. occidentalis began increasing during the 1880s in stands containing trees >130 yr old. Relatively steady establishment ensued into the 1950s and then began to progress at a geometric rate in the 1960s. / occidentalis encroachment into aspen stands began between 1910 and 1920. The largest proportion of juvenile trees established beneath Artemwia species in sagebrush-grassland communities./, occidentalis trees appeared to reach full reproductive potential at >50 yr of age. The ratio of male;female trees increased from 1.7 in scattered/, occidentalis stands to 3.8 in closed stands. The initiation of/, occidentalis encroachment during the late 1800s coincides with optimal climatic conditions for Juniperus beiTy production and establishment, reduced fire-retum intervals, and heavy livestock grazing. The accelerated increase in /. occidentalis expansion since 1960 may be due to the continued absence of fire, abundant woody plant cover, and the large increase in/, occidentalis seed production. Key words: western juniper, Juniperus occidentalis, expansion. Great Basin, intennountain shrub steppe, aspen, Populus tremuloides, succession. One of the most pronounced plant commu- nity changes in the 20th century has occurred in the juniper and pinyon-juniper woodlands, a major vegetation type characterizing the Intermountain Region. These woodlands, sometimes described as pygmy forests, cur- rently occupy 17 million ha throughout this region (West 1988). Juniperus occidentalis ssp. occidentalis Hook, (western juniper) is consid- ered the Northwest representative of the pin- yon-juniper zone in the Intermountain Region (Franklin and Dyrness 1973) and occupies over 1 million ha (Dealy et al. 1978) in eastern Oregon, southwestern Idaho, and northeastern California (Cronquist et al. 1972). This sub- species of/, occidentalis is found primarily north of the polar front gradient (Neilson 1987; parallel to the Oregon and Nevada border, latitude 42°) where temperatures are cooler, summer precipitation decreases, and winter precipitation increases (Mitchell 1976). Relict juniper woodlands, tree-age class dis- tribution, fire scars, and historical documents indicate presettlement pinyon-juniper and juniper woodlands were usually open, savan- nah-like (Nichol 1937, West 1988), or confined to rocky surfaces or ridges (Cottam and Stewart 1940, Barney and Frishknecht 1974, Hopkins 1979, Johnson and Simon 1987). /. occidentalis began increasing in both density and distribu- tion in the late 1800s (Burkhardt and Tisdale 1976, Young and Evans 1981, Eddleman 1987), invading Artemisia tridentata subsp. vaseyana (mountain big sagebrush), Artemisia arhuscula (low sagebrush), Populus tremuloides (quaking aspen), and riparian communities. Although /. occidentalis is long lived (Vasek 1966, Lanner 1984), less than 3% of the woodlands in Oregon are characterized by trees >100 years old (USDI-BLM 1990). In 1825, Ogden' observed only occasional /. occidentalis (reported as cedars) growing on hillsides while traveling through the Crooked River drainage in central Oregon (Rich et al. 1950). Today these hill- sides are covered by dense /. occidentalis woodlands. In a nearby area J. W Meldrum's 1870 survey notes describe a gently rolling landscape covered with an abundance of peren- nial bunchgrasses and a wide scattering of/. occidentalis trees (Caraher 1977). Today, / 'Eastern Oregon Agricultural Research Center, HC 71 4.51 Hwy 20.5, Bums, OR 97720. The Eastern Oregon Agricultural Research Center, including the Bums and Union stations, is jointly operated by the Oregon Agricultural E.\perinient Station of Oregon State University and USDA Agricultural Research Service. 37 38 CiHEAT Basin Natuiulist [Volume 55 occidentalin densities on this site range between 125 and 250 ha"^. In Silver Lake, Oregon, /. occidentalis density increased from 62 ha-1 in 1890 to over 400 ha-1 by 1970 (Adams 1975). On another site in central Oregon where trees were absent prior to 1880, /. occidentalis increased to 1018 ha~^ by 1980 (Eddleman 1987). Recent expansion is similar to increases in other Juniperus species throughout western United States (Ellis and Schuster 1968, Tausch et al. 1981, West 1984, Tausch and West 1988). The objectives of our study were to (1) describe the chronology of/, occidentalis expansion during the past several centuries in southeastern Oregon, (2) determine the effect of plant canopy and interspace on / occiden- talis seedling establishment and growth rates, and (3) determine the age when /. occidentalis reaches maximum reproductive potential. Methods Study Area The study area is located on Steens Moun- tain in southeastern Oregon, approximately 80 km south of Burns. This isolated volcanic fault-block, which lies in the extreme north- west Basin and Range Province (Fenneman 1931), is about 80 km long and oriented in a northeast direction (Baldwin 1981). The eleva- tion of Steens Mountain ranges from 1268 to 2949 m, with a steep east-facing escarpment and a gentle west-facing slope. Climate is cool and semiarid, characteristic of the northern Great Basin. Annual precipitation at the lower elevations averages 220-280 mm, increasing to > 700 mm at higher elevations (NOAA 1993). Most moisture is received as snow in Novem- ber, December, and January and as rain in March through June. /. occidentalis woodlands on Steens Moun- tain form a discontinuous belt between 1450 and 2100 m in elevation. Severe winter condi- tions probably restrict /. occidentalis from ex- panding into higher elevations (Billings 1954, Mehringer 1987). Limited distribution below 1500 m is possibly due to a combination of late spring frosts (Billings 1954) and limiting mois- ture. Tree canopy cover varies from open to 30% cover, except on mesic P. treniuloides sites where /. occidentalis cover approaches 100%. However, based on age structure and canopy leader growth, tree canopies are still actively expanding on the majority of sites measured. Early observations on Steens Mountain indi- cate the landscape contained only scattered stands of/ occidentalis (Griffiths 1902). Since 1900 the abundance of/, occidentalis pollen in the Steens Mountain area has increased five- fold (Mehringer and Wigand 1990). Plant communities characteiistic of/ occi- dentalis woodlands are Artemisia tridentata ssp. vaseyana/Festuca idahoensis (Idaho fes- cue), Artemisia arhuscula/E idahoensis, and P. tremuloides. P. tremuloides communities on Steens Mountain range in elevation from 1760 to 2400 m. At lower elevations, in the / occi- dentalis woodland belt, P. tremuloides stands form long, narrow communities along north aspects, which capture windblown snow and runoff. Plot Layout Plot locations were selected in an attempt to reflect sagebrush-grassland communities in different stages of/ occidentalis invasion on the west slope of Steens Mountain. Old stands on the rocky outcrops, which make up only a small percentage of present-day woodlands, were not measured. Sites selected support, or have the potential to support, sagebrush-grass- land communities. Currently these sites are occupied by varying numbers and sizes of/ occidentalis dominance, creating a woodland structure of dispersed, intermediate, and closed tree stands (Table 1). Twent\'-two 0.4-ha plots were located within the / occidentalis belt of Steens Mountain; tliey ranged from 1500 to 2000 m in elevation and were distributed 32 km along the mountain range. Plots were situ- ated along an elevation gradient representing communities from the lower- to upper-eleva- tion / occidentalis woodland belt. Dominant understory vegetation in the dispersed and intermediate plots was A. tridentata spp. vaseijana and Festuca idahoensis (13 stands), A. arbuscula and F idahoensis (4 stands), and a mosaic of A. arbuscula and A. tridentata ssp. vaseijana (2 stands). Understoiy vegetation in the closed stands (n = 3) comprised a few remnant deep-rooted perennial grasses, skele- tons of dead A. tridentata ssp. vaseijana, and 70% bare ground (EOARC data file). An additional six 0.1-ha plots were estab- lished in six separate P. tremuloides stands. Three stands were in advanced stages of/ occidentalis invasion with few to no adult P. 1995] Western Juniper Expansion 39 Table L Juniperus occidcntalis stand maturity classes in Artemisia communities (modified fi-om Blackburn and Tueller 1970). Closed Abundant adult trees generally >5 ni tall and usually several trees > 130 yr of age, with little understory, particularly on south slopes. Intermediate Abundant/ occidentalis of all age classes, with a more open tree canopy and an understory beginning to decline; trees > 130 yr of age are rare. Dispersed Abundant young trees <2 m tall, a few adult trees but old trees absent, and a well-developed understory. tremuloides trees and dead P. tremuloides trunks on the ground. The remaining three stands were characterized by a dominant P. tremuloides overstory and an understory of young/, occidentalis. Elevation for the/, occi- dentalis-P. tremuloides plots ranged from 1930 to 2000 m, all with a similar northeast aspect. Measurements Prior to sampling, string was stretched along the contour of each 0.4-ha plot at 1-m intervals to keep track of measured trees. / occidentalis density (trees ha~^) was recorded for trees <0.5 m tall, defined as adult, across the entire plot. Tree height, minimal and maximal crown diameters, and basal area just above the trunk swell at the stem base near the litter layer were recorded. Tree height was measured with a tape for trees < 2 m and a clinometer for trees >2 m tall. Tree canopy cover was estimated by adding crown area measure- ments of all trees for each plot. Similar mea- surements were recorded on juvenile trees (defined as trees < 0.5 m tall), but only those on the lower left quarter (0.1 ha) of each 0.4- ha plot. Current-year / occidentalis seedlings (any plant with cotyledons still attached) were not recorded. Establishment location of each juvenile tree was recorded: beneath the canopy of/ occidentalis, Artemisia, other shrubs, tus- sock grass, or in the interspace. Less than 1% of juveniles were located beneath other shrubs or grasses; therefore, only / occidentalis, Artemisia, and interspace are reported. / occidentalis is considered submonoecious (Vasek 1966). Male and female reproductive status was detemiined by estimating abundance of cones and berries for each tree. Abundance was ranked in four classes: (0) absent, (1) scarce, (2) obvious but not abundant, and (3) abundant. In each plot a 10-tree subsample was ran- domly selected for aging in each of four height classes: (1) <0.5 m, (2) 0.5-1.8 m, (3) 1.8-3 m, and (4) >3 m. In several of the dispersed plots, sample size for trees >3 m was smaller than 10, due to a lack of trees. We also sampled all old trees on plots when they occun^ed (n = 0-5 ha~l). Old trees were easily identified by their growth form, containing rounded tops and heavy limbs, and lacking strong terminal leader growth (Burkhardt and Tisdale 1969). A cross section was removed approximately 30 cm above ground level from each tree >0.5 m tall and at ground level for trees < 0.5 m, and then brought back to the lab for aging. Two radii from each cross section were polished, stained, and counted. Age was estimated by averaging both radii and adding 10 yr to cor- rect for the 30-cm base. Mean differences between radii were 4% for trees >50 yr and 1% for trees < 50 yr of age. Adams (1975) reported that growth-ring characteristics of/ occidentalis are useful in dendrochronological studies. The presence of false and missing rings was similar to that for Pinus ponderosa. Over 1200 trees were aged and approximately 14,000 counted and measured. In the six P. tremuloides stands, density of both /. occiden- talis and P. tremuloides and age and height for / occidentalis were measured across the entire 0.1-haplot. Evidence indicated minimal / occidentalis mortality has occurred on Steens Mountain during the past 120 years. We observed very few dead or dying trees for all age classes (excluding seedlings), except where individual / occidentalis trees had been cut or burned. Mortality of Jimiperus species rapidly declines following the seedling stage (Van Pelt et al. 1990). Juniperus has few pests that prove fatal to the tree (Lanner 1984). We avoided recently cut or burned stands, which constituted a small percentage of/, occidentalis-occupied stands. Where remains of dead trees were observed, we noted they persisted for a long period of time. By recutting several stumps adjacent to one of our plots and aging and matching ring widths with adjacent live trees, we determined these trees were harvested around 1920. Others have also observed the persistence of Juniperus stumps (Young and Budy 1979). 40 Great Basin Naturalist [Volume 55 Statistical Analysis Height growth data for adult trees were analyzed using a randomized complete block- design in PROG GLM of SAS (SAS 1986). Means were separated using Duncan's Multiple Range Test at /; < .05 level. A split- plot design was used in the analysis of juvenile height growth. Main plots were sites and sub- plots were location of establishment (interspace, Artemisia, J. occidentalis). A Duncan's Multiple Range Test was used to separate the means. Results Little change in /. occidentalis density appeared to occur between the early 1700s and the 1880s (Fig. 1). We encountered old trees (standing trees >130 years old, large stumps, and burned-out trunks) on several A. arhusciila flats and A. tridentata ssp. vaseyana communities. However, data indicated preset- tlement tree densities in these Artemisia com- munities were < 5 trees ha~^ suggesting very open /. occidentalis stands. The first evidence of an increase in tree densities occurred in the 1880s, with relatively steady establishment ensuing into the 1950s, similar to that observed by Tausch and West (1988). In the 1960s /. occidentalis establishment began occurring at a geometric rate. Glosed /. occidentalis stands, which once supported A. tridentata ssp. vaseyana, were characterized by an abundance of adult trees (> 3 m tall), a tree canopy cover of 18-28% (Table 2), and the presence of a few old trees (130+ yr; 2 to 5 ha~^). /. occidentalis densities began increasing in these stands between 1878 and 1890. In the intermediate/, occiden- talis stands, trees >130 yr were rare. Tree canopy cover ranged from about 8 to 16%, and densities of adult trees varied from 35 to 100 ha~l. Trees <3 m in height, particularly juve- niles, were abundant. /. occidentalis expansion in these sagebrush-grassland communities began between 1890 and 1910. In the dis- persed stands few trees were >60 yrs old, and we aged no trees > 100 yr. Tree canopy cover was usually <5% in the dispersed stands and densities of large adult trees <35 ha~l. Invasion of/, occidentalis into these sage- brush-grassland communities began after 1930. Greatest densities of/, occidentalis trees measured on Steens Mountain occurred in P. tremuloides sites (Table 3). In the late stages of /. occidentalis succession on these sites, tree canopy cover approached 100%. Live P. tremuloides occurred only on one of the three sites, and almost all trees were <0.5 m tall. In the remaining two stands only the remnants of large P. tremuloides trunks decaying in the understoiy were present. /. occidentalis inva- sion in these P. tremuloides sites began be- tween 1910 and 1920. No/, occidentalis trees T \ \ 1 1 1 r 1700 1720 1740 1760 1780 1800 1820 1840 1860 1880 1900 1920 1940 1960 1980 YEAR Fig. 1. Years of establisliiiient i'or J uuipcrus occidentalis trees on Steens Mountain, Oregon (n = 1200). 1995] Western Juniper Expansion 41 Table 2. General description of closed, inteniiediate, and dispt'rscd J unipcriis occidentalis stands on Steens Mountain in Artemisia tridentata ssp. vaseijami and A. arhuscula comminiities, and the percentage of juveniles located beneath / occidentalis, Ai-teinisia. and interspace. Canop\' cover, basal area, and density means are followed by range in paren- theses ( ). Establishment site % Canopy Basal area Density (#ha-l) (for juveniles) Adults JuNcniles # Site s cover '7i (ni-ha-^) >0.5 ni ht <0.5 III ht J \. occidentalis Aiiernisia Interspace A. tridentata ssp. vaseyana closed 6 22 (18-28) 5.2 (3.1-9.8) 296 (217-496) 580 (118-1226) 86^'l 9^ 5'' intermediate 8 6 (5-10) 1.8 (0..5-4.7) 95 (50-165) 815 (335-1423) 291' 58^ 13^- dispersed 2 2 (1-3) 0.4 (0.2-0.6) .52 (31-70) 188 (96-280) 3'' 50^ 47a A. arhuscula closed 3 15 (12-20) 3.5 (1.8-5.4) 158 (74-247) 99 (20-198) 27'' 67" 6'' intermediate 3 6 (4.5-6.7) 1.8 (0.9-3.2) 104 (77-153) 375 (167-790) 11'' 6P 28l' 'Sites of establishment means (%) folloux-d 1 .V similar limercast ■ IrttiTs are not simnl: le.mtiv iliflerent I.etNveei . establishment si les within / m rithnhdis stand maturity classes (p ■ i .05). >80 yr were encountered. In stands with a P. tretniiloides overstoiy, P. trcmiiloidcs density of small shoots was greater than that of/ occi- dentalis. However, P. tremuloides size classes between 0.5 m and large adults were absent, indicating a lack of P. tremuloides stand reju- venation. On these sites /. occidentalis inva- sion began between 1930 and 1940. Height growth for young /. occidentalis trees (<20 yr) across all sites averaged 2.9 cm yir^. Based on growth rates and height of trees between 10 and 20 yr of age (n = 200) across all Artemisia sites, 90% of trees 15 yr old were <1 m tall (64% were <0.5 m tall). Surprisingly, height growth rates of juvenile trees did not significantly differ between A. arhuscula and A. tridentata ssp. vaseyana communities. However, location of establish- ment within communities significantly influ- enced growth rates of young /. occidentalis trees (Table 4). Trees establishing beneath an Artemisia canopy grew faster than young trees growing in the interspace. Shrub and tree canopies also significantly influenced location of/, occidentalis seedling establishment in Artemisia communities. The largest proportion of juvenile trees was usually located beneath canopies of A. tridentata ssp. vaseyana or A. arhuscula and /. occidentalis (Table 2). Less than 20% of juveniles across all 22 Artemisia sites established in the interspace. On Steens Mountain, for trees >0.5 m tall, 32% expressed predominantly only male or only female characteristics, 38% both male and female, and 30% contained neither fnjits nor cones. /. occidentalis trees producing abun- dant crops of cones or berries were either male or female dominant. No trees were mea- sured which contained an abundant crop of both berries and cones. Sixty-five percent of/. occidentalis trees with an abundant crop of berries contained no male cones. The remain- ing 35% contained only a scarce number of cones. The majority of trees producing abun- dant crops of male cones contained only scarce numbers of berries. Approximately 75% of trees producing heavy crops of berries or cones were >50 yr old. Trees <20 yr old ex- pressing reproductive effort were rare and produced only a few cones or berries. The ratio of trees producing large crops of cones versus berries (cones :berries) increased from 1.7 in the scattered/, occidentalis stands to 3.8 in the closed stands. Discussion Low densities and limited distribution of/ occidentalis trees >130 yr and limited num- bers of dead trees or old stumps suggest /. occidentalis has greatly expanded on Steens Mountain during the past 100 yr. Distribution of old trees was generally limited to rocky ridges and A. arhuscula communities. Old trees were found only occasionally growing in deeper, well-drained soils such as A. tridentata ssp. vaseyana grassland communities and were absent in P. tremuloides communities. In northeastern California, Barbour and Major 42 Great Basin Naturalist [Volume 55 Table 3. Mean dcnsitifs (# ha"') follow i-d 1)\ rant^f in ( ) oi F()i)iilus tretnuloidcs nud Junipcriis occidentalis in P. treinuloides sites. Stage of succession P. tremuloides Adult Juvenile J- occidei Italia Adnlt Juvenile 1392 (929-2203) 1090 (632-1739) 9462 (4327-18,791) 2816 (622-5968) Late (n = 3) Intermediate (n=3) (0-50) 1060 (476-1670) 1316 (0-3952) 6553 (5266-9480) (1977) found a similar distribution of old and young /. occidentalis trees. A. tridentata ssp. vaseyana and A. arhtiscxda commimities, which contained a low density of/, occidentalis trees prior to settlement, were the earliest sites to initiate an increase in /. occidentalis. Dates of initial establishment of closed and intermedi- ate stands were similar to periods of early stand development reported by Young and Evans (1981) in northeastern California and Eddleman (1987) in central Oregon. Expansion of/, occidentalis coincides with Euro-American settlement in this portion of the Great Basin. Although no direct cause-and- effect relationship can be drawn, we hypothe- size that climate, altered fire frequencies, and grazing in the late 1800s were primary factors initiating the recent expansion of/, occidentalis. Following the end of the Little Ice Age in the mid 1800s (Biyson 1989), winters became more mild and precipitation increased above the present long-term average in the northern half of the Great Basin between 1850 and 1916 (Antevs 1948, Graumlich 1985). Mild, wet win- ters and cool, wet springs promote vigorous growth in /. occidentalis (Earle and Fritts 1986, Fritts and Xiangdig 1986). Presettlement fire-return intervals in A. tri- dentata ssp. vaseyana communities have been reported to vary from 15 to 25 yr (Houston 1973, Burkhardt and Tisdale 1976, Martin and Johnson 1979). Burkhardt and Tisdale (1976) concluded that fire-frequency intervals of 30-40 yr would be adequate to keep /. occi- dentalis from invading a sagebrush-grassland community. Following settlement, frequency of fire in sagebrush grasslands has greatly declined. The reduction of fine fuels by high densities of domestic livestock greatly reduced the potential for fire in the Intermountian Shrub Region (Burkhardt and Tisdale 1976, West 1988). Griffiths' (1902) observations of the overgrazed landscape on Steens Mountain support this hypothesis. Fires set by Native Americans also declined in the 19th century due to large reductions in their populations caused by European diseases (Thompson 1916, Grossman 1981) and relocation to reser- vations in the 1870s. The invasion of conifers into P. tremuloides communities is a common occurrence through- out the western U.S. However, conifers report- ed to typically invade P. tremuloides stands are species adapted to more mesic sites such as Piniis contorta (lodgepole pine), P. ponderosa, Pseudotsuga inenziesii (Douglas-fir), Abies concolor (white fir), Abies lasiocarpa (sub- alpine fir), Picea engelmannii (Engelmann spiTice), and Picea piingens (blue spnice) (Bartos 1973, Mueggler 1985). Invasion of the more drought-tolerant /. occidentalis into P. tremu- loides stands is not well documented. P. tremuloides is frequently considered a fire-induced species, replaced by less fire tol- erant conifers (Baker 1925, Daubenmire 1943, Mueggler 1976). Prior to settlement, lightning and human-set fires probably helped maintain many P. tremuloides communities. However, the occurrence of fire in P. tremuloides stands in the Rocky Mountains has been greatly reduced since the late 1800s (Jones and DeByle 1985). Mueggler (1985) suggested the combi- nation of fire suppression and heavy grazing in P. tremuloides communities may favor tlie estab- lishment of conifers. An increase in Artemisia cover may also enhance the invasion of/, occidentalis. As a sagebrush-grassland community shifts towards a greater dominance of shrubs, the number of safe sites for /. occidentalis seedling establish- ment increases. Others have also reported the majorit>' of/, occidentalis seedlings established beneath Artemisia canopies (Burkhardt and Tisdale 1976, Eddleman 1987). In west Texas, /. pinchotii frequently establishes beneath mesquite plants (McPherson et al. 1988). 1995] Western Juniper Expansion 43 Table 4. Mean growth rates for juvenile /w;ii;jf'n/.s occi- dentalis trees (2-30 yr old) in three different establish- ment sites. Establishment site cm yi Artemisia J. occidentalis Interspace 3.3A 2.7AB 2.4B Means followed by similar uppercase letters are not significantly different (/; < .05). Shading by nurse plants may benefit /. occi- dentalis seedlings (Johnsen 1962) by reducing summer surface temperatures by 45-57% of bare ground surface temperatures (Burkhardt and Tisdale 1976). Enhanced growth rates of young trees growing beneath A. tridentata ssp. vaseijana suggest microclimates beneath shrub canopies are more beneficial than conditions in the interspace. Burkhardt and Tisdale (1976) reported/, occidentalis seedling growth rates were correlated positively with Artemisia and correlated negatively with bare ground. /. occidentalis approached hill reproductive potential near 50 yr As /. occidentalis densities increased, the proportion of trees became pre- dominantly male across sites. Highly fecund female trees appeared to be most important in open stands where /. occidentalis was actively expanding. In central Oregon, Eddleman (1984) observed that trees in the interior woodlands were strongly dominated by male cone production while trees growing in the open produced more female cones. He also reported trees did not produce significant quantities of fruit until 50-70 yr of age. Conclusion Optimal climatic conditions around the turn of the century, reduced fire retuiTi inter- vals, and the indirect effect of livestock through the reduction of fine fuels and an increase in Artemisia cover are probably pri- mary factors that have contributed to the rapid expansion of/, occidentalis in southeast Oregon during the late 1800s and early 1900s. The accelerated increase in /. occidentalis density and invasion during the last 30 years into new communities is probably largely due to the continued absence of fire, abundant woody plant cover, and the large increase in / occidentalis seed rain. Acknowledgments This is Technical Report 10,494 of the Eastern Oregon Agricultural Research Center, Oregon State University. Literature Cited Ada.ms, a. W. 1975. A brief histoiy of juniper and shnib populations in southern Oregon. Wildlife Research Report 6, Oregon State Wildlife Commission, Corvallis. Antevs, E. 1948. Climatic changes and pre-white man. Pages 168-191 in The Great Basin, with emphasis on glacial and postglacial times. University of Utah Biological Service Bulletin 10(7). Baker, E S. 1925. Aspen in the central Rocky Mountain region. US DA Bulletin 1291. Baldwin, E. M. 1981. Geology of Oregon. Kendal/Hunt Publishing Co., Dubuque, lA. B,\RBOUR, M.G., AND J. Major. 1977. Terrestrial vegeta- tion of California. Wiley Interscience, New York, NY. Barney, M. A., and N. C. Frishknecht. 1974. Vegetation changes following fire in the pinyon-juniper type of west-central Utah. Journal of Range Management 74; 91-96. Bartos, D. L. 1973. A dynamic model of aspen succes- sion. Pages 1.3-25 in Proceedings, lUFRO Biomass Studies. lUFRO. Billings, W. D. 1954. Temperature inversions in the pinyon-juniper zone of a Nevada mountain range. Butler University Botanical Studies 12. Blackburn, W H., and P T. Tueller. 1970. Pinyon and juniper invasion in black sagebrush communities in east-central Nevada. Ecology 51: 841-848. Bryson, R. a. 1989. Late Quaternary volcanic modulation of milankovith climate forming. Theoretical Applied Climatology 39: 115-125. Burkhardt, J. W, and E. W Tisdale. 1969. Nature and successional status of western juniper vegetation in Idaho. Journal of Range Management 22: 264-270. . 1976. Causes of juniper invasion in southwestern Idaho. Ecology 76: 472-484. Caraher, D. L. 1977. The spread of western juniper in central Oregon. Pages 3-8 in R. E. Martin, J. E. Dealy, and D. L. Caraher, eds., Proceedings, Western Juniper Ecology and Management Workshop. US DA Forest Service, General Technical Report PNW-4. Cottam, W. E, and G. Stewart. 1940. Plant succession as a result of grazing and of meadow desiccation by erosion since settlement in 1892. Journal of Forestry 38: 613-626. Cressman, L. S. 1981. The sandal and the cave. Oregon State University Press, Corvallis. Cronquist, a., a. H. Holmgren, N. H. Holmgren, and J. L. Reveal. 1972. Intermountain flora: vascular plants of the Intermountain West, U.S.A. Volume 1. Hafiier Publishing Company, New York, NY. Daubenmire, R. F 1943. Vegetational zonation in the Rock>' Mountains. Botanical Review 9: 325-393. Dealy, J. E., J. M. Geist, and R. S. Driscoll. 1978. Western juniper communities on rangeland of the 44 Great Basin Natuiulist [Volume 55 Pacific Northwest. Pages 201-204 in D. K. ihiler, ed., Proceedings, First International liangeland Congress, Denver, CO. Earle, C. J., AND H. C. Fritts. 1986. Reconstructing river flow in the Sacramento Basin since 1560. Report. California Department of Resources, Agree- ment DWR B-55395. Laboratory of Tree-ring Research, University of Arizona, Tucson. Eddleman, L. E. 1984. Ecological studies on western juniper in central Oregon. Pages 27-35 in T. E. Bedell, ed.. Proceedings, Western Juniper Management Short Course. Oregon State University and Extension Service, Corvallis. . 1987. Establishment and stand development of western juniper in central Oregon. Pages 255-259 in R. L. Everett, ed.. Proceedings, Pinyon-Juniper Conference. US DA Forest Service, General Technical Report INT-215. Ellis, D., and J. C. Schuster. 1968. Juniper age and dis- tribution on an isolated butte in Garza County, Texas. Soutliwestem Naturalist 13: 343-.348. Fenneman, N. M. 1931. Physiography of the western United States. McGraw-Hill, New York, NY. Fr\nklin, J. F, and C. T. Dyrness. 1973. Natural vegeta- tion of Oregon and Washington. USDA Forest Service, General Technical Report PNW-8. Portland, OR. Fritts, H. C, and W Xiangdig. 1986. A comparison between response-function analysis and other re- gression techniques. Tree-ring Bulletin 46: 31^6. Gr.\UMLICH, L. 1985. Long-term records of temperature and precipitation in the Pacific Northwest derived from tree rings. Unpublished doctoral dissertation. University' of Washington, Seattle. Griffiths, D. 1902. Forage conditions on the northern border of the Great Basin. Bureau of Plant Industry', USDA, Bulletin 15. Hopkins, W. E. 1979. Plant associations of the Fremont National Forest. USDA Forest Service, Pacific Northwest Region, R6-ECOL-79-004. Houston, D. B. 1973. Wildfires in northern Yellowstone National Park. Ecolog\' 54: 1109-1117, JOHNSEN, T. N. 1962. One-seed juniper invasion of north- em Arizona grasslands. Ecological Monographs 32: 187-207. Johnson, C. G., Jr., and S. A. Simon. 1987. Plant associa- tions of the Wallowa-Snake Province, Wallowa- Whitman National Forest. USDA Forest Service, Pacific Northwest Region Report R6-ECOL-TP- 255B-86. Portland, OR. Jones, J. R. and N. V. DeByle. 1985. Fire. Pages 77-81 in N. V. DeByle and R. E Winokur, eds.. Aspen: ecology and management in the western United States. USDA Forest Service General Technical Report RM-119. Lanner, R. M. 1984. Trees of the Great Basin: a natural history. University of Nevada Press, Reno. Martin, R. E., and A. H. Johnson. 1979. Fire management of Lava Beds National Monument. Pages 1209-1217 in R. M. Linn, ed.. Proceedings, First Conference of Science and Research in the National Parks. USDI National Park Service, Transactions Proceedings Serial 5. McPherson, G. R., H. a. Wright, and D. B. Wester. 1988. Patterns of shrub invasion in semiarid Texas grasslands. American .Midland Naturalist 102: 391-397. .Vli'HRlN(;ER, R J., Jr. 1987. Late Holocene environments on the northern periphei-y of the Great Basin. Final report. Bureau of Land Management, Portland, OR. Mehringer, R J., Jr., .vnd P E. Wigand. 1990. Compari- son of late Holocene environments from woodrat middens and pollen: Diamond Craters, Oregon. Pages 294-325 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. Mitchell, V. L. 1976. The regionalization of climate in the western United States. Journal of Applied Meteorology 15: 920-927. Muegc;ler, W F 1976. Type variability and succession in Rocky Mountain aspen. Pages 16-19 in Proceedings, Utilization and Marketing Tools for Aspen Manage- ment in the Rocky Mountains. USDA Forest Service, General Technical Report RM-29. . 1985. Vegetation associations. Pages 4.5-55 in N. V. DeByle and R. P Winokur, eds.. Aspen: ecology and management in the western United States. USDA Forest Service, General Technical Report RM-119. Neilson, R. E 1987. On the interflice between current ecological studies and the paleobotany of pinyon- juniper woodlands. Pages 93-98 in R. Everett, ed.. Proceedings, Pinyon-Juniper Conference. USDA Forest Service, General Technical Report INT-215. NiCHOL, A. A. 1937. The natural vegetation of Arizona. University of Arizona Technical Bulletin 68. NOAA. 1993. National Climatic Data Center. Federal Building, Asheville, NC. Rich, E. E., A. M. Johnson, and B. R. Baker, eds. 1950. Peter Skene Ogden's Snake Country journals: 1824—25 and 1825-26. The Hudson Bay Society, London. SAS. 1986. SAS-STAT user's guide (release 603). SAS Institute, Inc., Gary, NC. Tausch, R. J., N. E. West, and A. A. Nabi. 1981. Tree age and dominance patterns in Great Basin pinyon- juniper woodlands. Journal of Range Management 34: 259-264. Tausch, R. J., and N. E. West 1988. Differential establish- ment of pinyon and juniper following fire. American Midland Naturalist 119: 174-184. Thompson, D. 1916. David Thompson's narrative. J. B. Tyrrel, ed. The Champlain Society, Toronto, Ontario, Canada USDI-BLM. 1990. The juniper resources of eastern Oregon. USDA, Bureau of Land Management Infor- mation Bulletin OR-90-166. Van Pelt N., R. Stevens, and N. E. West 1990. Sun'ival and growth of immature Juniperus osteospenna and Piniis edulis following woodland chaining in central Utah. Southwestern Naturalist 35: 322-328. Vasek, F C. 1966. The distribution and tiixonomy of three western junipers. Brittonia 18: 350-372. West, N. E. 1984. Successional patterns and productivity of pinyon-juniper ecosystems. Pages 1301-1332 in Developing strategies for range management. Westview Press, Boulder, CO. . 1988. Intermountain deserts, shrub steppes, and woodlands. Pages 209-230 in M. B. Barbour and W D. Billings, eds.. North American terrestrial vegetation. Cambridge University Press, Cambridge, MA. 1995] Western Juniper Expansion 45 Young, J. A., and J. D. Budy. 1979. Historical use of Received 7 February 1994 Nevada's pinyon-juniper woodlands. Journal of Forest Accepted 8 June 1994 History 23: 113-121. Young, J. A., and R. A. Evans. 1981. Demography and fire history of a western juniper stand. Journal of Range Management 34: 501-506. Great Basin Naturalist 55(1 j, © 1995, pp. 4(i-57 RANGELAND ALPHA DIVERSITIES: HARVEY VALLEY, LASSEN NATIONAL FOREST CALIFORNIAl Raymond D. Ratliff2 Abstract. — Monitoring diversib,- usually begins by estimating alpha diversity of a plant community on a specific- site. The objectives of this study were to provide alpha diversity benchmarks and to determine whether rangeland com- munity basal cover characteristics explained variation in diversity estimates. Plant and surface component cover per- centages were estimated on 51 plots (representing four vegetation types) on the Lassen National Forest, CA. Each plot was sampled with 30 random, 102 basal point transects. Jackknife procedures were used to compute means and standard errors for Margalef's diversity' index (D,„), which stresses species richness, and Simpson's index (Dj, which stresses species dominance. Within vegetation tvpes, D,„ and D, did not rank all plots in the same order Highest D^, values occurred with the most species. Highest D^ values occurred with comparatively few species but more uniform cover. With either index, average diversity declined from the meadow to grassland to open shrub-grass to timber-bunchgrass t>'pes. All possible subset regressions of diversity on the basal cover characteristics were computed. Portions of the vari- ance accounted for by the best models were too low to allow prediction of D,„ and D,. The relation of alpha diversity to rangeland health is discussed. Kei/ worch: ecology, plant communities, Margalef's index, Simpson's index, monitoring, basal cover Biological diversity (hereafter called diver- sity) involves ecological processes, structures, and functions and may occur at any spatial scale (Society of American Foresters 1992). Diversity refers to variety and abundance; it is variety or multiformity — of different forms or kinds (Stein and Urdang 1966). There are alpha, beta, and gamma diversities (Whittaker 1972). Alpha diversity is the variety that occurs with- in a plant community of a specific site. A site or stand is defined as an individual unit that is homogeneous in vegetation, soil, topography, microclimate, and history (West 1993). Beta diversity is the variety of communities along a gradient (e.g., topography, soil acidity, or mois- ture regime) or on a given site through time. Gamma or large-scale diversity is the variety of plant communities, or the total number of species present, or both in a specific geo- graphic area (e.g., grazing allotment or water- shed). Diversity has two components, richness and evenness (Ludwig and Reynolds 1988, Magurran 1988). Richness refers to variety (numbers) of species, for example. Evenness refers to equality (abundance or numbers) of species botanical composition, for example. Diversity may or may not follow traditional concepts of succession and increase from pio- neer to climax plant communities or decrease with rangeland deterioration. Over large areas diversity may be higher if communities are at several serai stages than if the entire area is at a single serai stage. Within specific sites phys- ical/chemical factors or intense competition or both may work to reduce diversity (Odum 1959). Absence of an expected species may be due to fi-equent disturbances, a low immigration potential, an immature soil, or an inhospitable moisture regime (del Moral and Wood 1988). Nevertheless, because it may change with the kind of management, diversity should be assessed as part of range health evaluations. Diversity indices provide information that may not be immediately apparent from basic measures of the plant community such as cover and composition. High diversity of plant species is important in maintaining processes and flow pathways for energy and nutrients within and among communities. Higher diver- sity implies a greater number of occupied niches (Whittaker 1972). Protecting or enhancing diversit); or both, are goals commonly set by policy or law. West 'This iirtit'li- was written and prepared by U.S. governnK'iit employees on official time; it is therefb ^Pacific- Southwest Research Station, USDA Forest Service, 2081 E. Sierra, Fresno, CA 93710. I tlie pubhc domain and not subject to copyright. 46 1995] Alpha DivERSiri' 47 (1993) gave four reasons for having diverse plant communities: a sense of moral obligation to living things, an aesthetic appreciation of nature, economic benefits possible from them (e.g., the gene pool for cultivated crops), and the important array of services they provide (e.g., maintaining oxygen levels and cycling nutrients). A major cause of rangeland deterioration is selective grazing of preferred plants and sites in similar patterns each year (Hormay 1970). Even with conservative grazing, populations of preferred plants on preferred sites may dis- appear, thereby reducing the overall diversity of vegetation. If such populations are ecotypes (Odum 1959), the ability of the species to recapture site resources is reduced. Because nature abhors a vacuum, other species may increase or invade as those pre- ferred by livestock decrease in abundance (Dyksterhuis 1949). As a result, plant species diversity may be higher rather than lower under grazing, at least initially. As preferred species decrease and less preferred ones in- crease, their abundances tend to become more even (Dyksterhuis 1949). With continued deterioration, species not previously able to compete tend to invade and become established and thereby increase species richness. The new plant community, though possibly comprising more species that are more evenly abundant, may cover less total area, and higher diversity may be associated with greater amounts of bare soil. Increasingly, land managers are asked to monitor and determine change in diversity. Monitoring diversity usually starts with an esti- mation of alpha diversity for plant communities on specific sites. Such estimates are rare for rangelands. To derive the greatest benefit from monitoring efforts, managers must know what constitutes high and low diversity in given situa- tions. They need to know how diversity changes when other commonly estimated properties of the site change (e.g., litter cover and amount of bare soil). Seldom will examples of pristine or climax plant communities be available for developing diversity guides. Current plant communities represent the sums of all past influences. Current vegetation and site characteristics, therefore, must serve as benchmarks from which to develop guides and evaluate future change. The objectives of this study were (1) to pro- vide local rangeland managers with indices of alpha diversity from plant communities to use as guides of expected diversity for similar sites, and (2) to question whether variation in basal cover percentages of common and important indicators of rangeland health could explain variation in diversity. Although the findings are specific to the study area, it is hoped they may assist others dealing with questions of plant species diversity on range- lands. Methods Study Plots During 1964 and 1965, 51 plots were estab- lished on the Harvey Valley and neighboring grazing allotments of the Lassen National Forest, CA (Radifif et al. 1972). The plots were either 0.1 ha or 0.2 ha and unevenly distrib- uted among meadow (8), open grassland (13), open shrub-grass (12), and timber-bunchgrass (18) vegetation types. These plots were used for evaluating range condition (health) at Harvey Valley relative to the neighboring allotments. Meadows ranged from ephemeral lake sites with hardpans to deep, organically rich soil of drainage bottoms. Open grasslands included those dominated by shorthair sedge {Carex exserta) and those where shorthair sedge had been replaced by grasses. Open shrub-grass areas included silver sagebrush {Artemisia cana), black sagebrush (A. arhuscida), big sage- brush (A. tridentata), and bitterbrush {Purshia tridentata) subtypes. The timber-bunchgrass types were all in second-growth ponderosa pine {Pinus ponderosa). Some of them had bit- terbrush and big sagebrush along with grasses in the understory. Data Collection Data used to estimate alpha diversity on each plot were actual point contacts (hits) with plant bases or soil surface components (gravel, litter, rock, bare soil, and large woody debris) and shrub crown area. A hit on a shrub was recorded when a point contacted the shrub crown or was within its projected crown area at the soil surface. For each plot 3060 hits were recorded, consisting of 102 points (in regularly spaced 3-point quadrats) on each of 30 ran- domly placed transects. Points in a quadrat 48 Great Basin Naturalist [Volume 55 were at 23-cni eenters and projected vertical- ly. Within transects, quadrat spacing was either 0.6 ni or 0.9 m, depending upon plot width. Basal cover percentages (proportions of the surface occupied by different plants and surface components) were calculated from the hits and summarized (Radiff et al. 1972). Diversity Indices Two indices of diversitv were used: (1) Margalef's {D„, = (S - l)/ln N}, where S is the number of species and N is the total num- ber of individuals (hits) for all species and (2) Simpson's s (D^ = 1/D), where D = I {nj(nj- - 1)/ i = l N(N - 1)} and n^- is the number of individuals (here the percentage cover) of the ith species (Magurran 1988). D,^^ was selected for its sim- plicity and because it stresses the species rich- ness component. D^. was selected because it is well known and stresses the species evenness (dominance) component. In addition, these indices were selected because they do not require testing assumptions regarding the underlying distributions of species abundance. An overall estimate of diversity was com- puted for each plot using each index. Then 30 new diversity estimates were computed using the jackknife procedure. This procedure con- sisted of deleting each transect in turn from the data set. From each new estimate and the overall estimate, a pseudovalue (related form) was computed. From the pseudovalues, means and standard errors for the two indices were derived for each plot. Use of the jackknife pro- cedure to improve estimates of diversity and piovide a way of calculating confidence inter- vals was suggested by Magurran (1988). Basal Cover Relationships Contributions of basal cover of various char- acteristics to the variance in estimates of alpha diversity were examined. Characteristics for each plot were basal covers of grasses, grass- like herbaceous plants, forbs, shrubs, and soil surface components. All possible subset re- gressions of D,,^ and D^^ on the characteristics were computed using the Mallow's-Cp criteri- on of the REG procedure (a multiple linear regression program) of the SAS Institute, Inc. (1982). Subset regression models explaining most variation in the indices were selected for study. The Pearson correlation matrix was com- puted, using the correlations (CORR) module of SYSTAT (Wilkinson 1989), to help assess the influence of individual characteristics on the indices. Results Alpha Diversity Indices Diversity indices and basal cover values are available for all 51 plots. Here, only those plots within each vegetation type ranking lowest Table 1. Numbers of species, dominant species and percentage composition, and jackknifed means and standard errors (SE) for Margalef's and Simpsons diversitv' indices^ for vegetation tvpe^ benchmarks in 1964-65, Eagle Lake Ranger District, Lassen National Forest, CA. Dominant Diversi ity index Margalef" s Simpson'; S Veg. No. of Composi tion type species Species percentage Mean SE Mean SE MD 6 Eleocharis paltt.stris 52 1.0 0.2 2.8 0.2 19 Deschampsia caespitosa 62 4.0 0.4 2.4 0.2 14 Juncus balticu,s 16 2.3 0.2 9.9 0.7 GR 9 Carex exserta 71 1.6 0.2 1.9 0.1 9 C. exserta 78 1.6 0.2 1.6 0.1 19 C. exserta 46 3.2 0.3 3.8 0.3 11 Festiica idiihocnsis 26 2.0 0.3 5.8 0.3 SG 5 Artemisia tridcntata 97 0.8 0.2 1.0 0.0 17 A. arhusruhi 60 3.6 0.5 2.6 0.4 11 Leptodactijlon pun- benchmarks: (a) Elencharis ixihislris-, (h) Deschampsia caespitosa-, and {c) Jiincus balticus -dominated plots; Eagle Lake Ranger District, Lassen National Forest, CA. 50 Great Basin Naturalist [Volume 55 and highest for D„j and D^. are specihcalK' dis- cussed. Those plots are considered diversity benchmarks for their vegetation types in and near the Haney Valley allotment. Meadow. — D,^ in the meadows was lowest on a plot with just six species and demonstrat- ed the effect of lack of richness (Table 1). The site was an ephcnicral lake meadow (Fig. la) where dominant species covered 3.3% of the surface. Among the meadow plots, percentage litter cover was lowest and percentage bare soil was highest (Table 2). D,„ was highest, but D^ was lowest on a meadow with 19 species. That finding demon- strated the effect of good variety with uneven abundance. The site was a basin meadow, pos- sibly an ancient lake (Fig. lb). There the domi- nant species covered 5.7% of the surface. Only one species, among the others, contributed as much as 5% to the composition. Percentages of litter and bare soil were higher and lower, respectively, than averages for tlie meadow plots (Table 3). D^ was highest on a plot with 14 species. The site was a groundwater-fed meadow (Fig. Ic). Evenness in species abundance with mod- erate variety was demonstrated. Four species (including the dominant) each constituted more than 10% of the composition but less than 1% of the basal cover Only one species, among the others, contributed less than 1% to the composition. Total live plant cover was below average, but percentages of litter and soil cover were well above and below the averages, respectively. Grassland. — Both D„^ and D^ were lowest on grassland plots, with nine species (Figs. 2a, 2b), respectively. Shorthair sedge was the main contributor to the composition. In the case of Dj^^, three species each contributed 5% or more, and five species each contributed 1% or less. In the case of D^, only one species, other than shorthair sedge, contributed as much as 5% of the composition. For the plot with low D„^ the evenness component was better, litter cover was higher, and bare soil cover was lower than for the plot with low D^,. D,^^ was highest on a plot with 19 species (Fig. 2c). Shorthair sedge, Idaho fescue {Festiica idahoensis, 20%), and Sandberg bluegrass {Poa sandbergii, 9%) were main contributors to the composition. Sixteen species contributed less than 5% each. Among the grassland plots, this plot had the highest live plant cover and was well above average in litter cover and well be- low average in percentage of bare soil. Idaho fescue dominated the plot with high- est D^ (Fig. 2d). Four of the other 10 species present each made up more than 10% of the composition; two species each made up about 6%. While the evenness component of diversi- ty was good and total live plant cover was Table 2. Percentages of liasal cover for plant gronps and surface components for vegetation type benchmarks 1964-65, Eagle Lake Ranger District, Lassen National Forest, CA. Perc entage basal cover Plant groups' Surface components- Vegetation t\'pe'^ ,^1" gl bl sh dp Ip Gr Li Ro So Wo Nieadow LI 4.6 0.6 t^ 6,2 47.1 — 46.6 — 6,1 2.2 0.9 — 0,7 9,2 0.1 80.2 — 9.8 — 1..3 2.6 1.2 — — 5,1 — 90.3 0.1 4.5 — Open grassland 1.5 5.4 0.5 0.2 15.0 7,6 3,2 .57,1 — 17.1 — 0.7 4.7 0,4 0.2 12.0 6,0 6.0 .50,6 0.5 24.9 — 4.8 5.3 0,6 0.3 7,6 11,1 1.3 60,1 t 19.9 t 4.4 2,5 0,3 3.0 1,7 10,2 11.3 26.1 2.6 48.1 — Open shrub-grass 0.7 — — 25.4 16,2 26,1 18.9 28.4 — 10.4 — 1.8 0.5 0,8 4.6 8,2 7.6 4.0 .38.0 t 42.1 — 3.6 3.4 1,1 4.5 3,1 12.5 8.6 21.0 1.2 53.7 — Timl)er-bunchgrass 1.2 0.6 — 21.6 11,8 23.3 1.2 ,54.0 0.5 6.5 2.7 1.0 0.4 0.1 4.5 5,6 5.9 2.9 76.2 4.0 4.6 0.7 I.O 0.6 — 0.6 3,1 2.1 2.8 62.3 9.4 13.8 6.6 'gr = grasses, gl = grasslike herbaceous plants, bl = hroadleaf heibaceous plants (forbs), sli = shrubs, dp + bl + sh). ^Gr = gravel, Li = litter, Ro = rock. So = soil. Wo = large woody debris, ^Types follow Ratliffet al. (1972), and plot order is (he s.une as in Table 1. ■'t = less than 0.1% of basal cover deail attached eo\er. Ip = li\e plant co\er (g,r + gl 1995] Alpha Diversity' 51 Table 3. Average diversiW indices and percentages of basal cover for plant groups, and surface components by vege- tation type. Eagle Lake Ranger District, Lassen National Forest, CA, 1964-65. Diversity inde.x' Percentage basal cover Plant groups^ Surface components'^ Vegetation type D,„ Ds ■•-i' gl bl sh dp Ip Gr Li Ro So Wo Meadow Grassland Open shrub-grass Tiniber-bunchgrass 2.29 2.28 2.28 L75 4.38 3.63 2.78 2.39 2.1 3.3 1.6 1.1 3.6 3.6 1.3 1.1 0.8 1.2 0.6 0.1 0.8 14.4 7.7 0.3 5.6 10.7 4.0 6.4 8.9 18.0 10.0 0.2 9.3 7.2 2.5 77.2 39.0 30.2 60.2 0.1 1.5 0.2 3.0 15.8 35.7 33.8 16.4 t 3.9 'D„, = Margalef s index, D^ = Simpson s index. ^gr = grasses, gl = grasslike herbaceons plants, hi = broadlcaf herbaceous plant,s (forbs). + bl + sh). ''Gr = gravel, Li = litter, Ro = rock. So = soil. Wo = wood. = shrubs, dp = dead attached coven Ip = live plant cover (gr + ] above average, Idaho fescue covered only 3% of the surface and htter cover was well below but bare soil was well above average. Shrub-grass. — Both indices were lowest on an open shrub-grass plot where big sage- brush contributed over 95% of the composi- tion (Fig. 3a). Only one other species, bottle- brush squirreltail {Sitanion hystrix), made up as much as 1%, and only five species occurred on that plot. This finding demonstrates the effects of both low variety and low evenness on diversity. Among the shrub-grass plots, this plot was second highest in total live plant cover (nearly all sagebrush), highest in gravel cover, and lowest in bare soil. This suggests soil loss and formation of pavement. Black sagebrush dominated the plot with highest D„^ (Fig. 3b). Of the 17 species on that plot, 12 of them each contributed less than 3% of the composition. The plot was above aver- age in both litter and soil cover, but lowest in total live plant cover The plot with highest D^, (Fig. 3c) had just 11 species and was dominated by false phlox {Leptodacttjlon ptmgens). Five other species combined contributed nearly 62% of the com- position. Among the shrub-grass plots, this plot was well below average in litter cover but highest in bare soil. TiMBER-BUNCHGRASS. — Both indices were lowest (Fig. 4a) on a timber-bunchgrass plot with seven species. Bitterbrush contributed over 80% of the composition. Three species contributed 2% or more and three species contributed less than 1% of the composition. While total live plant cover was above aver- age, litter was near average and bare soil was well below average; there were few species, and they were unevenly abundant. This plot was similar in diversity to the shrub-grass plot with D„, and D^. both low. D„^ was highest on a plot with 13 species (Fig. 4b). Nine of them contributed 1% or less of the composition, thereby demonstrating that high evenness is not required when vari- ety is the main component of diversity' consid- ered. Big sagebrush dominated the understory and covered 4.3% of the surface. Litter cover was well above and bare soil was well below average for the timber-bunchgrass plots. By contrast, D^ was highest on a plot with just six species (Fig. 4c). Ross sedge {Carex rossii) contributed most of the composition (0.6% of the surface cover), three species con- tributed 15-26% each, and two species con- tributed 3% each, thereby demonstrating that high variety is not required when evenness is the main component of diversity considered. Percentages of soil and litter cover were near average for the timber-bunchgrass plots. Beta Diversity Indices Statistical comparisons of diversity among communities and vegetation types were not made. Nevertheless, average values for both indices declined from meadow to grassland to open shrub-grass to timber-bunchgrass tvpes (Table 3). Relative plot ranking (high to low diversity) depends on the index used, and inconsistent ranking by D,^^ and D^ was expected. Among the open shrub-grass and timber-bunchgrass types only two plots ranked the same, those with lowest diversity by both indices. Rankings by D,.,.j and D^ were the same for 3 of the 8 meadow plots and 2 of the 13 open grassland plots. Basal Cover Relationships Meadow and grassland plots had higher average diversity indices tlian open shrub-grass 52 Great Basin Natuiulist [Volume 55 b Fig. 2. Open grassland diNeisity Ijenchniaiks: (a, h, c) Carex exserta- and (d) Festiica i^ffl/ioensis-dominated plots; Eagle Lake Ranger District, Lassen National Forest, CA. or timber-bunchgrass plots, but lower average percentages of live plant cover (Table 3). Total live plant cover was largely ci property of shrub cover because projected crown hits were in- corporated into the data base. Significant portions of variances in the diversity indices (all 51 plots included) were accounted for by variation in percentages of some basal cover characteristics. Forty-seven percent of the variation in D„-, and 27% of the variation in Dj. were explained by the best models (Table 4). Dn, = a + grbi + glb2 + shb3 + Grl)4 + Sob5 + Wobg + error) and D^ = a + shb^ + Grb2 + error, where a, gr, gl, sh, Gr, So, and Wo are explained in Table 4; and the (bj)'s are the coefficients. Although gravel and bare soil were includ- ed in the model for D„^, they did not signifi- cantly correlate with D„^. Also, while in the model for D^, gravel was not significantly cor- related with D^. Individually, correlation with D,^^ was posi- tive for grasses (r = .471) but negative for grass- like plants (r = -.014), shrubs (r = -.320), and wood (r = -.348). Correlation of Dj. with shrubs was negative (r = -.507), also. 1995] Alpha Diversity 53 d Fig. 2. Continued. Discussion Alpha Diversity Many diversity indices are available to the land manager. Although a particular diversity index may be preferred, it is generally best to use one that stresses species richness and one that stresses evenness (dominance), such as D„, and Dj, respectively. Doing so allows the man- ager to consider both components of diversity. The richness component of diversity may in- crease at the expense of the evenness compo- nent, or vice versa. Also, those indices that stress richness and those that stress evenness tend to be poorly correlated (Magurran 1988). Beta Diversity Data used in this study represent single- time samples and were not designed to esti- mate beta diversity. Testing for differences in diversity using such data was not considered reliable (West and Reese 1991). Nevertheless, diversity indices for different but closely similar plots or communities, when computed by the same methods, should be nearly equal. With time or different treatment, wade divergence of the indices may occur. 54 Great Basin Naturalist [Volume 55 Fig. 3. Open shrub-grass diversity- benchmarks: (a) Artemisia tridentata-. (b) A. arbuscuki-, and (c) Leptodactyh. pun^ens-domuvdh'd plots; Eagle Lake Ranger District, Lassen National Forest, CA. 1995] Alpha Diversity 55 'Jr*^^? Fig. 4. Timber-bunchgrass diversih lienchmarks: (a) Ptirshki trklentata- (b) Artemisia tridentata-, and (c) Carex rossii -dominated plots; Eagle Lake Ranger District, Lassen National Forest, CA. 56 Great Basin Naturalist [Volume 55 Table 4. Best model multiple linear regre.ssion coeffieients, tests of siKnificance (T), and pn)l)al)ilities of signifieance (P) for Margalef's and Simpson's diversit>- indices; Eagle Lake Ranger District, Lassen National Forest, CA, 1964-65. Symbol Diversity index .\largalef\s Simpson s Variable Coeff".' T P CoeO: T F Constant a 2.696 8.436 .000 3.584 12.372 .000 Grasses gr 0.238 3.522 .001 Grasslike plants gl -0.161 -3.285 .002 Shrubs sh -0.036 -2.722 .009 -0.102 -4.194 .000 Gravel Gr -0.026 -1.505 .139 0.033 0.970 .337 Bare soil So -0.007 -1.197 .238 Wood Wo -0.128 -3.173 .003 'Regression coefficient Permanent plots represent a resource for assessing beta diversity responses to land man- agement practices. Although sampling a site to include within- and between-season variation is desirable, doing so is seldom possible, given time and monetaiy constraints. As an alternative, one might restrict sampling to times when selected species indicators are in specific phe- nologic stages (e.g., budding or flowering). Basal Cover Because of the usual dominance of a single species and because that species tends to occupy high proportions of an area, reductions in diversity indices with increases in shrub cover may be expected. Both diversity indices may be related posi- tively or negatively to characteristics of basal cover or to soil properties. Nevertheless, D,^^ was related to a greater number of characteris- tics than D^, suggesting that D„, may be the more desirable index for comparing plant com- munities of different sites or plant communities present through time on a given site. Conclusions For similar communities we can expect plant species diversity to be highest in the meadow and lowest in the pine-bunchgrass types. High and low values of Margalef's and Simpson's diversity indices are available for benchmark plots of different vegetation types in and near the Harvey Valley allotment. Diversity' indices for and averages among 51 plots are available by vegetation types. The influence of species richness on D„^ was clearly evident. D„, tended to be highest with the greatest numbers of species. Frequently that occurred when one species was clearly dominant and the others contributed little plant cover The inff uence of evenness in abun- dance on D^ was clearly evident. D^ tended to be highest when species were more or less evenly abundant. Frequently that occun-ed with relatively few species. Few species with one contributing a high percentage of the compo- sition produced low values of both indices. Situations with many species, all contributing equally to the composition, were not encoun- tered, but such situations should give high val- ues of D,^^ and Dj,. Higher diversity did not necessarily mean greater plant cover or greater forage cover or more litter or less bare soil. While some rela- tionships between diversity and basal cover values were significant, coefficients of deter- mination were too low to allow either of the best models to be used to predict diversity'. Neither index should be relied on apart from other information for evaluating range- land health. Nevertheless, plants capture the sun's energy and pass it as food for other orga- nisms, and a high degree of plant diversity may equate with high diversity in other parts of the biotic community. Literature Cited DEL xMoiUL, R., and D. M. Wood. 1988. The higli elevation flora of Mount St. Helens, Washington. Madrofio 35: 309-319. Dvksterhuls, E. J. 1949. Condition and management of range land based on quantitative ecolog>-. Journal of Range Management 2: 104-115. HORNLAV, A. L. 1970. Principles of rest-rotation grazing and multiple-use land management. Training te.xt 4(2200). USDA, Forest Service, Washington, DC. 26 pp. Ludwk;, J. A., .\ND J. F REYNf)LDS. 1988. Statistical ecology. John Wiley & Sons, New York, NY. 1995] Alpha DivERSiTi' 57 Magurran, a. E. 1988. Ecological diversity and its mea- surement. Princeton University Press, Princeton, NJ. Odum, E. R 1959. Fundamentals of ecology. W. B. Saunders Co., Philadelphia, PA. Ratliff, R. D., J. N. Reppert, .vnd R. J. McConne.n. 1972. Rest-rotation grazing at Harvey Valley . . . range health, cattle gains, costs. USDA, Forest Service, Pacific Southwest E.xperiment Station. Research Paper PS\V-77. 24 pp. SAS Institute, In'c. 1982. SAS user's guide; statistics. 1982 edition. SAS Institute, Inc., Gary, NC. Society of American Foresters. 1992. Biological diver- sity in forest ecosystems, a position of the Society of American Foresters. Journal of Forestn 90: 42—43. Stein, J., and L. Urdang, eds. 1966. The Random House dictionary of the English language. Random House, New York, NY. West, N. E. 1993. Biodiversity of rangelands. Journal of Range Management 46: 2-13. West, N. E., and G. A. Reese. 1991. Gomparison of some methods for collecting and analyzing data on above- ground net production and diversity of herbaceous vegetation in a northern Utah subalpine context. Vegetatio 96: 145-163. Whittaker, R. H. 1972. Evolution and measurement of species diversity. Taxon 21: 213-251. Wilkinson, L. 1989. SYSTAT: the system for statistics. SYSTAT, Inc., Evanston, IL. Received 26 July 1993 Accepted 26 May 1994 Great Basin Naturalist 55(1), © 1995, pp. 58-65 EFFECTS OF SALINITY ON ESTABLISHMENT OF POPULUS FREMONTII (COTTONWOOD) AND TAMARIX RAMOSISSIMA (SALTCEDAR) IN SOUTHWESTERN UNITED STATES Patrick B. Shafroth', Jonathan M. Friedman', and Lee S. Ischinger^ Abstract. — The e.xotic shrub Tatnarix minosissimu (saltcedar) has replaced the nati\t' I'opulus jninontii (cottonwood) along many streams in southwestern United States. We used a controlled outdoor experiment to examine the influence of river salinity on germination and first-year survival of P. fremontii van wislizenii (Rio Grande cottonwood) and T. ramosissima on freshly deposited alluvial bars. We grew both species from seed in planters of sand subjected to a declin- ing water table and solutions containing 0, 1, 3, and 5 times the concentrations of major ions in the Rio Grande at San Marcial, NM (1.2, 10.0, 25.7, and 37.4 meq l-l; 0.11, 0.97, 2.37, and 3.45 dS m-i). Germination of P. fremontii declined by 35% with increasing salinity (P = .008). Germination of T. ramosissima was not affected. There were no significant effects of salinity on mortality' or above- and belowground growth of either species. In laboratory tests the same salini- ties had no effect on P. fremontii germination. P. fremontii germination is more sensitive to salinity outdoors than in cov- ered petri dishes, prolwbly because water scarcity resulting from e\aporation intensifies the low soil water potentials associated with high salinity. River salinity appears to play only a minor role in determining relative numbers of P. fre- montii and T. ramosissima seedlings on freshly deposited sandbars. However, over many years salt becomes concentrat- ed on floodplains as a result of evaporation and salt extrusion from saltcedar leaves. T. ramosissima is known to be more tolerant of the resulting extreme salinities than P. fremontii . Therefore, increases in river salinities could indirectly con- tribute to decline o{ P. fremontii forests by exacerbating salt accumulation on floodplains. Key words: exotic species, Tamarix ramosissima, Populus fremontii, river salinity, seedling estahlisliment, Rio Grande, riparian vegetation, Bosque del Apache National Wildlife Refuge. In the last century the exotic shi-ub saltcedar {Tamarix ramosissima Ledebour) has spread throughout southwestern United States, where it now dominates many riparian ecosystems (Bowser 1958, Robinson 1965). In many areas T. ramosissima has replaced stands dominated by the native Fremont cottonwood {Populus fremontii Wats.; Campbell and Dick-Peddie 1964, Ohmart et al. 1977), decreasing the habi- tat of Neotropical migrant birds (Anderson et al. 1977, Cohan et al. 1978) and altering fluvial processes (Graf 1978, Blackburn et al. 1982). Understanding the factors controlling estab- lishment of T. ramosissima and P. fremontii can aid in managing these species. Successhil invasion by Tamarix in the South- west has been attributed to many factors. Much of the early spread probably resulted from the coincidental timing of clearing of P. fremontii stands by early settlers and the availability of Tamarix seed (Campbell and Dick-Peddie 1964, Harris 1966, Horton and Campbell 1974, Ohmart et al. 1977). Subsequent spread resulted largely from effects of damming and channelizing southwestern watercourses. Reductions in the magnitude of high flows and associated reductions in channel movements decreased the formation of bare, moist alluvial bars, which provide ideal P. fremontii seedling habitat (Ohmart et al. 1977, Stromberg et al. 1991). Smaller peak flows have also reduced leaching of salts from floodplain soils (Busch and Smith in press), perhaps favoring the salt- tolerant Tamarix (Everitt 1980, Brotherson and Winkel 1986, Jackson et al. 1990). Flow regulations that have altered the historical timing of peak flows may have inhibited P. fre- montii regeneration because of its short period of seed dispersal and viability in early summer (Horton 1977, Everitt 1980), but they have enhanced Tamarix regeneration because of its abundant seed production throughout the growing season (Merkel and Hopkins 1957, Tomanek and Ziegler 1962, Wanen and Turner 1975, Horton 1977). Finally, successful inva- sion of T ramosissima has been attributed to its superior ability to resprout following fire (Busch and Smith 1993). 'National Biological Siii-vcy, Miclcoiitinenf Ecological Science Center, Fort Collins, CO 80525-3400. 58 1995] SALiNiTi' Effects on Populus and Tamarix 59 We conducted experiments to examine the influence of river salinity on germination, sur- vival, and growth of Popuhis fremontii var wis- lizenii (Rio Grande cottonwood) and T. ramo- sissima on freshly deposited alluvial bars, the principal habitat for seedling establishment of both species. Field observations have suggest- ed that P. fremontii is more negatively affected by high salt concentrations than T. ramosissi- ma (Brotherson and Winkel 1986, Anderson 1989). Laboratoiy studies have confirmed this difference by exposing seedlings and cuttings of these species to varying concentrations of NaCl and CaCl2 (Jackson et al. 1990, Siegel and Brock 1990). Two factors potentially con- found the relationship of results of laboratory studies to field conditions. First, the mix of salts found in riparian ecosystems typically includes many constituents other than Na, Ca, and Cl. In many plants, salinit>' effects result from toxicity of specific ions as opposed to osmotic stress (Greenway and Munns 1980). Second, moisture availability is lower and more variable in the field than in these labora- tory studies. This factor is important because low soil water potential caused by high salinity is exacerbated by low soil moisture content. We addressed these concerns by exposing T. ramosissima and P. fremontii seedlings to four different concentrations of a mix of salts designed to mimic ion concentrations in the Rio Grande. The experiment was conducted outdoors in planters subjected to a controlled water-table drawdown. Experimental condi- tions were designed to simulate alluvial bars along the Rio Grande in central New Mexico, where once-extensive P. fremontii forests have largely been replaced by T. ramosissima thick- ets (Campbell and Dick-Peddie 1964). Our outdoor experiments were supplemented by studies of germination under similar salinity treatments in the laboratoiy. Methods Seedling establishment experiments were conducted outdoors in 1993 near Fort Collins, CO, at latitude 40° 35' north, longitude 105° 5' west, and elevation 1524 m. Twelve 122 x 92- cm (diameter X depth) epoxy-lined steel tanks contained six 30 X 100-cm planters made of PVC pipe. Holes 1.26 cm in diameter were drilled into the lower 10 cm of each planter to allow water exchange, and the planters were filled to 92 cm with washed coarse sand (approximately 6% gravel [>2000 fim], 78% sand [> 300-2000 /am], 16% fine sand [> 75-300 ^im], and <1% silt and clay). Four salinity treatments were each replicat- ed in three tanks (12 tanks total). Each tank contained three planters of P. fremontii var. wislizenii and three of T. ramosissima. Thus, the experimental unit for each species was a group of three planters within a tank. To avoid pseudoreplication, responses were measured as the mean value of the three planters. The results for the two species were analyzed as separate, completely randomized experiments with four treatments and three replicates per treatment. The tanks were filled with water from the Cache la Poudre River (a snowmelt stream low in dissolved solids), and solutions containing multiples (0, 1, 3, and 5 times) of the mean con- centration of all major ions in the middle Rio Grande were made. These four solutions con- stitute treatments Ox, Ix, 3x, and 5x. Mean ion concentrations were derived from eight mea- surements from the conveyance channel at San Marcial, NM, between October 1989 and September 1991 (U.S. Geological Survey 1991, 1992). The following salts were added to make treatment Ix: 309.9 mg h^ CaS04*2H20; 302.4 mg 1-1 NaHCOg; 122.0 mg H MgCf2*6H20; 70.1 mg 1-1 NaCl; 13.9 mg l-l K2S04.''Because the coarse sand substrate was low in nutrients (c£ Segelquist et al. 1993), 15 mg 1-1 of Fisons Technigro fertilizer (16% N, 17% I^ 17% K) was added to every tank. At the time of planting and for 1 wk there- after, the water level was 10 cm below the soil surface. A 3.5-cm-week-l drawdown rate was applied for the remainder of the growing sea- son (17 June to late September). Water-table drawdowns are associated with summer declines in discharge along western streams. The 3.5-cm-week-l drawdown rate was select- ed because a previous study (Segelquist et al. 1993) indicated that it is within the optimal range for establishment and growth of plains cottonwood {Populus deltoides ssp. monilifera). Flowering panicles of T. ramosissima were collected on 17 May at the Bosque del Apache National Wildlife Refuge (latitude 33° 46' north, longitude 106° 54' west, elevation 1375 m). The panicles were air-dried for 48 h to enhance opening of seed capsules. Collected material was sifted through a series of soil 60 Great Basin Naturalist [Volume 55 screens until clean samples of seeds were ob- tained. Catkins of P. fremontii were collected at the Bosque del Apache on 1 June. The cat- kins were air-dried for 72 h to enhance open- ing of seed capsules. Capsules were placed between soil screens and seeds were separat- ed from the cotton and capsules using forced air. Seeds of both species were sealed in plas- tic containers and refrigerated at 5°C (Zasada and Densmore 1977). On 10 June, 100 P. fre- montii seeds were planted in each of three planters per tank, and 200 T. ramosissirna seeds were planted in each of the other three planters. Electrical conductivity (EC) and tempera- ture were measured using a Yellow Springs Instrument Co., Inc., Model 33 S-C-T meter, and pH was measured using a Corning 105 hand-held pH meter in conjunction with a Coming ATC temperature probe and a Coming general purpose combination electrode. EC was measured weekly in every tank begining 12 June (17 measuring dates). Whenever EC was measured, a representative water temper- ature for that day was determined by averag- ing the temperature values from hve randomly selected tanks. All EC measurements were corrected for temperature and reported at 25°C. Fourteen weekly measurements of pH were made beginning 30 June. On 16 June, 14 July, 18 August, and 17 September, water sam- ples from one randomly selected tank per treatment were analyzed to determine con- centrations of Ca, Mg, Na, K, CO3, HCO3, Cl, SO4, and NO3. Ca, Mg, Na, and K were deter- mined by inductively coupled plasma emis- sion spectroscopy (ICP; EPA method 200.0, United States Environmental Protection Agency 1983); CO3 and HCO3 were deter- mined by titration (EPA method 310.1, United States Environmental Protection Agency 1983); Cl, SO4, and NO3 were determined by ion chromatography. Concentrations are reported in meq 1~^ to facilitate comparison of our solutions to solutions in other studies and because meq 1~^ can be related easily to elec- trical conductivity, which is commonly report- ed in the context of salinity studies. On 29 September 1993 (day 112) we mea- sured the shoot length of every living seedling. We harvested all live seedlings in early October To harvest, we lifted a planter and laid it hori- zontally in a water-filled basin. The planter was then slowly lifted upside down, leaving the substrate column and seedlings in the basin. We gently separated seedlings from the sand and water and measured total length of every harvested seedling. Mean root lengths were determined by subtracting the mean shoot length for a planter from the mean total length in that planter Roots and shoots were separated for both species, and P. fremontii leaves were stripped from the stems. Roots, shoots, and leaves were dried at 60 °C for 72 h and weighed. One-way analysis of variance (SAS Institute, Inc. 1990) was used to assess the significance of treatment differences within the tvvo species for five variables: percent of planted seeds alive at the end of the experiment ("end-of-season survival"), shoot length, root length, per-plant aboveground biomass, and per-plant root bio- mass. For all variables the mean value of the three planters in a tank was the unit of analysis. The arcsine transformation was applied to end- of-season survival values to meet die equal vari- ance assumption (Snedecor and Cochran 1980). Data from the Colorado Climate Center were used to determine the difference be- tween precipitation and open-pan evaporation (adjusted with pan coefficient = 0.73) for the period 1 June-30 September 1993 in Fort Collins. Evaporation at Fort Collins exceeded precipitation by 26.2 cm during this period. The same calculation was made for the Bosque del Apache using data from the Western Regional Climate Center for the years 1975 through 1990. Precipitation data are from the Bosque del Apache National Wildlife Refuge, and open-pan evaporation data are from Socorro, NM (latitude 34°5' north, longitude 106° 53' west, elevation 1399 m; pan coefficient = 0.73). Growing-season evaporation at the Bosque del Apache exceeded precipitation by an average of 40.6 cm; n = 16, maximum = 51.0 cm, and minimum = 32.3 cm during these 16 years. We performed laboratoiy gemiination exper- iments in January 1994. Five 25-seed repli- cates of five salinity treatments were com- pletely randomized for both T. ramosissirna and P. fremontii. Seeds were sowed in 7.5-cm petri dishes containing a Whatman #3 filter and 7 ml of a treatment solution. Petri dishes were placed in a Percival Model 1-35 biological incubator after sealing the dish tops with Para- film. Temperature in the incubator was 20 °C throughout the experiment, and petri dishes were exposed to 16 h of light and 8 h of dark- ness each day. Four of the treatment solutions 1995] SALiNiTi' Effects on Populus and Taaiarix 61 were the same as those used in the estabhsh- ment experiment (0, 1, 3, and 5 times the con- centration of the Rio Grande at San Marcial, NM); the fifth sokition contained 7 times the concenti'ation of the Hio Grande. Genninants in every petri dish were counted after seven days. A seed was considered gemiinated if it exliibited expanded cotyledons and an elongated radicle. The arcsine transformation was applied to per- cent germination values to meet the equal variance assumption, and one-way analysis of variance was performed on the transformed values (SAS Institute, Inc. 1990). When germi- nation equaled 100%, the proportion was counted as (n - 0.25)/n, where n = the num- ber of seeds planted (Snedecor and Cochran 1980). Results EC and pH in the tanks varied little within treatments over the course of the experiment (Table 1). Mean temperature in the tanks was 21.7°C (standard en-or = 0.8, n = 17). Concen- trations of measured chemical constituents in different treatments did not increase propor- tionally to the quantities of salt originally added, indicating that salts (especially CaC03) precipitated at higher concentrations (Table 1). Nevertheless, concentrations increased across treatments, with total concentrations ranging from 0.7 meq h^ (0.11 dS m"^) in treatment Ox to 37.4 meq \~^ (3.45 dS m~l) in treatment 5x (Table 1). For P. fremontii there was a significant treatment effect (F = .003) on end-of-season sui-vival, but not on any of the four measured growth variables (Table 2). End-of-season sur- vival was negatively associated with increasing salinity: survival was greatest in treatment Ox and lowest in treatment 5x. Because the end- of-season sui"vival variable combines germina- tion and mortality, we analyzed the arcsine- transfomied number of seedlings 7 d after plant- ing (germination), and the arcsine-transformed difference between germination and end-of- Table 1. Chemical analysis of tank water for four treatments in the outdoor establishment experiment in Fort Collins, CO. For ion concentrations (n = 4), minimum and ma.\imum values are presented in parentheses below treatment means. For electrical conductivity' {n = 51) and pH (n = 42), means ± 1 standard error are presented. Treatment Factor Ox Ix 3x 5x Ca (mmol h^) 0.36 (0.20, 0.52) 1.82 (1.71, 2.00) 4.02 (3.49, 4.83) 4. .54 (3.02, 7.02) Mg (mmol 1-1) 0.11 (0.08,0.16) 0.60 (0.46, 0.75) 1.65 (1.47, 1.97) 2.62 (2.28, 2.97) Na (mmol l^) 0.17 (0.09, 0.28) 4.85 (4.41,5.11) 13.87 (11.91, 15.49) 22.24 (19.33, 24.65) K (mmol 1-1) 0.08 (0.06, 0.09) 0.26 (0.20, 0.34) 0.51 (0.44, 0.,55) 0.79 (0.72, 0.90) HCO3 (mmol 1-1) 1.04 (0.62, 1.44) 3.92 (3.24, 4.44) 8.34 (7.29, 9.96) 9.60 (5.87, 15.74) CI (mmol 1-1) 0.10 (0.07,0.14) 2.47 (1.88, 2.82) 7.10 (6.96, 7.31) 12.12 (10.88, 13.21) SO4 (mmol 1-1) 0.04 (0.04, 0.05) 1.66 (1.32, 1.86) 5.06 (4.76, 5.32) 7.73 (7.13, 8.33) NO3 (mmol 1-1) 0.03 (0.002, 0.08) 0.03 (0.006, 0.09) 0.05 (0.01, 0.08) 0.06 (0.02,0.15) Total cations (meq l-i) 1.2 (0.7, 1.6) 10.0 (9.2, 10.8) 25.7 (23.8, 26.7) 37.4 (34.5, 41.5) EC (dS m-i) 1.09 ±0.03 0.97 ±0.11 2.37 ±0.23 3.45 ± 0..39 pH 7..54 ± 0.03 8.10 ±0.02 8.29 ± 0.02 8.05 ± 0.03 62 Great Basin Naturalist [Volume 55 T.'VBLE 2. Survival and growth oi P(>))iilus frcmontii and Tainarix raiiio.sis.siiim se( ty treatnient.s for one growing sea.son outdoors in Fort Collins, CO. Higli and low ses below the treatment means in = 3). Treatment effeets were anaKzed 1)\ con Survival .\.\0\'A was performed on arcsine-transformed data. 'dlings e.\posed to four different salini- replicate means are given in parenthe- ipleteK randomized one-way ANOVA. Species Variable Trc; :itinent (K 1.x 3x 5.\ Cottonwood Survival (% of planted seed) 57.0 (50.0, 63.0) 49.3 (45.7, 54.0) 46.6 (41.0,51.0) 29.0 (20.7, 35.0) Shoot height (mm) 33.9 ■ (32.8, 34.5) 36.3 (.34.5, 38.5) 39.6 (.36.5, 43.9) .38.3 (34.7, 40.8) Hoot length (mm) 239.2 (227.1,258.4) 280.9 (257.8, 309.3) 286.9 (253.6,311.7) 247.4 (206.3, 274.6) Per-plant shoot biomass (mg) 14.1 (13.7, 14.4) 14.6 (11.2, 16.6) 21.4 (18.9, 25.8) 19.8 (14.3, 25.9) Per-plant root biomass (mg) 26.8 (21.2, 35.5) 19.6 (16.4,21.3) 31.8 (21.6, 43.2) 31.2 (17.2, 42.9) Saltcedar Sunival (% of planted seed) 42.3 (29.5, 51.6) 37.8 (33.8, 42.0) 37.3 (31.8, 40.8) 29.5 (22.8, 35.2) Shoot height (mm) 18.1 (17.3, 18.8) 17.7 (15.5, 19.8) 18.2 (15.6, 22.2) 18.3 (18.2, 18.3) Root length (mm) 174.4 (166.4, 184.9) 173.6 (154.8, 192.9) 179.0 (128.1,243.6) 162.0 (147.2, 169.5) Per-plant shoot biomass (mg) 5.5 (4.8, 6.2) 5.5 (4.1,6.4) 6.3 (4.3, 9.6) 6.2 (5.8, 6.4) Per-plant root biomass (mg) 7.7 (7.1. 8.9) 7.3 (5.5, 9.2) 9.9 (7.0. 14.7) 9.5 (7.6. 12.1) 11.4 2.6 2.1 2.8 1.1 1.6 0.04 0.15 0.22 0.74 .003 .13 .17 .11 .41 .26 .99 .92 .88 .56 season sui-vival (mortality). There was a signif- icant treatment effect on germination (F = .008), but not on mortality [P = .45), indicat- ing that the effect on end-of-season survival was predominantly due to lower germination at higher salt concentrations. For T. ramosimma there were no significant treatment effects (Table 2). Although P. fremontii germination in out- door tanks was significantly decreased at high salinity, laboratory germination was not simi- larly affected even at seven times the salinity of the Rio Grande, total concentration 48.4 meq \~^ (4.56 dS m"^; Table 3). There was a sig- nificant positive effect of increasing salinity on T. ramosissima germination (P = .03) (Table 3). Discussion The absence of a negative effect of salinity on P. fremontii germination in the laboratory at concentrations as high as 48.4 meq 1~^ (4.56 dS m~l) is consistent with results of earlier studies. Jackson et al. (1990) found that P fre- montii germinated in the laboratoiy at salini- ties of 0, 27, and 106 meq 1~^ using a mixture of NaCl and CaCU, but not at 319 meq h^ or above. Siegel and Brock (1990) observed high- er percent germination of P. fremontii in the laboratoiy in NaCl solutions of 0, 25, and 50 meq h^ than at 100 meq \~^ and above. There- fore, P. fremontii is no more sensitive to the mix of salts present in the Rio Grande than to NaCl and CaCl2 solutions of equal strength. Tests at higher salinities with the same ionic ratios were not possible with our Rio Grande mix because of low solubilities of some of the constituent salts. The decrease in T. ramosissi- ma germination at low salinit}' in the laborato- ry (Table 3) is consistent with the finding by Jackson et al. (1990) that germination increas- es between 0 and 106 meq 1~^. Our results indicate that a given water salin- ity may negatively affect germination of P. 1995] SALiNiTi' Effects on Populus and Tamarix 63 Table 3. Percent germination of Populus fremontii and Tamarix ramosissitna seedlings exposed to five salinity treat- ments in covered petri dishes. High and low replicate values are given below the treatment mean {n = 5). Treatment effects were analyzed by completely randomized one-way ANOVA using arcsine-transformed data. Species Treatment Ox Ix 3x Cottonwood Saltcedar 90.4 (80.0, 100.0) 69.6 (60.0, 88.0) 96.0 (92.0, 100.0) (56.0, 80.0) 96.0 (92.0, 100.0) 78.4 (68.0, 92.0) 92.8 (84.0, 96.0) 84.8 (76.0, 92.0) 96.0 (92.0, 100.0) 84.0 (76.0, 92.0) 1.2 3.3 .35 .03 fremontii seeds under ambient conditions but not under laboratory conditions. This may have resulted from an interaction between the effects of salinity and soil moisture content, or from vapor-pressure deficit differences. In outdoor planters, but not laboratoiy petri dish- es, evaporation of water may have resulted in lower soil moisture and higher salt concentra- tion at the soil surface. These factors would both tend to reduce soil water potential, thereby increasing plant water stress. Because the difference between evaporation and pre- cipitation is somewhat greater at the Bosque del Apache than in Fort Collins, the effect of salin- ity might be stronger at the Bosque, especially in dry years. Finally, greater vapor-pressure deficits in the field relative to the laboratory may have exacerbated plant water stress. Salinity appears to be a relatively minor fac- tor regulating numbers of P. fremontii and T. romosissinia seedlings on freshly deposited sandbars along the Rio Grande. The only signif- icant effects of increasing salinity were a small decrease in P. fremontii germination in out- door planters and a small increase in T. ramo- sissima germination in the laboratory. There were no significant effects on survival after germination or above- or belowground growth for either species, even at water salinities sev- eral times that of the Rio Grande. The presence of abundant seedlings of P. fremontii and T. ramosissima on sandbars along the Rio Grande in most years is consistent with our results. Although salinity may play only a minor role in the colonization of newly deposited alluvial bars by T. ramosissima and P. fremontii, this factor can become more important over time. Over many years salt becomes concentrated on some floodplains as a result of evaporation and salt extrusion from T. ramosissima leaves. EC readings as high as 10.0 dS m~^ have been reported in floodplain sediment at the Bosque del Apache (John Taylor, Bosque del Apache National Wildlife Refuge, personal communica- tion), and soil salinity levels as high as 60,000 mg 1~1 occur on floodplain sites along the lower Colorado River (Jackson et al. 1990). Soil EC above 2.0 dS m"^ can reduce the growth of P. fremontii pole plantings (Anderson 1989). T. ramosissima has been shown to be less susceptible than P. fremontii to many of the negative effects of higher salinities (Brotherson and Winkle 1986, Jackson et al. 1990). Tamarix species avoid harmful effects of salts through extrusion from leaves and cellu- lar compartmentation (Berr\' 1970, Kleinkopf and Wallace 1974, Waisel 1991). Our results could be applied to efforts to revegetate riparian areas from seed. Riparian revegetation in the Southwest has largely con- sisted of planting poles or potted shoot cut- tings. Although these approaches have been successhil in some areas (Anderson et al. 1990), they can cost up to $10,000 per hectare (Ohmart et al. 1988). Furthermore, they require the destruction of parts of existing trees, and often entire trees or stands. Finally, these approach- es may require importing cuttings or poles adapted to different site conditions. One alter- native is regeneration of native cottonwoods and willows using natural seedfall (Friedman 1993, John Taylor personal communication). This approach generally involves clearing and irrigating an area so that seeds from nearby trees can colonize it. Our results suggest that water as saline as 37.4 meq 1~^ (EC 3.45 dS m~l) can be used to grow P. fremontii from seed on sand (Tables 1, 2). However, care must be taken to prevent long-term salt accumula- tion through evaporation (e.g., through period- ic flooding to flush salts) and to avoid sites with preexisting high salinity. Use of water with low salinity can help prevent negative effects on P. fremontii and may decrease the 64 Great Basin Naturalist [Volume 55 germination rate of T. ramosis.mna (Table 3). However, in a restoration effort along the Cache la Poudre Rixer, T. ratn(>.sissi)na became established in large numbers along with P. del- toides in spite of use of water of low salinity (Douglas Gladwin, National Biological Survey, personal communication). Therefore, low salin- it\' will not prevent establishment of T. ramo- sissima from seed when moisture, a bare sedi- ment, and a seed source are present. Acknowledgments G. T Auble, D. E. Busch, and an anonymous reviewer provided constructive comments on the manuscript. We thank E. R. Auble, G. T Auble, J. Back, E. D. Eggleston, M. Jordan, and M. L. Scott for invaluable assistance with the experiments. D. Smeltzer, B. Upton, and the Colorado Division of Wildlife generously provided access to the Bellvue-Watson Fish Rearing Unit where the outdoor experiment was conducted. T Kem and P Soltanpour pro- vided useful advice regarding the salinity treatments. Concentrations of ions in solutions were measured by the Soil, Plant and Water Testing Laboratoiy at Colorado State University, Fort Collins, CO.' Literature Cited Anderson, B. W. 1989. Research as an integral part of revegetation projects. Pages 413—119 in D. L. Abell, technical coordinator. Proceedings of the California Riparian Systems Conference: protection, manage- ment, and restoration for the 1990s. USDA Forest Service General Technical Report PSW-110. Berkeley, CA. ANDER.SON, B. VV., A. HiGGINS, AND R. D. Ohmart. 1977. Avian use of saltcedar communities in the lower Colo- rado River Valley. Pages 128-136 in R. R. Johnson and D. A. Jones, technical coordinators. Importance, Presei^vation, and Management of Riparian Habitat: a symposium. USDA Forest Service General Technical Report RM-43. Fort Collins, CO. Anderson, B. VV, E. R. Miller, and J. E Washlngton. 1990. Revegetation on the Kern River preserve 1986-1989. Research report prepared for The Nature Conservancy and California Department of Fish and Game. Revegetation and Wildlife Vlanage- ment Center, Inc., Blythe, CA. 19 pp. Berry, W. L. 1970. Characteristics of salts secreted by Tamarix aphylhi. American Journal of Botany 57: 1226-1230. Blackburn, W. H., R. W. Knight, and J. L. Schuster. 1982. Saltcedar influence on sedimentation in the Brazos River. Journal of Soil and Water Conserva- tion, 37: 298-301. Bowser, C. W 1958. Introduction and spread of the un- desirable tamarisks in the Pacific southwestern sec- tion oi the United States and comments concerning the plants influence upon the indigenous vegetation. Pages 12—16 in Symposium on Phreatophytes. Ameri- can Geophysical Union, Sacramento, CA. Brotherson, J. D., AND V. WiNKEL. 1986. Habitat rela- tionshijis o{ saltcedar {Tamarix ramosissiina) in cen- tral Utah. Great Basin Naturalist 46: 53.5-541. BUSGH, D. E., AND S. D. Smith. 1993. Effects of fire on water and salinity relations of riparian woody taxa. Oecologia 94: 186-194. . In press. Decline, persistence, and competition of woody ta.xa in riparian ecosystems of the southwest- em U.S. Submitted to Ecological Monographs. Campbell, C. J., and VV. A. Dick-Peddie. 1964. Compari- son of phreatophyte communities on the Rio Grande in New Mexico. Ecology 45: 492-502. Cohan, D. R., B. W Anderson, and R. D. Ohmart. 1978. Avian population responses to salt cedar along the lower Colorado River. Pages 371—382 in R. R. Johnson and J. F McCormick, technical coordinators. Strategies for Protection and Management of Flood- plain Wetlands and Other Riparian Ecosystems: pro- ceedings of the symposium. USDA Forest Service General Technical Report WO-12. Washington, DC. EvERITT, B. L. 1980. Ecology of saltcedar — a plea for research. Environmental Geolog}' 3: 77-84. Friedman, J. M. 1993. Vegetation establishment and channel narrowing along a Great-plains stream fol- lowing a catastrophic flood. Unpublished doctoral dissertation. University of Colorado, Boulder. Graf, W L. 1978. Fluvial adjustments to the spread of tamarisk in the Colorado Plateau region. Geological Society of America Bulletin 89: 1491-1501. Greenway, H., and R. Munns. 1980. Mechanisms of salt tolerance in nonhalophytes. Annual Review of Plant Physiolog>'31: 149-190. Harris, D. R. 1966. Recent plant invasions in the arid and semi-arid southwest of the United States. Annals of the Association of American Geographers 56: 408-422. HORTON, J. S. 1977. The development and peipetuation of the permanent tamarisk type in the phreatophyte zone of the Southwest. Pages 124-127 in R. R. Johnson and D. A. Jones, technical coordinators. Importance, Presei'vation, and Management of Ripar- ian Habitat: a symposium. USDA Forest Service General Technical Report RM-43. Fort Collins, CO. HoRTON, J. S., and C. J. Campbell. 1974. Management of phreatophyte and riparian vegetation for maximum multiple use values. USDA Forest Service Research Paper RM-1 17. 23 pp. Jackson, J., J. T. Ball, and M. R. Rose. 1990. Assessment of the salinity tolerance of eight Sonoran desert riparian trees and shrubs. Find report, U.S. Bmeau of Reclamation Contract No. 9-CP-30-0717(). Biological Sciences Center, Desert Research Institute, Univer- sity of Nevada System, Reno. 102 pp. Kleinkopf, G. E., and A. Wallace. 1974. Physiological basis for salt tolerance in Tamarix ramosissima. Plant Science Letters 3: 157-163. Merkel, D. L., and H. H. Hopkins. 1957. Life histoiy of salt cedar {Tamarix gallica L.). Transactions of the Kansas Academy of Science 60: 360—369. Ohmart, R. D., B. VV. Anderson, and W C. Hunter. 1988. The ecolog>' of the lower Colorado River from Davis Dam to the Mexico-United States Interna- tional Boimdan': a commimitv' profile. U.S. Fish and Wildlife Service Biological Report 85(7.19). 296 pp. 1995] Salinity Effects on Populus and Tamarix 65 Ohmart, R. D., W. O. Deason, and C. Burke. 1977. A riparian case history: the Colorado River. Pages 35-47 in R. R. Johnson and D. A. Jones, technical coordinators, Importance, Preservation, and Man- agement of Riparian Habitat: a symposium. USDA Forest Service General Technical Report RM-43. Fort Collins, CO. Robinson, T. W. 1965. Introduction, spread and areal extent of saltcedar {Tamarix) in the western states. United States Geological Survey Professional Paper 491-A. 13 pp. SAS Institute, Inc. 1990. SAS/STAT user's guide, ver- sion 6, 4th edition. SAS Institute, Inc., Cary, NC. Segelquist, C. A., M. L. Scott, and G. T. Auble. 1993. Establishment of Populus deltoides under simulated alluvial groundwater declines. American Midland Naturalist 130: 274-285. SlEGEL, R. S., AND J. H. Brock. 1990. Gemiination require- ments of key Southwestern woody riparian species. Desert Plants 10: 3-8, 34. Snedecor, G. W., and W. G. Cochran. 1980. Statistical methods. Iowa State University Press, Ames. 507 pp. Stromberg, J. C, D. T. Patten, and B. D. Richter. 1991. Flood flows and dynamics of Sonoran riparian forests. Rivers. 2: 221-235. Tomanek, G. W., and R. L. Ziegler. 1962. Ecological studies of salt cedar. Unpublished report. Division of Biological Sciences, Fort Hays Kansas State College, Hays. 128 pp. United States Environmental Protection Agency. 1983. Methods for chemical analysis of water and wastes. Publication identification: EPA-600 4-79-020. Environmental Monitoring and Support Laboratory, Office of Research and Development, Cincinnati, OH. United States Geological Survey. 1991. Water re- sources data. New Mexico, water-year 1990. U.S. Geo- logical Survey water-data report NM-90. . 1992. Water resources data. New Mexico, water- year 1991. U.S. Geological Survey water-data report NM-9I. Waisel, Y. 1991. The glands oiTaiiuirix aphylla: a system for salt secretion or for carbon concentration? Physiologia Plantarum 83: 506-510. Warren, D. K., and R. M. Turner. 1975. Saltcedar {Tama- rix chinensis) seed production, seedling establishment, and response to inundation. Journal of the Arizona Academy of Science 10: 135-144. Zasada, J. C, AND R. Dens.more. 1977. Changes in Sali- caceae seed viability during storage. Seed Science and Technology 5: 509-517. Received 14 March 1994 Accepted 12 August 1994 Great Basin Naturalist 55(1), © 1995, pp. 66-73 NAMES AND TYPES OF HEDYSARUM L. (FABACEAE) IN NORTH AMERICA Stanley L. Welsh ^ Abstract. — The names and t\pes of Hedysaruiii L., sensu stricto, for North America are included, alon^ with biblio- graphic citations, type information and place of deposit of types, and all synonyms. Lectotypes are designated for Hechjsarum auriculatum Eastvv., H. carnulosum Greene, H. marginatum Greene, H. pabulare A. Nels., and H. truncatum Eastw'. Key wards: Hed\ sarum, types, nomcnclatttrc. The following list of names and types in Hedysannn L. was prepared preliminary to submittal of a summary revision to the Flora North America Project. The genus Hedysannn L. as here inteipreted for American taxa extends from the Bering Strait to Newfoundland and Vermont, and from the Polar Sea and the Cana- dian Arctic Archipelago south through the mountains and plains of western North Anerica to Oklahoma, New Mexico, Arizona, and Nevada. Excluded from this treatment are those taxa originally included in Hedysannn, which are now interpreted as belonging to other genera, i.e., to Desmodiinn. The genus in the restricted sense consists of two complexes, i.e., those with leaflets thickened and veins obscured (the boreale complex) and those with relatively thin leaflets in which the veins are rather readily apparent (the alpinum complex). The earliest taxon within Hedysanim alpinum complex is that by Michaux (1803), who estab- lished the trinomial Hedysannn alpinum ameri- canum Michx. The boreale complex was initi- ated by Nuttall (1818) with the publication of H. boreale. Taxa in the two complexes demonstrate re- markable morphological and geographical par- allelism. Each consists of additional taxa sepa- rable generally into two geographical sub- groups juxtaposed at or near the 50th parallel of longitude (somewhat north of the Canada- U.S. boundary). North of that parallel lies most of H. boreale ssp. mackenzii (Richards.) Welsh, and most of H. alpinum sensu stricto. To the south occurs H. occidentale Greene, most of H. sulphurescens Rydb., limited extensions of H. alpinum L., and most of H. boreale ssp. bore- ale. Glacial events during the Pleistocene have been suggested as having separated the sub- sets, allowing them to achieve the degree of morphological and genetic integrity of the present populations. The present juxtaposition is suggested to have resulted by expansion of the respective entities into areas previously occupied by glaciers. The rather large number of names involved in the genus is indicative of variation inherent in the various taxa. Flower size, plant size, leaflet size, and pubescence are features vari- able in both complexes. Apparent correlation of two or more of these features has served as justification for several names. Indeed, when one observes dwarf, large-flowered plants in either complex, there appears to be a compel- ling need for their recognition. However, much, if not all, of the variation is haphazard, or the attempt at segregation devolves to use of a single characteristic, such as presence or absence of pubescence, which fails also. There are few truly diagnostic characteristics once the two complexes are separated. The taxono- mist ultimately must rely on a series of varying features to identify a particular specimen. Fortunately, the taxa are, with some notable exceptions, disjunct from each other. If the disjunction is not apparent from examination of a distribution map, it is often apparent in the field where the plants grow in different habitats. For example, the range of yellow- flowered H. sulphurescens apparently overlaps that of pink-purple-flowered H. occidentale in large part; yet, they seldom occur together, and only an exceptional intermediate is known. 'Department ol Botany ami Life Science Mnscuni, Brii^hani Younf; University-. Provo. UT 84602. 66 1995] North American Hedysarum 67 There are, in spite of gross similarities of the taxa within the respective complexes, few recorded intermediates. Adding to the difficulties of interpretation of the North American materials is the inter- rupted circumboreal distribution of H. alpinum, a species with several close allies in Siberia. The initial interpretation by Michaux of North American H. alpinum as being taxo- nomically different {'H. alpinum: americanwn) from that of the Old World has paraded appar- ition-like through most subsequent treatments of the genus. Unpublished work by Northstrom (1974) refuted the claim to difference between Siberian and American phases of the species, at least as far as broad categories were con- cerned. The claim that North American mate- rials constitute a separate entity is likewise re- futed by comparison of specimens from Siberia and North America in the present study. Other workers have asserted that large-flowered, low-growing plants of the species are identical with substantial Asian taxa [i.e., H. hedysari- oides (L.) Schinz & Thellung {Astragalus hedy- sarioides L.)]. Such claims were investigated by Northstrom (1974), who determined that tliere is little basis for such assertions. Evidence to support the conspecific nature of the sup- posed entities is apparent when localities of such supposed taxa in western Alaska are examined and plants with larger flowers are found to occur within populations having small flowers, and that flower size within the species in a broad sense forms a continuum. And, occasional tall specimens within the alpinum complex also bear large flowers. Another factor leading to the creation of a large number of synonyms was the early mis- inteipretation of specimens of H. alpinum under the name H. boreale. This switching of names, while not uniquely a problem in this genus, became of great importance to those workers who encountered the genus piecemeal and treated the variants as though they had not already been named. It was not helpful, per- haps, that the most ardent authors of western American plant names should be involved with the genus (i.e., Edward L. Greene, Per Axel Rydberg, and Aven Nelson). Greene, as the record indicates, was prone to name the same species several times in this and other genera, not recognizing, or possibly not caring, that he was renaming the same taxon. Still another trend resulting in the forma- tion of inconsequential names was the well- intentioned effort to provide epithets for spec- imens differing in insubstantial ways, i.e., the naming of white-flowered or teratological speci- mens as formae. The following list is thought to be exhaus- tive for Hedysarum names in North America. Pertinent types have been received on loan through the kindness of curators of herbaria cited with the specimens. Abbreviations for herbaria are those standard ones cited in Index Herbariorum. Type information is pre- sented below in dual fomiat for some taxa, with type information (type locality) as recorded with the protologue cited first and label data of the type specimen (type) cited second where there is a substantial difference in the two accounts. Hedysarum albiflorum (Macoun) Fedtsch., Acta Hort. Petrop. 19: 252. 1902. Basionym: H. boreale var. ulhiflorum Macoun = H. sulphurescens R\db. Hedysarum alpinum L., Sp. Pi. 750. 1753. Type locality: "Habitat in Siberia" (Linnaeus I.e.). Tvpe: Possible lectotvpe 921.54 LINN (microfiche BRY!). Hedysarum alpinum var. americanum Michx., Fl. Bor. Anier. 2: 74. 1803. = H. alpinum L. Synonyms: H. alpinum ssp. americanum (Mich.x.) Fedtsch., Acta Hort. Petrop. 19: 2.55. 1902, in part; H. americanum (Mich.x.) Britt., Mem. Torrey Bot. Club 5: 201. 1894. Type locality': "In borealibus Canadae, et in cataractis montium alleghanis." Type: "Hedys. J ii p. 74-75. Herb. M.x" (isotype NY!). There is a mounted half herbarium sheet at NY Torrey! bearing a large portion of a stem with a leaf and mature, strigose fiaiit of H. boreale var. boreale. This specimen is apparentlx' superfluous (probably having been added later when additional material be- came available to Dr Torrey from western American collections), but more pertinent to the present case the sheet also has an attached fragment envelope on which is written the type information noted above. The envelope contains a portion of an inflo- rescence, a flower, and several immature loment segments. The segments are glabrous, have a defi- nite winged margin, and are identifiable as H. alpinum L. It is probable that the specimen from which the fragments were removed is with the Michaux herbarium at R Continued recognition of the American materi- als of H. alpinum at any infraspecific rank is fraught with difficulties; there are no diagnostic features known that will allow segregation of the American specimens from the Asiatic ones. 68 Great Basin Naturalist [Volume 55 Hedysarum alpinum van americanuni f. alhiflorum (Standi.) Fern., Rhodora35: 275. 1933. BasioiiNin; //. americanum f. alhiflorum Staiidl. = H. alpiiniin L. The publication by Fernald (1933) recognizes white-flowered plants from Newfoundland. Hedysarum alpinum van grandiflorum Rollin.s, Rhodora 42: 233. 1940. Type: "Newfoundland, Pistolet Bay, Mo.ssy and turfy trap cliffs and talu.s, An.se aux Sauvages, M. L. Fernald, K. M. Wiegand and Bayard Long 28625, August 11, 1925"; holotype GH! Paratype: "New Fonndland, Region of Port a Port Bay, No. 10849. In humus or turf on the limestone table- land, altitude 200-300 m.. Table .Mountain, M. L. Fernald and H. St. John, July 16 6c 17. 1914" (CAN!; BM!). This name is based on large-flowered (about 16 mm long), low-growing specimens from Newfound- land. There are other similar plants scattered through much of the distribution of H. alpinum in North America, but they are more consistently rep- resented in frigid or other inhospitable arctic or subarctic sites. Even in the type series cited with the protologue there is considerable variation. The paratype cited above differs significantly from the holotype specimen; it is much taller and has flowers of a size intermediate with those of specimens more usual for var alpinum in a more strict sense. Indeed, the low-growing, larger-flowered phase appears to be a phenotypically recurring recombinant form within a complex exhibiting much variation in flower size and other features. However, size of flower is not always conelated with plant height or flower number. All possible combinations of flower size, flower number, and plant height are represent- ed in the species as a whole. It is possible to write a key that will separate these plants, but it seems that such a key will not then be segregating natural ta.\a. Hedysarum alpinum ssp. philosocia (A. Nels.) Love & Love,Ta.\on31:.347. 1982. Basionym: H. philosocia A. Nels. = H. alpinum L. Hedysarum alpinum var. philosocia (A. Nels.) Rollins, Rhodora 42: 224. 1940. Basionym: H. philosocia A. Nels. = H. alpinum L. Hedysarum americanum (Michx.) Britt., Mem. Torrey Bot. Clul) 5: 202. 1894. Basionym: H. alpinum var americanum Michx. = H. alpinum L. Hedysarum americanum f. alhiflorum Standi., Incld Mus. Pub. Bot. 8: 15. 1930. = H. alpinum L. Synonym: H. alhiflorum (Macoun) Fedtsch. Type: "Alaska: Davidson Glacier July 4, 1929, William S. Cooper & Frances E. Andrews 95 (Herb. Field Mus. No. 598,264, type)"; holotype F! White-flowered specimens occur sporadically through populations of taxa with generally pink- purple flowers. Their recognition at any taxonomic rank is probably moot, and the publication of the tiixon by Standley (1930) is therefore inconsequential. Hedysarum americanum van mackenzii (Richards.) Britt., Mem. Torrey Bot. Club 5: 202. 1894. Basionym: H. mackenzii Richards. — H. horeale ssp. mackenzii (Richards.) Welsh Hedysarum auriculatum Eastw., Bot. Gaz. 33: 205. 1902. = H. alpinum L. Type: Alaska, Cape Nome, Blaisdell s.n. sununer 1900 (lectotype selected here; GH!, isolectotype US!). Specimens on which H. auriculatum is based were distributed from the California Academy of Sciences herbarium with collection information recorded on labels of that institution. The only known specimens in contemporary collections are those at GH and US. The two specimens consist of almost identical branches of H. alpinum, with both flowers and fruit, although that at GH is designated on the label as a duplicate of the type, which was presumably at GAS prior to the San Francisco earthquake early in this centuiy Fire resulting from that devastating tragedy destroyed much of the early Academy herbarium. Hedysarum bakeri Greene ex Rydb., Bull. Agric. Exper Sta. Colorado, 100: 215. 1906. pro syn. = H. horeale Nutt. var horeale Intended type: "Flora of Colorado. Plants the Gunnison Watershed, Cimarron, June 28. Stems in large clus- ters 8 in. to 1 1/2 ft. on dr>' open slopes. Collected in 1901 by C. F Baker, No. 274" (NDG!). Evidently the name was never published by E. L. Greene but was cited as a synonym of H. pabulare A. Nelson by Rydberg in his Flora of Colorado. The intended type has three mounted stems showing flowers and maturing fruit; they are strigose both on herbage and on the loments. The plants differ in no material way from a great many specimens from Colorado. Perhaps Greene also realized as much. Hedysarum horeale Nutt., Gen. N. Amer Pi. 2: 110. 1818. Type locality: North Dakota, "around Fort Mandan, on the banks of the Missouri," Nuttall (I.e.). Type: "Hedysanmi horeale — Sources of the xMissouri," Nuttall (probably late June) 1811; holotype BM! The name H. horeale was early transfeired to the concept of H. alpinum, and part of the synonymy reflects attempts by various authors to resolve the apparent lack of a name for this wide-ranging and highly variable species. Nuttall (Torrey and Gray 1838) named the species a second time, as H. canescens, based on specimens from along the Snake River in present Idaho taken in 1834. He was in the vicinity of Fort Hall, Idaho, from 14 July to 6 August 1834 (McKelvey 1955: 602). Whether he noted the similarity between earlier- and later- named materials is not known. It seems likely that 1995] North American Hedysarum 69 he did not have authentic material of the earher- named taxon at hand for comparison with his col- lections on the 1834 Wyeth expedition. The t\'pe at BM, a solitary' flowering stem, is mounted with sev- eral flowering stems of H. alpinum of unknown col- lector. Hedysarum boreale var. alhiflorum Macoun, Cat. Canad. PI. 1; 510. 1S84. noni. nud. = H. sulphurescens R\db. Syn: H. alhiflorum (Macoun) Fedtsch. T>pe localib,': "This form is peculiar to the foothills and drier mountain slopes, and is abundant from the Kananaskis through the Rocky Mountains to the Columbia valley at Donald, Lat. 51° (Macoun). Eastern summit of the North Kootanie Pass, Rocky Mountains" (I.e.). Type; "Geological and Natural History of Canada. No. 1111.5390. Hedysarum boreale var alhiflorum. Dry soil. East summit of North Kootanie Pass, R. Mts. Dawson. July 29th 1883" and "Geological and Natiual Histor\' Survey of Canada. No. 533, 5389. Hedysarum boreale Nutt. var. alba. Macoun. Mountain slopes. Kananaskis. Rocky Mts. Macoun. June 24th 1885"; syn txT^es CAN! There is no description aside fiom the designation "alhiflorum" proposed as an epithet. The only other information provided hy Macoun aside from that related with the locality data is the statement: "This fine plant is closely related to H. boreale, but is cer- tainly distinct." The name is regarded as a nomen nudum. The collection by Dawson is, nevertheless, an excellent flowering example of H. sulphurescens, and the Macoun sheet consists of two plants with both flowers and immature to mature fruits, both also H. sulphurescens. The indication by Macoun of relationship of var. albiforum to H. boreale reflects the general misapplication by many American botanists of H. boreale to the alpinum complex in North America, of which H. sulphurescens is a por- tion. Macoun used the number 533 for several col- lections of Hedysarum taken from 1883 to 1885. Hedysarum boreale var. cinerascens (Rydb.) Rollins, Rhodora 42; 234. 1940. Basionym; H. cinerascens Rydb. et H. canescens Nutt. in seq. = H. boreale Nutt. var boreale Hedysarum boreale var. cinerascens f. album Boivin, Naturaliste Canad. 87: 34. 1960. = H. boreale Nutt. van boreale Type: "Canada, Saskatchewan, Maple Creek District, Eastend, hillside along river valley, 19 July 1950, R. C. Russell S 5075" (holotype at DAO!). Hedysarum boreale var. flavescens (Coult. & Fisher) Fedtsch., Bull. Herb. Boiss. 7: 256. 1899. Basionym; H. flavescens Coult. & Fisher = H. sulphurescens Rydb. Hedysarum boreale var. gremiale (Rollins) Northstrom & Welsh, Great Basin Nat. 30; 125. 1970. Basionym; H. gremiale Rollins Hedysarum boreale var. leucanthum (Greene) M. E. Jones, Proc. Calif Acad. Sci. 5; 677. 1895. Basionym; H. mackenzii var. leucanthum Greene = H. boreale ssp. mackenzii (Richards.) Welsh Hedysarum boreale ssp. mackenzii (Richards.) Welsh, Great Basin Nat. 28; 152. 1968. Basionym; H. mackenzii Richards. Hedysarum boreale var. mackenzii (Richards.) C. L. Hitchc, Vase. Pi. Pacific N. W 3; 275. 1961. Basionym; H. mackenzii Richards. Hedysarum boreale var. mackenzii f. niveum (Boivin) Boivin, Naturaliste Canad. 93; 433. 1966. Basionym; H. mackenzii var. mackenzii f. niveum Boivin = H. boreale ssp. mackenzii (Richards.) Welsh Hedysarum boreale var. obovatum Rollins, Rhodora 42: 235. 1940. = H. boreale Nutt. van boreale Type: Nevada, Elko County, Thorpe Creek, E of Lamoile, 25 July 1928, H. H. Price 168 (holotype photo RM!). Hedysarum boreale f. proliferum (Dore) Boivin, Naturaliste Canad. 94; 630. 1967. Basionym; H. mackenzii f proliferum Dore = H. boreale ssp. mackenzii (Richards.) Welsh Hedysarum boreale var. utahense (Rydb.) Rollins, Rhodora 42; 2.35. 1940. Basionym; H. utahense Rydb. — Hedysarum boreale Nutt. var boreale Hedysarum canescens Nutt., in Torr. & Gray, Fl. N. Amer 1: 357. 18.38. Not H. canescens L. Basionym; H. cinerascens Rydb.; H. boreale var. cin- erascens (Rydb.) Rollins = H. boreale Nutt. var boreale Type locality; Idaho, "Plains of the Rocky Mountains, particularly near Lewis's River," Nuttall (I.e.) Tvpe; "HedysaiTim * canescens. H. mackenzii? Hook. R. Mts. Lewis [Snake] R." Nuttall s.n. (probably in July) 1834 (holotype PH!; isotypes GH!, BM!, 2 sheets?). The specimen at PH (which is mounted on a sheet with two other superfluous collections) bears the date "July 12," with the incorrect year date 1833 obviously added later. Nuttall was with the Wyeth Expedition in 1834, and on 12 July was a short distance east of where Fort Hall would be constructed subsequently. Despite the existence of the earlier-named H. boreale, with which H. canescens is synonymous, this name or its substi- tutes would be featured prominently in 19th-centu- ry accounts of the genus in the American West. There are two of Nuttall's specimens on the sheet at GH, each provided with a label — both with flow- ers and both representing the same taxon. The label information consists of the following: "Hedysarum * canescens. H. mackenzii? R. Mts.," and Hedysarum * canescens R. Mts." Since no additional locality 70 Great Basin Naturalist [Volume 55 iiiforniation or date accompanies the labels, the sta- tus as exact dupheates is unknown. It seems hkely that both were included within the concept of H. canescens by Nuttall, and both can be regarded as isotypes. There is a second possible isotype of H. canescens at BM, "Hedysarum mackenzii? Fort Hall. Prairie, common. Aug." It lacks the * usual for Nuttall's labels, and his name is not in evidence, but the handwriting appears to be his. Hedijsarum carnulosum Greene, Pittonia 3: 212. 1897. = //. horcdic \iitt. \ ar. borcale Type localit\'; "Common in claye\' soil about the mouth of the Canon of the Arkansas, in southern Colorado' (Greene I.e.). Type: Colorado, Fremont Co., "Plants of Colorado, Canon Cit>', 8 Sept. 1896, Edw. L. Greene" (lecto- tvpe here designated: NDG!, 2 isolectotypes also NDG!). While no specimens were cited with the original description, the three specimens so named in Greene's handwriting at NDG are most certainly type material. All bear the same date and locality in- formation. The specimen bearing the "Greeneanum Herbarium ' number 35686 is here chosen as lecto- type; the others, 35687 and 35688, are considered isolectotypes. The lectotype has both flowers and fruit; the other two are in fruit and flower (with immature fruit), respectively. All have strigose herbage and foments. Usual flowering time for the species is April to late July. Is it possible that the species flowered again following late summer rains at Canon City in September 1896? Hedysarum cinerascens Rydb., Mem. N. Y. Bot. Card. 1: 257. 1900. nom. no\. pro H. canescens Nutt. — H. horeale Nutt. var boreale Basionym: H. canescens Nutt. Syn: H. boreale var cinerascens (Rydb.) Rollins This material was retained liy Northstrom (1974) at varietal rank. The taxon stands on the sole char- acter of pubescence, and a plotting of the distribu- tion of hairy versus glabrous plants demonstrates much overlap. The specimens can be separated, but do they represent taxa? Hedysarum flavescens Coult. & Fisher, Bot. Gaz. 18: 300. 1893, non Regel & Schmalh. Basionym of: H. sulphurescens Rydb. Type: Montana, near Helena, May 1892, F D. Kelsey s.n. (holotype F!). It is unfortunate that the epithet /7r/r('.sf(?n.s was occupied; it fits well the description of flower color in this ta.xon. xMany of the Kelsey collections are in the U.S. National Museum (Elisens 1985), but the type of H. flavescens is at F, where Coulter's herbarium is deposited. Hedysarum gremiale Rollins, Rhodora 42: 230. 1940. = H. boreale var gremiale (Rollins) Northstrom & Welsh Type: Utah, Uintah County, "ca 14 mi \V. of Vernal, 16 June 1937," R. C. Rollins 1733 (holot\pe GH!, iso- types RMI, US!, CAS!, UTC!, MONTU!, PH!, F!). This taxon stands on the feature of lateral spines on the foment segments; it is otherwise indistin- guishable from plants of var horeale by which it is suiTOunded (Northstrom and Welsh 1970). Hedysarum lancifolium Rydb., Mem. New York Bot. Card. 1: 256. 1900. = H. occidentale Greene Type: "Mountain woods near head waters of Jocko River, Montana, — flowers pale purple, W. M. Canby 93, July 15, 1883" (holotype NY!). The type specimen consists of a folded plant some 75 cm tall bearing leaves and flowers, and a stem fragment bearing immature fruit. Mature flowers are about 16 mm long, on the short side of the variation in H. occidentale, but the fruit, even though immature, is of the size and form of that species. Hedysarum leucanthum (Greene) Greene, Pittonia 3: 213. 1897. Basion\ni; H. mackenzii \ar leucanthum Greene = H. boreale ssp. )nackenzii (Richards.) Welsh Hedysarum mackenzii Richards., in Franklin, 1st Journey Bot. Append. 745. 1823. — H. boreale ssp. mackenzii (Richardson) Welsh Type: Canada, Barren grounds. Point Lake to the Arctic Sea, Richardson s. n. [holot\pe BM (photo CAN!), isotype NY Columbia! "Hedysarum macken- zii. Franklin's Journey. Dr Hooker isotype PH!; "Astr (crossed out) Hedys. Mackenzii Richard. N. W Amen Fn Exp. Dr Ho." and "Herb. A. Gray. H. Mackenzii. Torn & Gray, Fl. N. Amer Richardson (B.D.G.)," isotypes GH!]'. Evidently Dr William Jackson Hooker sent rep- resentative material obtained b\' John Richardson, botanist with the Franklin expedition, to the Philadelphia Academy and to John Torrey and Asa Gray. Collections from the Franklin expedition demonstrate the variability represented in a rather large set of specimens, each matched by modern collections of the ta,xon. The second sheet cited at GH is doubly moimted with a mere fragment pre- sumed to have come from the Franklin expedition in the lower portion and a second fragmentar)' col- lection by Burke (apparently a phase of H. boreale) from the Rocky Mountains. The latter material is not a portion of the type of H. nuickenzii. Hedysarum mackenzii var. fraseri Boivin, Canad. Field- Nat. 65: 20. 1951. = H. boreale Nutt. var boreale Type locality: Canada: "Saskatchewan: W. R Eraser, Langham, river valle\', June 12 and 26, 1938 " (I.e.) Type: "Hedysarum Mackenzii Richards. River valley, Langham, Sask., W.RF[raser]., June 12 & 26, 1938" (holot\pe DAO!). Boivin (I.e.) cites the revisionary treatment of Rollins (1940) as indicating that H. mackenzii var u}ackenzii has flowers 18-21 mm long, but with 1995] North American Hedysarum 71 some 16-20 mm. Prairie plants fiom Saskatchewan and Alberta, however, have flowers 13-15 mm long. These latter plants are the basis of his \Ar. fraseri. The type of var. fraseri represents H. horeale ssp. horeale var. horeale, which is present along the plains and foot slope of the Rocky Mountains, and has flowers of the size indicated for the variety. Hedysarum mackenzii var. leticanthum Greene, Pittonia 2: 294. 1892. Basionym of; H. leucanthum (Greene) Greene — H. horeale ssp. maekenzii (Richards.) Welsh Type locality: "On the Porcupine River, northern Alaska, Mr. J. J. Turner" (Greene I.e.). Type: "Plants of Alaska, collected on the Porcupine River, 1891, Mr J. Turner" (holotv'pe NDG!). Greene (1892) notes that this is "far more than an albino state of H. Mackenzii; perhaps identical with some asiatic species; but the plants were just coming into flower when gathered, in that there is no trace of the loment." He later (1897) elevated it to species rank. However, except for white flowers, which occur with some frequency in the species, the plant differs in no respect from numerous other plants fi'om the arctic range of the ta.\on generally. Hedysarum mackenzii var. mackenzii f. niveum Boivin, Canad. Field-Nat. 65: 20. 1951. Basionym of: H. horeale var. mackenzii f. niveum (Boivin) Boivin = H. horeale ssp. mackenzii (Richards.) Welsh Type: "Yukon Territory: J. W. Abbott 17a, Pine Creek, sandy land, June 7, 1946" (holotype DAO!). The collection consists of five flowering stems of H. horeale ssp. mackenzii, all with white flowers. The condition of white flowers is occasional throughout the subspecies and hardly worthy of taxonomic consideration. Hedysarum mackenzii var. pahulare (A. Nels.) Kearney & Peebles, J. Wash. Acad. Sci. 29: 485. 1939. Basionym: H. pahulare A. Nels. = Hedysarum horeale Nutt. van horeale Hedysarum horeale var. leucanthum (Greene) M. E. Jones, Proc. Calif. Acad. II. 5: 677. = H. horeale ssp. mackenzii (Richards.) Welsh Syn: H. leucanthum (Greene) Greene Hedysarum mackenzii f. proliferum Dore, Canad. Field- Nat. 73: 151. 1959. Basionym of: H. horeale f. proliferum (Dore) Boivin — H. horeale ssp. mackenzii (Richards.) Welsh Type: "Plants of Alaska. A single clump in shallow soil over broken rock (growing beside common petalif- erous plant, cf 4983). North Side of Tanana River, Mile 277, Richardson Highway, 64°10'N, 145°52'W, W J. Cody & T J. M. Webster 4984, June 3, 1951" (holotvpe DAO!). This name is based (Dore 1959) on a teratologi- cal specimen of H. horeale ssp. mackenzii, a recur- ring variant induced by a pathogen, likely a smut- fungus. Teratology occurs in several if not all boreal legumes native to Alaska. In certain of those in- stances the inflorescence typically elongates, flow- ers become erect on attenuated pedicels, petals are deformed, and the ovary is typically exserted from the flower. In some specimens at least the ovary is filled with black spores. The type specimen of f. proliferum exhibits another variant than that typi- cally encountered. The inflorescence is shortened and modified flower buds are in tight clusters. Whether elongate or compact, specimens on which such aberrations are based are not taxa, and the need to name them is therefore moot. Hedysarum macquenzii f. canescens (Nutt.) Fedtsch., Acta Hort. Petrop. 19: 272. 1902. Basionym: H. canescens Nutt. = H. horeale Nutt. var. horeale Hedysarum marginatum Greene, Pittonia, 4: 138. 1900. = H. occidentale Greene Type locality: "Mountains above Cimarron, southern Colorado, collected by the writer, 30 Aug. 1896; also near Pagosa Springs, Colo., 26 July, 1899, C. E Baker" (Greene I.e.). Type: Colorado, "Plants of Colorado, Near Cimarron, 30 Aug. 1896, Edw. L. Greene" (lectotv^je NDG!, here chosen); "Plants of Southern Colorado, Pagosa Springs, 26 July 1899, C. E Baker" (syntypes NDCl, NY!. RM!, GH!, F!). The lectotype collected by Greene is in fruit; syntypes at NDG and NY have both flowering and fruiting branches. In both the loment articles are markedly winged and strigose; herbage is strigose also (see Greene 1900). Hedysarum occidentale Greene, Pittonia 3: 19. 1896. Type locality: "Olympic Mountains, Washington, 1890, C. V. Piper" (Greene I.e.). Type: Washington, "Olympic Mts., C. V' Piper 905, flowers 11 August 1890, fruit 30 Sept. 1890" (holo- type NDG!). Greene (1896) provides a description and a short note: "Plant like H. horeale when in flower, though with broader leaflets and widely different fruit." A second sheet fi-om the Olympic Mountains at NDGl, Piper 2227 (August 1895), has the epithet "occiden- tale" in Greene's hand, but it was not cited by him. It is much better material than the type. For a long time the name H. horeale was included within the concept of H. alpimnn. It is likely that Greene was under a similar misconception. The general aspect of H. occidentale (i.e., conspicuously veined leaflets and large loments with prominently reticulate venation), which occurs from Vancouver Island, British Columbia, and the Olympic Peninsula, Washington, disjunctly eastward to northern and eastern Idaho, western Montana, western Wyoming, northeastern Utah, and montane southern Colorado, is that of H. alpinum; and it differs generally in the manner indicated by Greene. The most distinctive feature separating most, if not all, specimens of H. occidentale from H. 72 Great Basin Naturalist [Volume 55 alpintirn is the much larger, rather conspicuously wing-margined loment segments. Flowers are gen- erally larger, often much larger. However, plants from the Absaroka Range of northwestern Wyoming approach H. alpinum in occasionally hav- ing small flowers, but when collected at maturity, the fruit is that of H. occidentale . Additional collec- tions might demonstrate that H. alpinum per se is indeed in the Absarokas. Large-flowered plants of H. alpinum, mainly of frigid sites in the arctic, approach the size of flowers of some H. occidentale specimens, but the fruit there is that of H. alpinum. H. occidentale has loments very similar to those of the closely allied H. sulphur escens. Hedysariim occidentale Greene var. canone Welsh, Great Basin Nat. 38: .314. 1978. Type: Utah, Carbon County, "ca 14 mi due ENE of Helper, Soldier Creek,' 30 June 1977, Welsh & Taylor 15256" (holotype BRY!; isotype at NY!). The syndrome of characters associated with this taxon is shared individually elsewhere within the species as a whole. However, specimens from Duchesne, Carbon, and Emeiy counties, Utah, and Gunnison County, Colorado, are recognizable by their large, thick, ovate to ovate-lanceolate, yellow- green leaflets, and large pale flowers. Plants are known from rather xeric sites in pinyon-juniper and mountain brush communities, whereas plants of the type variety are mainly of more mesic sites. Although the taxon is segregated on weak diagnos- tic features, it seems to be at least a trend worthy of taxonomic recognition. It has long been known in collections. Hedysarum pabiilare A. Nels., Proc. Biol. Soc. Wash. 15: 185. 1902. = H. horeale Nutt. var. horeale Type: Wyoming, Wind River, Dubois, A. Nelson 752, 1894 (lectotype here designated RM!). This name is based on several Wyoming, Colorado, and Utah syntypes: i.e., M. E. Jones 5592, Soldier Summit, Utah, in 1894, POM?, BM!, F!; Snake River, Wyoming, A. Nelson 3496, 19 August 1899 RM!; Wyoming, Natrona Co., Bates Creek, L. N. Goodding 201, 5 July 1901, RM!, F! Hedysarum palndare var. rivulare L. O. Williams, Ann. Missouri Bot. Card. 21: 344. 1934. — H. horeale Nutt. van horeale Type: Wyoming, Teton County, along the Snake River, 31 July 1932, L. O. Williams 975 (holotype RM!; isotypes GH!, CAS!). Hedysarum philoscia A. Nels., Proc. Biol. Soc. Wash. 15: 185. 1902. Basionyin: H. alpinum var. philoscia (A. Nels.) Rollins; H. alpinum ssp. philoscia (A. Nels.) Love & Love = H. alpinum L. Type: Wyoming, Albany County, Head of Crow Creek, Laramie Mountains, 1896, A. Nelson 2034; holotype RM! Material from the Black Hills of South Dakota and from southeastern Wyoming is morphologically similar and has been recognized as belonging to a taxon that sui-vived south of the major glacial events of the Pleistocene. The main diagnostic criterion is, however, loment pubescence. That feature is incon- stant within the southern material and often is pres- ent in plants far beyond its supposed range (which has been plotted to include plants as far north as the 50th parallel). Recognition of plants at any taxo- nomic rank is, therefore, problematical. Hedysarum roezlianum Prantl, Ind. Sem. Hort. Wirceb. 8. 1873. = H. horeale Nutt. var. horeale? I have been unable to find any reference to this taxon aside from its citation by Rollins (1940). Hedysarum sulphurescens Rydb., Bull. Torrey Bot. Club. 24:251.1897. Basionyin: H. flavescens Coult. & Fisher, not Regel & Schmalh. Yellow to yellowish flowers easily distinguish this entity, which shares the peculiar loment fea- tures of H. occidentale. The species ranges from the southern British Columbia-Alberta Rockies south through north central Washington, northern Idaho, western Montana, and northwestern Wyoming. Hedysarum truncatum Eastw., Bot. Gaz. 33: 205. 1902. = H. alpinum L. Type: Alaska, Nome, Dr E E. Blaisdell s.n. summer 1900 (lectotype NY! here designated; isolectotype GH!). Type material is low, about 2-2.5 dm tall, has mature flowers about 12 mm long, and has fruit the size and conformation of H. alpinum. It is identical for all practical purposes with material named by Eastwood simultaneously as H. auriculatum and taken at the same place and time by the same col- lector in 1900. Hedysarum uintahense A. Nels., Proc. Biol. Soc. Wash. 15: 186. 1902. = H. occidentale Greene Type: Wyoming, "In draws of the foothills, Evanston, A. Nelson 7198, 14 June 1900" (holotype RM!; iso- types RM!, NY!, GH!). Type sheets uniformly bear thick, lanceolate to lance-ovate leaflets similar to var. canone, but with flowers of typical H. occidentale. Plants from south- west Wyoming are not uniformly of the uintahense type, but vary from one population to another, with most being similar to traditional H. occidentale. Hedysarum utahense Rydb., Bull. Torrey Bot. Club 34: 424. 1907. — H. horeale Nutt. var. horeale Type: Utah, Salt Lake County, "vicinity of Salt Lake City Utah," Leonard 55, 26 May 1883 (holotype NY!). The type consists of two complete stems and a fragmentary branch; it is typical of the material 1995] North American Hedysarum 73 growing through much of Utah and elsewhere in the West. References DORE, W. G. 1959. Some inflorescence form.s in clovers and other legumes. Canadian Field Naturalist 73; 147-154. Elisens, W. J. 1985. The Montana collections of Francis Duncan Kelsey. Brittonia37: 382-391. Fedtschenko, B. 1902. A revision of the Genus Hedy- sarum. Acta Hortus Petropolitani 19: 185-325. Fernald, M. L. 1933. Recent discoveries in the Newfound- land Flora. Brittonia 35: 26.5-283. Gree.ne, E. L. 1892. New or noteworthy species. — XIV. Pittonia 2: 293-298. . 1896. New or noteworthy species. — XV Pittonia 3: 12-28. . 1897. New or noteworthy species. -XIX. Pittonia 3: 212-230. . 1900. A fascicle of new Papilionaceae. Pittonia 4: 132-139. McKelvey, S. D. 1955. Botanical e.xploration of the Trans- Mississippi West, 1790-1850. Arnold Arboretum, Jamaica Plains. 1144 pp. NoRTHSTROM, T. E. 1974. The Genus Hedysarum in North America. Unpublished master's thesis, Brigliam Young University, Provo, UT NoRTHSTROM, T. E., AND S. L. WELSH. 1970. Revision of the Hedysarum horeale complex. Great Basin Naturalist 30: 109-130. NUTTALL, T 1818. The genera of North American plants. Volume 2. D. Heartt, Philadelphia, PA. MiCHAUX, A. 1803. Flora Boreali-Americana. 2: 1-340. Rollins, R. C. 1940. Studies in the genus Hedysarum in North America. Rhodora 42: 217- 239. Standley, P C. 19.30. Studies of American plants — III. Field Museum of Natural Histor\' Botanical Series 8: 3-73. TORREY, J., AND A. GRAY. 1838. Flora of North America. Volume 1. Wiley & Putnam, New York, NY. Received 28 February 1994 Accepted 3 June 1994 Great Basin Naturalist 55(1), © 1995, pp. 7 '4-77 WHIPWORM (TRICHURIS DIPODOMYS) INFECTION IN KANGAROO RATS {DIPODOMYS SPE): EFFECTS ON DIGESTIVE EFFICIENCY James C. Munj^^er' and Todd A. Slicliterl Abstract. — To determine whether infections by whipworms (Trichuris dipodornijs [Nematoda: Trichurata: Trichuritlae] ) might affect digestixe efficiency and therefore energ>' budgets of two species of kangaroo rats {Dipodomijs micrups and Dipodonujs ordii [Rodentia; Ileteromyidae]), we compared the apparent dry matter digestibility' of three groups of hosts: those naturally infected with whipworms, those naturally uninfected with whipwoiTns, and those origi- nally naturally infected but later deinfected by treatment with the anthelminthic Ivermectin. Prevalence of T. dipodotnys was higher in D. tnicrops (53%) than in D. ordii (14%), Apparent dr\' matter digestibility was reduced by whipworm infection in D. microps but not in D. ordii. Although a statistically significant effect was shown, its small mag- nitude indicates that whipworm infection is unlikely to have a biologically significant impact on the energy budgets of host kangaroo rats. Key words: parasite, digestive ejficieney, whipwonn, kangaroo rat, Trichuris, Dipodomys, energy budget. Parasites inhabiting the gastrointestinal tract of a host may reduce the efficiency of the organs they inhabit either through direct com- petition for nutrients or through damage to absoqDtive surfaces. Because decreased diges- tive efficiency may reduce the rate of energy input into a host, gastrointestinal parasites have the potential to cause a change in host energy allocation (e.g., reduced activity or reduced reproduction), and thereby impact the ecology of the host (Munger and Karasov 1989). Tapeworm infections have a measurable effect on digestive efficiency, but a biologically unimportant effect on the energy budget of host white-footed mice [Peromyscus leucopus; Munger and Karasov 1989). The present study was designed to determine if infection by a nematode, the whipworm Trichuris dipodomys, has a substantial effect on one aspect of the energy budget, digestive efficiency, of host kangaroo rats {Dipodomys microps and D. ordii). Materials and Methods Our study site, located 2 km N of Muq^hy, Owyhee County, ID, is in desertscrub habitat with sandy loam substrate. Primary shrub species of the study area are Artemisia spinescens, Artemisia tridentata, Atriplex canescens, Atriplex confertifolia, Atriplex spin- osa, and Chrysothatnniis nauseosus. Six rodent species were captured at tlie site, Ammospermo- philus leucurus, Neotoma lepida, Perognathus flaviis, Peromyscus manicuhitus, and two species of kangaroo rats, Dipodomys ordii and Dipodomys microps. Dipodomys ordii ranges from 42 to 72 g and consumes a diet consisting primarily of seeds (Zeveloff 1988). Dipodomys microps is larger, 72-91 g, and is unique among kangaroo rats in that it relies heavily on leaves of Atriplex confertifolia for forage (Kenagy 1972, Zeveloff 1988). Both species are liable to infection by the whipworm Trichuris dipodo- mys, a nematode that inhabits the cecum of infected hosts (Giimdmann 1957, Whitaker et al. 1993). On the study site we established a 13 X 13 grid of 169 Sherman live traps baited with millet and placed at 15 m intervals. During two trapping sessions, 14-22 June and 15-18 August 1990, kangaroo rats (30 individuals of D. microps and 85 of D. ordii) were captured and brought into the laboratory. Fecal speci- mens from each animal were anaK zed for the presence of parasite eggs by standard centrifu- gal flotation techniques using saturated sucrose solution (Pritchard and Kruse 1982). Six in- fected but untreated animals from the June experiment were included in the pool of ani- mals used in the August experiment. The few animals that failed to thrive in the lab were removed from the experiment; data from a 'Department of UiologN, Boise State University, 1910 University Drive, Boise, ID 83725. 74 1995] Whipworms in Kangaroo Rats 75 total of 29 D. microps individuals and 56 D. ordii were analyzed. Each month's set of captures was subjected to the following protocol: (1) Kangaroo rats were acclimated to a diet of millet seed for 3-11 d. (2) A pretreatment feeding trial was per- formed: Animals were placed in wire-bottomed cages with a measured amount of whole millet seed. At the end of 5 d, fecal pellets were sep- arated from spilled food and dried >24 h at 50° C. Initial digestive efficiency of each ani- mal was measured as apparent dry matter digestibility (i.e., the proportion of mass con- sumed but not lost as waste), which was calcu- lated as (Mpo -Mpe) / Mpo, where Mpo and Mpp are the mass of food consumed and feces produced, respectively. (3) Half of the infected animals were then injected subcutaneously with a solution of Ivermectin (a svstemic anthelminthic; Ivomec brand, from MSD AG VET, Rah way, NJ). Figure I gives sample sizes of treatment groups. June captures received, on each of two con- secutive days, a 0.2-cc injection of Ivermectin in 40% glycerol formal and 60% propylene glycol; each injection delivered ca 350 /xg Ivermectin / kg body mass. Controls received equal-volume injections of the glycerol for- mal-propylene glycol carrier. This dosage had little effect on the presence of whipworm eggs in feces of injected animals. Therefore animals received 8 d later a second set of two injec- tions, each of 0.15 cc and delivering ca 2 mg Ivermectin / kg body mass; control animals received the carrier August captures received, on each of two consecutive days, an injection of 0.15 cc volume delivering ca 2 mg Ivermectin / kg body mass. Control animals received the carrier. To control for possible side effects of Ivermectin, half of the uninfect- ed animals captured in August were also injected with a solution of Ivermectin. (4) Two days after each set of injections a posttreatment feeding trial was conducted using techniques in (2) above. Only results of the pretreatment feeding trials and feeding trials following the 2-mg Ivermectin / kg body mass injection will be presented below. Results and Discussion Adult worms (seven of each gender) taken from a Dipodomys microps at our site were identified as Trichuris dipodomys. Although m a .02- m Deinfecled • Infected o Uninfected ,01 • _ ..s 0 - «>9 I ) 40 J - ,01 • r 6,0 l|7 110 — np. [_ P. microDS D. ordii Fig, 1. Effects of variation in parasite load on propor- tional change in dry matter digestibility. Means ± SE. Numbers represent sample sizes. some minor morphological differences from the original species description (Read 1956) do exist, perhaps as a result of geographical variation, the specimens most closely match Read's description of T. dipodomys (A. Shostak personal communication). Measurements of several key morphological characters are as follows (X ± SD): total length: S 25.6 ± 0.8 mm, 9 41.3 ± 2.9 mm; hindbody length: 6 12.7 ± 0.4 mm, 9 23.7 ± 1.9 mm; spicule length: 850 ± 85.1 yitm; egg length: 64.8 ± 5.0 fxm; egg width: 33.5 ± 1.0 /xm. Voucher speci- mens were deposited with the University of Alberta Parasite Collection (#'s UAPC 11464 and UAPC 11465). Although we did not identi- fy whipworms from D. ordii, we are confident they are T. dipodomys; the type host for T. dipodomys is D. ordii, and T. dipodomys is known only from D. ordii and D. microps (Whitaker et al. 1993). Prevalence in Host Species. Trichuris dipodomys occurred at substan- tially higher prevalence in D. microps than in D. ordii (Table 1), a result similar to that of Grundmann (1957). We can speculate as to three possible explanations for this pattern. The first is that eggs produced by adult worms in D. microps may become embryonated more easily than those in D. ordii. Freshly produced fecal pellets of D. microps appear moister than those of D. ordii (Munger personal observa- tion), probably because of the higher amount of green or leafy vegetation in the diet of D. microps. If moisture is necessary for embiyona- tion of the eggs (as is implied by Parry 1968), 76 Great Basin Naturalist [Volume 55 Table 1. Infection oltwo species of kangaroo rat with the nematode Trichuris dipodornijs. D. microps D. ordii Infected Uninfected Infected Uninfected June trapping 10 5 August trapping 6 9 5 39 7 34 moister feces may lead to higher embryonation rates and therefore higher prevcilence among D. microps. The second explanation is that social and burrow use behavior may differ between these species. For example, perhaps D. microps individuals visit one another's burrows (and thereby become exposed to parasite eggs) at a substantially higher fiequency tlian do D. ordii. Also, D. microps inhabits a mound up to 2 m in diameter while D. ordii inhabits less sub- stantial individual holes. Studies of another system of two species of kangaroo rats has shown that the larger, mound-inhabiting D. spectahilis uses its burrow system for pro- longed periods, while the smaller D. merriami rotates among several burrows (Jones 1989). This latter behavior would tend to reduce reinfection of individuals; it would be interest- ing to see if behaviors differ similarly between D. microps and D. ordii. The third explanation is that resistance to infection may differ between these two host species. Effects on Digestive Efficiency Apparent dry matter digestibility (ADMD) of millet seed was quite high, >95% on aver- age (Table 2), a figure comparable to that found by Schrieber (1979) for granivorous rodents. Injection of Ivermectin did not appear to affect ADMD of animals uninfected by whipworms, an effect that might occur through the removal of other symbionts, or through some direct effect (proportional change in ADMD, X ± SE; untreated: -0.0043 ± 0.0035, treated: -0.0058 ± 0.0037). Therefore, in the following analyses all natu- rally uninfected animals are combined into one class. The effect of whipworm removal on ADMD was analyzed with a two-way analysis of vari- ance (AN OVA). One factor analyzed was the treatment: deinfected (naturally infected but treated with Ivermectin) vs. infected (naturally infected but not treated with Ivermectin) vs. naturally uninfected. The other factor was Table 2. Effects of whipworm infection on apparent dry matter digestibility (ADMD). Standard errors are in parentheses. Figures on change between initial and final feeding trials, as well as sample sizes, are in Figure 1. See text for a description of treatments. Treatment Deinfected Infected Uninfected Dipodomys microps Initial ADMD .9.56 .96.5 .9.55 (.0051) (.0029) (.010.3) Final ADMD .961 .9.50 .953 (.0039) (.0026) (.00.52) Dipodomijfi ordii Initial ADMD .967 .957 .961 (.0107) (.0076) (.0022) Final ADMD .9.55 .9,58 .9.57 (.0034) (.0037) (.0014) species. Experimental period (July vs. August) was included as a blocking factor. The depen- dent variable in the analysis was proportional change between pretreatment and posttreat- ment ADMD ([post-pre]/pre); this measure should be more sensitive than posttreatment ADMD in expressing treatment effects because it takes account of initial differences in ADMD among hosts. Although there were no statistically signifi- cant main effects of treatment or species on ADMD, there was a significant interaction between these factors (Table 3), indicating that the two host species differ in their response to treatment. This difference between species was explored using a separate AN OVA for each species, which revealed that treatment with Ivermectin had a significant effect on change in ADMD in D. microps, but not in D. ordii (Table 4, Fig. 1). A Tukey's a posteriori multiple sample test revealed that, within D. microps, the change in ADMD of the deinfected group differed significantly from the change in ADMD of both the infected group and the uninfected group. These results can be inter- preted as showing that the deinfected group had 1.9% higher ADMD than the other two groups. Of interest is the lack of effect Trichuris causes in D. ordii. This may be due to what appears to be a higher intensity of infection (more parasites per infected host) in D. microps: fecal floats of D. microps in general contained more eggs than did floats of D. ordii J^D. microps X = 254, SE = 115.2; D. ordii X = 63.5, SE = 21.0; Mann-Whitney U-test, U = 79, P = .1). If fewer worms were present in D. 1995] Whipworms in Kangaroo Rats 77 Table 3. F values and probability values (P) from three-way analyses of variance on effects of species, month, and treatment (deinfected, infected, or uninfected) on apparent dry matter digestibility (ADMD). Dependent variable Proportion change df Initial ADMD Final ADMD in ADMD Source F P F P F P Treatment 2 .15 .86 .72 .49 .47 .63 Species 1 .46 .50 .82 .37 .33 .57 Treatment * Species 2 1.33 .27 1.78 .18 4.74 .012 Block (= Month) 1 .51 .48 9.11 .003 .00 .95 Error 77 Table 4. Results from one-way analyses of variance on the effect of treatment (deinfected, infected, and uninfect- ed) on % change in dry matter digestibility in D. microps and D. ordii. Species Source df MS D. microps D. ordii Treatment 2 .00106 4.64 .019 Error 27 .000229 Treatment Error 2 52 .00034 .01442 1.21 .31 ordii, the effect of eradicating those worms would have been less apparent. One might question the biological impor- tance of the slight, albeit statistically significant, decrease in ADMD caused by Trichuris infec- tion. Munger and Karasov (1989) showed an effect of similar magnitude resulting from tape- worm infection [Hijrnenolepis citelli) in white- footed mice {Perotnyscus leucopus). They argued that hosts can easily compensate for such small effects by slight increases in food consumption or decreases in expenditures, or by changes in gut morphology (Mettrick 1980), and concluded that such effects on ADMD are unlikely to affect host energy budgets or to translate through to population-level effects. The same conclusion is likely to apply to the kangaroo rat-whipworm system. Acknowledgments We thank Allen Shostak of the University of Alberta's Parasitology Museum for measuring specimens of the parasite and for its identifica- tion. Kay Kesling helped both in the field and in the lab. Sara Murray and Aaron Munger helped in the field. Discussion with Mary Price was helpful, as were comments from anony- mous reviewers. This research was supported by an Intramural Faculty Research Grant from Boise State University. Literature Cited Grundmann, a. W. 1957. Nematode parasites of mam- mals of the Great Salt Lake Desert of Utah. Journal of Parasitology 43: 105-112. Jones, W. T. 1989. Dispersal distance and the range of nightly movements in Merriam's kangaroo rats. Journal of Mammalogy 70: 27-34. Kenagy, G. J. 1972. Saltbush leaves: excision of hypersaline tissue by a kangaroo rat. Science 178: 1094—1096. Mettrick, D. E 1980. The intestine as an environment for Hijrnenolepis diminuta. Pages 281-356 in H. P Arai, ed.. Biology of the tapeworm Hijrnenolepis diminuta. Academic Press, New York, NY. Munger, J. C., and W. H. K.\r\sov. 1989. Sublethal para- sites and host energy budgets: tapeworm infection in white-footed mice. Ecology 70: 904-921. Parry, J. E. 1968. Transmission studies of nematodes with direct life histories in selected Utah mammals. Unpublished doctoral dissertation, University of Utah, Salt Lake City. Pritchard, M., and G. Kruse. 1982. The collection and preservation of animal parasites. University of Nebraska Press, Lincoln. 141 pp. Read, C. P 1956. Trichuris dipodomijs, n. sp., from Ord's kangaroo rat. Proceedings of the Helminthological Society of Washington 23:119. Schrieber, R. K. 1979. Coefficients of digestibility and caloric diet of rodents in the northern Great Basin desert. Journal of Mammalogy 60:416-420. Whitaker, J. O., Jr., W. J. Wrenn, and R. E. Lewis. 1993. Parasites. In: H. H. Genoways and J. H. Brown, eds.. Biology of the Heteromyidae. American Society of Mammalogists Special Publication 10. Zeveloff, S. I. 1988. Mammals of the Intermountain West. University' of Utah Press, Salt Lake City. 365 pp. Received 27 July 1993 Accepted 20 June 1994 Great Basin Naturalist 55(1), © 1995, pp. 78-83 LOCAL DISTRIBUTION AND FORAGING BEHAVIOR OF THE SPOTTED BAT {EUDERMA MACULATUM) IN NORTHWESTERN COLORADO AND ADJACENT UTAH Jay F Stoiz^ Abstract. — This study investigated local distribution and foraging behavior of the spotted bat {Eudenna maculatum) in Dinosaur National Monument, Colorado-Utah, by monitoring audible echolocation calls. The occurrence of this species was verified in a variety of habitat types in canyon bottoms and other relatively low elevation sites, indicating that the animals are widely distributed and locally common in the area. Foraging spotted bats concentrated flight activi- ty in the open-air space above meadows and occasionally exploited near-canopy habitat (within 8 m of foliage). Bats began to forage shortly after dark, and activity levels were relatively constant throughout the night. Foraging spotted bats attacked airborne prey every 2.15 min on average. Consistent with published observations, spotted bats maintained exclusive foraging areas. Distinct vocalizations indicating agonistic encounters occurred when a bat encroached on the foraging area of a conspecific. Key words: spotted bat, Euderma maculatum, Colorado, Utah, Dinosaur, National Monument, foraf:,ing, habitat me, attack rates, echolocation. The spotted bat {Eudenna maculatum) is widely distributed across western North America and apparently exists in low popula- tion numbers throughout its range (Fenton et al. 1987). The species is rare in collections, and viable populations have been documented in only a few widely separated localities (Watkins 1977, O'Fan-ell 1981). Findings presented here and those of Navo et al. (1992) indicate that E. maculatum is locally common in canyon bot- toms and other low-elevation sites in Dinosaur National Monument, Colorado-Utah, and occurs throughout a diverse range of habitat types. Population studies (e.g., Leonard and Fenton 1983) in south central British Columbia have demonstrated that foraging spotted bats exhibit considerable habitat specificity; radiotracking in this same area (Wai-Ping and Fenton 1989) has demonstrated that individuals are faithful to specific sites over several consecutive nights. However, no clear association between forag- ing activity and any specific habitat conditions is apparent. In British Columbia, spotted bats forage over clearings in ponderosa pine {Pinus ponderosa) forests, open fields, and marshes (Leonard and Fenton 1983, Wai-Ping and Fenton 1989). There is little information about foraging habitat throughout the remainder of the geographic range of E. maculatum. The purpose of this study was (1) to investi- gate local distribution of E. maculatum by monitoring echolocation calls, (2) to identify and describe foraging habitat, and (3) to make a preliminary examination of spatial and tem- poral patterns of habitat use by spotted bats in the study area. Methods This study was conducted in the canyon bottoms and other relatively low elevation sites in Dinosaur National Monument (109°W, 40°31'N), northwestern Colorado and north- eastern Utah, from 17 May to 9 June 1993. Navo et al. (1992) provided a description of the physiography and vegetation of Dinosaur National Monument. In each study site where spotted bats occurred, I monitored movement patterns and foraging behavior by listening to the low-frequency (15-9 kHz; Leonard and Fenton 1984) echo-location calls of this species, which are clearly audible to the unaid- ed human ear (Woodsworth et al. 1981). As reported previously (Navo et al. 1992), E. maculatum is readily identifiable because it has the lowest-frequency echolocation calls of 'Environmental, Population, and Organisniic Biologv', Universit>' of Colorado. Boulder. CO 80309-0.334. Present address: Department ofBiolog>'. Boston University. Boston. M.\ 0221.5. 78 1995] Spotted Bats in Colorado-Utah 79 any bat species in the study area. Nyctinomops macrotis and Idionycteris phyllotis also pro- duce orientation sounds that are partly audible to humans, with frequencies of 25-17 kHz for N. macrotis (Fenton and Bell 1981) and 24-12 kHz for I. phyllotis (Simmons and O'Farrell 1977). These two species inhabit southern parts of the Colorado Plateau and the Great Basin (Milner et al. 1990, Tumlison 1993), but neither is known to occur as far north as Dinosaur National Monument, extralimital records of N. macrotis notwithstanding (Milner et al. 1990). To further ensure conect call identi- fication, I referred to recordings of known E. maciilatum calls. I also visually identified free- flying individuals (based on conspicuous white venter and large ears) at close range in the beam of a high-intensity flashlight after locat- ing the animals by listening to orientation sounds. It should be noted that /. phyllotis is often buff-colored ventrally and therefore could be visually misidentified as E. maculatum in areas of sympatry. Sampling Locations To investigate the ecological distribution of E. maculatum, I sampled 15 sites at 12 loca- tions representative of common low-elevation habitat types in the area (Fig. 1). Riparian sites (Jenny Lind Rock, 1603 m; Echo Park, 1553 m; Split Mountain Gorge, riverbank and sand- bar, 1439 m) are characterized by wide chan- nels and reaches of calm water bounded by steep sandstone cliffs. Isolated stands of box- elder {Acer negundo) and cottonwood {Poptdiis fremontii) line the riverbanks along with thick- ets of tamarisk {Tamarix sp.). Orchid Draw (1484 m) and Red Wash (1537 m) are dry desert washes characterized by rab- bitbrush {Chrysothamnus nauseosus), sage- brush {Seriphidium tridentata), greasewood [Sarcobatus vermiculatus), and shadscale {Atri- plex confertifolia), with tamarisk dominating drainage bottoms. Echo Park Meadow (1548 m) and Pool Creek (1635 m) are both open meadows with domi- nant ground cover of cheatgrass {Anisantha tectortim), various bunchgrasses, and isolated clumps of boxelder. Echo Park Meadow en- compasses an area of ca 18 ha, bounded by the Green River to the west and high (150-230 m) sandstone cliffs on remaining sides. The mead- ow at Pool Creek (ca 8 ha) is situated at the mouth of a narrow canyon; boxelder and cotton- wood form a dense, continuous canopy over much of the adjacent creek. Remaining locations consist of a moist mead- ow (Hog Canyon, 1635 m), open sagebrush shrublands (Rainbow Park, 1488 m; Island Park, 1512 m), and a narrow canyon with thick riparian vegetation (Jones Hole, 1585 m). Sampling Methods At all locations I remained at a single site during each night of sampling. By pacing from a boxelder, which sei"ved as a focal point of bat foraging activity in Echo Park Meadow, I esti- mated that calls of E. maculatum were detect- able at a distance of roughly 100 m. Therefore, the area sampled at each site is here defined as the air space within a hemisphere of radius 100 m. On several nights periods of high wind and/or rain reduced this range of detectability, with an attendant underestimation of bat activ- ity. Furthermore, sites differed slightly in levels of background noise from nearby streams, the amount of obstructive vegetation, and various atmospheric conditions such as relative humidity, all of which affect the propagation of sound (Lawrence and Simmons 1982). Study sites were situated either at the mouths of canyons or draws or in the middle of open areas where movement patterns of bats could best be assessed and the range of detect- ability was maximized. In locations character- ized by expansive terrain (open meadows or shrublands), I monitored two different sites separated by >300 m on consecutive nights to assess uniformity of activity levels over large areas. All sites were monitored from 2000 to 0200 h with the exception of Echo Park Meadow, which was monitored from 2000 to 0400 h for seven consecutive nights (19-26 May) to assess temporal patterns of foraging activity. At locations where I observed high levels of foraging activity (e.g., Echo Park Meadow and Pool Creek), bat activity was quantified by timing the duration of individual foraging ses- sions and recording the number of feeding buzzes (the increased rate of echolocation pulse repetition associated with attacks on air- borne prey; Griffin et al. 1960). Following Leonard and Fenton (1983), the occurrence of feeding buzzes indicates foraging activity, and a foraging session is defined as the time during which a single spotted bat hunted continuous- ly within the study site. To permit comparison 80 Great Basin Naturalist [Volume 55 1 1 1 1 N A 5 km — 40° 35' N -^. 1 Jones ' ' Hole t Echo Park K ^~ Island fe^ Park Rainbow^,^ Park^^ • / ^ 1 Pool Creek 1 Jenny Lind Rock Echo Park C. Meadow 7 Orchid R^d Draw ^^sh • • C_ Split Mtn. Gorge (sandbar) Yampa River R f|> Split Mtn. Gorge (riverbank) 'In. Hog Canyon 8 1 109° W ~~~^^ UTAH COLORADO c V 1 Green River 1 1 1 _1 Fig. 1. Map showing sampling locations for monitoring activity oi Euderma maculatum in Dinosaur National Monument in late spring 1993. Circles = locations at which transient occurrences of commuting or foraging bats were recorded; triangles = foraging areas (see te.xt for details). of relative levels of activity throughout the night, the time spent by spotted bats in the study site was totaled for every 15-min period sampled. Sampling periods during which heavy rain occurred were not considered. To assess spatial patterns of habitat use, I described the foraging flights of spotted bats into a minicassette recorder and noted flight patterns and use of available foraging space relative to a near-canopy habitat zone (within 8 m of tree canopies) and an open-area zone (the clutter-free air space over the open mead- ow). These habitats con^espond to habitat zones 1 and 4, respectively, proposed by Aldridge and Rautenbach (1987). I recorded the dura- tion of foraging activity occurring within each zone as well as the number of bats simultane- ously present within the study site and inter- actions between them. I recorded the number of feeding buzzes heard during each foraging session for each night of observation at Echo Park Meadow and Pool Creek for the purpose of calculating attack rates (feeding buzzes/min) of foraging spotted bats. I considered only those foraging sessions of duration >3 min during which all feeding buzzes produced by a single individ- ual within the study site could be counted accurately. At other locations where I observed only transient occurrences of foraging or commut- ing spotted bats, activity was quantified by recording the number of bat passes (sensu Fenton 1970) per 15-min sampling period. Results and Discussion I observed spotted bats in 13 of 15 sites sampled (Table 1). At 8 of these locations I observed only commuting bats. Passes of com- muting spotted bats occurred sporadically throughout the night. At locations where two separate sites were monitored on consecutive nights, the number of passes remained fairly constant (passes/night: Echo Park, 5, 4; Hog Canyon, 5, 6; Island Park, 6, 10), and direc- tions of travel appeared similar for bats on both nights. Availability of cliff roosting sites has been suggested as a limiting factor in the distribu- 1995] Spotted Bats in Colorado-Utah 81 Table 1. Number of passes of Eiiderma maciilatiim per 15-min sampling period between 2000 and 0200 h at Dinosaur National Monument (16 May-8 June 1993). See text for general description of habitat types. Number of Number sampling Sampling location of nights periods n X Range Jenny Lind Rock 1 24 0 0 — Echo Park 2 48 9 .19 0-2 Hog Canyon 2 48 11 .23 0-3 Orchid Draw 1 24 18 .75 0-6 Red Wash 1 24 32 1.33 0-1 Split Mountain Gorge (sandbar) 1 24 1 .04 0-5 Split Mountain Gorge (riverbank) 1 24 8 .33 0-8 Rainbow Park 1 16 4 .25 0-1 Island Park 2 48 16 .33 0-2 Jones Hole 1 24 0 0 — All sites 13 280 99 .35 0-8 tion of E. maculatiun (Easterla 1973). The abundance of high ehfifs in Dinosaur National Monument as well as transient occurences of commuting bats throughout a variety of wide- ly separated low-elevation sites (Navo et al. 1992, this study) suggests that suitable roost- ing habitat is widespread throughout the area. However, information about microclimate re- quirements of this species is needed to fully assess actual availability of suitable roost sites. I obsei-ved foraging spotted bats by sight and sound at five locations, three of which (Echo Park, Orchid Draw, Red Wash) involved only transient occurrences of bats that were observed executing steep dives and other abrupt flight maneuvers coincident with feed- ing buzzes as they passed through the area. I observed a single spotted bat foraging over a sand-and-gravel bar at Echo Park, but activity levels at this location were lower than those reported by Navo et al. (1992), who sampled this same site previously. Fairly high levels of activity occurred at Orchid Draw and Red Wash (Table 1), and I heard three feeding buzzes at each site. However, because spotted bats apparently capture prey opportunistically while commuting to specific foraging sites (Wai-Ping and Fenton 1989), observations of foraging bats passing through an area cannot be considered indicative of habitat preferences. At Echo Park Meadow, spotted bats first arrived at the study site at 2123 h ± 11 min Mountain Daylight Time (n = 6 rain-free evenings), always after dark, and remained active throughout the night (Fig. 2). Spotted bats foraged within the study site for 6.22 ± 2.40 min out of every 15-min sampling period between 2100 and 0400 h {n = 2490 min; Fig. 2), and foraging sessions lasted 5.48 ± 2.74 min {n = 187). At Pool Creek, spotted bats hunted within the study site for 6.82 ± 5.03 min out of every 15-min sampling period between 2100 and 0200 h (n = 525 min), and foraging sessions lasted 8.97 ± 8.78 min (n = 30). These activity levels offer strong evidence that open meadows represent important forag- ing habitat for E. maculatum in this area. Comparatively low levels of activity were recorded at riparian sites adjacent to Echo Park Meadow (Echo Park, Jenny Lind Rock). Because no physiographic barriers are present that might restrict accessibility to the bats, it appears that open water courses do not repre- sent foraging areas of choice. These observa- tions agree with those of Leonard and Fenton (1983), who reported that in British Columbia spotted bats foraged in forest clearings and open fields to the exclusion of a nearby river The temporal pattern of foraging activity in Dinosaur National Monument is similar to that reported from British Columbia (Leonard and Fenton 1983), where spotted bats were active throughout the night. Because radio- tracking (Wai-Ping and Fenton 1989) has demonstrated that individual spotted bats hunt on the wing >300 min per night, reports of apparent peaks in nightly activity (which have been especially pronounced in mistnet- ting studies, e.g., Easterla 1973) are likely arti- facts related to the proximit)' of sampling sites to diurnal roosts and/or drinking sites. At Echo Park Meadow and Pool Creek, for- aging spotted bats typically flew in large circu- lar or elliptical orbits at heights of 10-30 m above the ground. In 1088.8 min of observa- tion of foraging spotted bats at Echo Park 82 Great Basin Naturalist [Volume 55 15.0- •- 10.0 E 5.0- Time (h) Fig. 2. Foraging activity patterns ol Euderma nuwiihi- tum at Echo Park Meadow (19-26 May 1993). Bars repre- sent mean time (+ SD) spent by bats in the study site per 15-min sampHng period from 2000 to 0400 h (n = 6 for each 15-min period in the inter\'al 2000-0345 h, n = 4 for 0345-0400 h). Meadow, 81.5% of activity occurred over the open meadow, which constituted roughly 85% of the site, while 18.5% of activity occurred within 8 m of the foliage of fully leafed box- elders at mid- to upper-canopy level. Such activity consisted of bats circling closely above and around individual trees or isolated clumps of trees. I rarely obsei"ved bats within 0.5 m of the canopy, and I never observed hovering flight or other evidence of foliage gleaning. In 290.8 min of obser\'ation of foraging spotted bats at Pool Creek, all activity occurred over the open meadow, although a much larger percentage of the study site area comprised canopies of boxelder and cottonwood than at Echo Park Meadow. The predilection of £. maciilatiim for forag- ing over open terrain in Dinosaur National Monument agrees with the pattern observed in previous studies (e.g., Woodsworth et al. 1981, Leonard and Fenton 1983). Low-fre- quency echolocation calls and long intercall intei-vals suggest that spotted bats use a forag- ing strategy based on long-range prey detec- tion and high-level flight (Simmons and Stein 1980, Woodsworth et al. 1981, Barclay 1986). This strategy likely is best suited to open areas (Neuweiler 1984). Although 1 never directly observed these bats gleaning prey from foliage during this study, observations of near-canopy foraging contrast with those of some other workers (e.g., Wai-Ping and Fenton 1989) who have reported that this species never attacked insects near foliage or any other type of sur- face. Information about individual variability in foraging behavior is needed before drawing conclusions about variabilit)' between popula- tions related to different ecological conditions. At both Echo Park Meadow and Pool Creek, there were 118 instances in which two or three E. macidatum were present within the study site simultaneously. Leonard and Fenton (1983, 1984) estimated that spotted bats in British Columbia maintain a distance of at least 50 m between adjoining foraging areas and suggested that this spacing is accomplished through a combination of mutual avoidance and active monitoring of encroachments by con- specifics. This same system appears to be oper- ating at foraging areas in Dinosaur National Monument. Consistent with observations of Leonard and Fenton (1983), foraging spotted bats often produced agonistic vocalizations when the 50-m buffer zone was breached by an intruding bat. Such vocalizations sounded qualitatively different from feeding buzzes and occurred only during close-range encounters between conspecifics. Information about known individuals and resource availability is needed to elucidate the role of agonistic inter- actions in the foraging ecology of E. niacidatum. During this study I heard a total of 247 feed- ing buzzes, and never more than one per min from the same individual. In a sample of 37 foraging sessions, spotted bats attacked an insect eveiy 2.15 min on average (0.466 ± 0.294 attacks/min, range 0.16-0.94; n = 152 feeding buzzes). These rates generally agree with val- ues reported in previous studies (Leonard and Fenton 1983, Wai-Ping and Fenton 1989), fur- ther confirmation that this species attacks prey at a rate much lower than is typical of bats that forage from continuous flight (Barclay 1985, Hickey and Fenton 1990). Density of clutter in an environment im- poses differential constraints on the maneuver- ability and perceptual capacities of bats, there- by determining the accessibility of different habitats b\' influencing foraging efficiency (Neu- weiler 1984, Aldridge and Rautenbach 1987, Fenton 1990). Spotted bats appear to forage preferentially in open areas, which may be 1995] Spotted Bats in Color.\do-Utah 83 related to the use of a long-range foraging stiat- egy (Barclay 1986), and the ability to exploit edge situations may reflect a measure of behavioral flexibility in this regard. Because spotted bats are obviously not greatly restrict- ed in foraging habitat with regard to vegeta- tion associations (Wai-Ping and Fenton 1989, Navo et al. 1992), structural features of the environment related to density of clutter may be more predictive of habitat suitability and the use of available foraging space. However, information on individual variability is needed before drawing conclusions about the foraging strategy of this species. Acknowledgments I am grateful to D. M. Armstrong, K. W. Navo, M. B. Fenton, M. L. Leonard, M. A. Bogan, C. E. Bock, J. A. Gore, and G. T. Skiba for advice regarding study site locations and sampling methods. I thank the personnel of Dinosaur National Monument, and especially S. J. Petersburg, for cooperation and for shar- ing knowledge of the area. Critical comment on the manuscript fiom D. M. Amistrong, M. B. Fenton, R. M. Timm, M. J. O'Farrell, and two anonymous reviewers was much appreciated. Funding was provided by the Undergraduate Research Opportunities Program, University of Colorado at Boulder. Literature Cited Aldridge, H. D. J. N., AND I. L. Rautenbach. 1987. Morphology, echolocation, and resource partitioning in insectivorous bats. Journal of Animal Ecology 56: 763-778. Barclay, R. M. R. 1985. Long- versus short-range forag- ing strategies of hoary {Lasiunis cinereus) and silver- haired {Lasioiujcteris noctivagans) bats and the con- sequences for prey selection. Canadian Journal of Zoology 63: 2507-2515. . 1986. The echolocation calls of hoaiy {Lasiunis cinereus) and silver-haired [Lasiomjcteris noctiva- gans) bats as adaptations for long- versus short-range foraging strategies and the consequences for prey selection. Canadian Journal of Zoology 64: 2700-2705. Easterla, D. a. 1973. Ecology of the 18 species of Chiroptera at Big Bend National Park, Te.\as. North- west Missouri State University- Studies 349: 1-165. Fenton, M. B. 1970. A technique for monitoring bat activity with results obtained from different environ- ments in southern Ontario. Canadian Journal of Zoology 48: 47-51. . 1990. The foraging behaviour and ecologx' of ani- mal-eating bats. Canadian Journal of Zoology 68: 411-422. ' Fenton, M. B., and G. P Bell. 1981. Recognition of species of insectivorous bats by their echolocation calls. Journal of Mammalogy 62: 233-243. Fenton, M. B., D. C. Tennant, and J. Wyszeckl 1987. LJsing echolocation calls to measure the distribution of bats: the case of Euderma maculatum. Journal of Mammalogy 68: 142-148. Griffin, D. R., F A. Webster, and C. R. Michael. 1960. The echolocation of flying insects by bats. Animal Behavior 18: .5.5-61. HiCKEY, M. B. C, and M. B. Fenton. 1990. Foraging by red bats [Lasiunis boreaUs): Do intraspecific chases mean tenitoriality? Canadian Journal of Zoology 68: 2477-2482. Lawrence, B. D., and J. A. Simmons. 1982. Measurements of atmospheric attenuation at ultrasonic frequencies and the significance for echolocation by bats. Journal of the Acoustical Society of America 71: 585-590. Leonard, M. L., and M. B. Fenton. 1983. Habitat use by spotted bats {Euderma maculatum, Chiroptera: Vespertilionidae): roosting and foraging behavior Canadian Journal of Zoologv' 61: 1487-1491. . 1984. Echolocation calls of Eudenna maculatum (Chiroptera: Vespertilionidae): use in orientation and communication. Journal of Mammalogy' 65: 122-126. Milner, J., C. Jones, and J. K. Jones, Jr. 1990. Nijcti- nomops macrotis. Mammalian Species 351: 1—1. Navo. K. W, J. A. Gore, and G. T Skiba. 1992. Obsewa- tions on the the spotted bat, Eudenna maculatum, in northwestern Colorado. Journal of Mammalogy 73; 547-551. Neuweiler, G. 1984. Foraging, echolocation, and audi- tion in bats. Natunvissenschaften 71: 446—455. O'Farrell, M. J. 1981. Status report: Eudenna maculatum (J.A. Allen). United States Fish and Wildlife Sei-vice, Office of Endangered Species. 29 pp. Simmons, J. A., and M. J. O'Farrell. 1977. Echolocation by the long-eared bat, Plecotus phijllotis. Journal of Comparative Physiology 122: 201-214. Simmons, J. A., and R. A. Stein. 1980. Acoustic imaging in bat sonar: echolocation signals and the evolution of echolocation. Joumal of Comparative Physiology 135: 61-84. Tumlison, R. 1993. Geographic variation in the lappet- eared bat, Idiomjcteris phijllotis, with descriptions of subspecies. Joumal of Mammalogy 74: 412^21. Wai-Plng, v., and M. B. Fenton. 1989. Ecology of spot- ted bats {Eudenna maculatum): roosting and forag- ing behavior. Journal of Mammalogy 70: 617-622. Watkins, L. C. 1977. Eudenna maculatum. Mammalian Species 77: 1-4. Woodsworth, G. C, G. R Bell, .-vnd M. B. Fenton. 1981. Obsei-vations of the echolocation, feeding behavior, and habitat use of Eudenna maculatum (Chiroptera: Vespertilionidae) in southcentral British Columbia. Canadian Journal of Zoology- .59: 1099-1102. Received 7 Febnianj 1994 Accepted 20 June 1994 Great Basin Naturalist 55(1), © 1995, pp. 84-88 THE CHRYSOTHAMNUS-ERICAMERIA CONNECTION (ASTERACEAE) Ijoran C. Anderson ^ AlJSTfUCT. — The geniis Chrysothdiiuiiis (Asteraceae) contains 16 species. Recently, 4 species were transferred to Ericameria, and the remaining 12 were left in Chnjsothamnus. The remaining species are now transferred to Ericameria as £. albida, E. depressa, E. eremobia, E. graminea, E.filifolia (formerly C. greenei). E. hwnilis, E. linifolia, E. molesta, E. pulchella, E. pulclielloides (a fossil species), £. spathulaia, E. vaseyi, and E. liscidiflora. Section alignments are given, and some infraspecific combinations are also made. Key words: Chr>sothamnus, Ericameria, rahhithnish, nomenclature transfers. The Asteraceae are a relatively young group, and yet they have experienced rapid evohition into a great number of species. One result is that many taxa appear more distant moq^holog- ically (phenotypically) than they actually are genetically, and, conversely, some taxa may appear more closely related than they are. These situations have created havoc amongst taxonomists in their attempts to circumscribe genera. This is particularly evident in the tribe Astereae. In 1894, E. L. Greene stated: In North America the Astereae are excessively numerous, and no natural assemblage of plants has seemed to present such difficulties to the systema- tist; and the widest conceivable diversities of opin- ion as to the limits of genera have found expression among botanists when undertaking to classify them. The situation continues a century later The genus Haplopapptis was thought to be an unnatural, polyphyletic assemblage by many (e.g., Shinners 1950, Anderson 1966, Johnston 1970, Turner and Sanderson 1971, Clark 1977, Urbatsch 1978). Nevertheless, because there was no suitable taxonomic reor- ganization of the group, I continued to describe new taxa in Haplopappiis (Anderson 1980a, 1983b), even though the species would probably be placed in some other genera at a later date. Recently, additional data have contributed to a clearer understanding of the relationships in this and related groups (Morgan and Simpson 1992), and several genera have been recog- nized for North American Haplopappi. In a 1976 presentation at national meet- ings, I discussed the close affinity of Chnjso- thamnus with woody elements of Haplopappiis and suggested that the Asiris-Ericamcria- Macronema complex of Haplopappiis probably should be included in Chnjsothamnus. But, given the state of knowledge at that time, I de- ferred. In 1990, Nesom reorganized Ericameria as a genus to include Asiris and Macronema. Recently, based on occurrences of intergeneric hybrids (Anderson and Reveal 1966, Anderson 1970) and DNA data (Morgan and Simpson 1992), Nesom and Baird (1993) transferred four species of Chnjsothamnus into Ericameria (C. nauseosus and C. parnji of section Nauseosi and C paniculatus and C. teretifolius of section Piinctati). They continued to recognize Chnjso- thamnus as a distinct (but smaller) genus and gave arguments for separating the two. A problem in separating Ericameria and Chnjsothamnus (sensu Nesom and Baird) is the occurrence of hybrids (Anderson 1970, 1973) between C. nauseosus (their Ericameria) and C. alhidiis (their Chnjsothamnus). After study- ing a specimen of only one of the three collec- tions involved, Nesom and Baird (1993) deval- ued the connection by stating that "the plant in question [is] characteristic of C. nauseosus, and we identify it as C. nauseosus, finding no strong reason to implicate C. albidus in its parentage." They stated that achenes of C. albidus are linear and consistently producing 10 slightly raised nerves, whereas those of C. nauseosus are nar- rowly obovate with 5-7 nerves. Actually, ach- enes of both species can be characterized as being narrowly cylindrical. The number of vascular bundles (associated with the nerves) in the achenes averages approximately 7 and 'Dcpartuifiit ol Biological Science, Florida State University. Tallahassee, FL 32306-2043. 84 1995] Chrysothminus-Ericameria Connection 85 ranges from 5 to 10 (but mostly 6-8 in Ash Meadows) for C. alhidus (Anderson 1970, 1973), whereas achene bundle number in C. nauseosus ranges from 5 to 12 (but is restrict- ed to 5 for those in Ash Meadows). The interspecific hybrid examined by Nesom and Baird {Beatleij 11894, KSC) was studied anatomically by Anderson (1973); its hybridity is indicated by low pollen fertility and by mor- phological intermediacy between the two species in its revolute leaves, in vascular bundle number in the ovary wall, in corolla lobe length, and in anther appendage length. It has secretoiy canals in the ovaiy wall and glandular trichomes on the corolla tube (like C. nauseo- sus, unlike C. alhidus) and ovaiy wall (unlike C. nauseosus, like C. alhidus). Further, proge- ny from one of my C. alhidus garden plants also has low pollen fertility and looks interme- diate between its seed parent and C. nauseo- sus (Anderson 1970). Its flowers have secreto- ry canals in the ovar>' wall and glandular tri- chomes on the corolla tube but lack glandular trichomes on the ovary wall; those three fea- tures are characteristic of C nauseosus but not of C. alhidus (the seed parent), clearly suggest- ing hybridity. If existence of interspecific hybrids is used to justify transferring C. nau- seosus to Ericameria, then this feature also argues for bringing the remainder of Chryso- thamnus into Ericameria. The warranted position of Chnjsothamnus teretifolius in Ericameria is taken by Nesom and Baird (1993: 80) because, like many Eri- cameria species (sensu strictum), that species has the tendency for the "resiniferous ducts that are almost always distinctly associated with the phyllaiy midvein to expand near the apex of the phyllary." This characteristic also occurs in many species of Chnjsothamnus (sensu Nesom and Baird) as illustrated for C. vaseiji (Anderson 1963: 660) and cannot be used to distinguish the two groups. I have ob- served adjacent populations of C. viscidiflorus subsp. puherulus in which plants of one had prominently enlarged resin ducts at the phyl- lary tips and plants of the other did not. With the transferral of four species from Chnjsothamnus to Ericameria, Nesom and Baird (1993) separate the two newly struc- tured genera with six criteria. (1) Leaves 3- nerved for Chrysothamnus and 1-nerved for Ericameria — but many of their Ericameria have prominently 3-nerved leaves. Hall and Clements (1923) used nei"ve number to distin- guish C. nauseosus ssp. graveolens from spp. consimilis (so the character is variable even within a species). Many of the latter group, such as C. alhidus, C. greenei, and some forms of C. viscidiflorus, appear to have 1-nerved leaves. Actually, all species of Ericameria and Chrysothamnus have trilacunar, 3-trace nodal anatomy (personal obsei^vation); thus, the char- acter of 1 versus 3 nerves is a matter of per- ception, not of fact. (2) Leaf margins ciliate in the former and never in the latter — but C. alhidus, C. eremohius, and C. viscidiflorus subsp. planifolius of the former have entire leaf margins; also in that group, C. pulchellus subsp. pulchellus has entire leaf margins, whereas subsp. haileyi has ciliate leaf margins, and some populations of C. gramineus and C. vaseyi have entire leaf margins, but others do not. Ericameria (sensu Nesom 1990) has sev- eral species that have leaves with ciliate leaf margins, fairly prominent in E. cooperi and less so in several other species (e.g., E. cervina, E. nana, E. ophitidis, and E. zionis). (3) Corollas more or less abruptly broadened from the tube into the throat with long, recurving or coiling lobes in the former and corollas tubu- lar with short, erect or spreading lobes in the latter — but corollas of C. spathulatus (of the latter) have relatively broad tubes that lack noticeably flaring throats, C. humilis (of the former) has tubular corollas with short, erect lobes (Anderson 1964: 226), and C. nauseosus ssp. ceruminosus (of the latter) has corollas that are abruptly broadened from the tube into the throat with long, spreading lobes. (4) Style appendage collecting hairs merely papil- late in the former, whereas they are long and sweeping in the latter — but C. alhidus, C. molestus, C. pulchellus, and certain popula- tions of C. viscidiflorus (all of the former) have style appendages with moderately long, sweep- ing hairs. Diversity in collecting hairs is greater in Chrysothamnus (sensu Anderson 1986) than Nesom and Baird (1993) imply and does not fall into two groups. Collecting hair length may be correlated with other floral features; namely, the corollas, style lengths, and pollen volumes of the former group (Anderson 1966) are generally smaller than those of the latter (5) Involucral bracts in vertical files in the for- mer (caveat noted) and usually not in vertical files in the latter — but, perhaps the most strongly aligned bracts occur in C. nauseosus 86 Great Basin Natur^vlist [Volume 55 ssp. arenarius (of the latter). (6) Achenes glan- dular with nonresinous nerves in the former and eglandular (with duplex hairs) and resinous nerves in the latter — but only five species of the former have glandular achenes (in some they are hidden by duplex hairs) and the other seven do not, having either glabrous achenes or achenes with duplex hairs exclu- sively (Anderson 1970, 1983a), and many have resin canals associated with the bundles of the achenes, admittedly fewer than in those of the latter but well developed in C. molestus of the former. Also, C. paniculatus (of the latter group- ing) lacks resin canals in its achenes (Anderson 1970). None of these six sets of characteristics can be used to consistently separate the two groups. Clearly, Chrysothamnus (sensu Anderson 1986, not Nesom and Baird 1993) is fairly homogeneous and should not be dismem- bered. If some are to go into Ericameria (and DNA data suggest they should), then all should go into Ericameria. Therefore, the remaining 12 species of Chrysothamnus are transfened to Ericameria, and new combinations are made here. 1. Ericameria albida (M. E. Jones ex A. Gray) L. C. Anders., comb. nov. Basionyni: Bigelovia albi- da M. E. Jones ex A. Gray, Proc. Amer. Acad. Aits 17: 209. 1882. Chrysothamnus albidus (M. E. Jones ex A. Gray) E. Greene, Eiythea 3: 107. 1895. 2. Ericameria depressa (Nutt.) L. C. Anders., comb. nov. Basionym: Chrysuthainnus deprcssus Nutt., Proc. Acad. Nat. Sci. Philadelphia 4: 19. 1948. Linosyris depressa (Nutt.) Ton., in Stigreaves, Kept. Exped. Zuni & Colorado Rivers 161. 1853. Bigelovia depressa (Nutt.) A. Gray, Proc. Amer. Acad. Arts 8: 643. 1873. 3. Ericameria eremohia (L. C. Anders.) L. C. Anders., comb. nov. Basionym: Chrysothamnus eremobius L. C. Anders., Brittonia 35: 2.3. 1983. 4. Ericameria graminea (H. M. Hall) L. C. Anders., comb. nov. Basionym: Chrysothamnus gramineus H. M. Hall, Muhlenbergia 2: 342. 1916. Petradoria discoidea L. C. Anders., Trans. Kansas Acad. Sci. 66: 676. 1964. 5. Ericameria filifoUa (Rydb.) L. C. Anders. comb. nov. Basionym: Chrysothamnus filijolius Rydb., Bull. Toney Bot. Club 28: 503. im\. Bigelovia greenei A. Gray, Proc. Amer. Acad. Arts 11: 75. 1876 [not Ericameria greenei (A. Gray) Nesom]. Chrysothamnus greenei (A. Gray) E. Greene, Erythea 3: 94. 1895. Chrysothamnus pumilus var acuminatus A. Nels., Bot. Gaz. 28: 376. 1899. Chrysothamnus scoparius Rydb., Bull. Torrey Bot. Club 28: 504. 1901. Chrysothamnus laricinus E. Greene, PittoniaS: 110. 1903. 6. Ericameria humilis (E. Greene) L. C. Anders., comb. nov. Basionym: Chrysotlunnnus humilis E. Greene, Pittonia 3: 24. 1896. 7. Ericameria linifolia (E. Greene) L. C. Anders., comb. nov. Basionym: Chrysothamnus linifolius E. Greene, Pittonia 3: 24. 1896. 8. Ericameria molesta (Blake) L. C. Anders., comb. nov. Basionym: Chrysoihamnus viscidiflorus var. molestus Blake, J. Wash. Acad. Sci. 30: 368. 1940. Chrysothamnus molestus (Blake) L. C. Anders., Madroiio 17: 222. 1964. 9a. Ericameria pulchella (Gray) L. C. Anders., comb. nov. Basionym: Linosyris pulchella A. Gray, Pi. Wright. [Smidis. Contr. Know!.] 3(5): 96. 1856. Bigelovia pulchella (A. Gray) A. Gray, Proc. Amer Acad. Arts 8: 643. 1873. Chrysothamnus pulchellus (A. Gray) E. Greene. Enthea 3: 107. 1895. 9b. Ericameria pulchella subsp. baileyi (Woot. & Standi.) L. C. Anders., comb. nov. Basionym: Chrysothamnus baileyi Woot. & Standi., Contr. U.S. Nati. Herb. 18: 181. 1913. 9c. Ericameria pidchella subsp. pulchella var. elatior (Standi.) L. C. Anders., comb. nov. Basionym: Chrysothamnus elatior Standi., Proc. Biol. Soc. Wash. 26: 118. 1913. This variety with uniformly pubes- cent leaves occurs sporadically in a few populations of the typically glabrous-leaved subspecies pulchel- lus and does not warrant a higher taxonomic status than this quadrinomial affords. 10. Ericameria spathulata (L. C. Anders.) L. C. Anders., comb. nov. Basionym: Chrysothamnus spathulatus L. C. Anders., Madroiio 17: 226. 1964. Chrysothamnus viscidiflorus var. ludens Shinners, Sida 1: 374. 1964. 11. Ericameria vaseyi (A. Gray) L. C. Anders., comb. nov. Basionym: Bigelovia vaseyi A. Gray, Proc. Amer. Acad. Arts 12: 58. 1876. Chrysothamnus vaseyi (A. Gray) E. Greene, Erythea 3: 96. 1895. Chrysothamnus bakeri E. Greene, Pittonia 4: 152. 1900. 12a. Ericameria viscidiflora (Hook.) L. C. Anders, comb. nov. Basionym: Crinitaria viscidiflo- ra Hook., Fl. Bor. Am. 2: 24. 1834. Chrysothanmus viscidiflorus (Hook.) Nutt. Trans. Amen Philos. Soc. 11, 7: 324. 1840. Bigelovia douglasii A. Gra\', Proc. Amer Acad. Arts 8: 645. 1873. Chrysotha)nnus dou- glasii (A. Gray) Clements & Clements, Rocky Mtn. Els. 226. 1914. Chrysothamnus pumilus Nutt., Trans. Amer Philos. Soc. II, 7: 323. 1840. Linosyris serrulata Torr, Stansbuiy Rep. 1: 389. 1851. Chryso- thamnus serrulatus (Torn) Rydb., Bull. Tonxy Bot. Club 33: 152. 1906. Chrysothanmus tortifolius E. Greene, Fl. Fran. 368. 1897. Chrysodianmus Icuco- cladus E. Greene, Pittonia 5: 59. 1902. Chrysodiam- nus stenolepis Rydb., Bull. Torrey Bot. Club 37: 131. 1910. 12b. Ericameria viscidiflora subsp. viscidiflora var. latifolia (D. C. Eaton) L. C. Anders., comb, nov. Basionxni: Linosyris viscidiflora van latifolia 1995] Chrysothamnus-Ericameria Connection 87 D. C. Eaton, Bot. King Expl. 157. 1871. Chnjsotluiin- mts latifolins (D. C. Eaton) Rydh., Bull. Toirey Bot. Club 33: 152. 1906. 12c. Ericameria viscidiflora subsp. viscidiflora van stenophylla (A. Gray) L. C. Anders., comb. nov. Basionym: Bigelovio doiiglasii var. stenophylla A. Gray, Proc. Amer Acad. Arts 8: 646. 1873. Chryso- thamnus stenophyllus (A. Gray) E. Greene, Eiythea 3: 94. 1895. These quadrinomials (12b and 12c) identify sporadic but rather distinctive morpho- types that occur in the northern regions of this sub- species (for conceptual distinction between sub- species and variety, see Anderson 1980b) 12d. Ericameria viscidiflora subsp. axillaris (Keck) L. C. Anders., comb. nov. Basionym: Chryso- thainnii.s (ixillaris Keck, Aliso 4: 104. 1958. 12e. Ericameria viscidiflora subsp. lanceolata (Nutt.) L. C. Anders., comb. nov. Basionym: Chryso- thamnus lanceolatus Nutt., Trans. Amer Philos. Soc. II, 7: 324. 1840. Chrysothainiuts elegons E. Greene, Erythea 3: 94. 1895. Bigelovia doiiglasii var spathii- lata Jones, Proc. Calif Acad. Sci. II 5: 690. 1895. Chrysothamnus glaticus A. Nels., Bull. Torrey Bot. Club 25: 377. 1898. Chrysothamnus pumilus var latus A. Nels., Bot. Gaz. 54: 413. 1912. 12f. Ericameria viscidiflora subsp. planifolia (L. C. Anders.) L. C. Anders., comb. nov. Basionym: Chrysothamnus viscidiflorus subsp. planifolius L. C. Anders., Madrono 17: 223. 1964. 12g. Ericameria viscidiflora subsp. puberula (D. C. Eaton) L. C. Anders., comb. nov. Basionym: Linosyris viscidiflora var puberula D. C. Eaton, Bot. King Expl. 158. 1871. Chrysothamnus puherulus (D. C. Eaton) E. Greene, Erythea 3: 93. 1895. Chrysothamnus marianus Rydb., Bull. Torrey Bot. Club 37: 131. 1910. The following sections in Ericameria are proposed to accommodate these species trans- fers: Ericameria section Chrysothamnus (A. Gray) L. C. Anders., comh. nov. Basionym: Bige- lovia section Chrysothamnus A. Gray., Proc. Amer. Acad. Arts 8: 641. 1873. This section in- cludes E. albida, E. filifolia, E. humilis, E. lini- folia, E. spathulata, and E. viscidiflora. Ericam- eria section Gramini (L. C. Anders.) L. G. Anders., comb. nov. Basionym: Chrysothamnus section Gramini L. G. Anders., Proc. Symp. Biology o{ Artemisia and Chrysothamnus 29. 1986. This section includes E. eremobia and E. graminea. Ericameria section Pulchelli (Hall & Clements) L. C. Anders., comb. nov. Basionym: Chrysothamnus section Pulchelli Hall & Clements, Carnegie Inst. Publ. 326: 175, 193. 1923. This section includes E. depressa, E. molesta, E. pulchella, and E. vaseyi. Additionally, there is a fossil species, Eri- cameria pulchelloides (L. G. Anders.) L. G. Anders., comh. nov. Basionym: Chrysothamnus pulchelloides L. G. Anders., Great Basin Naturahst 40: 351. 1980. Nesom and Baird (1993) suggest the Chnj.so- thamnus taxa that I have just transferred to Ericameria should be placed in a restructured genus to include elements of Hesperodoria, Petradoria, and Vanclevea. They conclude that chloroplast DNA data (Suh 1989) show Petra- doria to be integrally related to the Solidago lineage and far removed from Ericameria. However, they note that neither Suh (1989) nor Morgan and Simpson (1992) sampled any taxa o( Chrysothamnus sensu Nesom and Baird. These taxa need DNA profiles determined because they certainly do not make a morpho- logically compatible grouping with Petradoria or Vanclevea. For example, Petradoria (Anderson 1963) has radiate heads with disk flowers that lack stigmatic areas on the style branches and have abortive ovaries, and Vanclevea (Anderson and Weberg 1974) has large turbinate heads with many phyllaries, many flowers, and a tardily deciduous pappus of paleacous awns — none of these conditions are found in Chryso- thamnus sensu Nesom and Baird. The cohe- siveness of Chrysothamnus sensu Anderson is further illustrated in that C. spathulatus twigs emit odor similar to that of C. nauseosus (Anderson 1964: 227). Two alternate taxonomies are now available: one for Chrysothamnus as a genus (Anderson 1986) or as a component of Ericameria (Nesom and Baird 1993, and here); both are preferable to merging some elements of Chrysothamnus with Petradoria or Vanclevea. Acknowledgments James Reveal and Arnold Tiehm offered constructive comments on the manuscript. Literature Cited Anderson, L. C. 1963. Studies on Petradoria (Compos- itae): anatomy, cytology, and ta.xonomy. Transactions of the Kansas Academy of Science 66: 632-684. . 1964. TiL\onomic notes on the Chrysothamnus vis- cidiflorus complex (Astereae, Compositae). Madroiio 17; 222-227. . 1966. Cytota.xonomic stndies in Chrysothamnus (Astereae, Compositae). American Journal Botany 53: 204-211. . 1970. Floral anatomy of Chrysothamnus (Astereae, Compositae). Sida 3: 466-503. . 1973. Unique Chrysothamnus hybridizations in Ash Meadows, Nevada. Bulletin of the ToiTe\^ Botanical Club 100: 171-177. 88 Great Basin Naturalist [Volume 55 . 1980a. ll(i})lopa})pus iil}Hmis (Asteraceae): a new species fVorii Nevada. Great Basin Naturalist 40; 73-77. . 19S0h. Identity of narrow-leaved ahnjsofhaiiiiius viscidifloni.s (Asteraceae). Great Basin Naturalist 40: 117-120. . 1983a. CJinjsothammt.s ereinobius (Asteraceae): a new species troiu Nevada. Brittonia 35: 23-27. . 19831). Haplopappus crispii.s and //. zionk (Aster- aceae): new species from Utali. (Ireat Basin Naturalist 43: 358-364. 1986. An overview of the genus Chnjsotliatnnua (Asteraceae). Pages 29-45 iii E. D. McArtliur and B. L. Welsh, eds.. Proceedings, Symposium on the Biol- ogy of Artemma and Chnjsothamniis. USDA, Forest Service, Intermountain Research Station, Ogden, UT 398 pp. Anderson, L. C., .and J. L. Reveal. 1966. Chnjsothainnus bolanderi, an intergeneric hybrid. Madrono 18: 225-233. Anderson, L. C., and R S. Weberg. 1974. The anatomy and taxonomy of Vanclevea (Asteraceae). Great Basin Naturalist 34: 151-160. Clark, W. D. 1977. Chemosystematics of the genus Hazardia (Compositae). Journal of the Arizona Academy of Science 12: 16. Greene, E. L. 1894. Observations on the Compositae. IV. Erythea 2: 53-60. Hall, H. M., and E E. Clements. 1923. The phylogenet- ic method in tiLxonomy: the North American species of Artemisia, Chrijsothammis, and Atriplex. Carnegie Institute Publication 326: 1-355. Johnston, M. C. 1970. Compositae. Pages 152.3-1744 in D. S. Correll and M. C. Johnston, Manual of the vas- cular plants of Texas. Texas Research Foundation, Renner, TX. 1881 pp. .Morgan, D. R., and B. B. Slmpson. 1992. A systematic study of Machaeranthera (Asteracaee) and related groups using restriction site analysis of chloroplast DNA. Systematic Botany 17: 511-531. Nesom, G. L. 1990. Nomenclatural summary of Ericaineria (Asteraceae: Astereae) with the inclusion of Haplopappus sects. Asiris and Macronema. Ph\ tol- ogia 68: 144-155. Nesom, G. L., and G. I Baird. 1993. Completion of Ericaineria (Asteraceae: Astereae), diminution of Chnjsothainnus. Phytologia 75: 74-93. Shinners, L. II. 1950. Notes on Texas Compositae, IV, V. Field and Laboratory 18: 25^2. SuH, Y. 1989. Phylogenetic studies of North American Astereae (Asteraceae) based on chloroplast DNA. Unpublished doctoral dissertation. University of Texas, Austin. Turner, B. L., and S. Sanderson. 1971. Natural hybridi- zation between the Composite "genera" Machaeran- thera and Haplopappus (sec. Blepharadon). American Journal of Botany 58: 467. Urbatsch, L. E. 1978. The Chihuahuan Desert species of Ericameria (Compositae, Astereae). Sida 7: 298-303. Received 7 Felniranj 1994 Accepted 2 June 1994 Great Basin Naturalist 55(1), © 1995, pp. 89-91 REPRODUCTIVE BEHAVIOR IN MERRIAM'S CHIPMUNK {TAMIAS MERRIAMI) Stephen B. Compton^ and J. R. Callahan^ Key words: Tamias, Eutamias, Neotamias, chipmunk, copulation, olfaction. The literature contains little information regarding mating chases and copulation in any of the western chipmunks {Tamias, subgenus Neotamias). Callahan (1981) reported mating chases for Merriam's {T. merriami) and dusky chipmunks {T. ohscurus), but noted (unpub- lished) that both copulating pairs were partly concealed by fohage. Larson (1981) described two copulations for Merriam's chipmunk, but a careful reading suggests that one of these was a mounting attempt by an immature male and the other was observed from a consider- able distance. Best and Granai (1994) found no references on this subject other than Callahan (1981) and Larson (1981). There has been some speculation and dif- ference of opinion regarding reproductive iso- lating mechanisms in parapatric species of western chipmunks. Blankenship and Brand (1987) reported differences in vocal behavior between Tamias merriami and T. ohscunis at Black Mountain (Riverside County, CA) and noted a possible role in reproductive isolation. One of us (JRC), however, had previously con- ducted a more extensive study of vocal behavior in these two cryptic species at Black Mountain from 1975 to 1980. Vocalizing individuals were collected to confirm species identity, and sono- grams were prepared and measured; yet no statistically significant vocal differences were found (Callahan 1981, and in preparation). Ecological, olfactory, and mechanical barriers to hybridization also have been suggested (Callahan 1977, 1981, Patterson 1984). These hypotheses cannot be tested without more data on chipmunk reproductive behavior. Accordingly, this note provides the first detailed description of western chipmunk cop- ulation that has been published, to the best of our knowledge. Comparative data for other western chipmunk species would be of interest. The obsei-vation was made 1 April 1994 in a wooded residential area in Idyllwild, Riverside County, CA (elevation 1590 m), between 1000 and 1130 h. The habitat is mixed-conifer forest dominated by incense cedar {Calocedrus decurrens), yellow pine {Pimis ponderosa), live oak {Quercus chrysolepis), and black oak {Q. kelloggii), with a sparse understory of chapar- ral shrubs. When the observer arrived at 1000 h, six to seven Merriam's chipmunks, many of them males, were running over, around, and through a large woodpile while performing conspicuous leaping maneuvers. No agonistic interaction was observed. It was not possible at this stage to identify the female(s) or to tell in which direction the "chase " was headed. The overall effect was somewhere between a Sciurus-hke mating chase (e.g., Thompson 1977), in which several males follow one female, and a lek, involving male display. The chase covered an area 13-15 m in diameter but centered on the woodpile and a nearby heap of smaller pine branches. After about 20 min, one chipmunk (later identified as female) ran up on one of the piled branches. A second chipmunk approached and they ran around for a few seconds. The female stopped on a branch and the other chipmunk, a male, ran up beside her. His entire right side was in contact with her left side for about 1 sec, during which he made a nuzzling motion with the right side of his face on the rear left portion of her face. The expected nasal/genital contact was not observed, but the pair had been out of sight for a short time previously and this could have occurred. The female then jumped to another branch, which was 5 cm in diameter and 20 cm above the ground, sloping at a 25° angle so that the female was facing downhill. Copulation then occurred only 2 m from the observer (who was inside a parked vehicle). 'Send reprint requests to Box 3140, Hemet, CA 92546. ^Museum of Southwestern Biology, University of New Mexico, Albuquerque, NM 871.31. 89 90 Great Basin Natufl\list [Volume 55 The female crouched on the branch as if resting. The male jumped to that branch and quickly grasped the female from the rear with his forelimbs around her pectoral region. The female's tail was deflected to the side and slightly raised, and the male's tail was extend- ed to the rear. Copulation consisted of four series of pelvic thrusts. Each series (except the last) lasted about 4 sec and comprised an esti- mated 12-24 thrusts, at a rate of 3-6 per sec. Each series of rapid thrusts was followed by a short resting period, during which the male stopped thrusting and brushed his face (mouth, nose, and chin) from side to side 2-4 times against the back of the female's neck. The fourth and last series of thrusts was short- er than the first three. The male then released his grip on the female, dismounted, and ran off into the woodpile. The female, who had remained motionless during the act, remained on the branch about 1 sec and then also ran to the woodpile. The entire copulation lasted about 18 sec. Although several male chipmunks partici- pated in the chase, none of them approached the copulating pair. No chipmunks were heard vocalizing during the mating chase or copula- tion. We did not note any pre-mating vocal dis- play or Lockrufe by the estrous female (Callahan 1981), but we were not present on the days when the display (if any) would have taken place. The Tamias vocal display has been reported for a few species of chipmunks by Callahan (1981), Blake (1992), and others. It is not clear whether this vocal display is universal or occurs only at low population densities, when the female benefits by attract- ing more distant males. No further copulations were seen, but as many as seven male chipmunks continued to mn around the same woodpile for another hour. The level of activity appeared to decrease, and there were none of the prodigious leaps seen earlier. The group then gradually dispersed as individuals headed for an adjacent area where other chipmunks were heard giving occasional "chipper" vocalizations (not the long series of chips that characterizes the Lockrufe). The behavior described above suggests that scent glands play a key role in reproductive be- havior of this species. Larson (1981) and others have noted that male chipmunks have scent glands near the chin and angle of the jaw (oral glands) that become enlarged during the breeding season. Scent marking is prevalent in sciurids, but usually this means marking the ground or a branch, not marking another ani- mal. The "nuzzling" and "brushing" behavior of the male Merriam's chipmunk, before and during copulation, suggests that he was scent marking the female. Conspecific marking has been described for various mammals, such as rabbits (Mykytowycz 1965), but not for sciurids. Gurnell (1987) describes "face-wiping" behavior by various tree squirrels, but only in the context of sub- strate marking and (in Paraxenis) self-groom- ing; his description of copulation in Sciiirus and Tamiasciurus says nothing about the male marking the female. With reference to olfacto- ly communication in ground squirrels, Halpin (1984) wrote that "there is no experimental evidence that conspecific marking . . . actually occur[s] among the sciurids." Our obsei-vation indicates that conspecific marking does occur in Merriam's chipmunk as a component of reproductive behavior Without experimental data, it is not possible to deter- mine the significance of this marking. Pair bonding comes to mind, but there is no good evidence of long-term pair bonding in Merriam's or any other species of western chipmunk, despite many years of field obser- vation. Other possibilities include the follow- ing: (1) the marking induces some required physiological state in the female; (2) the mark- ing tells other males that the female has already mated (before the copulation plug forms and the message becomes redundant); or (3) the marking reinforces a short-term pair bond to ensure that subsequent copulations (if any) on the day of estrus will be with the same male. Larson (1981) indicated that the same estrous female sometimes copulates more than once. Mortality from all causes is higher for male than for female chipmunks (Smith 1978), per- haps due in part to the dispersal and exposure associated with the breeding season (Callahan 1981). After incurring the risk of predation and expending considerable energy on the mating chase, it should be to the male's advan- tage to ensure that his genes are passed to all the female's offspring of the season. Literature Cited Best, T. L., and N. J. Granai. 1994. Tamias meniami. Mammalian Species 476: 1-9. 1995] Notes 91 Blake, B. H. 1992. Estrous calls in captive Asian chip- munks, Tamias sibiriciis. Journal of Mannnalosy 73: 597-603. Blankenship, D. J., AND L. R. Brand. 1987. Geographic variation in vocalizations of California chipmunks Tamias obscunis and T. merriami. Bulletin of the Southern California Academy of Sciences 86: 126-135. Callahan, J. R. 1977. Diagnosis of Eutainias ohscurus (Rodentia: Sciuridae). Journal of Mammalogy 58: 188-201. Callahan, J. R. 1981. Vocal solicitation and parental investment in female Eutainias. American Naturalist 118: 872-875. Cornell, J. 1987. The natural histoiy of squirrels. Ricts on File, New York. Halpin, Z. T. 1984. The role of olfactoiy communication in the social systems of ground-dwelling sciurids. Pages 201-225 in J. O. Murie and C. R. Michener, eds.. The biology of ground-dwelling squirrels. University of Nebraska Press, Lincoln. Larson, E. A. 1981. Merriam's chipmunk on Palo Escrito in the Santa Lucia Mountains of California. Part L Regimen with recorded episodes of natiualistic be- havior. Enid A. LarsonAVacoba Press, Big Pine, CA. Myk'VTOWTCZ, R. 1965. Fiuther obsei-vations on the terri- torial function and histology of the submandibular cutaneous (chin) glands in the rabbit, Orijctolagus cimiculiis (L). Animal Behaviour 13: 400-412. R'VTTERSON, B. D. 1984. Geographic variation and ttixonomy of Colorado and Hopi chipnumks (genus Eutainias). Journal of Mammalogy 65: 442^56. Smith, S. E 1978. Alarm calls, their origin and use in Eutainias sonomae. Journal of Mammalogy 59: 888-893. Thompson, D. C. 1977. Reproductive behavior of the grey squinel. Canadian Journal of Zoology 55: 1176-1184. Received 11 April 1994 Accepted 1 9 October 1 994 Great Basin Naturalist 55(1), © 1995, pp. 92-94 ADDITIONAL RECORDS OF FLEAS (SIPHONAPTERA) FROM UTAH James R. Kucera' Key words: Si^hunaptcrci. Jleas, Utah. Mcgabothris asio megacolpus, Euhoplops>lliis glacialis Knx. Subsequent to the important work of Stark (1959), few publications have given flea collec- tion records from Utah. These include Jellison and Senger (1976) and Kucera and Haas (1992); but most effort in this area has been that of Egoscue (1966, 1976, 1977, 1988, 1989). Herein is presented information for 10 species of Siphonaptera for Utah. A number of important records were obtained from the flea collection at the Monte L. Bean (MLB) Life Science Museum, Brigham Yoinig University, Provo, UT. Catalog numbers of host specimens deposited in the University of Utah Museum of Natural Histoiy (UU) mammal collection and flea specimens in the MLB Museum (BYU) insect collection are given in parentheses when available. Unless indicated otherwise, speci- mens were collected by me and are retained in my personal collection. Carter ett a chivata Good 1942 Washington Co.: west slope Beaver Dam Mts., vie. Welcome Spring, 1220 m, 20 March 1988, 1 9 ex Chaetodipus formosus. "Beaver Dam," 23 Februaiy 1952, 1 9 (BYU #3462) ex Perognathus [ = Chaetodipiis] formosus, coll. C. L. Hayward. Beaver Dam Wash, 17 April 1952, Ic? (BYU #3607) ex Peromysciis truei, coll. Grace Grant et al. Few collections of this species are known from Utah (Tooele County: Stark 1959 [Id], Egoscue 1976 [1 specimen, sex unknown]; Washington County: Jellison and Senger 1976 [2d (5, 2 9 9]). It has also been collected in Clark County, NV (the type locality, Good 1942), and Mohave County, AZ (Augustson and Dur- ham 1961). It is likely a nest flea o{C. formosus. Nearctopsylla brooksi (Rothschild 1904) Utah Co.: Provo, 21 August 1951; M, 19 (BYU #1366 & 1365) e\ Mustek frenata, coll. D. Brown. Provo Canyon, 16 June 1959; 26 6 , 79 9 ex Spilogale gracilis, coll. D E. Beck. This species was previously known in Utah from a single collection in Sevier County (Stark 1959). It is usually found on weasels {Mustela spp.). NearctopsijIIa hi/rtaci (Rothschild 1904) Salt Lake Co.: Wasatch Mts., Big Cotton- wood Canyon, vie. Redman campground, 2560 m (spruce-fir), 21 October 1990, 19 ex Sorex monticolus (UU #29163). Same locality, 28 October 1990, 19 ex Sorex monticolus (UU #29164). Stark (1959) reported this species from Cache County. My collections extend the known range of this species further south in Utah along the Wasatch Cordillera. It is foimd on shrews [Sorex spp.) and Mustela spp. Delotelis telegoni (Rothschild 1905) Salt Lake Co.: Wasatch Mts., Big Cotton- wood Canyon, vie. Redman campgroimd, 2560 m (spruce-fir), 21 October 1990, Id ex Clethrionomys gapperi. Same date & locality, 1 9 ex Tamiasciurus hudsonicus nest. [The nest, about 3 m above ground level, also con- tained many red squirrel fleas {Orchopeas c. caedens). The squirrel probably carried this vole flea to its nest.] Same locality, 15 September 1991, 19 ex Peromyscus manicula- tus. Same locality, 5 October 1991, Id ex Cleth rionomys gapperi. 'Associated Regional and Hinversih' Palliologists, liie . Salt Uke City, UT 84108. Address for correspondence: .59.30 S. Siillan Circle, Murra\'. UT 84107-H930. 92 1995] Notes 93 Delotelis telegoni has rarely been found in Utah; single specimens have been collected in Sanpete County (Stark 1959) and in Utah County (Egoscue 1988). It is significant that these collections were made in summer months (August 1951 and July 1985, respectively). Many more collections will likely be made if this species is searched for during the cooler months of the year Also, it presumably would be profitable to search nests of Microtus and Clethrionomys for this species. Meringis shannoni (Jordan 1929) Eads et al. (1987) listed two collections of this species from Utah. The specimens are present in the BYU collection. However, the records are erroneous because the collection locality (Douglas County) does not exist in Utah. All other cited records of this species are from the states of Washington and Oregon, except a lone locality record in Humboldt County, NV (Lewis et al. 1988). Stenistomera hubbardi Egoscue 1968 This rare species was listed by Tipton and Saunders (1971) as occurring in Utah although no specific records were cited. Egoscue (per- sonal communication) knows of no records from Utah, and no specimens were present in the MLB Life Science Museum, the main reposi- tory of Tipton's Utah collections. In addition to the type specimens from Oregon (Egoscue 1968), the only other published record of S. hubbardi is that of Lewis et al. (1988), also from Oregon. It is unlikely that the species has been collected in Utah. Megarthroglosstis becki Tipton & Allred 1951 Salt Lake Co.: Wasatch Mts., mouth of Little Cottonwood Canyon, 1676 m (scrub oak), 3 December 1989; 1 c? , 1 9 ex Neotoma cinerea nest. The species has been collected only in Utah (Kane, Piute, Utah, and Wayne counties; Tipton et al. 1979) and Arizona (Augustson and Durham 1961). This is the northernmost record known, some 37 km north of the type locality in the Wasatch Mountains. Megarthroglossus becki is a nest flea of woodrats, principally the bushy-tailed woodrat Neotoma cinerea. Megabothris asio megacolpus (Jordan 1929) Rich Co.: Laketown, 22 August 1952; nSS, 26$ $ (BYU #5097-5099, 5101-5103, 5105-5119, 5121-5131, 5135, 5137, 5138, 5737, 5738) ex Microtus montanus nests [3 ex- amined], coll. D E. Beck & L. Beck. Laketown, 26 June 1953; Id, 49 9 (BYU #7823-7827) ex Microtus nests [3 examined], coll. Beck et al. Sevier Co.: Fish Lake [south end], 5 August 1952, 19 (BYU #5622) ex Microtus sp., coll. Coffey & Killpack. Monroe Mt., 7 mi. [= 11 km] W of Koosharem, 30 July 1958; M, 29 9 ex Microtus sp. [3 examined], coll. unknown. The range of this boreal vole flea extends deep into south central Utah. Only two speci- mens are known from Idaho, including one from Bear Lake County adjacent to Rich County (Baird and Saunders 1992). Collections from Ravalli and Beaverhead counties, MT, are documented (Holland 1950), and speci- mens from Ravalli County are present in the Natural History Museum, London (T. M. Howard personal communication). A point- mapped record roughly on the Utah-Wyoming border given in Haddow et al. (1983) is evi- dently meant to be Laketown because four specimens with the same collection data as the 22 August 1952 series are present in the Natural History Museum, London (Howard personal communication). Chaetopsylla stewarti Johnson 1955 Utah Co.: Wasatch Mts., "near summit of Alpine Loop" [American Fork or Provo Can- yons], 24 November 1965; 666,699 ex Mustela frenata, coll. D. Andrews. Summit Co.: Uinta Mts., 1/2 mi. [0.8 km] E Bald Mt., 8 August 1957, 16 ex Martes sp., coll. D. Allred & M. Killpack. These specimens are the only ones known, other than the type series (from Cache County, Lewis and Lewis 1994). Weasels seem to be the preferred host. Euhoplopsyllus glacialis hjnx (Baker 1904) Salt Lake Co.: Wasatch Mts., Big Cotton- wood Canyon, vie. Redman Campground, 2560 m, 17 August 1988; 39 9, 16 ex Lepus ameri- canus (UU #28674). Big Cottonwood Canyon, 2280 m, 30 August 1988, 1 9 ex Lepus ameri- canus. Big Cottonwood Canyon, vie. Butler 94 Great Basin Natur.\list [Volume 55 Fork trailhead, 2182 m, 19 May 1991, 26 6 ex Lepus atnericanus. Prc'Niously unknown in Utah, the nearest published records are for Ravalli County, MT (Kohls 1940), more than 570 km to the north. The t\pe lociilitv' is Moscow, ID, about 790 km to the northwest (Baker 1904). This flea is con- sistently found on the snowshoe hare [Lepus americanus) and its predator, the hnx {Lynx canadensis). The foim E. glacialis affinis is com- mon in Utah and surrounding states on rabbits and jackrabbits (Sylcilagiis spp. and Lepus spp. other than L. americanus). Questionable records of E. g. lynx from the states of Tamaulipas and Veracruz, Mexico (ex Sylvilagus floridanus and unidentified Sylvilagus sp.), are listed by Ayala et al. (1988). Acknowledgments Comments by Glenn E. Haas, B. C. Kondra- tieff, and an anonymous reviewer improved the manuscript. Flichard W. Baumann, Curator of Insects at the M. L. Bean Life Science Museum, Brigham Young University, kindly allowed me to examine specimens kept there. Harold J. Egoscue confirmed identification of the E. g. lynx. Theresa M. Howard of The Natural History Museum, London, sent data on specimens in the Rothschild Collection. Literature Cited AuGUST.soN, G. E, AND E E. Durham. 1961. Records of flea.s (Siphonaptera) from northwestern Arizona. Southern CaHfornia Academy of Sciences Bulletin 60: 100-105. Ayala, R., J. C. Morales, N. Wilson, J. E. Llorente, and H. E. Ponce. 1988. Catalogo de las pulgas (Insecta; Siphonaptera) en el Museo de Zoologia, Ricultad de Ciencias Universidad Nacional Autonoma de Me.xico 1: Coleccion Alfredo Barrera. Serie Catalogos del Museo de Zoologia "Alfonso L. Herrera" Catalogo No. 1. 102 pp. Baird, C. R., and R. C. Saunders. 1992. An annotated checklist of the fleas of Idaho (Siphonaptera). Idaho Agricultural E.xperiment Station, Bulletin 148. Baker, C. E 1904. A revision of the American Siphonap- tera, or fleas, together with a complete list and bibli- ograph>' of tlie group. Proceedings of die U.S. National Museum 27: 365-469. Eads, R. B., E. G. Campos, and G. O. Maupin. 1987. A review of the genus Meringis (Siphonaptera: Hystri- chopsyllidae). Jouj^nal of Medical Entomology 24: 467-476. EgoscuE, H. J. 1966. New and additional host-flea associ- ations and distributional records of fleas from Utah. Great Basin Naturalist 26: 71-75. . 1968. A new species of the genus Stenistomcra (Siphona])tera: Ilystrichopsyllidae). Southern Cali- fornia Academy of Sciences Biflletin 67: 1.38-142. . 1976. }""lea exchange between deer mice and some associated small mannnals in western Utah. Great Basin Naturalist 36: 475-480. . 1977. The sagebrush vole flea, Megahothri.s cliin- toni princeu in western Utah, with comments on the distribution o{ Megabothris in the Bonneville Basin. Great Basin Naturalist 37: 75-76. . 1988. Noteworthv flea records from Utah, Nevada, and Oregon. Great Basin Naturalist 48: 530-.532. . 1989. A new species of the genus Traiibella (Sipho- naptera: Ceratophyllidae). Southern California Academy of Sciences Bulletin 88: 131-134. Good, N. E. 1942. Carteretta carteri clavata, a new sub- species from Nevada, and notes on s>Tionymy (Sipho- naptera). Annals of the Entomological Society of America .35: 110-113. Haddovv, J., R. Traub, and M. Rothschild. 1983. Dis- tribution of ceratophyllid fleas and notes on their hosts. Pages 42-163 in R. Traub, M. Rothschild, and J. E Haddow, The Rothschild collection of fleas — the Ceratophyllidae: keys to the genera and host relation- ships with notes on their evolution, zoogeography and medical importance. 288 pp. [Privately published.] Holland, G. P. 1950. Notes on Megabothris asio (Baker) and M. calcurifer (Wagner) with the description of a new subspecies (Siphonaptera: Ceratophyllidae). Canadian Entomologist 82: 126-133. Jellison, W. L., and C. M. Senger. 1976. Fleas of west- ern North America e.xcept Montana in the Rocky Mountain Laboratory collection. Pages 55-136 in H. C. Taylor, Jr., and J. Clark, eds.. Papers in honor of Jerry Flora. Western Washington State College, Bellingham. Kucera, J. R., and G. E. Ha.\s. 1992. Siphonaptera (fleas) collected from small mammals in montane southern Utah. Great Basin Naturalist 52: 382-384. Kohls, G. M. 1940. Siphonaptera — a study of the species infesting wild hares and rabbits of North America north of Mexico. National Institute of Health Bulletin 175. Lewis, R. E., and J. H. Lewis. 1994. Siphonaptera of North America north of Mexico: Vermips>'llidae and Rhopalopsyllidae. Journal of Medical Entomology 31: 82-98. ' Lewis, R. E., J. H. Lewis, and C. Maser. 1988. The fleas of the Pacific Northwest. Oregon State Universit>' Press, Coi^vallis. 296 pp. Stark, H. E. 1959. The Siphonaptera of Utah. U.S. Department of Health, Education and Welfare, Communicable Disease Center, Atlanta, GA. 239 pp. Tipton, V J., and R. C. Saunders. 1971. A list of arthro- pods of medical importance which occur in Utah with a review of arthropod-bome diseases endemic in the state. Brigham Young University' Science Bulletin, Biological Series 15: 1-31. Tipton, V. J., H. E. Stark, and J. A. Wildie. 1979. Anomiopsyllinae (Siphonaptera: Hystrichops> llidae), n. The genera CullistopsijUus, ConorhinopsyUa, Mcgarthroglossiis, and Sti'ni.sto)nera. Great Basin Naturalist 39: 351-418. Received 25 May 1994 Accepted 10 August 1994 INFORMATION FOR AUTHORS The Great Basin Naturalist welcomes previously unpublished niauuscripts pertaining to the biologi- cal natiual histor\ of western North America. Pref- erence will be given to concise manuscripts of up to 12,000 words. Simple species lists are discouraged. SUBMIT MANUSCRIPTS to Richard W. Baumann, Editor, Great Basin Naturalist, 290 MLBM, PO Box 20200, Brighani 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, 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, phone: (312) 616- 0800, F.w: (312) 616-0226; $24). TYPE AND DOUBLE SPACE all materials, including literatiu'e 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. Distiu-bance and patch dynamics on rockv 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 number, and E-mail address • 3 copies of the manuscript and WordPerfect diskette • Conformity with instructions • Photocopies of illustrations (ISSN 001 7-3614) GREAT BASIN NATURALIST Vol 55, no l, January 1995 CONTENTS Articles Life histories of stoneflies (Plecoptera) in the Rio Conejos of southern Colorado R. Edward DeWalt and Kenneth W. Stewart 1 Polhnator sharing by three sympatric milkvetches, inchiding the endangered species Astragalus montii S. M. Geer, V. J. Tepedino, T. L. Griswold, and W. R. Bowhn 1 9 Factors affecting selection of winter food and roosting resources by porcupines in Utah Dave Stricklan, Jerran T. Flinders, and Rex G. Cates 29 Historic expansion of Jiiniperus occidentalis (western juniper) in southeastern Oregon Richard F Miller and Jeffery A. Rose 37 Rangeland alpha diversities: Hai-vey Valley, Lassen National Forest, California ... Raymond D. Ratliff 46 Effects of salinity on establishment of Populus fretnontii (cottonwood) and Tamarix ramosissima (saltcedar) in southwestern United States Patrick B. Shafroth, Jonathan M. Friedman, and Lee S. Ischinger 58 Names and types of Hedijsarum L. (Fabaceae) in North America Stanley L. Welsh 66 Whipwonii {Trichiihs dipodomys) infection in kangaroo rats {Dipodomys spp.): effects on digestive efficiency James C. Munger and Todd A. Slichter 74 Local distribution and foraging behavior of the spotted bat {Eudertna maculatum) in northwestern Colorado and adjacent Utah Jay F Storz 78 The Chrysothamnus-Ericameria connection (Asteraceae) Loran C. Anderson 84 Notes Reproductive behavior in Merriam's chipmunk {Tamias merriami) Stephen B. Compton and J. R. Callahan 89 Additional records of fleas (Siphonaptera) from Utah James R. Kucera 92 H E GREAT BASIN NATURALIST VOLUME 55 NO 2 — APRIL 1995 BRIGHAM YOUNG UNIVERSITY GREAT BASIN NATURALIST Editor Richard W. Baumann 290 MLBM PO Box 20200 Brigham 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. Bowers Blandy Experimental Farm, UniversiU' of Virginia, Box 175, Boyce, VA 22620 J. R. Callahan Museum of Southwestern Biology, Universit) of New Mexico, Alhuquerque, 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 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, Botany and Range Science. All are at Brigham Young University. Ex Officio Editorial Board members include Steven L. Taylor College of Biology and Agriculture; H. Duane Smith, 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 1995 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 © 199.5 by Brigham Young University Official publication date: 21 April 1995 ISSN 0017-3614 4-95 750 13821 MCZ The Great Basin Natfli-alist Published AT Provo, Utah, BY (j j jv^'vy5 Brigham Young University ISSN 0017-3614 H A R V A ^"^ ^^ UNiVER- Volume 55 30 April 1995 No. 2 Great Basin Naturalist 55(2), © 1995, pp. 95-104 DIETS OF YOUNG COLORADO SQUAWFISH AND OTHER SMALL FISH IN BACKWATERS OF THE GREEN RIVER, COLORADO AND UTAH Robert T. Muthl and Barrel E. Snyderl Abstract. — We compared diet of young-of-year Colorado squawfish {Pfijchocheihis hiciits), an endangered cyprinid, with diets of other fish <75 mm total length (TL) collected fiom backwaters of the Green River between river kilome- ters 555 and 35 during summer and autumn 1987. Species included native Wiinichthys osciiliis, Catostomus discobolus, and C. latipinnis, and nonnative Cyprinella hitrensis, Notropis stramineiis, Pimephales promelas, Ictalunis pimctatus, and Lepomis cijanellus. For each species, diet varied with size and between upper and lower river reaches but not between seasons for fish of similar size. Larval chironomids and ceratopogonids were principal foods of most fishes. Copepods and cladocerans were important in diets of E lucius <21 mm TL and L. cijanellus <31 mm TL. Catostomus discobolus was the only species that ate moderate amounts of algae. Fish (all larvae) were in digestive tracts of only 10 P. lucius (21-73 mm TL), about 1% of P. lucius analyzed. High diet overlap occuired between some size-reach groups of P. lucius and C. hitrensis, R. osculus, C. lotipinnis, I. punctatus, and L. cijanellus. Potential for food competition between young- of-year P. lucius and other fishes in backwaters appeared greatest with the ver>' abundant C. hitrensis. Key words: Pt>'chocheilus lucius, CNTDrinella lutrensis, nonnative fishes, young-of-yean diets, diet overlap, backwaters. Green River Wild populations of federally endangered Colorado squawfish (Nesler et al. 1988, Haines Colorado squawfish {Ptychocheilus lucius) per- and Tyus 1990, Tyus and Haines 1991). Ichdiyo- sist only in the upper Colorado River basin, fauna of these backwaters is dominated by They are most abundant in the Green and nonnative fishes, especially red shiner {Cypri- Yampa rivers of eastern Utah and northwest- nella hitrensis; Tyus et al. 1982, Haines and em Colorado (Tyus 1991a). Decline of this and Tyus 1990). This observation has led to a hy- other native fishes in the Colorado River basin pothesis that nonnative fishes adversely affect has been attributed to habitat alterations survival of young Colorado squawfish through caused by water development and introduc- competition or predation. Stanford (1993) sug- tion and proliferation of nonnative fishes gested that strong food-web interactions be- (Carlson and Muth 1989, Minckley 1991). tween native and nonnative fishes probably Backwaters of the Green River below its occur, but dietary relationships have not been confluence with the Yampa River are impor- adequately documented (Haines and Tyus tant nursery areas for young-of-year (YOY) 1990, Ruppert et al. 1993). Our objectives 'Lanal Fish Lal)orator\, Department of Fishery and Wildlife BioloKV". Colorado State University, Fort Collins, CO 80523. 95 96 Great Basin Naturalist [Volume 55 were to (1) describe diets of YOY Colorado squawfish and other small fish in backwaters of the Green River and (2) examine diet oxerlap and potential for competition with Colorado squawfish. Methods Samples of small fish were provided by the U.S. Fish and Wildlife Service Field Station at Vernal, UT. These were collected from back- waters of the Green River during summer (30 June-27 August) and autumn (22 September- 10 Decembei) 1987. The study area extends from confluence of the Green and Yampa rivers in Echo Park, Dinosaur National Monu- ment, CO, to Turks Head in Canyonlands National Park, UT — river kilometer (RK) 555 to 35 above confluence with the Colorado Fliver. Upper and lower reaches are divided at Sand Wash (RK 346), UT, a convenient access point just above Desolation Canyon. Each river reach began with a rocky, high-gradient (1.3-2.1 m/km) segment and continued with a sand- and silt-laden, low-gradient (0.2-0.4 m/km) segment known for relatively high catches of YOY Colorado squawfish (Haines and Tyus 1990, Tyus and Haines 1991). The river was further divided into 8-km sections starting from a random location within each reach to help assure an even distribution of collection sites. Back-waters were defined as shallow (typi- cally <0.5 m maximum depth), ephemeral embayments with negligible water velocity. Substrates consisted of silt and sand or silt and mud, sometimes overlaying or interspersed with gravel or cobble. Backwaters had little or no rooted aquatic vegetation, but some had dense mats of algae. Two backwaters were sampled weekly in each 8-km section during daylight (1000-1800 h) using l-m^ seines (0.8- mm^ mesh) in summer and 1-m X 3-m seines (3.2-mm X 4.8-mm mesh) in autumn. Fish were killed and fixed in 10% formalin immedi- ately after collection. Up to five specimens < 20 mm total length (TL) and five >20 mm TL of each fish species, representing graded size series, were selected from each sample. Each digestive tract (from esophagus to vent) was removed, opened, and visually assessed for percent fullness. Food items were identified to lowest practical taxon, and a visual estimate was made of percentage contributed by each taxon to total \'olume of food in each digestive tract (Larimore 1957, Mathur 1977). For diet analyses, food-item taxa (total of 124) were grouped into 20 family, order, or liroader-based categories, sometimes divided according to habitat (e.g., aquatic or terrestrial). Data for each fish species were stratified according to length (10-mm TL or larger inter- vals) by season (summer or autumn) within river reach (upper or lower). Only subsets with at least six fish containing food were included in analyses. Diet measures calculated for each subset were (1) mean percentage each food categoiy contributed to total volume of food in each digestive tract (mean of volume percent- ages) and (2) percentage of all digestive tracts in which each food category occurred (per- centage of occurrence). Wallace (1981) evalu- ated several diet measures and concluded that mean of volume percentages is the best mea- sure for calculating overlap. However, per- centage of occurrence is useful for describing general variations in diet (Wallace 1981, Bowen 1983). Similarities in diet by subset between Colo- rado squawfish and other fishes were evaluat- ed by Schoener's (1970) resource-overlap index: n a = 1-0.5(I|P.17-F?//|), /=1 where n is the number of food categories, Pxi is the proportion of food category / (expressed as mean of volume percentages) in the diet of species x (Colorado squawfish), and Piji is the proportion of food category / in the diet of species y (other fishes). Values range from 0.0 (no overlap) to 1.0 (complete overlap). When data on resource availability are absent, Schoener's index is one of the best indices available for calculating resource overlap (Hurlbert 1978, Linton et al. 1981, Wallace 1981). Diet overlap is useful in helping to elu- cidate food relationships among species and has been considered "biologically important when values exceed 0.60 (Zaret and Rand 1971, Matthews and Hill 1980, Galat and Vucinich 1983). Results Digestive tracts from 2554 fish represent- ing 15 species were examined for food items; 1995] Diets of Fishes in Backwaters 97 <3% were empty, mostly from fish < 13 mm TL. After subsets with <6 specimens contain- ing food were ehminated from the data set, 2297 specimens representing nine species remained for diet analyses. Native fish includ- ed 972 Colorado squawfish (7.5-73.0 mm TL, mean = 19.1), 35 speckled dace {Rlunichthys osciihis\ 23.1-39.8 mm TL, mean = 28.1), 42 bluehead sucker [Catostomus discoholus; 23.0-58.9 mm TL, mean = 35.9), and 21 flan- nelmouth sucker (C. latipinnis\ 32.0-64.3 mm TL, mean = 47.9). Nonnative fish included 729 red shiner (11.3-74.5 mm TL, mean = 29.1), 92 sand shiner {Notropis stramineus; 22.2-53.2 mm TL, mean = 31.0), 330 fathead minnow {Pimephales promelas; 11.0-65.9 mm TL, mean = 32.5), 58 channel catfish {Ictaluriis pimctatiis; 22.5-70.0 mm TL, mean = 42.9), and 18 green sunfish {Lepomis cyanellus; 20.7-56.8 mm TL, mean = 39.6). Characterization of Diets No major or consistent seasonal differences in diet measures were obsei^ved within species for fish of similar size. Accordingly, summer and autumn data were combined for species and lengths by river reach. Trends in values of proportional importance of each food categoiy were similar between the two diet measures for all fishes; therefore, only means of volume percentages are reported. Diets consisted mostly of insects, zooplank- ton, algae, seeds, and organic and inorganic debris; but relative importance of these food categories varied among fishes or subsets within species (Table 1). Based on total num- ber of food categories included in the diet of each fish species, diets of Colorado squawfish and red shiner were the most varied (18 and 17 food categories, respectively), followed by speckled dace (15), fathead minnow, channel catfish, and green sunfish (12 each), sand shin- er (11), flannelmouth sucker (9), and bluehead sucker (6). Variety of food consumed was greater in the lower than upper reach for red shiner, Colorado squawfish, flannelmouth sucker, channel catfish, and green sunfish, whereas diets of sand shiner, fathead minnow, and speckled dace were more varied in the upper reach (diet of bluehead sucker was ana- lyzed for fish from the upper reach only). Diet variety relative to fish length was greatest in red shiner, sand shiner, fathead minnow, Colorado squawfish, speckled dace, and blue- head sucker 21-30 or 31-40 mm TL and in flannelmouth sucker, channel catfish, and green sunfish >40 mm TL. Mean percent full- ness of digestive tracts was highest in fish 21-30 or 31—40 mm TL for all species. Aquatic insects were a principal part of diets for all fishes except fathead minnow and bluehead sucker. Of identifiable insects, immature dipterans (especially larval chirono- mids) were predominant in digestive tracts. Larval chironomids were represented by at least 21 genera, the most common being Chironomiis followed by Wieotanytarsii^, Eukiefferiella, Polij- pedilum, Tanytarsus, Cricotopus, and Microp- sectra. Representative families of other imma- ture dipterans were (in order of importance) Ceratopogonidae, Simuliidae, Dolichopodidae, Empididae, Muscidae, and Tipulidae. Propor- tional contribution of immature dipterans to diets of red shiner, sand shiner, speckled dace, and flannelmouth sucker was higher in the lower than upper reach. Relative importance of immature dipterans in diets of red shiner, sand shiner, and speckled dace decreased and utilization of other insects increased as fish length increased. Conversely, relative impor- tance of immature dipterans in diets of Colorado squawfish and channel catfish increased or remained high with increasing fish length. Corixids, lai"val and adult aquatic coleopterans (predominantly Dytiscidae, Elmi- dae, Haliplidae, and Hydrophilidae), trichopter- an lai^vae (mainly Hydropsychidae and Hydrop- tilidae), and ephemeropteran nymphs (pre- dominantly Baetidae and Heptageniidae) were minor components of diets for all fishes (<10% of food volume) except larger red shiner, speckled dace, and green sunfish. Red shiner and sand shiner ate more semi- aquatic or terrestrial insects than other fishes. Semiaquatic insects consumed were primarily larval and adult coleopterans (predominantly Heterocercidae and Staphylinidae) and adult hymenopterans (Scelionidae). Terrestrial insects consumed were primarily hemipterans and formicids. All fishes ate zooplankton, but it was partic- ularly important in diets of Colorado squaw- fish <31 mm TL (especially <21 mm TL), green sunfish <31 mm TL, and, to a lesser extent, red shiner and channel catfish <31 mm TL and flannelmouth sucker Cladocerans (many identified as Daphnia, Eurycercus, and Macro- thrix) and especially cyclopoid copepods 98 Great Basin Naturalist [Volume 55 Table 1. I^iets by tot; sure is mean percentage il-Iengtli intervals (mm) of nine fish species collected during sunnner and autunm 1987 from contributed by each food categoiy to total volume of food in each digestive tract (mean of vol- Colorado Red si liincr Sa lid shiner Fatheai 1 minnow squawfish Food categon 1 1-20 21-30 31-40 >40 21-30 31-40 >40 11-20 21-30 31-10 >4() <11 11-20 - Upper n eaclv — Insects Unidentitial)lc parts 11 9 25 37 3 9 22 2 1 1 Semiaeiuatic or terrestrial 1 4 6 12 5 <1 Diptera imnuitiires 30 27 29 13 25 19 3 7 4 4 13 70 Chirononiidac adults 1 3 3 10 9 4 9 Anisoptera nymphs Aquatic Coleoptera 3 10 Corixidae <1 1 10 Trichoptera larvae 1 1 1 1 Ephenieroptera uyniplis 1 2 Zooplankton Cladocera and Copepoda 7 6 3 3 <1 <1 26 16 Rotifera 6 <1 <1 <1 <1 16 1 Ostracoda I 1 Ganiniaridae <1 Hydracarina 2 <1 Invertebrate eggs 1 1 o\e confluence of tlie Green and Colorado i Monument, CO. to Sand Wash, UT (RK .346); lower reach = Sand Wash to Turks Head. Canyonlands National Park. UT (RK 35). in Echo Park, Dinosaur National 1995] Diets of Fishes in Backwaters 99 backwaters in two reaches of the Green River below its confluence with tlie Yampa River, Colorado and Utah. Diet mea- ume percentages). Colorado Fliinnelniouth Green squavvTish Speckled dace Bluehead sucker sucker Channel catfish sunfish 21-30 3I-K) >4() 21-30 31^0 21-.30 31-40 >40 31-40 >40 21-.30 31-40 >40 21-,30 >40 1 1 1 1 61 66 1 4 <1 Upper reach" 10 34 4 2 1 72 52 54 31 5 1 28 22 70 66 8 13 3 12 4 2 <1 8 13 1 1 <1 14 2 6 6 5 25 29 8 13 10 10 3 46 12 11 18 61 50 57 73 14 6 8 78 0 6 71 18 13 1 3 <1 5 <1 4 1 2 6 30%) in all fishes except speckled dace and green sunfish. It was over 80% of gut content in fathead minnow and bluehead sucker. Debris consisted of fibrous particles of vascular plant tissue usually mixed with large amounts of clay particles and sand grains, suggesting bot- tom feeding. Seeds (many identified as tama- risk [Tarnarix gallica]) were eaten by all fishes, especially red shiner <31 mm TL. Two obsenations were unique to Colorado squawfish. Fish larvae were found in digestive tracts of 10 Colorado squawfish (about 1% of total examined); 1 was 21 mm TL, 8 were 36-48 mm TL, and 1 was 73 mm TL (probably a yearling). No fish were detected in digestive tracts of other species. Of the 18 fish larvae found, most were too digested for species identification or accurate length measurement, but all were cypriniforms (mostly cyprinids) and probably < 10 mm TL. Six fish larvae (6-9 mm TL) were identified as red shiner, and one (about 8 mm TL) as fathead minnow. Interest- ingly, the smallest Colorado squawfish had four prey fish (all red shiner), whereas only one or two fish were found in digestive tracts of the others. Gut contents of six Colorado squawfish, 36-48 mm TL, and the 73-mm-TL specimen were exclusively fish; those for the remaining specimens were 70-80% fish. Digestive tracts of six Colorado squawfish contained 2-6 cestode parasites (probably Proteocephalus ptychocheilus; Flagg 1982); cestodes were not found in guts of other fish- es. Colorado squawfish infested with cestodes were larger than 27 mm TL and were collect- ed from both river reaches in autumn. Diet Overlap Degree of diet overlap between YOY Colo- rado squawfish and other fishes was influenced mainly by zooplankton and especially imma- ture dipterans (Table 2). Within each reach, diet overlap for all length intervals of Colorado squawfish generally decreased as lengths of other species increased. Degree of diet overlap among fish of similar size was generally greater in the lower than upper reach. Overlap values were <0.60 (range = 0.10-0.59) for most comparisons; generally, values were lowest for comparisons with fat- head minnow and bluehead sucker (range = 0.10-0.44). Biologically important overlap (values >0.60) occurred only between Colorado squawfish > 10 mm TL and some size-reach groups of native speckled dace and flannelmouth sucker and nonnative red shiner, green sunfish, and especially channel catfish. These higher overlap values were primarily attributed to high proportions of larval chi- ronomids in diets and, secondarily, especially for diet overlap with green sunfish >40 mm TL (upper reach) and 21-30 mm TL (lower reach), to proportions of zooplankton. Degree of diet overlap was greatest with channel cat- fish and green sunfish. Discussion Comparisons among food-habits investiga- tions are difficult because of differences in study design, location, and season. However, our observations on diets of native and nonna- tive fishes in back"waters of the Green River generally agree with results of prior studies in the upper Colorado River basin (e.g., Vanicek and Kramer 1969, Jacobi and Jacobi 1982, McAda and Tyus 1984) and reported food habits of the nonnative species within their native ranges (e.g., Carlander 1969, 1977, Pflieger 1975, Harlan et al. 1987). Larger YOY or yearling red shiner, sand shiner, speckled dace, flannelmouth sucker, channel catfish, and green sunfish eat mainly immature aquatic insects. Diets of larger YOY or yearling fathead minnow and bluehead sucker consist mostly of algae and organic debris. Diet of YOY Colo- rado s(|uawfish consists primarily of zooplank- ton and immature insects (especially chirono- mid larvae) and occasionally includes fish. Reported size at which wild Colorado squawfish shift to a more piscivorous diet 1995] Diets of Fishes in Backwaters 101 varies, but generally fish become an important food item after Colorado squawfish attain a length of >40 mm. Osmundson and Kaeding (1989) suggested that slower growth and poor- er condition of YOY and especially yearling Colorado squawfish in grow-out ponds with lower densities of appropriate-size forage fish might have been caused by higher reliance on insect forage. Identifiable fish reported in digestive tracts of YOY Colorado squawfish here and by McAda and Tyus (1984) and Grabowski and Hiebert (1989) were either red shiner or fathead minnow larvae. These non- native species are short-lived fractional spawn- ers (Gale and Buynak 1982, Gale 1986) and are typically present in high numbers and at appropriate forage sizes in back-waters of the Green River throughout summer and autumn (Tyus et al. 1982, Karp and Tyus 1990). Kaip and Tyus (1990) suggested that although the abundance of small nonnative prey fishes in the Green River might benefit growth of young Colorado squawfish, the benefit might be countered by the aggressive nature of some nonnative fishes, which could have negative effects on growth and survival of young Colorado squawfish. In their laboratory exper- iments on behavioral interactions, Karp and Tyus observed that red shiner, fathead min- now, and green sunfish shared activity sched- ules and space with Colorado squawfish and exhibited antagonistic behaviors toward small- er Colorado squawfish. We could not effectively evaluate competi- tion for food between YOY Colorado squaw- fish and other fishes because study design did not provide for estimation of resource abun- dance and availability, intraspecific diet selec- tivity, and effects of interspecific use of impor- tant resources. Direct evidence for interspecific competition should be determined through experiments demonstrating that shared use of a limited resource negatively affects one or more of the species (Schoener 1983, Under- wood 1986, Wiens 1992). Additionally, we assume gut contents represented iood con- sumed in the backwaters of capture, but this might not always have been the case. Tyus (1991b) observed that although young Colo- rado squawfish in the Green River were found mostly in backwaters, some moved to or from other habitats during 24-h periods. We found that diet overlap for most comparisons with Colorado squawfish was below the level gen- erally considered biologically important (Table 2). Although not conclusive, these compar- isons suggest either general resource parti- tioning or differences in diet preferences. Diet overlap values were considered biologically important only for comparisons with certain size-interval, river-reach groups of five fishes. Because interspecific demand for resources might not exceed supply, Bowen (1983) noted that even extensive diet overlap is not conclu- sive evidence for competition. Accordingly, McAda and Tyus (1984), who also used Schoener's index to examine diet overlap between YOY Colorado squawfish and nonna- tive fishes in the Green River, suggested that high diet overlap they observed between Colorado squawfish 22-40 mm TL and chan- nel catfish 19-55 mm TL (overlap value = 0.60) and especially red shiner 15-69 mm TL (overlap values 0.70-0.80) might reflect shared use of abundant resources, primarily imma- ture dipterans, rather than competition. The same may be true for higher diet overlaps we obsei-ved. Ward et al. (1986) reported that chi- ronomids, the principal food category result- ing in high diet overlap, were among the more common benthic invertebrates in the Colo- rado River basin. We observed that overlap values were gen- erally higher and, for most fishes, diet variety was greater in the lower than upper reach, perhaps because food resources were more abundant and diverse in backwaters of the lower reach. Based on observations during summer and autumn 1979-1988, Haines and Tyus (1990) found that backwaters in the upper and lower reaches were similar in mean surface area, but that those in the lower reach were shallower and warmer, conditions that may favor higher productivity. Also, within the upper reach, Grabowski and Hiebert (1989) noted that during summer and autumn 1987-88 concentrations of backwater nutri- ents, particulate organic matter, phytoplank- ton, zooplankton, and benthic macroinverte- brates (particularly chironomid larvae) in- creased progressively downstream. They sug- gested this trend was due to attenuation of flow releases from Flaming Gorge Reservoir (located near the Wyoming-Utah border) at downstream sites that reduced the degree of water exchange between the main channel and backwaters and allowed for greater backwater warming and stability. 102 Great Basin Naturalist [Volume 55 Tahle 2. Diet overlap by total-length (TL) intei-vals (mm) beUveen yoiiiiK-of'-year Colorado squavvfish and eight other confluence with the Yampa River, Colorado and Utaii. Overlap \'alues were calculated using Schoener's (1970) index asterisk (*). Red sliiiKT Saiul sliiiu Fathfad iiiiiiiKm Ippcr reach' LouiT reacii' Upper Upper TL of Colorado squawfish 11-20 21-30 31-40 >40 11-20 21-30 31-40 >40 21-30 31-40 >40 21-30 31-40 >40 21-30 31-40 >40 ^71 aiij 054 040 0.43 0.42 0~49 oio 047 0~46 047 035 053 055 052 O40 037 0.38 11-20 0.49 0.45 0.43 0.31 0.63* 0.53 0.47 038 0.37 0.31 0.14 0.49 045 0.38 0.18 0.15 0.15 21-30 0.55 0.51 053 0.41 0.74* 0.57 0.51 0.42 0.42 0.43 0.27 0.49 0.47 0.38 0.23 0.19 0.23 .31_1() 0.40 0.39 0.40 0.27 0.73* 0.57 0.52 0.42 0.35 0.29 0.12 0.50 0.45 0.38 0.17 0.13 015 >40 039 0.37 0.39 0.35 065* 047 0.44 0.37 0.34 0.28 0.12 0.53 0.47 037 0.17 0.13 0.14 "Upper reach = eonfliience of Green and Yampa rivers at RK 555 (river kilometers above confluence of Green and Colorado ri\ers) in Echo Park, Dinosaur National Moiinriicii CO, to Sand Wash. UT (RK 346); lower reach = Sand Wash to Turks Head, Canyonlands National Park. UT (RK 35) Alternatively, greater diet overlap and vari- ety in the lower reach might have been a reflection of a difference in backwater avail- ability between the upper and lower reaches. Tyus and Haines (1991) reported about 150% more backwaters per kilometer in the upper than lower reach. Fishes in the lower reach might have been more crowded in available backwaters, resulting in greater shared use and broader intraspecific use of available food. McAda and Tyus (1984) attributed reduc- tions in diet overlap between Colorado squaw- fish >40 mm TL and red shiner or channel catfish to decreased consumption of immature dipterans and increased consumption of fish by Colorado squawfish. However, Ruppert et al. (1993) reported fish larvae in digestive tracts of 15% of adult red shiner (36-79 mm TL) from ephemeral shoreline embayments near confluence of the Green and Yampa rivers. Unlike our study, they sampled on a diel basis and killed fish with an overdose of anesthetic before preservation to minimize possible regurgitation. Their results suggest that high diet overlap between young Colorado squawfish > 40 mm TL and red shiner might reoccur or continue with larger, piscivorous red shiner. Although we docu- mented high diet overlap between young Colorado squawfish >10 mm TL and other fishes in backwaters of the Green River, espe- cially channel catfish (Table 2), only red shiner, because of its extreme abundance (Haines and Tyus 1990), is likely to be a serious competitor for food with young Colorado squawfish. Red shiner has often been implicated in decline of native fishes of the American Southwest (e.g.. Minckley 1973, Greger and Deacon 1988, Rinne 1991). Competition might also be a factor between smaller specimens of both Colorado squawfish and other fishes. Few specimens <21 mm TL, other than red shiner and fathead minnow 11-20 mm TL, were available for comparisons with Colorado squawfish. However, as for smaller Colorado squawfish, zooplankton would likely be an important component of their diets (Joseph et al. 1977), and corre- sponding overlap values would be high, espe- cially for specimens <11 mm TL. Although dense populations may develop in backwaters, zooplankton may be limited under certain conditions because plankton communities in rivers are subject to dramatic spatial or tempo- ral fluctuations in abundance and diversity (Hynes 1970, Welcomme 1985, Ward 1989). In support of this generalization, Grabowski and Hiebert (1989) reported that zooplankton den- sities were higher in back-waters than in main- channel habitats within the upper reach and documented both spatial and temporal fluctu- ations in zooplankton abundance. They also observed higher concentrations of zooplank- ton in more confined backwaters than those with a broad connection to the river and sug- gested that densities were influenced by extent of water exchange between backwaters and the main river. In conclusion, we found high diet overlap between YOY Colorado squawfish and several small size groups of other fish species in Green River backwaters. Because of the extreme abundance of red shiner, we speculate that diet overlap could result in food competition 1995] Diets of Fishes in Backwaters 103 fish species collected during suniiiier and aiitiunn 1987 from backwaters in two reaches of the Green River below its with mean of volume percentages as the diet measure; values >0.60 (biologically important overlap) are marked with an Fathead minnow Flannelmouth Speckled dace Bluehead sucker sucker C>'liannel cattish Lower Upper Lower Upper Upper Lower Upper Lower Green sunHsh Upper Lower 11-2021-30 31-40 >40 21-30 21-30 31-1() 21-30 31-40 >40 31-40 >40 >40 >40 21-30 31-40 >40 >40 21-30 >40 0.44 0.43 0.37 0.33 0.21 ().3(i 0..34 0.39 0.3.5 0.34 0.52 0.57 0.59 0.34 0.40 0.34 0.30 0.21 0.56 0.25 0.22 0.21 0.15 0.11 0.37 0.59 0.57 0.16 0.12 0.11 0.46 0.45 0.65* 0.83* 0.72* 0.61* 0.57 0.73* 0.91* 0.26 0.21 0.21 0.15 0.12 0.52 0.76* 0.61* 0.20 0.16 0.19 0.42 0.36 0.69* 0.89* 0.81* 0.75* 0.75* 0.73* 0.75* 0.27 0.21 0.21 0.14 0.10 0.42 0.78* 0.61* 0.14 0.11 0.11 0.38 0.31 0.69* 0.68* 0.81* 0.79* 0.82* 0.61* 0.69* 0.27 0.21 0.20 0.14 0.10 0..35 0.74* 0.58 0.14 0.10 0.10 0.38 0.31 0.63* 0.64* 0.69* 0.89* 0.77* 0.68* 0.57 0.24 and might have a negative impact on Colorado squawfish growth, condition, or survival. Studies are needed to better assess the type and strength of interactions between native and nonnative fishes in backwater food webs under present regulated flow regimes and to define factors affecting these interactions. Acknowledgments H. Tyus, C. Karp, and S. Lanigan initiated this study and provided samples and field data. H. Copeland, J. Piccolo, and E Sikoski assisted with analysis of gut contents. H. Tyus and C. Karp reviewed data analyses. K. Bestgen, D. Beyers, J. Deacon, G. Haines, J. Hawkins, C. Karp, H. Tyus, and R. Valdez reviewed drafts of the manuscript. This proj- ect was funded by the Recovery Implemen- tation Program for Endangered Fish Species in the Upper Colorado River Basin. The pro- gram is a joint effort of the U.S. Fish and Wildlife Service, U.S. Bureau of Reclamation, Western Area Power Administration, states of Colorado, Utah, and Wyoming, upper basin water users, and environmental organizations. This paper is Contribution No. 75 of the Colo- rado State University Larval Fish Laboratory. Literature Cited BowEN, S. H. 1983. Quantitative description of diet. Pages 325-336 in L. A. Nielsen and D. L. Johnson, editors, Fisheries techniques. American Fisheries Society', Bethesda, MD. Carlander, K. D. 1969. Handbook of freshwater fishery' biology. Volume 1. Iowa State University Press, Ames. 752 pp. . 1977. Handbook of freshwater fishery biology. Volume 2. Iowa State University Press, Ames. 431 pp. Carlson, C. A., and R. T. Muth. 1989. The Colorado River: lifeline of the American Southwest. Pages 220-239 in D. P Dodge, editor. Proceedings of the International Large River Symposium. Canadian Special Publication of Fisheries and Aquatic Sciences 106. Flagg, R. 1982. Disease survey of the Colorado River fishes. Pages 177-184 in Colorado River fishery proj- ect final report. Part 3, Contracted studies. U.S. Fish and Wildlife Service and Bureau of Reclamation, Salt Lake City, UT Galat, D. L., and N. Vuginich. 1983. Food partitioning between young of the year of two sympatric tui chub morphs. Transactions of the American Fisheries Society 112:486-497. Gale, W. F 1986. Indeterminate fecundity and spawning behavior of captive red shiner — fractional, crevice spawner. Transactions of the American Fisheries Society' 115: 429-437. Gale, W. F, and G. L. Buynak. 1982. Fecundity and spawning frequency of the fathead minnow — frac- tional spawner. Transactions of the American Fisheries Society' 111: 35-40. Grabowski, S. J., AND S. D. HiEBERT. 1989. Some aspects of trophic interactions in selected backwaters and the main channel of the Green River, Utah. Final report of U.S. Bureau of Reclamation, Research and Laboratory Services Division, Applied Sciences Branch, Environmental Sciences Section, Denver CO, for U.S. Bureau of Reclamation, Upper Colorado Regional Off"ice, Salt Lake City, UT 131 pp. Greger, P D., and J. E. Deagon. 1988. Food partitioning among fishes of the Virgin River. Copeia 1988: 314-323. Haines, G. B., and H. M. Tius. 1990. Fish associations and environmental variables in age-0 Colorado squaw- fish habitats. Green River, Utali. Journal of Freshwater Ecology 5: 427-435. Harlan, J. R., E. B. Speaker, and J. Mayhew. 1987. Iowa fish and fishing. Iowa Department of Natural Resources, Des Moines. 323 pp. Hurlbert, S. H. 1978. The measurement of niche overlap and some relatives. Ecology 59: 67-77. 104 Great Basin Naturalist [Volume 55 Hynes, H. B. N. 1970. The ecoloj^y of rumung water. University of" Toronto Press, Ontario, Canada. 555 pp. JacOBI, G. Z., .\nd M. D. J.\cobi. 1982. Fish stomach con- tent analysis. Pages 285-324 in Colorado River fish- ery project final report. Part 3, Contracted studies. U.S. Fish and Wildlife Service and Bureau of Reclamation, Salt Lake City, UT. Joseph, T. W., J. A. Sinning, R. J. Behnke, and R B. IIoi.DEN. 1977. An evaluation of the status, life histo- ry, and habitat reciuirements of endangered and threatened fishes of the Upper Colorado River System. U.S. Fish and Wildlife Service, FWS/OBS- 77/62. 169 pp. K.\RP, C. A., AND H. M. Tvus. 1990. Behaxioral interac- tions between young Colorado squawtish and six fish species. Copeia 1990: 25-34. Larimore, W R. 1957. Ecological life histoi-y of the war- mouth (Centrarchidae). Illinois Natural History Survey Bulletin 27: 1-83. Linton, L. R., R. W Davies, and E J. Wrona. 1981. Resource utilization indices: an assessment. Journal of Animal Ecologv' 50: 283-292. Mathur, D. 1977. Food habits and competitive relation- ships of the bandfin shiner in Halawakee Creek, Alabama. American Midland Naturalist 97; 89-100. Matthews, W. J., and L. G. Hill. 1980. Habitat partition- ing in the fish community of a southwestern river Southwestern Naturalist 25: 51-66. McAda, C. W, and H. M. Tvus. 1984. Resource overlap of age-0 Colorado squawfish with other fish species in the Green River, fall 1980. Proceedings of the Bonneville Chapter American Fisheries Society 1984: 44-54. MiNCKLEY, W. L. 1973. Fishes of Arizona. Arizona Came and Fish Department, Phoeni.x. 293 pp. . 1991. Native fishes of the Grand Canyon region: an obituaiy? Pages 124-178 in Colorado River ecolo- gy and dam management. National Academy Press, Washington, DC. Nesler, T. R, R. T. Muth, and A. E Wasowicz. 1988. Evidence for baseline flow spikes as spawning cues for Colorado squawfish in the Yampa River, Colorado. American Fisheries Society Symposium 5: 68-79. Osmund.son, D. B., and L. R. K,\eding. 1989. Colorado squawfish and razorback sucker grow-out pond studies as part of conservation measures for the Green Mountain and Ruedi Reservoir water sales. Final report of U.S. Fish and Wildlife Sei^vice, Grand Junction, CO. 57 pp. Pflieger, W. L. 1975. The fishes of Missouri. Missouri Department of Conservation, Jefferson City. 343 pp. Rlnne, J. N. 1991. Habitat use by spikedace, Meda fulgida (Pices: Cyprinidae), in southwestern streams with reference to probable habitat competition by red shiner, Notropis lutrensis (Pices; CyiDrinidae). South- western Naturalist 36: 7-13. Ruppert, J. B., R. T. Muth, and T. R Nesler. 1993. Predation on fish larvae by adult red shiner, Yampa and Green rivers, Colorado. Southwestern Naturalist 38: 397-399. SCHOENER, T. W. 1970. Non-synchronous spatial overlap of lizards in patchy habitats. Ecology 51: 408-418. . 1983. Field experiments on interspecific competi- tion. American Naturalist 122: 240-285. Stanford, J. A. 1993. Instream flows to assist the recov- ery of endangered fishes of the upper Colorado River basin: review and synthesis of ecological infor- mation, issues, methods, and rationale. Final report of Flathead Lake Biological Station, University of Montana, Poison, for U.S. Fish and Wildlife Service, Region 6, Denver, CO. 89 pp -I- appendices. Tvus, H. M. 1991a. Ecology and management of Colorado squawfish. Pages 379-402 in W. L. Minckley and J. E. Deacon, editors. Battle against extinction. University of Arizona Press, Tucson. . 1991b. Movements and habitat use of young Colo- rado squawfish in the Green River, Utah. Journal of Freshwater Ecology 6: 43-51. Tvus, H. M., AND G. B. Haines. 1991. Distribution, habi- tat use, and growth of age-0 Colorado squawfish in the Green River basin, Colorado and Utah. Transactions of the American Fisheries Society 120: 79-89. Tvus, H. M., B. D. Burdick, R. A. Valdez, C. M. H.-wnes, T. a. Lvtle, and C. R. Berrv. 1982. Fishes of the upper Colorado River basin: distribution, abundance, and status. Pages 12-70 in W. H. Miller, H. M. Tyus, and C. A. Carlson, editors. Fishes of the upper Colorado River system: present and future. Western Division of the American Fisheiy Society, Bethesda, MD. Underwood, T. 1986. The analysis of competition by field experiments. Pages 240-268 in J. Kikkawa and D. J. Anderson, editors, Commiuiity ecology: pattern and process. Black-well Scientific Publications, Oxford, England. Vanicek, C. D., and R. H. Kramer. 1969. Life histoiy of the Colorado squawfish, Ptijchocheilus liicius, and the Colorado chub, Gila robusta, in the Green River in Dinosaur National Monument, 1964-1966. Transactions of the American Fisheries Society 98: 193-208. Wallace, R. K., Jr. 1981. An assessment of diet-overlap indexes. Transactions of the American Fisheries Society 110: 72-76. Ward, J. V. 1989. Riverine-wetland interactions. Pages 385-400 in R. R. Sharitz and J. W Gibbons, editors. Freshwater wetlands and wildlife. U.S. Department of Energy Symposium Series 61. U.S. Department of Energy Office of Scientific and Technical Information, Oak Ridge, TN. Ward, J. V, H. J. Zimmerman, and L. D. Gline. 1986. Lotic zoobenthos of the Colorado system. Pages 403-422 in B. R. Davies and K. E Walker, editors. The ecology of river systems. Dr W. Jimk, Dordrecht, The Netherlands. WiENS, J. A. 1992. The ecology of bird communities. Volume 2. Cambridge University Press, New York, NY 316 pp. Welcomme, R. L. 1985. River fisheries. FAO Fisheries Technical Paper 262. 330 pp. Zaret, T. M., and a. S. R\nd. 1971. Competition in tropi- cal stream fishes: support for the competitive exclu- sion principle. Ecology 52: 336-342. Received 21 April 1994 Accepted 15 September 1994 Great Basin Naturalist 55(2), © 1995, pp. 105-116 INVERTEBRATE FAUNA OF WASTEWATER PONDS IN SOUTHEASTERN IDAHO Karen L. Cieminskil'2 and Lester D. Flake^"^ Abstract. — Water column invertebrates were sampled with .3.8-L activity traps in 15 sewage, industrial, and radioactive wastewater ponds at the Idaho National Engineering Laboratoiy in southeastern Idaho. One collection was made per pond, per month, during all months the ponds were ice-free from June 1990 through July 1991. In addition, nutrient and selected heavy metal concentrations in pond water were determined in July 1991. Arsenic, barium, boron, lead, selenium, and mercuiy were detected in ponds. Sewage ponds generally had higher nitrogen and phosphorus lev- els than industrial and radioactive ponds. Of the .30 aquatic invertebrate taxa collected, the most ubiquitous were Rotifera, Daphnidae, Eucopepoda, Ostracoda, Acari, Baetidae, Corixidae, Notonectidae, Dytiscidae, and Chironomidae. Activity trap samples from sewage ponds contained more Rotifera, Daphnidae, and Notonectidae, whereas industrial ponds yielded more Chydoridae, Acari, and Baetidae. Numbers of Oligochaeta, Eucopepoda, Ostracoda, Corixidae, Dytiscidae, and Chironomidae collected were not significantly different between sewage and industrial ponds. Compared with natural systems, these ponds had fewer taxa, but a greater number of individuals of most taxa. The high number of invertebrates collected is attributed to the lack of fish in wastewater ponds and the high levels of nitrogen and phosphorus, particularly in sewage ponds. Key words: aquatic invertebrates, sanitarij wastewater, industrial wastewater, Idaho National Engineering Laboratory. Constructed ponds have been a common tool in wastewater treatment for decades (Gloyna et al. 1976). Wastewater ponds are constructed in a variety of manners and used in various treatment procedures, from settling ponds to ponds with various aquatic macro- phytes that enhance removal of nutrients and break down organic materials (Brix 1993). Recently, constructed wetlands have also been incorjDorated into many wastewater treatment systems associated with municipalities and industry (Task Force on Natural Systems 1990, Moshiri 1993). Wastewater ponds and wet- lands are also associated with federal research sites such as the Idaho National Engineering Laboratory (INEL) in southeastern Idaho and the Hanford Site in south central Washington. Wastewater ponds at INEL receive sani- tary, industrial, and radioactive waste pro- duced at the facility. Other than wildlife watering cisterns and ephemeral rain pools, waste disposal ponds are usually the only sur- face water at INEL and, as such, attract wildlife (Halford and Millard 1978, Howe and Flake 1989, Millard et al. 1990, Cieminski 1993). Migrating and resident waterfowl, shore- birds, blackbirds, and swallows use the ponds heavily, feeding partially or exclusively on aquatic invertebrates, and on invertebrates that have emerged from the ponds (Millard et al. 1990, Cieminski 1993). Most studies of macroinvertebrates, espe- cially insects, in conjunction with waste treat- ment have been limited to studies of benthic invertebrate assemblages in streams receiving raw sewage or effluent from sewage treatment plants (e.g., Klotz 1977, Kownacki 1977, Duda et al. 1982, Kondratieff and Simmons 1982, Kondratieff et al. 1984, Chadwick et al. 1986, Lewis 1986, Crawford et al. 1992). Literature on plankton and nekton in constructed ponds focuses mainly on pathogens, and microscopic flora and fauna important in waste decomposi- tion, such as bacteria, protozoa, and algae (Goulden 1976, Task Force on Natural Systems 1990). Because the invertebrate fauna of waste- water ponds attracts wildlife, it is important to understand invertebrate communities of the ponds, as well as if and how they differ from natural communities. Our objectives were to (1) provide baseline data on invertebrate 'Department of Wildlife and Fisheries Sciences, South Dakota State Universitv", Box 2140B, Brookings, SD 57007 ^Present address: National Park Service, 1302.5 Riley's Lock Road, Pooles\ille, MD 208.37. ■^Address reprint requests to this author 105 106 Great Basin Naturalist [Volume 55 resources available to migrating birds in con- structed waste ponds and (2) determine if nutrients and selected heax')' metals in ponds influence invertebrate populations. Study Site The 231,600-ha INEL lies in Butte, Bonneville, Bingham, Clark, and Jefferson counties, ID, on the western edge of the Snake River plain near the foothills of the Lost River, Lemhi, and Bitterroot mountain ranges (Fig. 1). Topography at INEL is flat to rolling, with elevation ranging from 1463 m to 1829 m. Big Lost River, Little Lost River, and Birch Creek drainages terminate in playas on or near INEL; flow is intermittent and largely diverted for agriculture. During this study no surface water flowed onto INEL. Plant com- munities are dominated by big sagebrush {Artemisia tridentata), low sagebrush (A. arbiis- cula), and three-tipped sagebrush (A. triparti- ta) (McBride et al. 1978). INEL lies in a semiarid, cold desert. Annual temperatures range from -42 °C to 39 °C. Average annual precipitation is 19.1 cm, 40% of which falls fiom April through June (Clawson et al. 1989). Precipitation levels are lowest in July. Snowfall averages 71.3 cm per year, and snow cover can persist fi-om December through March. Wastewater ponds on INEL contained san- itary waste (eight ponds), industrial waste (four ponds), or radioactive waste (three ponds) (Fig. 1). Because two radioactive ponds also contained industrial waste, in most analy- ses radioactive ponds were grouped with in- dustrial ponds (as "industrial ponds") for com- parison with sewage ponds. Ponds were grouped around INEL facili- ties, which were 4-36 km apart. Generally, each facility had between one and four sewage ponds and an industrial waste pond. Sewage ponds ranged from 0.04 to 2.20 ha and were 0.6—2 m deep. Industrial waste ponds ranged from 0.20 to 2.24 ha and were 0.3-4.5 m deep. Seven of the sewage ponds and one industrial pond were lined to prevent infiltration into surrounding soil. Four ponds (all industrial and/or radioactive) supported emergent plant growth. A more thorough description of the ponds can be found in Cieminski (1993). Methods Water samples were collected at ponds in July 1991 and analyzed for nutrients (nitrogen and phosphorus) and selected heavy metals (arsenic, barium, ber\'llium, boron, lead, sele- nium, and mercury) that could influence pres- ence of invertebrates. Water pH was taken once at each pond at the same time water samples were collected. Further heavy metal and nutrient sampling was prohibitively expensive and time consuming. Water samples were analyzed at the U.S. Geological Sui-vey's National Water Qualit\' Laboratoiy at Ai-vada, CO. Collection and analysis methods were as per Brown et al. (1970) and Fishman and Friedman (1989). Data on heavy metals for pond ANLi (acronyms and names of pools are included in Tables 1 and 5) were taken from analyses conducted in 1988. Benthic samples were not taken because most ponds had lined bottoms, or because sediment sampling was not pennitted for other reasons. We collected water column inverte- brates once each month to obtain gross esti- mates of invertebrate populations. Additional collections and identification were time- and cost-prohibitive, given our concunent collection of bird and mammal count data at these ponds for a related project. Nevertheless, we felt that invertebrates influenced bird use of ponds, thus the need for estimates of invertebrate abundance. Water column invertebrates were collected at all nonradioactive ponds in months the ponds were ice-free from June 1990 through May 1991. Because of restricted access to radioactive waste ponds, they were sampled only once during July 1991. Invertebrates were collected in 3.8-L activity traps (Ross and Murkin 1989) suspended horizontally 5.3 cm under the water surface for approximately 24 h. Modifications on the technique of Ross and Murkin (1989) were necessary' since most ponds had artificial liners; therefore, jars could not be suspended from a pipe driven in the pond bottom. Instead, jars were suspended from floats and attached to a 50- to 300-cm- long piece of PVC pipe anchored on the pond's shore. The first sample was taken at the southeast corner of each pond. Subsequent monthly sample locations were chosen ran- domly based on a single-digit number of paces 1995] Invertebrates in Wastewater Ponds 107 Bitteroot Rangoy Containment Test Facility disposal pond (CTFi) Technical Support Facility disposal pond (TSFir) Naval Reactors Facility industrial waste ditch (NRFi) sewage pond (NRFs) Argonne National Laboratory - West m. — secondary sewage ' pond (ANLs2) primary sewage pond (ANLsl) industnal waste {>ond (ANIi) -north cold waste pond(TRAi2) ~ south cold waste pond(TRAil) ■ ■- east percolation y^ pond (CPPirl) west percolation pond (CPPirl) Fig. 1. Map of the Idaho National Engineering Laboraton; indicating location of facilities and wastewater ponds where invertebrate fauna was sampled. Waste type is indicated by lowercase letter in the pond code: s = sewage, i = industrial, r = radioactive. counterclockwise from the previous sample site. Where dense emergent vegetation cov- ered the near-shore zone, the activity trap was placed in the nearest open water. Activity trap contents were strained tiirough a 75-/x.m (No. 200) sieve and preserved in 80% propanol. In the laboratory, macroinverte- brates were removed first. Samples from shal- low ponds with unlined bottoms often con- tained sediment. To these, rose bengal stain was added to aid in sorting microinvertebrates (Mason and Yevich 1967). Samples in which zooplankton was estimated to exceed 300 indi- viduals were subsampled. To subsample. 108 Great Basin Naturalist [Volume 55 samplers were diluted to 500 or 1000 ml and stirred while 1% of the volume was drawn out with 1- and 2-ml Henson-Stemple pipettes. Invertebrate fauna were counted and iden- tified to family, with the exception of the orders Oligochaeta, Acari, Araneae, Eucope- poda, Ostracoda, and Lepidoptera, and the phyla Nematoda and Rotifera. Invertebrates were identified using keys in Pennak (1989) for non-insects, Merritt and Cummins (1984) for aquatic insects, and Borror and DeLong (1971) for terrestrial insects. B. McDaniel (Plant Science Department, South Dakota State University, Brookings) identified terres- trial invertebrate families and verified other identifications. Because data were not normally distrib- uted, nonparametric analysis methods were used. A median test was conducted on the dozen most common invertebrate taxa to determine if their abundance in sewage ponds differed from that in industrial ponds. For each taxa, numbers of individuals collected in each sample were used in analysis. Data were pooled over all ponds, years, and months with- in each of the two groups: sewage ponds and industrial ponds. Pooling samples for years and ponds allowed ample sample size for com- parison of gross invertebrate population differ- ences between pond types. A median test was also run on the total number of species collect- ed per pond during the entire sampling period to determine if species richness was greater at sewage ponds or industrial ponds. A third median test was conducted to compare inver- tebrate numbers between ponds with heavy metal concentrations greater than EPA criteria and those with heavy metal concentrations within EPA chronic exposure standards. Data were again pooled over all ponds, years, and months. Radioactive waste ponds were elimi- nated from median tests because only one sample was taken from them. Results Water Chemistry Heavy metal concentrations in most ponds were below criteria established by the EPA (U.S. Environmental Protection Agency 1987) (Table 1). Mercury was the only metal found in concentrations that might affect aquatic life (ponds TRAr and NRFi). However, in TRAr and NRFi mercury concentration was below the acute value of 2.4 ^tg/L (U.S. Environ- mental Protection Agency 1987). Sewage ponds had higher nitrogen and phosphorus concentrations than industrial and radioactive ponds (Table 2). Ammonia (NH4-N) concentrations in most ponds were within the range found in unpolluted surface water (Wetzel 1983); however, NH4-N con- centrations at ICPP sewage ponds were well above those usually found in eutrophic lakes. Nitrite (NO2-N) concentrations indicated high organic pollution at all sewage ponds except NRFs, which was the only sewage pond where NO2-N concentrations did not exceed those of industrial and radioactive Table L Selected heav\' metal concentrations (p-g/h) in wastewater ponds at INEL, Idiilio, August 1991, and EPA criteria''. Fond'' Criteria Metal ANLi^ CPPir2 TR.\r TR\il NRFi CTFi TSFir (Mg/L) Arsenic 9.4 0 <1'1 <1 3 5 2 190'" Barium 71 < 100 <100 <100 <100 <100 100 50,000 Bervllium <5 <10 <10 <10 <10 <10 <10 5.3 Boron — 30 .50 70 120 90 10 5000 Lead <2.1 3 3 3 2 3 2 3.2f Selenium <2 1 <1 1 2 1 1 35 Mercun <2() <0.1 0.2 <0.1 1.4 <0.1 <0.1 0.0 12« ''Concentrations at or below these le\els should have no adverse effects on freshwater systems. Naval Reactor Facilities officials suggested die following clarifica- tion: "The criteria in the last column have questionable applicabilitv' to die NRFi. The EPA maximum contaminant level for mercur\- in public community- drink- ing water systems is 2.0/ig/L." "ANLJ = Argonne National Laboratory-west industrial waste pond, CPPir2 = Idaho Chemical Processing Plant east percolation pond (industrial and radioactive), TRAr = Test Reactor Area warm waste pond (radioactive), TRAil = Test Reactor .\rea south cold waste pond (industri;d), NRFi = Naval Reactors Facility industrial waste ditch, CTFi = Containment Test Facility disposal pond (industrial), TSFir = Technical Support Facilitx disposal pond (industrial and radioactive). •^ANLi water sample tested at Envirodyne Engineers. St. Louis. MO, February 1988. "< symbol means water sample contained less than the detection level, which follows the < svnibol. ''Arsenic (III) 'At wafer hardness of 100 nig/L. Value is 1.3 at water hardness of 50 mg/L. KMercury (II) 1995] Invertebrates in Wastewater Ponds 109 Table 2. Nutrient concentrations in wastewater ponds at INEL, Idaho, August I99I.'* Nitrogen PliosphoiTiS pH'i (mg/Las N) (nig/L as P) Pond NH4+ NO2- NO2-+NO3 NO3 N03:NH4+ POj-^ Sewage ponds ANLs2 9.02 0.19 0.17 0.46 0.29 1.50 1.20 CPPsl 7.52 11.00 2.20 4.60 2.40 0.21 4.00 CPPs2 7.23 17.00 0.69 2.40 1.71 0.10 4.80 CPPs3 7.33 17.00 0.15 0.46 0.31 0.02 6.40 CPPs4 7.43 17.00 0.14 0.43 0.29 0.02 6.10 TRAs 6.87 0.41 0.13 5.10 4.97 12.12 0.79 NRFs 9.90 0.40 0.02 0.14 0.12 0.30 3.00 Nonsewage ponds AN Li 7.42 0.97 0.09 0.74 0.65 0.67 1.40 CPPir2 8.80 0.04 0.05 1.30 1.25 30.49 0.01 TRAil 7.60 0.01 0.06 1.10 1.04 104.00 0.07 TRAr 8.43 0.15 0.01 0.27 0.26 1.73 0.01 NRFi 7.42 0.01 O.OI 1.60 1.59 159.00 0.40 CTFir 9.97 0.01 0.01 0.45 0.44 44.00 0.09 TSFir 9.75 0.04 0.02 0.11 0.09 2.17 0.12 -'Samples were collected between 0800 and 1400 h, Mountain Standard Time. ''Water pH values fluctuate readilv. According to the INEL Industrial Waste Management Information System, 1989 effluent pH ranges and numbers of months pH was sampled ( ) were as follows: ANLsl, 7.8-9.8 (7); CPPsl-4, 7.5-8.6 (12); TRAsl-2. 7.1-8.0 (10); NRFs, 7.4-11.0 (12); TRAil-2, 7.5-8.0 (6); TRAr 6.3-6.8 (2); NRFi, 6.9-7.5 (12); TSFir, 7.1-7.9 (12). ponds. Nitrate (NO3-N) concentrations were not noticeably different between sewage ponds and industrial/radioactive ponds, and NO3-N levels of all ponds were within ranges commonly found in unpolluted freshwater (Wetzel 1983). The N03-N:NH4-N ratio is an indication of organic pollution, a lower number indicat- ing greater pollution (Wetzel 1983). The N03-N:NH4-N ratio was <1 at all sewage ponds except ANLs2 and TRAs, and >1 at all industrial and radioactive ponds except ANLi. However, only in ICPP sewage ponds were ratios small enough to be considered organi- cally contaminated (Wetzel 1983). Phosphorus concentrations at most sewage ponds were much higher than the concentration in the highest industrial/radioactive pond. Compared with maximums in uncontaminated surface waters, phosphorus concentrations in sewage ponds were 4-30 times greater, but of the industrial and radioactive ponds only concen- trations in ANLi and NRFi were substantially greater (7 and 2X) (Wetzel 1983). Invertebrate Fauna Forty-nine taxa of invertebrates were col- lected from waste ponds, of which 30 were aquatic (Table 3). Most nonaquatic forms were found in small numbers. Collembola, however, were found regularly and were probably on the water surface or shaken from emergent vegetation in the collection process. In order of decreasing abundance, the main taxa col- lected were Rotifera, Daphnidae, Ostracoda, Eucopepoda, Chydoridae, Corixidae, Chirono- midae, Oligochaeta, Baetidae, Psychodidae, Acari, Dytiscidae, and Notonectidae. The above taxa were also the most ubiquitous, except Chydoridae, Oligochaeta, and Psycho- didae, which were found in large numbers but in few samples. The number of invertebrate taxa collected per pond ranged from 5 to 22. Excluding ter- restrial taxa, the number of aquatic taxa col- lected ranged from 4 to 16 per pond. Radioactive ponds were sampled only in July, but the number of taxa collected was almost identical to July samples from nonradioactive industrial ponds (Table 4). Statistical analyses were not performed on radioactive ponds because only one activity trap sample was col- lected. Industrial (ANLi, TRAil and 2, NRFi, and CTFi) and sewage ponds had similar (P = .11) numbers of taxa per sample. Within most taxa, the number of individu- als collected varied greatly from pond to pond (Table 5). A median test revealed that activity trap samples from sewage ponds contained more Rotifera (P < .01), Daphnidae (F < .01), and Notonectidae (P = .04), whereas industri- al ponds yielded more Chydoridae (P < .01), no Great Basin Naturalist [Volume 55 Table 3. Invertebrate taxa and mean number collected from 15 wastewater ponds at INEL, Idaho, I99()-91-'. x/24 h TiLxa in = 96) Phylinn Rotifera 1471.14 Phylum Nematoda 0.05 Phylum Annelida Class Oligocliacta (ac|uatic earthworms) 6.32 Class Iliiiidinea (leeches) ( )rder Rh\ nchohdellida FamiK (Mossiphoniidae 0.02 Pin lum Arthropoda Class Crustacea Order Cladocera (water fleas) Famih' Daphnidae 1351.26 FamiK Ch\ doridae 102.88 Family Sididae 0.09 Order Eucopepoda (copepods) 151.45 Order Ostracoda (seed shrimps) 317.17 Order Amphipoda (scuds) FamiK' Talitridae 0.45 Class Arachnoidea Order Acari (mites) 1.51 Order Araneae (spiders)'' 0.04 Class Insecta Order Collembola (springtails) Family Entomobiyidae'' 0.57 Family Onychimidae'' 0.30 Order Ephemeroptera (mayflies) Family Baetidae 5.71 FamiK' Caenidae 0.01 Order Odonata Suborder Anisoptera (dragonflies) Family Aeshnidae 0.01 Suborder Zygoptera (damselflies) Family Coenagrionidae 0.31 Order Thysanoptera (thrips)'' Family Thripidae (common thrips)'' 0.11 Family Aeolothripidae (l:)anded thrips)'' 0.02 Order Hemiptera (tnie bugs) Family Corixidae (water boatmen) 39.76 Family Notonectidae (backswimmers) 0.53 Order Homoptera Family Aphidae (aphids)'' 0.05 Family Cercopidae (spittlebugs)'' 0.01 Family Cicadellidae (leaflioppers)'' 0.03 Family unidentified'' 0.25 Order Coleoptera (beetles) Family Chr\'somelidae (leaf beetles) 0.03 Family Coccinellidae (ladybird beetles)'' 0.01 Family Dytiscidae (predaceous diving beetles) 0.65 Family Elmidae (riffle beetles) 0.0 1 Family Gyrinidae (whirligig beetles) 0.01 FamiK Haliplidae (crawling water beetles) 0.02 Family Hydrophilidae (water scavenger beetles) 0.02 Family Ptiliidae (feather-winged beetles) 0.01 Family Staphylinidae (rove beetles) 0.02 Order Trichoptera (caddisflies) Family Leptoceridae 0.05 Order Lepidoptera (Ijuttei-flies and moths)'' 0.02 Order Diptera (flies) Family Ceratopogonidae (biting midges) 0.01 Family Psychodidae (moth flies and sand flies) 1.68 Family Chironomidae (midges) 11.52 Family Tipulidae (crane flies) 0.02 Family imidentified, adults'' 0.80 Family imidentified, pupae 0.99 Order Ilymenoptera Family Formicidae (ants)'' 0.03 Family Platygasteridae'' 0.01 Family Braconidae'' 0.01 Rimily Encyrtidae'' 0.01 Rimily Pteromalidae'' 0.01 Family Scelionidae'' 0.01 Family Sphecidae (sphecid wasps)'' 0.01 "Iiix'ertebrates were collected in 3.8-L activitv' traps suspended in the water column tor 24 h, one per pond, per month. Collections were June-October 1990 and March-May 1991 for 12 ponds, and July 1991 for .3 radioactive ponds. "Individuals found were mosth or e.\clusiveK' terrestrial. Acari (P = .01), and Baetidae (P = .01). Numbers of Oligochaeta (P = .44), Eucope- poda (P = .50), Ostracoda (P = .09), Corixidae (P = .08), Dytiscidae (P = .54), and Chirono- midae (P = .70) collected were not significant- ly different between sewage and industrial ponds. Invertebrate numbers in pond NRFi, which had a high mercury content, were compared to those in the remaining industrial ponds, where mercury was not detected. Samples from NRFi contained more Chironomidae (P = .02) and Oligochaeta (P < .01), and fewer Chydoridae (P = .03) and Ostracoda (P = .03) than ponds ANLi, TRAi, and CTFi. Numbers of Rotifera (P = .10), Daphnidae (P = .10), Eucopepoda (P = .10), Acari (P = .15), Baeti- dae (P = .55), Cori.xidae (P = .07), Notonectidae (P = .45), and Dytiscidae (P = .07) were simi- lar between the pond with mercuiy and those without. Discussion Wastewater ponds at INEL were nutrient- rich, especially sewage ponds. Organic enrich- ment may be the cause of high abundance and low number of invertebrate taxa found. Species richness at sewage ponds was similar to that at industrial ponds. However, species composi- tion differed between sewage and industrial ponds. Differences were probably due to the greater organic enrichment in sewage ponds. Activity trap samples from INEL ponds contained fewer invertebrate taxa than compa- rable samples fi-om natural waters (Gordon et al. 1995] Invertebrates in Wastewater Ponds 111 Table 4. Number of aquatic invertebrates per collec- tion (activity trap set for 24 h) from radioactive waste ponds at INEL. Idaho, Jul> 199 b>. CPPir2l' TIUi TSFir TlLxa (h = 1) (n = 1) (/i = I) Daphnidae 94 1 59 Chydoridae 0 0 129 Eucopepoda 35 0 818 Ostracoda 5 0 1620 Amphipoda 0 0 1 Baetidae 2 0 0 Corixidae 1 5 0 Dytiscidae 0 6 4 Chironomidae 7 0 18 ■'Data troin iadioacti\e vxaste ponds were not anal\ zed \\ itli tlioie troni sewage and industrial ponds because onK one sample was taken from radioactive ponds. ''CPPir2 = Idalio Chemical Processing Plant east percolation pond (industrial and radioactive), TRAr = Test Reactor Area warm waste pond (radioactive), TSFir = Technical Support Facilitv- disposal pond (industrial and radioactive). 1990, Neckles et al. 1990). Dominant taxa col- lected from study ponds were similar to domi- nant taxa collected in activity traps at natural wetlands in Nebraska (Gordon et al. 1990) and Manitoba (Neckles et al. 1990), with the exception of Culicidae, Turbellaria (Neckles 1990), and Gastropoda (Gordon et al. 1990, Neckles et al. 1990), which were not collected from wastewater ponds. In our study fewer taxa per sample were collected compared to activity trap samples from seasonal wetlands (Cowardin et al. 1979, Neckles et al. 1990); seasonal wetlands, like organically enriched systems of sewage ponds, tend to have low invertebrate taxa diversity (Wiggins et al. 1980). The reduced number of taxa in wastewater ponds may be due to lack of emergent vegeta- tion in most ponds. Odonate families Libelluli- dae and Lestidae, which were collected by Gordon et al. (1990) but not from wastewater ponds, are commonly associated with vascular hydrophytes (Merritt and Cummins 1984). Vegetation has been found to be correlated with macroinvertebrate species richness (Gilinsk-y 1984). Another possible cause of low species rich- ness in wastewater ponds is high organic waste content. Streams and wetlands receiv- ing organic waste typically exhibit low inverte- brate taxa diversity (Olive and Dambach 1973, Brightman and Fox 1976, Kondratieff and Simmons 1982, Kondratieff et al. 1984, Victor and Dickson 1985, Pearson and Penridge 1987). Hilsenhoff (1988) assigned arthropod families from streams in the Great Lakes region a tol- erance value from 0 (lowest tolerance to organic pollution) to 10 (highest). Eleven of the families for which Hilsenhoff (1988) presented tolerance values were found in INEL ponds, and only 2 had tolerance values of less than 4. Those 11 families and tolerance values are as follows: Aeshnidae and Tipulidae (3), Baetidae, Elmidae, and Leptoceridae (4), Ceratopogon- idae (6), Caenidae (7), Chironomidae and Talitridae (8), Coenagrionidae (9), and Psycho- didae (10). The two families with a 3 tolerance rating were represented by only single speci- mens in INEL wastewater ponds. Low invertebrate diversity in industrial ponds may be caused by organic or chemical constituents. Although nutrients in industrial waste ponds were within ranges found in nat- ural waters, most industrial ponds at INEL would be considered eutrophic (Wetzel 1983). Additional organic enrichment in sewage ponds did not affect species richness compared to industrial ponds; however, species composi- tion (%) was different between the two pond types. Metal and saline pollution has also been found to decrease aquatic invertebrate diversi- ty (Savage and Rabe 1973, Seagle et al. 1980, Euhss 1989). In most instances, the seven heavy metals tested did not occur in concentrations great enough to affect aquatic life. Only mercury was found at concentrations over chronic exposure levels. At concentrations below chronic levels, freshwater organisms should show no chronic toxic effects (U.S. Environmental Protection Agency 1987). Chydoridae and Ostracoda were scarcer, and Chironomidae and Oligochaeta more abundant, in samples from pond NRFi, wherein mercury was detected. Other toxins may occur in the water, and no other ponds with elevated mercury concentrations were available for comparison. Therefore, we do not know if mercury caused tlie difference detected. Although species richness of INEL ponds was low, comparison with natural wetlands (Gordon et al. 1990, Neckles et al. 1990) revealed that study ponds exhibited high invertebrate abundance. Of the taxa that waste- water pond and Nebraska wetland collections had in common, wastewater pond samples con- tained higher densities of all except Gyrinidae, Ceratopogonidae, and Hirudinea (Gordon et al. 1990). Gyrinidae and Ceratopogonidae were collected in almost identical amounts, and Hirudinea were more abundant in Nebraska 112 Great Basin Naturalist [Volume 55 O --I o a: " 7. 3 Z < 3 ^'i U 3 U 3 o-i II ^ in CO iiij —I lO lO o o o r^ CO ^^ Lo m ic i7 M ^ ^ -i< U 3 -^ II ^ t~- ^ CO ,— , O CO )n. .-', ^ O — I o n T2 ,u _= 5 Q Z 1j '-^- ^ CLi o >. _C -C O W O U D U U < CQ tl ■= II .2 H ■:r ~ t^ -= - "c p- T ^ f- = 1^ -a fe o s,^ A 1 2- ^ J: S J c z 11. ill = z 'S £ z tt- T 1 -P O 5i lO -^ T3 -o -a ^, in C C C .OJ Q- O O C i, c H a. -^ M tX 61 5-5 ■II s "S ■« ' II O C; 0; P u 2 ^ _aj _ai f U o.'g 1 "g "g ^z o < a; Z > 3 »i K' " t; t; - p <"S "S "I 1 cl = S S S 1995] Invertebrates in Wastewater Ponds 113 wetlands, compared to our study ponds (Gordon et al. 1990). Also, in our study, more Cladocera and Ostracoda were collected compared to activity trap samples from seasonal wetlands (Neckles et al. 1990), which tend to have a high invertebrate abundance (Wiggins et al. 1980). Nutrient-polluted natural waters also have invertebrate communities containing many individuals of a few species (Brightman and Fox 1976, Lubini-Ferlin 1986); Brightman and Fox (1976) attribute this partially to a reduction in competition from pollution-intol- erant forms. High invertebrate growth and abundance have been associated with high algal produc- tivity (Wallace and Merritt 1980, Richardson 1984), which in turn has been associated with high phosphorus and nitrogen concentrations (Liao and Lean 1978, Wetzel 1983). Most INEL wastewater ponds were eutrophic or highly eutrophic (Wetzel 1983). Therefore, wastewater ponds, which are higher in nutrients than nat- ural wetlands, would be expected to produce more invertebrate biomass. The absence of fish in study ponds proba- bly also contributed to high invertebrate den- sities. Fish have been shown to decrease aquatic invertebrate densities (Gilinsky 1984). For most taxa, collections from industrial ponds also had more individuals than collections from natural systems (Gordon et al. 1990, Neckles et al. 1990), even though industrial ponds were not as nutrient-rich as sewage ponds. In certain systems a large abundance of in- vertebrates has also been attributed to a paucity of insect predators (Brightman and Fox 1976, Williams 1985, Dodson 1987). However, sev- eral predaceous taxa were collected from waste ponds, most notably Dytiscidae and Notonectidae. Because these taxa were col- lected in greater numbers from wastewater ponds than from natural wetlands (Gordon et al. 1990), and because Notonectidae were most numerous in sewage ponds where many prey taxa were also most numerous, we surmise the large number of invertebrates collected from waste ponds resulted mostly from a reduction in competition from pollution-intolerant taxa, high algal productivity, and the absence of fish, rather than from lack of invertebrate pre- dation. Comparison of our results on water column invertebrates with other studies of sewage ponds is limited due to a scarcity of published papers. Porcella et al. (1972) noted large popu- lations of Daphnia in a reservoir fed mostly by treated sanitary wastewater. Daphnidae, Rotifera, and Notonectidae were more com- mon in INEL sewage ponds than in industrial ponds. All three species, as well as Oligochaeta, Eucopepoda, Ostracoda, and Corixidae (Sinclair 1975), are common inhabitants of sanitary wastewater Oligochaeta, Eucopepoda, Ostra- coda, Corixidae, and Chironomidae were abundant in sewage ponds, but not more so than in industrial ponds. Cladocera, Euco- pepoda, Ostracoda, Corixidae, and Chironomi- dae were also common in evaporation ponds in California, which contain salts and heavy metals (Euliss et al. 1991). Invertebrate communities in INEL sewage ponds differed from those in organically pol- luted streams. However, in making these com- parisons we note that our sampling methods did not target benthic organisms. In nutrient- enriched stream reaches, oligochaetes and chironomids are dominant (Duda et al. 1982, Pearson and Penridge 1987, Crawford et al. 1992), but we found no difference in numbers between sewage and industrial ponds. Some chironomid species (Kownacki 1977) and oligochaete families (Lewis 1986) are charac- teristic of clean waters, and it is possible the species inhabiting sewage ponds differed from those in industrial ponds. Ostracoda have also been described as pollution tolerant (Kownacki 1977), but we found no difference in their numbers at the .05 level of significance; at the .10 level, sewage pond samples contained more ostracods. Baetidae may be either pollution tolerant (Savage and Rabe 1973, Victor and Dickson 1985) or intolerant (Kownacki 1977) depending upon the species. We found more Baetidae in industrial ponds, indicating they, as well as Chydoridae and Acari which were also more abundant in industrial pond sam- ples, may be less tolerant of low oxygen con- centrations than the other common taxa. Taxa found in greater abundance in sewage ponds than in industrial ponds were those that could take advantage of the unique and difficult living conditions. Eutrophic waters typically exhibit lower dissolved oxygen concentrations and greater fluctuations in dissolved oxygen and pH than less organically enriched waters. Some cladoceran species can form hemoglo- bin when dissolved oxygen concentrations are 114 Great Basin Natuiulist [Volume 55 low; thus, oxygen levels are rarely a limiting factor (Pennak 1989). The same is true of" rotifers; certain genera are capable of with- standing anaerobic conditions for a short time and \ery low ox\'gen concentrations for extended periods (Pennak 1989). Since Notonectidae breathe at the water surface (Merritt and Cummins 1984), they are unaf- fected by dissolved oxygen concentrations. Most Cladocera are less affected by pH fluctu- ations than some taxa because they typically occur over a wide pH range (Pennak 1989). If pH levels are too high or too low, Cladocera and Rotifera can withstand temporarily unfa- \'orabIe environmental situations by producing resting eggs that are resistant to adverse chemical conditions. Under more favorable- conditions, Cladocera and Rotifera life cycles allow them to respond (quickly to improving conditions (Pennak 1989). Regarding the feeding habits of taxa that were more abundant in sewage ponds, Notonectidae were possibly taking advantage of the reduced competition from other preda- tors. Both rotifers and Daphnia are omnivo- rous and feed on any suitable-sized food parti- cle; therefore food was abundant for them in sewage ponds (Sinclair 1975). Daphnia can alter their body structure in response to algal concentrations, which is thought to be a mechanism for sui-viving algal blooms (Pennak 1989). Thus, while conditions in sewage ponds are hostile to many species, those that can tol- erate the conditions flourish due to an abun- dant food supply and the absence offish. In summary, wastewater ponds had low invertebrate diversity, which we attribute to lack of vegetation and inability of many species to withstand the environmental condi- tions. Wastewater ponds also had high inverte- brate abundance, which we attribute to reduc- tion of competing taxa, organic enrichment, and absence of vertebrate predators. There was no indication that heavy metal concentra- tions were high enough to reduce water column invertebrate concentrations in most ponds. High invertebrate concentrations in INEL wastewater ponds provided an abimdant food source for many bird species, migrator) and resident, which used INEL wastewater ponds. Bacteria, protozoa, and algae are important in waste treatment because they reduce the organic load of wastewater and convert waste into a form useable by organisms in the receiv- ing water body (Goulden 1976). In systems like some at INEL where water loss is through evaporation, all waste processing occurs in the pond. Zooplankton are also important in waste elimination and transfer (Goulden 1976, Patrick 1976, Bogatova and Yerofeyeva 1980). Other aquatic invertebrates that consume algae or bacteria, or feed on zoo- plankton, and are then eaten by birds also influence the reduction and transformation of organic waste and its dissipation out of the system. Acknowledgments We thank O. D. Markham for suggestions from initiation through project completion. We appreciate the assistance of L. Knobel and R. Bartholomay of the U.S. Geological Survey, which provided water chemistiy analysis. We thank W L. Tucker, Experiment Station statis- tician. South Dakota State University, for pro- viding statistical advice, and B. McDaniels and W G. Duffy of South Dakota State Uni- versity for assisting in invertebrate identifica- tion. W G. DuffV, O. D. Markham, and R. C. Morris reviewed the manuscript. Field and lab assistance was provided by L. Maddison, N. Anderson, P Saffel, S. Allen, and C. Birkelo. This research is a contribution from the INEL Radioecology and Ecology Program and was funded by the New Production Reactor Office, Idaho Field Office, and the Office of Health and Environmental Research, U.S. Department of Energy. Literature Cited Bogatova, I. B., and Z. I. Yerofeyeva. 1980. The use of container-reared cultures of Cladocera in polishing fish farm effluents. Hydrobiological Journal 16: ,56-61. BoRROR, D. J., AND D. M DeLong. 1971. An introduction to the study of insects. 3rd edition. Holt, Rinehart and Winston, New York, NY. 812 pp. Brightman, R. S., AND J. L. Fox. 1976. The response of benthic in\ertehrate populations to sewage addition. Pages 29.5-308 in Third annual report on cypress wetlands. Florida University, Center for Wetlands, Gainesville. Brix, H. 1993. Wastewater treatment in constructed wet- lands: system design, removal processes, and treat- ment performance. Pages 9-22 in G. A. Moshiri, editor. Constructed wetlands for water quality im- provement. Lewis Publishers, Ann Arbor, Ml. Brown, E., M. W Skougstad, and M. J. Fishman. 1970. Methods for collection and analysis of water samples for dissolved minerals and gases. Techniques of 1995] Invertebrates in Wastewater Ponds 115 water-resources investigations of the United States Geological Survey, Book 5, Chapter Al. U.S. Goxern- ment Printing Office, Washington, DC. 160 pp. Ch,'\dvvick, J. W., S. E Canton, .^nd R. L. Dent. 1986. Recovery of benthic invertebrate communities in Silver Bow Creek, Montana, following improved metal mine wastewater treatment. Water, Air, and Soil Pollution 28: 427-438. CiEMiNSKi, K. L. 1993. Wildlife use of wastewater ponds at the Idaho National Engineering Laboratory. Unpublished master's thesis. South Dakota State University, Brookings. 311 pp. Cl.wson, K. L., G. E. Start, and N. R. Ricks. 1989. Climatography of the Idaho National Engineering Laboratory. 2nd edition. U.S. Department of Com- merce, National Oceanic and Atmospheric Adminis- tration, Idaho Falls, ID. DOE-ID-12118. 155 pp. Cow ardin, L. M., V Carter, E C. Golet, and E. T LaRoe. 1979. Classification of wetlands and deepwater habi- tats of the United States. U.S. Fish and Wildlife Service, Office of Biological Sei^vices, Washington, DC. FWS/OBS -79/31. Cr.\wford, C. C, D. J. Wangsness, and J. D. Martin. 1992. Recover)' of benthic-invertebrate communities in the White River near Indianapolis, Indiana, LISA, following implementation of advanced treatment of municipal wastewater. Archi\' fiir H\drobiologie 126: 67-84. Dodson, S. I. 1987. Animal assemblages in temporary desert rock pools: aspects of the ecology oi Dasijhelea sublettei (Diptera: Ceratopogonidae). Journal of the North American Benthological Society 6: 65-71. Duda, a. M., D. R. Lenat, and D. L. Penrose. 1982. Water quality in urban streams — what we can e.xpect. Journal of the Water Pollution Control Federation 54: 1139-1147. EULISS, N. H., Jr. 1989. Assessment of drainwater evapo- ration ponds as waterfowl habitat in the San Joaquin Valley, California. Unpublished doctoral dissertation, Oregon State Universit\, Conallis. EuLiss, N. H., Jr., R. L. Jarvis, .\nd D. S. Gilmer. 1991. Feeding ecology of waterfowl wintering on exapora- tion ponds in California. Condor 93: 582-590. FisH.\iAN, M. J., AND L. C. Friedman, editors. 1989. Methods for determination of inorganic substances in water and fluvial sediments. 3rd edition. Tech- niques of water-resources investigations of the United States Geological Survey, Book 5, Chapter Al. U.S. Government Printing Office, Wishington, DC. 545 pp. GlLlNSKT, E. 1984. The role offish predation and spatial heterogeneity in determining benthic community structure. Ecology 65: 455-468. Gloyna, E. F, J. F Malina, Jr., .\nd E. M. D.wis, edi- tors. 1976. Ponds as a wastewater treatment alteiTia- tive. Center for Research in Water Resources, Unixer- sity of Te.xas at Austin. 447 pp. Gordon, C. C, L. D. Flake, and K. F Higgins. 1990. Aquatic invertebrates in the Rainwater Basin area of Nebraska. Prairie Naturalist 22: 191-200. GouLDEN, C. E. 1976. Biological species interactions and their significance in waste stabilization ponds. Pages 57-67 in E. F Gloyna, J. F Malina, Jr., and E. M. Davis, editors. Ponds as a wastewater treatment alternative. Center for Research in Water Resources, University' of Texas at Austin. 447 pp. Halford, D. K., and J. B. Millard. 1978. Vertebrate fauna of a radioactive leaching pond complex in southeastern Idaho. Great Basin Naturalist 38: 64-70. HiLSENHOFK, W L. 1988. Rapid field assessment of organ- ic pollution with a family-level biotic index. Journal of the North American Benthological Society 7: 65-68. Howe, F P, and L. D. Fl\ke. 1989. Mourning dove use of man-made ponds in a cold-desert ecosystem in Idaho. Great Basin Naturalist 49: 627-631. Klotz, L. 1977. The effects of secondarily treated sewage effluent on the Willimantic/Shetueket River. Uni- versity of Connecticut Institute of Water Resources, Storrs, Report 27. 85 pp. KONDR/VTIEFF, R F, R. A. MATTHEWS, AND A. L. BUIKEMA, Jr. 1984. A stressed stream ecosystem: macroinver- tebrate community integrity and microbial trophic response. Hydrobiologia 111: 81-91. KONDR.^TIEFF, P F, AND G. M. SiMMONS, JR. 1982. Nutrient retention and macroinvertebrate community structure in a small stream receiving sewage efflu- ent. Archiv fiir Hydrobiologie 94: 83-98. KovvNACKl, A. 1977. Biocenosis of a high mountain stream under the influence of tourism. 4. The bottom fauna of the stream R\'bi Potok (the High Tatra Mts.). Acta Hydrobiologica 19: 293-312. Lewis, M. A. 1986. Impact of a municipal wastewater effluent on water quality, periphyton, and inverte- brates in the Little Miami River near Xenia, Ohio. Ohio JouiTial of Science 86: 2-8. LlAO, C. F-H., AND D. R. S. Lean. 1978. Nitrogen trans- formations within the trophogenic zone of lakes. Journal of Fisheries Research Board of Canada 35: 1102-1108. Lubini-Ferlin, V. 1986. The influence of sewage treat- ment plant effluents on benthic invertebrates in Lake Zurich Switzerland. Schweitzerische Zeit- schrift fiir Hydrologie 48: 53-63. Mason, W. T, Jr., and P P Yevich. 1967. The use of phloxine B and rose bengal stains to facilitate sorting benthic samples. Transactions of the American Microscopical Societ\' 86: 221-223. McBride, R., N. R. French, A. H. Dahl and J. E. Detmer. 1978. Vegetation tv'pes and surface soils of the Idaho National Engineering Laboratory Site. lDO-12084. U.S. Department of Energy, Idaho Operations Office, Idaho Falls. 29 pp. Merritt, R. W, and K. W. Cummins. 1984. An introduc- tion to the aquatic insects of North America. 2nd edition. Kendall/Hunt Publishing Co., Dubuque, lO. 722 pp. Millard, J. B., F W Whicker, and O. D. Markha.m. 1990. Radionuclide uptake and growdi of barn swal- lows nesting by radioactive leaching ponds. Health Physics 58: 429-439. MOSHIRI, G. A., EDITOR. 1993. Constructed wetlands for water quality improvement. Lewis Publishers, Ann Arbor, MI. 632 pp. Neckles, H. a., H. R. Murkin, and J. A. Cooper. 1990. Influences of seasonal flooding on macroinverte- brate abundance in wetland habitats. Freshwater Biology 23: 311-322. Olive, J. H., and C. A. Dambach. 1973. Benthic macroin- vertebrates as indexes of water quality in whetstone Creek, Monow Count); Ohio (Scioto River Basin). Ohio Journal of Science 7.3: 129-149. 116 Great Basin Naturalist [Volume 55 Patrick, R. 1976. The effect of a stabilization pond on the Sabine Estuary. Pages 33-55 in E. F Gloyna, J. E Mahna, Jr., and E. M. Davis, editors. Ponds as a wastewater treatment alternative. Center for Research in Water Resources, University of Texas at Austin. 447 pp. Pearson, R. G., and L. K. Penridce. 1987. Effects of pol- lution by organic sugar mill effluent on the macro- invertebrates of a stream in tropical Queensland, Australia. Journal of Environmental Management 24: 205-215. Pennak, R. W. 1989. Fresh-water invertebrates of the United States: Protozoa to Mollusca. 3rd edition. John Wiley & Sons, Inc., New York, NY. 628 pp. PoRCELLA, D. B., P H. McGauhey, and G. L. Dugan. 1972. Response to tertiai-y effluent in Indian Creek Reservoir. Journal of the Water Pollution Control Federation 44: 2148-2161. Richardson, J. S. 1984. Effects of seston quality on the growth of a lake-outlet filter-feeder. Oikos 43: 386-390. Ross, L. C. M., AND H. R. Murkin. 1989. Invertebrates. Pages 35-38 in E. J. Murkin and H. R. Murkin, edi- tors. Marsh ecology research program: long term monitoring procedures manual. Delta Waterfowl Wetlands Research Station, Technical Bulletin 2. Savage, N. L., and E W. Rabe. 1973. The effects of mine and domestic wastes on macroinvertebrate commu- nity structure in the Coeur d'Alene River Northwest Science 47: 159-168. Seagle, H. H., Jr., A. C. Hendricks, and J. Cairns, Jr. 1980. Does improved waste treatment have demon- strable biological benefits? Environmental Manage- ment 4: 49-56. Sinclair, R. M. 1975. Fresliwater biology and pollution ecology training manual. EPA-430/1 -75-005. National Technical Information Service, Springfield, VA. Task Force on Natural Systems. 1990. Natural systems for wastewater treatment. Water Pollution Control Federation, Alexandria, VA. 270 pp. United States Environmental Protection Agency. 1987. Quality criteria for water 1986. EPA 440/5-86- 001. Office of Water Regulations and Standards, U.S. Government Printing Office, Washington, DC. 1987/1302-M/60645. Victor, R., and D. T. Dickson. 1985. Macrobenthic invertebrates of a perturbed stream in southern Nigeria. Environmental Pollution (Series A) 38: 99-107. Wallace, J. B., and R. W Merritt. 1980. Filter-feeding ecology of aquatic insects. Annual Review of Entom- ology 25: 103-132. Wetzel, R. G. 1983. Linniology. 2nd edition. Saunders College Publishing, Chicago, IL. 767 pp. Wiggins, G. B., R. J. Mackay, and I. M. Smith. 1980. Evolutionaiy and ecological strategies of animals in annual temporary pools. Archiv fiir Hydrobiologie Supplement 58: 97-206. Williams, W. D. 1985. Biotic adaptations in temporary lentic waters, with special reference to tliose in semi- arid and arid regions. Hydrobiologia 125: 85-110. Received 14 January 1994 Accepted 7 September 1994 Great Basin Naturalist 55(2), © 1995, pp. 117-123 GROWTH AND REPRODUCTION IN AN ALPINE CUSHION PLANT: ASTRAGALUS KENTROPHYTA VAR. IMPLEXUS Wayne R. Owenl Abstract. — A two-year field experiment was conducted to investigate factors hypothesized to affect the reproduc- tive potential of Astragalus kentwphyta van implexus and to test the importance of trade-offs between growth and repro- duction in this species. Levels of mineral nutrients, water, herbivory, and competition were manipulated. Seed output and growth of individuals in treatment groups were compared against control plants. Neither water nor mineral nutri- ents alone were shown to affect growth or reproduction. Herbivoiy was shown to be similarly unimportant in affecting growth and reproduction. Competition with other species influenced growth but not reproduction. No significant trade- offs between growth and reproduction were detected within \'ears. However, there did appear to be a trade-off between these major fitness components when compared between years. Key words: Astragalus, alpine, competition, fecundity, trade-ojf. White Mountains. The impact of resource availability on the reproductive output of plants is well estab- lished (Harper 1977, Schoener 1983, Fowler 1986, Welden and Slausen 1986). Plants may experience resource limitation as a result of competition (inter- or intraspecific) or poor habitat quality. Resource limitations can also occur when a portion of a plant's photosyn- thetic organs are removed (e.g., by herbivory), damage which clearly interferes with the plant's ability to provision its offspring (Marquis 1991). A number of authors (Cody 1966, MacArthur and Wilson 1967, Harper 1977, Grime 1979, Tilman 1982, Weiner 1988, 1990) have considered the ecological consequences of resource limitation for individuals and pop- ulations and have described various strategies that plants might be expected to pursue to optimize the allocation of limited resources. This study tests whether the availability of resources limits the fecundity of Astragalus kentrophyta Gray var. implexus (Canby) Barneby (hereafter, simply A. kentrophyta) and to what extent trade-offs between growth and reproduction might influence patterns of reproduction observed in this species. A. ken- trophyta is an alpine cushion plant indigenous to high elevations throughout the Intermoun- tain West of North America (Barneby 1964). Many lines of evidence suggest that repro- duction in A. kentrophyta might be resource limited. Experiments involving other organ- isms from this habitat have shown that avail- ability of resources influences the competitive ability and distribution of species (Wright and Mooney 1965, Mooney 1966, Marchand 1973), though this is not generally true of all alpine habitats (Korner 1989). Second, standing bio- mass and percent cover are substantially lower on dolomitic soils than on adjacent sandstone- and granite-derived substrates, suggesting that plants on the dolomite barrens might be rela- tively resource limited (Mooney 1966, Owen 1991). Third, A. kentrophyta plants routinely abort the majority of flowers they produce each year (Owen 1991), a pattern that has been attributed to resource limitations in a broad spectrum of species (Lovett Doust and Lovett Doust 1988). An experiment was designed (1) to test whether there are resource constraints on the reproduction and growth of A. kentrophyta and (2) to assess the interactions between two major components of fitness (i.e., growth and reproduction) under different regimes of resource availability. To do this, a factorial field experiment was established in which sep- arate groups of plants would receive either (1) water or (2) nutrient supplements, (3) protec- tion from herbivory, or (4) relief from the potentially competitive influence of neighbors. 'University of California, Davis, and White Mountain Research Station. University- of California, Los .\ngeles. Present address; Boise National Forest. 17.50 Front Street, Boise, ID 83702. 117 118 Great Basin Naturalist [Volume 55 Study Area The study was condueted on the alpine dolomite barrens of Sheep Mountain Pass above the Patriareh Grove bristlecone pine forest, in the White Mountains of Mono County, CA. Elevations at the site range from 3535 m (11,600 ft) to 3660 m (12,000 ft), and topographie relief of the site is minimal. In the White Mountains A. kentrophyta occurs only on dolomitic soils (Lloyd and Mitchell 1973, Hall 1991). Weather data were obtained from the White Mountain Research Station, Mt. Barcroft Laboratory, located 6 km north of the study site at an elevation of 3800 m. Soils on the dolo- mite barrens have a high cation exchange capacity and are depauperate in nitrogen, phosphorus, and potassium (Mooney et al. 1962, Wright and Mooney 1965, Brayton and Mooney 1966, Mooney 1966, Marchand 1973, 1974). The moisture-holding capacity of dolomite-derived soils is equivalent to that of adjacent granitic soils (Mooney et al. 1962, Wright and Mooney 1965, Marchand 1973). Vegetation of the White Mountains is general- ly xerophytic; this trend is especially prevalent on the dolomite barrens (Lloyd and Mitchell 1973). Materials and Methods In June 1989, 195 healthy A. kentrophyta plants were selected randomly from within an area of approximately 0.2 ha. Decadent (senes- cent) plants were disqualified from inclusion in this experiment. The specific location of the site was chosen for its apparent homogeneity with respect to soil physical characteristics, vegetation, and topographic profile. Plants were randomly allocated to five treatment regimes: (1) 50 plants were provided with three separate 1-L applications of water dur- ing the 1989 growing season. Plants were watered during the driest part of the summer (4 July, 2 August, and 19 August) to maximize the beneficial impact of the treatment. Water was applied slowly (to maximize infiltration) in a radius of 12.5 cm around each plant. This treatment supplied 6.1 cm of moisture to each plant. Expected precipitation for the three- month growing season is 8.7 cm (Pace et al. 1968). The 1989 summer precipitation was 1.1 cm. This treatment group will be referred to as "Water. " (2) Another 50 plants received sup- plemental nutrients. These plants were given approximately 17 g of a balanced general-pur- pose fertilizer (Scott's All- Purpose Builder, 12:10:12 N:P:K), providing each plant with 2.0 g N (in the form of ammoniacal nitrogen, ureas, and water soluble nitrogen), 1.7 g P (from phosphoric acid, P2O5), and 2.0 g K (from sol- uble potash, K2O). These quantities are equiv- alent to application rates of 13.8, 11.7, and 13.8 kg ha~^ respectively. A balanced fertilizer was chosen because experiments by Chambers et al. (1987) and Shaver and Chapin (1980) have shown that plants in cold environments re- spond most vigorously to resource augmenta- tion with fertilizer containing a balance of essential nutrients. The diy fertilizer was scat- tered in an approximately 2-cm-wide ring around the perimeter of each test plant. Summer seasonal precipitation in 1989 was apparently sufficient to solubilize the fertilizer and deliver it to the soil profile, as the granules had completely disappeared from the surface in approximately one month. This treatment group will be referred to as "Fertilized." (3) A third treatment was designed to protect plants from herbivoiy and predation on flowers and young fruits. Two locally common insects ha- bitually consume the reproductive parts of A. kentrophyta. The more common of these in- sects, a darkling beetle (Tenebrionidae: Coleop- tera), consumes flowers. Larvae of a locally com- mon Lycinid butterfly species (Lycaenidae: Lepidoptera) occasionally consume immature A. kentrophyta fruits. "Tangle-foot" brand sticky-trap was applied in a circle around each of 25 plants to exclude potential herbivores. Tanglefoot barriers were repaired as needed. This treatment group will be called "No Predation." (4) The fourth treatment sought to relieve a group of 20 A. kentrophyta plants from neighborhood competition. A 0.25-m- radius circle around a central target A. kentro- phyta plant was cleared of all other plants by cutting them off at ground level. This method minimized ground surface disturbance. Clearings were 0.2 m^ in area. The average number of neighbors (ramets) removed was 63 (mostly tillers of Poa rupicola), covering an average of 15% of the ground surface. Excavations of A. kentrophyta plants show diat its roots grow straight downward into the soil with minimal lateral root spread (Owen 1991). Roots of the target plants were therefore 1995] Astragalus Growth and Reproduction 119 thought to be well isolated from interactions with actively assimilating roots of other plants. Plants clipped in the cleared areas were trimmed if they resprouted. Plants in this treat- ment group are referred to as the "Target " group. (5) A final group of 50 unmanipulated plants was marked as a "Control" group. Size of the experimental groups was based on an analysis of expected variances in responses to the treatments; lower expected variances re- quire smaller necessarv samples (Sokal and Rohlfl981). Plant sizes (cushion area) were measured and recorded on 23 Jime 1989, shortly after initiation of growth for the season. Treatments were initially applied on 4 July 1989. In Sep- tember 1989 all plants were remeasured, and the entire fruit and seed crop produced by each of the 195 plants was han'ested. Since A. kentropJnjta forms a tight cushion that never exceeds 1 cm in height and seeds are not released from the plant before the end of the growing season, there was great confidence that the entire seed crop of each individual was retrieved. In early June 1990 I again mea- sured the area of all plants just as they were initiating growth for the season. Fertilized and Water treatments were not repeated in 1990 so as to evaluate the potential for lags in the effectiveness of resource supplementation. Tanglefoot barriers were maintained during 1990 to test for interannual variation in the effects of herbivores and predators. Clear zones around Target plants were maintained in 1990. All plants were allowed to grow through the season, and in September 1990 all 195 plants were remeasured and all fruits and seeds har- vested. No attempt was made to quantify' flower production, but previous experience (Owen 1991) had shown that seed production is a sig- nificant function of flower production (Owen 1991). Flowers, when aborted, are dropped at a very early age (Owen 1991) and probably represent a minimal per-unit cost in resources to the plant (Bookman 1983, Stephenson 1984). Therefore, the cost of flowers should be proportional to a plant's seed output and can safely be disregarded for the purpose of this work. Fruits and seeds were cleaned and sepa- rated in the laboratoiy, counted, and weighed. Results Weight of individual reproductive struc- tures (seeds and fruits) was independent of total numbers of those items produced per plant in both years (Table 1). Average seed and fruit weights were significantly correlated [R = .429 in 1989, R = .443 in 1990). There were no significant differences between treatment groups for the weight of individual seeds or fruits (results not presented). Because seed production is well correlated with other possi- ble measures of fitness in A. kentrophyta and weights of those seeds are independent of the numbers of reproductive structures produced on a plant (Table 1), seed output was used as an index of total reproductive effort. In a comparison of slopes of regression analyses, growth was a significant function of plant size in both 1989 and 1990 (Table 2), though the relationship was weaker in 1990. The weight of individual seeds and fruits was independent of seasonal growth (Table 2). The amount of growth across years was significant- ly but poorly correlated. Table 1. Con-elation matrix for selected demographic traits. Values above the diagonal are conelation coefficients (R) based on 1990 data; those below the diagonal are derived from 1989 data. Seed Seed Fruit Fruit Reproductive Seeds weight weight Fniits weight weight weight produced (a\erage) (total) produced (ax'erage) (total) (total) Seeds produced** 1 .003 .976* ,964* .143* .920* .966* Seed weight (average) .042 1 .139* -.001 .433* .081* .115* Seed weight (total) .977* .200* 1 .94,5* .229* .937* .987* Fruits produced** .963* .024 .033* 1 .106 .963* .968* Fruit weight (average) .136* .429* .215* .074 1 .289* .260* Fruit weight (total)** .943* .120* .949* .9,52* .284* 1 .981* Total reproductive weight** .973* .16,56* .989* .9,54* .249* .985* 1 ♦Kendall Rank Correlation is significant at P < .0,5. **Treatment differences noted with one-way .\NOVA. Tlies ■ diflcrences do not affect the magnitude of significance of the coiTelations. 120 Great Basin Naturalist [Volume 55 Table 2. Slopes of regressions for selected demographic traits on growth in 1989 and 1990 using the total data set (i.e., not partitioned by treatment). Where the overall regressions are not significant, there were also no treat- ment differences. Ciowth in 1989 Growth in 1990 Table 3. Result of an ANCOVA on seed production and growth h\' treatment group. The covariate is plant size. The treatments are those listed in the text (see also Table 4). Growth in 1990 .168* — Plant size .340* .110* Seed weight -.038 -.054 Fruit weight .035 .036 'Regressions are significantK positise (P < .0.5). f)ne-wa\- .WOVAs suggest differences between treatment groups for values ol these traits (P < .0.5). Seed production (square root transformed) was a positive linear function of plant size. Overall values of R- for regressions of seed production on plant size were .206 in 1989 and .182 in 1990. Slopes of individual regres- sions for each treatment for seed production on plant size did not differ from the slope for control plants. Plant size was a minor but important factor influencing both growth and reproduction in A. kentrophyta and indicates that size should be considered as a covariate in an analysis of variance of treatment effects in this experi- ment. Analyses of covariance (ANCOVA) and experimental results are presented in Tables 3 and 4, respectively. Plant size was a significant covariate in three of four analyses. There were no differences among treatment groups in seed production (reproduction) for either year Growth did not differ among treatment groups in 1989, but there was a significant difference between groups in 1990 (P = .047). A protect- ed least-significant-difference (LSD) test indi- cates that growth in the Target group was greater than that of individuals in other treat- ment groups (Table 4). Table 5 gives the results of two-tailed t tests comparing mean reproduction and growth across years within treatment groups. There were no significant differences for seed pro- duction among treatment groups between 1989 and 1990. Average size for plants in 1990 was consistently significantly greater than the size of the same plants the previous year (i.e., on average, plants grew larger over the course of the experiment). The No Predation treat- ment grew significantly less in 1990 than 1989, whereas plants in the Target group grew significantly more in 1990. There were no sig- nificant differences in growth across years for plants in the Control, Fertilized, or Water groups. Covariate Treatment 1989 Seed production 37.164 <.001 1.358 .25 1990 Seed production ,39.818 <.001 1.854 .12 1989 Growth 27.207 <.001 0.822 ..583 1990 Growth 0.893 .346 2.453 .047 A series of simple linear regressions was used to compare seed production with growth to test for the presence of a trade-off between these two primary components of fitness. When the data are corrected for the fact that larger plants are inherently more capable of producing more flowers and fruits, the analy- sis finds no significant differences among treatment groups (by virtue of overlapping 95% confidence intervals); and, therefore, no trade-off between growth and reproduction within a given year was detected. To compare trade-offs across years, the ratio of 1990 to 1989 data was used (Table 5). This provides a number >1.0 when 1990 data values exceed 1989 values; the converse is true when results are <1.0. Seed production was greater in 1990 than in 1989 regardless of treatment group. In contrast, growth in 1990 was less than that experienced in 1989 with the notable exception of Target plants. The results can be interpreted as evidence for a trade-off between growth and reproduction. They indicate that, in general, increased seed production is associated witli decreased growth. Furthermore, plants may be relieved of trade- off constraints by removing competitors, which should increase availability of mineral resources to the remaining (target) plant. Discussion Resource supplementation or alleviation of resource competition did not significantly influence the reproductive output of A. kentro- phyta. Instead, seed production was more close- ly related to the individual s past record of seed output (Tables 1, 3, 5). Plants that produced many seeds in 1989 tended to produce many seeds in 1990, regardless of treatment. Growth, while similarly unresponsive to the addition of single resources, increased significantly when potential competitors were removed (Tables 4, 1995] Astragalus Growth and Reproduction 121 Table 4. Treatment means (SD) in both 1989 and 1990 for important demographic traits. Control No bugs Fertilized Water Target 1989 Seed production 1990 Seed production 25.8 (25.2 32.2 (32.28) 16.1 (11.8) 20.5 (16.7) 30.6 (24.5) 39.7 (37.3) 25.1 (22.2) 30.7 (27.9) 44.2 (41.4) 54.5 (58.4) 1989 Plant size 1990 Plant size 5997.1 (2851.7) 7247.3 (3128.8) 4594.6 (1871.8) 5596.3 (2156.6) 6833.9 (2892.7) 7934.0 (3242.6) 6333.2 (2891.4) 7418.2 (3627.4) 7683.2 (3683.8) 8393.0 (4159.9) 1989 Growth 1990 Growth* 1478.4 (1329.7) 1156.1 (1529.9) 1530.0 (987.7) 808.4 (1000.4) 1772.1 (1634.2) 1587.8 (2044.5) 1797.9 (1486.9) 1395.0 (1760.3) 1503.1 (988.6) 2433.2 (1749.0) *Groutli in 1990 \aried significantK among treatments (see Table 3). The Target groups grew more, on a\erage, tlian did plants in an\ other treatment group. No other differences were significant. 5). These results differ from those of Wright and Mooney (1965), Mooney (1966), and Marchand (1973), which show that mineral nutrients were the primary factors limiting other species that occur on dolomite in the White Mountains {Artemisia tridentata, two Erigeron species, and Liipinus argenteus, respectively). Korner (1989) reports that the effect of fertilization on the growth of species from nutrient-poor environments is often diffi- cult to detect. He does not cite studies that address the relationship between growth and reproduction in nutrient-supplementation experiments. The addition of mineral nutrients or water alone may have been insufficient stimuli for A. kentrophyto to increase either reproduction or growth if both factors were limiting. Multiple limiting factors have been reported in a vari- ety of species (Harper 1977) and are specifi- cally predicted by Tilman's (1980, 1982) mod- els of optimal resource consumption. That there may be multiple resource limits to A. kentrophyta growth and reproduction is sup- ported by the response of A. kentrophyta to the removal of competitors in this study. Tanglefoot barriers were very effective at excluding ground-moving herbivores and predators. This was evidenced by the lack of foliar damage or partially eaten fruit and the capture of many insects in the traps. Flowers of A. kentrophyta are produced in sufficient excess to buffer individuals against the levels of flower and fruit predation observed in this population. Growth in A. kentrophyta, as has been re- ported for a number of species from arid regions throughout the world (Fonteyn and Mahall 1981, Robberecht et al. 1983, Ehleringer 1984, Parker and Salzman 1985, Shaw 1987, Manning and Barbour 1988, and Chapin et al. 1989), is most sensitive to the pro.ximity of its neighbors. It is unclear, however, why repro- duction among such species is rarely similarly influenced (as is the case with A. kentrophyta). The buffering of fitness components against environmental stochasticity is characteristic of density-vague demographics as described by Strong (1986). Under density-vague condi- tions, selection favors demographic functions with indeterminate functional thresholds. That is, current allocation decisions are only loosely linked to current environmental conditions (Strong 1986). Trade-offs between growth and reproduc- tion within years were not observed in this experiment under any conditions. A weak trade-off between growth and reproduction was identified in most treatment groups when data were compared across years (Table 5). It is of great interest that the Target group alone experienced an increase in both seed produc- tion and growth in 1990 compared to 1989 val- ues (and thus did not experience a trade-off). The absence of well-defined trade-offs between primaiy components of fitness could be due to one of several reasons. Lack of a discernible trade-off would be noted if resources were not truly limiting. It may also be that growth and reproduction are not co-limiting for this species in this environment. If this were true, factors that influence growth and reproduction are likely to be independent (e.g., one fitness component might be canalized and the other dependent on environmental conditions). Finally, a trade-off between growth and repro- duction would not be detected if a resource other than one provided in this experiment were limiting. Adult A. kentrophyta mortality at the Sheep Mountain study site is low, juvenile mortality is extremely high (even though germination 122 Great Basin Naturalist [Volume 55 Table 5. Cross-year comparisons of fitness components. 1990 \ akies represented as a fraction of 1989 trait valnes. Values of t and tiie associated prol)al)ilities (P) represent results of two-tailed / tests for differences in values between years. Refer to Table 4 lor raw tlata. Control No Inigs Fertilized Water Target Seed production 90/89* 1.25 1.16 1..32 1.15 1.18 t 1.41 1.71 1.80 1.39 0.71 P .17 .10 .08 .17 .49 Plant size t 7.06 5.02 4.90 5.05 3.50 P <.01 <.01 <.01 <.01 <.01 (wowth 90/89* 0.85 0.98 0.56 0.86 2.07 t 1.13 2.50 0.40 1.42 2.12 P .26 .02 .70 .16 .05 *\'alues listed represent the ratio ol 1990 trait \ allies to those ol 19S9 tests under controlled conditions show seed viability of greater than 95%), and recruitment is low (Owen 1991). These demographic attri- butes would certainly favor a strategy that routes resources away from the risky business of reproduction toward growth. The small but consistent portion of A. kentrophyta's annual accumulation of biomass allocated to repro- duction guarantees that each plant will proba- bly produce at least a few seeds each year while being able to dedicate most of each sea- son's accumulated resources to growth and survival. That the allocation of resources to reproduction, but not growth, in this species is constant over a broad range of resource avail- abilities is consistent with a bet-hedging life- history strategy (Kozlowski and Stearns 1989, Philippi and Seger 1989, Stearns 1989). Resource limitations on organisms are rarely simple or solitary. While fruit and flower pre- dation can be an important limit on fecundity, such an effect was not noted here. Similarly, the reproductive output of plants growing on the Sheep Mountain dolomite barrens would appear to be resource limited, although single resource augmentation had no direct effect on seed production. In combination, however, resources can influence the amount of realized growth that in subsequent years will affect reproduction. Acknowledgments I would like to thank the White Mountain Research Station for providing logistic and financial support for this project, especially the crew at the Mt. Barcroft Laboratory. T. Holmes, E. Nagy, A. Fitter, and two anony- mous reviewers made significant improve- ments on earlier drafts of this manuscript. Literature Cited B.^RNEBY, R. C. 1964. Atlas of North American Astragalus. Memoirs of the New York Botanical Garden 13: 1-1187. Bookman, S. S. 1983. Costs and benefits of flower abscis- sion and fruit abortion in Asclepias speciosa. Ecology 64; 264-273. Brayton, R., and H. a. Mooney. 1966. Population vari- ability of Cercocarpus in the White Mountains of Cali- fornia as related to habitat. Evolution 20: 383-391. Chambers, J. C, J. A. MacMahon, and R. W. Brown. 1987. Response of an early serai dominant alpine grass and a late serai dominant alpine forb to N and P availability. Reclamation and Revegetation Research 6: 219-234. Chapin, E S., J. B. McGraw, and G. R. Shaver. 1989. Competition causes regular spacing of alder in Alaskan shrub tundra. Oecologia 79: 412-416. Cody, M. L. 1966. A general theory of clutch size. Evolution 20: 174-184. Ehleringer, J. R. 1984. Intraspecific competitive effects on water relations, growth, and reproduction in Encelia farinosa. Oecologia 63: 153-158. EONTEYN, P J., AND B. E. Mahall. 1981. An e.xperimental analysis of structure in a desert plant community. Journal of Ecolog>' 69: 883-896. Fowler, N. 1986. The role of competition in plant commun- ities ill arid and semiarid regions. Annual Review of Ecology and Systematics 17: 443-464. Grime, J. P 1979. Plant strategies and vegetation process- es. John Wiley and Sons, New York, NY. Hall, C. A., editor. 1991. Natural histoiy of the White- Inyo Range. University of California Press, Berkeley. Harper, J. L. 1977. Population biolo,g\ of plants. Academic Press, New York, NY. KoRNER, C. 1989. The nutritional status of plants from higher altitudes. Oecologia 81: 379-391. Kozlowski, J., and S. C. Stearns. 1989. Hypotheses for the production of excess zygotes: models of bet- liedging and selective abortion. Exolution 43: 1369-1377. Lloyd, R. M., and R. S. Mitchell. 1973. A flora of the White Mountains, Ctilifornia and Nevada. University of California Press, Berkeley. Lovett Doust, J., and L. Lovett Dolst. 1988. Plant reproductive ecolog>'. O.xford Universit)' Press, New York, NY 1995] Astragalus Growth and Reproduction 123 MacArthur, R. H., and E. O. Wilson. 1967. The theow of island biogeography. Princeton University Press, Princeton, NJ. Manning, S. J., and M. G. Barboi r. 1988. Root systems, spatial patterns, and competition for soil moisture between desert subshrubs. American Journal of Botany 75: 885-893. Marchand, D. E. 1973. Edaphic control of plant distribu- tion in the White Mountains, eastern California. Ecolog>- 54: 233-250. . 1974. Chemical weathering, soil development, and geochemical fractionation in a part of the White Mountains, Mono and Inyo counties, California. uses Professional Paper 352-J. Marquis, R. J. 1991. Evolution of resistance in plants to herbivores. Evolutionar\' Trends in Plants 5: 23-29. MooNEV, H. A. 1966. Influence of soil t}'pe on the distribu- tion of Kvo closeK' related species oi Erigeron. Ecolog\' 47:950-958. MooNEY, H. A., G. St. Andre, and R. D. Wright. 1962. Alpine and subalpine vegetation patterns in the White Mountains of California. American Midland Naturalist 68: 257-273. Owen, W R. 1991. The reproductive ecology of an alpine legume: A. kentrophijfa var iinplexiis. Unpublished dissertation, Universit\' of California, Davis. 226 pp. Pace, N., D. W Kiepert, and E. M. Nissen. 1968. Clima- tological data summaiy for die Crooked Creek Labora- tory, 1949-1967, and the Barcroft Laboraton; 1953- 1967. University of California, White Mountain Research Station Publication, Berkeley. Parker, M. A., and A. G. Salzman. 1985. Herbivore exclosure and competitor removal: effects on juve- nile survivorship and growth in the shrub Giitierrezia microcephaUi. Journal of Ecology 73: 903-913. Philippi, T, and J. Seger. 1989. Hedging one's evolution- ary bets, revisited. Trends in Ecolog}- and Evolution 4:41-44. Robberecht, R., B. E. Mahall, and R S. Nobel. 1983. E.xperimental removal of intraspecific competitors — effects on water relations and productivity of a desert bunchgrass Hilaria rigida. Oecologia 60: 21-24. ScHOENER, T. W. 1983. Field experiments on interspecific competition. American Naturalist 122: 240-285. Shaver, G. R., and E S. Chapin. 1980. Response to fertil- ization by various plant growth forms in an Alaskan timdra: nutrient accumulation and growth. Ecology 61: 662-675. Sh.wv, R. G. 1987. Density dependence in Salvia Iijrata: experimental alterations of densities of established plants. Journal of Ecology 75: 1049-1063. SOKAL, R. R., and E J. Rohlk 1981. Biometry. W.H. Freeman and Company, New York, NY. Stearns, S. C. 1989. Trade-offs in life-histor\- evolution. Functional Ecolog>' 3: 259-268. Stephenson, A. G. 1984. The cost of over-initiating fruit. American Midland Naturalist 112: 379-386. Strong, D. R. 1986. Density vagueness: abiding the vari- ance in the demography of real popidations. Pages 257-268 in J. Diamond and T J. Case, editors, Communitv ecologv. Harper and Row Publishers, New York, NY TiLMAN, D. 1980. Resources, a graphical-mechanistic approach to competition and predation. American Naturalist 116: 362-393. . 1982. Resource competition and communit\' stnic- ture. Princeton University Press, Princeton, NJ. Weiner, J. 1988. Variation in the performance of individ- uals in plant populations. Pages 59-81 in A. J. Davey, M. J. Hutchings, and A. R. Watkinson, editors. Plant population ecolog\'. Black-well Scientific Publications, London. . 1990. Resource competition and commimit>- stmc- ture. Trends in Evolution and Ecology 5: 360-364. Welden, C. W, and W L. Slausen. 1986. The intensity of competition versus its importance: an overlooked distinction and some implications. Quarterly Review ofBiology 61: 23-44. Wright, R. D., and H. A. Mooney. 1965. Substrate-ori- ented distribution of bristlecone pine in the White Mountains of California. American Midland Naturalist 73; 257-284. Received 21 January 1994 Accepted 28 October 1994 Great Basin Naturalist 55(2), © 1995, pp. 124-134 CALILEUCTRA, A NEW GENUS, AND TWO NEW SPECIES OF STONEFLIES FROM CALIFORNIA (PLECOPTERA: LEUCTRIDAE) W. D. Shepard' and R. W. Baumann^ Abstract. — Calileuctra is proposed as a new genus in the family Leuctridae, with Calileuctra ephemera designated as the type species. All stages of Calileuctra ephemera are described. Calileuctra dohnji is described in the male and female stages. Both species inhabit the Mediterranean climatic region of California. A phylogenetic analysis of the gen- era in the family Leuctridae is given, which places Calileuctra near the genus Perlomijia. Key words. — Insecta, Plecoptera, Leuctridae, Calileuctra, description, distribution, plnjlogemj. Both of us have been collecting stoneflies from streams all across California. Several years ago one of us (WDS) collected a small and poorly sclerotized stonefly nymph from an inter- mittent Napa Valley stream. The male adult that was reared from the nymph could not be determined using existing keys by WDS. The specimen was then given to RWB for identifi- cation. His identification kept us collecting at the same site for nine years. The single male specimen was first thought to be a new species in the Asian genus Rhopalopsole. However, recent work indicates that the male represents a new genus in the family Leuctridae. Despite extensive searching in surrounding areas, only the Napa Valley population has been found. A few years after discovery of the first new species, RWB found, in the Natural History Museum of Los Angeles County, a small series of an interesting new leuctrid from the San Cabriel Mountains. Later, two additional fe- males of this species were collected in the Santa Ana Mountains. However, we decided that fresh male specimens were needed before a description could be undertaken. Keith Dobry, who was doing fieldwork in the Los Angeles area, was encouraged to look for additional specimens of this leuctrid species. He was successful in locating two additional populations, one in the San Cabriel Moun- tains, the other in the Santa Monica Mountains. This species is known from only four popula- tions, all from mountains surrounding the Los Angeles basin. Calileuctra, new genus Type species. — Calileuctra ephemera, new species Adults. — Body brownish, weakly sclero- tized; setation sparse, except for abundant tiny setae, "clothing hairs" (Figs. I, 10). Wings macropterous or brachypterous; venation as illustrated (Fig. 3). Prosternum with prester- num separate, furcasternum fused to base of triangular basisternum; meso- and metaster- num similar except basisternum rectangular (Fig. 2). Male. — Tergum IX with posterior border heavily sclerotized and irregularly serrate or dentate; tergum X with posterolateral corners, each with one or two elongate horns project- ing posteriorly (Figs. 4, 11); sternum IX pro- jecting posteriorly to cover base of paraprocts, with vesicle broadening posteriorly (Figs. 6, 13); paraprocts fused into a complex, T- shaped, subanal probe, with two ventromedial projections off subanal probe (Figs. 8, 9, 13). Female. — With weak abdominal scleroti- zation; sternum VII completely sclerotized; sternum VIII largely membranous; sternum IX completely sclerotized; subgenital plate poorly produced; sternum X incompletely sclerotized (Figs. 7, 14). Cerci one-segmented; elongate in male, poorly sclerotized on sides. ' Depaitnu-iit of Entomology, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94nf). Mailing address: 6824 Linda Sue Way, Fair Oaks. CA 9.5628. -Department of Zoology, Monte L. Bean Life Science Musenni, Brigliam Young University, Provo, UT 84602. 124 1995] Calileuctra New Genus 125 apically flat and membranous (Figs. 4, 11); simple in female (Figs. 7, 14). Nymph. — Mature nymph weakly sclero- tized; body elongate; setation scarce. Abdominal segments I-VII with membranous pleural fold. Mesosternal Y-ridge with double stem; arms meeting furcal pits at posterior ends. Paraprocts fused basally with no visible suture; sparse setation. Cereal segments each with apical fringe of 10-15 setae; setae approx- imately one-half length of cereal segments. Distribution. — Napa Valley and Los Angeles basin, CA. Diagnosis. — Males are best characterized by their unique elongate, flat-topped cerci. Females are characterized by sternum VIII being incomplete, and the lack of a posteriorly projecting subgenital plate. Nymphs are char- acterized by abdominal segments I-VII hav- ing a membranous pleural fold, the subanal lobes having basal fusion but no distinct fusion line, and the cereal segment setae being one- half length of the cereal segment. Eri'MOLOGY. — The prefix Cali- was select- ed to denote California, the origin of the spec- imens. The suffix -leuctra was selected to indi- cate placement of the genus in the family Leuctridae. Gender of the name is neuter. Key modification. — Modifications are given for the following identification keys for Nearctic leuctrid genera: Hai-per and Stewart (1984) — nymphal and adult keys; Stewart and Stark (1988) — nymphal key. Wording, style, and figure citations are as presented in the original keys. Haiper and Stewart (1984) — Nymphal Key 38 (37) Body robust, length less than 8 times width; body conspicuously clothed with hairs about one-fifth the length of middle Ab segment; subanal lobes of mature male a fused, strongly keeled plate, much pro- duced with no posterior notch (fig. 13.44) Megaleiictra 38' Body more elongate, Hue hair pile incon- spicuous, appearing naked; subanal lobes of mature male fused one-half to two- thirds length, leaving a notch at tip 38A 38A (38) Subanal lobes fused but with complete suture; apical setae on cereal segments usually less than one-half length of cer- eal segments Peiiomijia 38A' Subanal lobes basally fused, no suture in basal half; apical setae on cereal segments one-half length of cereal segments Calileuctra Harper and Stewart (1984)— Adult Key 58 (56) In hind wing, Cuj not forked . . . Calileuctra 58' In hind wing, Cuj forked 58A 58A (58') In hindwing, m-cu joining Cuj beyond fork of Cui ' 59 58A' In hindwing, m-cu joining Cuj before fork of Cui . 60 Stewart and Stark (1988)— Nymphal Key 5. Pronotum with no long, marginal setae (Fig. 8.14A); paraprocts of both sexes fused basally, with no distinct medial line of separation 5A Pronotum with 2-4 long hairs on anterior and posterior margins (Fig. 8.2A, 8.12A); paraprocts of both sexes fused with distinct medial line of separation or slightly separated medially (Fig. 8.2H,I; 8.12H,I) 6 5A. Abdominal segments 1-6 divided by ventro- lateral membrane; ENA and SW Zealeucfra 5A' Abdominal segments 1-7 divided by ventro- lateral membrane; WNA Calileuctra Key to Adults o( Calileuctra la Wings macropterous; male with epiproct bifurcate dorsally, tergite IX posteriorly emarginate with two large heavily sclerotized teeth, cerci with api- cal tooth, tergite X with one tooth on each postero- lateral corner, sternite IX with posterior projection broadly rounded; female sternite VII broad with a posteriorly projecting rectangular lobe, sternite VIII membranous C. dobryi lb Wings brachypterous; male with epiproct with one dorsal hook, tergite IX with single sclerotized pos- terior plate bearing numerous teeth, cerci without apical tooth, tergite X with two teeth on each pos- terolateral corner, sternite IX with posterior pro- jection broadly angulate; female sternite VII elon- gate with lateral constrictions, sternite VIII with two arcuate sclerotized plates C. ephemera Calileuctra ephemera, new species Figs. 1-9 Male. — General color brown; dark brown pattern as illustrated (Fig. 1). Length of body 4.5 mm. Brachypterous, length of forewing 2.5-3.0 mm; wings light brown, venation simi- lar to the genus Perlomyia (Fig. 3). Frothoracic basisternum triangular in shape (Fig. 2). Abdominal tergum IX with posterior border complete, projecting and serrate; tergum X 126 Great Basin Naturalist [Volume 55 ->^ "imtm-^-r-^ 'm^^jjp Fig. 1. Calileuctra ephemera. Habitus. incomplete medially, posterolaterally with two elongate projections (Figs. 4, 5). Sternum IX with basal pear-shaped vesicle; posterior bor- der extending to base of subanal probe (Fig. 6). Cerci extending beyond genitalia, with api- cal membranous area expanded and flattened (Figs. 4, 5, 6). Epiproct small and hook-shaped (Fig. 5). Subanal probe large, elongate, both membranous and sclerotized, expanded near apex (Figs. 8, 9). Female. — General color and wing vena- tion similar to male. Brachypterous, length of forewing 3.5-4.0 mm. Sternum VII constrict- ed laterally, projecting slightly over sternum VIII; sternum VIII reduced to 2 small arcuate sclerotized plates (Fig. 7). 1995] Calileuctra New Genus 127 Figs. 2, 3. Calileuctra ephemera: 2, ventral view of thorax; 3, wings. Nymph. — Body lightly sclerotized; light in color; setation sparse except on labrum, legs, and cerci; size small — 7.2 mm long. Head slighth' broader than thoriix; color pattern faint. Mouthparts of the herbivorous/detritivorous type [Type I (Stewart and Stark 1988)]. Labrum and clypeus with numerous long setae. Man- dible typical for Leuctridae: 4 dorsal cusps (2 large, 2 small), and 1 small ventral cusp on side of first large dorsal cusp; bowl-shaped molar region, with transverse ridges in the "bowl," and with a pectinate scraping ridge on the ventromedial edge. Maxillary palpi 5-seg- mented. Labial palpi 3-segmented; glossae and paraglossae short and subequal in size (paraglossae slightly larger). Pronotum quad- rangular; transverse anterior and posterior sclerotized bands; median longitudinal suture unsclerotized; color pattern weak. Mesonotum with two sclerites; anterior sclerite transverse and roughly trapezoidal; posterior sclerite roughly U-shaped. Metanotum like mesono- tum. Wing pads three or more times as long as wide; posterior wing pads a little shorter than anterior wing pads; longitudinal axes of wing pads almost parallel to axis of body. Prostenium naked and membranous except for two small sclerites between the coxae; sclerites forming a posteriorly directed U-shape. Mesosternum narrowest anteriorly, widest by coxae; mem- branous except for weak sclerotization of the furcal pits and the Y-ridge; Y-ridge with faintly sclerotized double stem, arms connecting to posterior ends of furcal pits; transverse ridge connecting anterior ends of furcal pits. Metastemum similar to mesosternum; sclero- tization only in a transversely rectangular area limited by the furcal pits, a transverse ridge connecting the anterior ends of the furcal pits, and the area between the pits and the ridge. All legs similar but increasing in size posteri- orly; setation consists of abundant ver>^ small setae ("clothing hairs") and sparse longer setae; tibiae and femora with setal fringes; apex of tibiae with a pair of spines; tarsi 3-seg- mented, first segment short and conical, sec- ond veiy short, ringlike with apex cleft, third elongate and cylindrical, suture between first and second tarsomeres very narrow and hard to see; tarsomeres with ventral pad of numer- ous fine setae; tarsal claws slender. Abdominal terga very weakly sclerotized; setation sparse except on end of tenth segment. Abdominal fold present on segments I-VIL Subanal lobes incompletely fused. Cereal segments with api- cal fringe of 10-15 setae; setae about one-half length of the segments. 128 Great Basin Naturalist [Volume 55 M!>%M.W'§ Figs. 4-7. Calileuctra ephemera: 4, male terminalia, dorsal view; 5, male terminalia, lateral view; 6, male terminalia, ventral view; 7, female terminalia, ventral view. 1995] Calileuctra New Genus 129 Figs. 8, 9. Calileuctra ephemera: 8, male subanal probe, right lateral view; 9, male subanal probe, ventral view. Egg. — Shape oval; size uniform, 0.133 mm in length, 0.095 mm in width. Surface coarsely rugose; large, coarse punctures present in an irregular distribution. Type locality. — California: Napa Co., 3.36 km (2.1 mi) N on Hw\' 128 from the inter- section of Hwy 128 and Hwy 121, unnamed tributary to Capell Creek (ca 300 m [275 ft] elevation). This intermittent stream has water present only a few months each year; some years there is no water (i.e., 1987 and 1990). When water is present, it flows down a small, steep canyon, across a grassy flat, under Hwy 128, and down a short cliff into Capell Creek. The stream course appears to be less than 350 m (1000 ft) long. All specimens have been collected in the grassy flat or just downstream. The stream course has a substrate of either bedrock or rocks on a clay soil. There is no obvious hyporheic zone. Detrital input is usually leaves from trees (mainly live oak), grass, and star-thistle. Type specimens. — Holotype: male, type locality, 19 II 1983, WDS-A-160, reared from nymph. To be deposited in the entomology collection at the California Academy of Sciences, San Francisco, CA. Allotype: female, type locality, 25 II 1984, WDS-A-240. To be deposited with the holotype. Paratypes: 1 male, type locality, 18 II 1984, WDS-A-234, reared from nymph (deposited at Monte L. Bean Life Science Museum, Brigham Young University, Provo, UT); 3 females, type locality, 18 II 1984, WDS-A-234 (deposited with male paratype). Additional specimen: 1 nymph, type locality, 27 II 1988, WDS-A-527 (deposit- ed with the holotype and allotype). Etymology. — The trivial name was select- ed to indicate the temporary nature of the stream at the type locality, and also to indicate the difficulty encountered when trying to col- lect specimens. Biology. — All specimens were collected during the last two weeks of February, when the stream was flowing. Streams in this area of the coastal mountains experience a Mediter- ranean climate with a December-to-Februaiy rainy season. Local intennittent streams usually have surface flow only from January through March. When first collected, all specimens were either late-instar nymphs (3) or adults (4). All field-collected adults (4 females) were swept from vegetation overhanging the stream. Two of the three nymphs collected were held in styrofoam containers until they molted to the adult stage. Both individuals were males. The bowl- shaped molar region of the man- dibles is similar to molar modifications found in beetle lawae that feed on fungal tissues, in general, and fungal spores, in particular (Lawrence 1977, Lawrence and Hlavac 1979, Lawrence and Newton 1980). Since fungal 130 Great Basin Naturalist [V'olume 55 tissues are high in protein (Martin 1987), use of them as a food would aid the fast growth and development of nymphs. Present information suggests that Calileuc- tra ephemera has a faeultati\ely long egg cha- pause, very fast nymplial deveh)pment, and short stadium for liotli nymplis and adults. The high protein content of fungal tissues (Martin 1987) may aid in the fast growth and development of nymphs of this unique species. These characteristics are similar to those of Zealeiictra (Snellen and Stewart 1979), an eastern North American genus and another inhabitant of intermittent streams. Calileuctra dohryi, new species Fi^js. 10-14 Male. — General color brown; dark pattern as illustrated (Fig. 10). Length of body 4.0-6.0 mm. Macropterous, length of forewing 4.5-5.5 mm, wings light brown, venation similar to the genus Perlomijia. Tergum IX with membra- nous median band dividing tergum into two sclerotized halves, each half bearing a small, nipplelike projection and a large, earlike pos- terior projection. Tergum X also divided into two halves, each half with a gently rounded, knoblike lobe and an enlarged, lateral posteri- or lobe which ends in a sclerotized prong that extends about one-third the way up the cercus (Figs. 11, 12). Sternum IX broadly rounded posteriorly, extending only to base of subanal probe, large vesicle present at median anterior margin, vesicle with truncate apex (Fig. 13). Cerci enlarged, elongate, extending beyond genitalia posteriorly, sclerotized on lateral margins, apex rounded, ending in a sclerotized lateral prong (Figs. 11, 12, 13). Epiproct with narrow bifurcate apex (Fig. 11). Subanal probe large, elongate, broadest medially, apex pointed (Fig. 13). Female. — General color and wing vena- tion similar to male. Length of body 5.0-6.0 mm. Macropterous, length of forewing 5.0-6.0 mm. Abdominal sternum VII enlarged, expanded slightly over VIII; posteromedial area formed into a narrow, medially roiuided lobe. Sternum VIII small and only lightly scle- rotized (Fig. 14). Nymph. — Unknown. Egg. — Unknown. Type locality. — California: Los Angeles Co., South Fork Elsmere Canyon, San Gabriel Mountains. The type locality is a very small headwater tributaiy of Elsmere Creek. It has an extremely steep gradient and is hard to access. Thus, the habitat has been presented more than the surrounding drainage area. Type specimens. — Holotype: male, tvpe locality, 22 IV 1991, K. F Dobry. To' be deposited in the entomology collection at the California Academy of Sciences, San Francisco, CA. ALLOTi'PE: female, same data as holotype. To be deposited with holotype. Paratypes: 1 male and 1 female, same data as holotype; 2 males and 1 female, CA: Los Angeles Co., Santa Monica Mountains, East Fork Arroyo Sequit, 5 mi NW Pacific Coast Highway off Mulholland Highway 28 II 1992, K. F Dobiy; 2 females, CA: Orange Co., Santa Ana Moun- tains, Trabuco Canyon, 1300', 11 1 1988, R. W. Baumann, B. J. Sargent, B. C. Kondratieff, and C. R. Nelson; 3 males and 1 female, CA: Los Angeles Co., San Gabriel Canyon, 23 IV 1960, D. Gibbo (LACM). Remaining paratypes to be deposited at Monte L. Bean Life Science Museum, Brigham Young University, Provo, UT Etymology. — The trivial name honors Keith E Dobiy, Los Angeles, CA, who collect- ed many of the specimens. Biology. — Specimens were collected as adults between January and April. All popula- tions are from the Los Angeles basin and experience a Mediterranean climate. Phylogeny Leuctrid phylogeny has been examined from a cladistic point of view only two times. The two studies (Ricker and Ross 1969, Nelson and Hanson 1973) are somewhat contradictoiy. However, examination of the analysis given in both studies shows Calileuctra to possess many character states that are termed primitive or ancestral. Following Nelson and Hanson's more comprehensive analysis, the character states present in Calileuctra are as follows: 1-0, 2-0, 3-0, 4-0, 5-0, 6-0, 7-1, 8-0, 9-0, 10-2, 11-0, 12-0, 13-0, 14-2, 15-0, 16-0, 17-0, 18-0,19-0, 20-0, 21-1, 22-0, 23-0, 24-1, 25-1, 26-1, 27-1, 28-1, 29-0, 30-1, and 31-2 (first number = character; second number = character state). See Nelson and Hanson (1973) for a key to the characters and character states. Character states for Calileuctra and those cited in Nelson and Hanson (1973) for other leuctrid genera were run through the PAUP 3.1.1 program 1995] Calileuctra New Genus 131 Fig. 10. Calileuctra dobryi. Habitus. 132 Great Basin Naturalist [Volume 55 Figs. 11-14. Calileuctra dohnji: 11, male terminalia, dorsal view; 12, male terminalia, lateral view; 13, male terminalia, ventral view; 15, female terminalia, ventral view. 1995] Calileuctra New Genus 133 15 Ancestor Moselia Leuctra Pachyleuctra Despaxia Paraleuctra Zealeuctra Rhopalopsole Perlomyia Calileuctra Tyrrhenoleuctra Megaleuctra Fig. 15. Phylogeny of the Leuctridae. using the branch and bound algorithm with all characters unordered. This analysis found one minimum-length tree (Fig. 15) with a length of 68, a consistency index of 0.82, and a retention index of 0.80. This new tree is not considerably different from that given by Nelson and Hanson (1973). It differs only in the collapse of the sister group relationship between Rhopalopsole and Zealeuctra and the exclusion of Euleiictra and "Leuctra" divisa from consideration. The sta- bility of diis tree with Calileuctra added is taken as evidence of the consistency of this data set and the overall stability of this new classifica- tion. It is heartening to find the cladogram of Nelson and Hanson (1973) stable despite the previous "extinction" (i.e., absence) o( Calileuc- tra. In tliis particular case, an "extinct" taxon did not particularly influence the overall topology of the cladogram. Hence, there is hope in our search for relationships among living taxa despite "known" extinction events. In this tree, Calileuctra is a the sister-taxon of the group containing Perlomyia, RJiopalop- sole, Zealeuctra, Paraleuctra, Despaxia, Pachy- leuctra, Leuctra, and Moselia. In leuctrid phy- logeny, Calileuctra occupies a near basal posi- tion and as such gives an important addition to our knowledge of the group. Acknowledgments Many thanks go to Charles H. Nelson, Uni- versity of Tennessee at Chattanooga, for run- ning phylogeny programs for us and for his many helpful comments. C. Riley Nelson, Uni- versity of Texas, Austin, provided a review and made valuable suggestions. Boris C. KondratiefF, Colorado State University, also offered many helpful suggestions as well as helped collect specimens. The late Charles L. Hogue kindly loaned specimens fiom the Los Angeles County Museum (LACM). Keith F. Dobry helped greatly in the collection of additional speci- mens. Jean A. Stanger made the many excel- lent illustrations. Literature Cited Harper, P. P., and K. W. Stewart. 1984. Chapter 13. Plecoptera. In: R. W. Menitt and K. W. Cummins, editors. An introduction to the aquatic insects of North America. Kendall/Hunt Publishing Co., Dubuque, lO. 722 pp. Lawrence, J. E 1977. The family Pterogeniidae, with notes on the phylogeny of the Heteromera. The Coleopterists' Bulletin 3i: 25-26. Lawrence, J. E, and T. E Hlavac. 1979. Review of the Derodontidae (Coleoptera: Polyphaga) with new species from North America and Chile. The Coleopterists' Bulletin 33: 369-414. Lawrence, J. E, and A. E Newton. 1980. Coleoptera associated with the fruiting bodies of slime molds (Myxomycetes). The Coleopterists' Bulletin 34: 129-143. Martin, M. M. 1987. Invertebrate-microbial interactions. Cornell University Press, Ithaca, NY. 148 pp. Nelson, C. H., and J. E Hanson. 1973. The genus Per- lomyia (Plecoptera: Leuctridae). Journal of the Kansas Entomological Societ>' 46: 187-199. RiCKER, W. E., and H. H. Ross. 1969. The genus Zealeuctra and its position in the family Leuctridae 134 Great Basin Naturalist [Volume 55 (Plecoptera: l^uctridaej. Canadian Jonnial ofZoology STEWART, K. W., AND B. P Stark. 1988. Nymphs of North 47: 1113-1127. American stonefly genera (Plecoptera). Thomas Say Snellen, R. K., and K. W. Stewart. 1979. The life cycle Foundation, Entomological Society of America 12: and drumming behavior of Zealeuctra claasseni 1-460. (Prison) and Zealeuctra hitei Ricker and Ross (Plecoptera: Leuctridae) in Te.xas, USA. Aquatic Received 27 Septe)nber 1994 Insects 1:6.5—89. Accepted 17 January 1995 Great Basin Naturalist 55(2), © 1995, pp. 135-141 CARBON ISOTOPE DISCRIMINATION IN THE C4 SHRUB ATRIPLEX CONFERTIFOLIA ALONG A SALINITY GRADIENT Darren R. Sandquistl and James R. Ehleringer^ Abstract — Carbon isotope discrimination (A) was measured for leaves ofAtriplex confertifolia along a salinity gradi- ent in northern Utah. Over this gradient, the variation of A values was high for a C4 species, and the A values were posi- tively correlated with salinity in both years of the study. Of the possible explanations for this pattern, the A results are consistent with tlie notion that salinity' induces an increase in the bundle sheath leakiness of tliese C4 plants. Key tvords: carbon isotope ratio, salt stress, bundle sheath leakiness. halophyte. desert ecology. The analysis of carbon isotope ratios (I'^C/l^C) has become a useful tool for under- standing various integrated aspects of plant metabolism, including numerous investiga- tions of plant-environment interactions. The impact of environmental factors on carbon iso- tope discrimination (A) by plants with C3 pho- tosynthesis has been well studied; however, only a limited number of studies have exam- ined variation of A values in C4 plants (O'Leary 1988, Farquhar et al. 1989, Peisker and Henderson 1992). In part, this disparity stems from C4 plants having much smaller variation of A values than C3 plants. Additionally, A values in C3 plants have been correlated with water-use efficiency, and this has lead to an emphasis on applying carbon isotope analyses to breeding programs (Farquhar et al. 1989, Ehleringer et al. 1993). However, a few recent studies have demon- strated that variation of A values in C4 plants may reflect environmental influences on phys- iological function (Bowman et al. 1989, Meinzer et al. 1994). In this study we exam- ined variation of A values in a C4 perennial shrub, Atriplex confertifolio (Torr. & Frem.) Wats., and its relationship to natural condi- tions of soil salinity. The A value of a C4 plant integrates two factors that can impact productivity: (1) the ratio of intercellular to ambient CO2 concen- tration (Cj/cJ, which can reduce photosynthe- tic activity when low, and (2) bundle sheath leakiness (0), which reduces photosynthetic efficiency when high. Farquhar (1983) mod- eled the relationship between these factors and carbon isotope discrimination in C4 plants as A = o + (Z?4 + Z?30 - a) c^/c.^^. (1) where a (4.4%c) is discrimination against the heavier ^'^C02 molecule relative to the lighter 1^C02 based on differential rates of diffusion, Z?3 (29%c) is the discrimination due to a greater affinity for 12CO2 relative to ^'^C02 by ribulose bisphosphate carboxylase (Rubisco), and b^ (typically = -5.2%c) is discrimination based on the steps leading to, and including, CO2 fixation by phosphoenol pyruvate car- boxylase (PEPC) after atmospheric CO2 enters the leaf The b^ term varies slightly as a function of temperature and is negative (greater proportion of l'^C02) due to fractiona- tion associated with the hydration of CO2 to HC03~ (Mook et al. 1974). The discrimination terms of Equation 1 [a, b^, and b_^) are con- stants, for the most part, and thus differences among A values are the result of changes in 0 and/or c^/c.^ during CO2 assimilation. In C4 plants, COg is'^initially fixed by PEPC in the mesophyll cells, transported and decar- boxylated in the bundle sheath cells, and then refixed by Rubisco. However, before the assimi- lation by Rubisco a fraction of the CO2 may diffuse out through apoplastic portions of the bundle sheath cells. This is known as "leaki- ness" and is thought to be reduced by suber- ization of bundle sheath surfaces (Farquhar 1983). This leakiness, however, may be in- creased by environmental stresses, such as salinity (Bowman et al. 1989), and an increase ' Department of Biology, UniversiU' of Utah, Salt Lake Cit\-, UT 84112 135 136 Great Basin Naturalist [Volume 55 in leakiness represents an energetic cost to the plant as a result of incomplete carbon assimi- lation or overcycling (Ehleringer and Pearcy 1983, Jenkins et al. 1989, Henderson et al. 1992). Leakiness affects A because it causes the bundle sheath cell to become an open system and therefore allows expression of discrimina- tion by Rubisco (^3). The proportion of CO2 that leaks out of the bundle sheath cell (0) modifies the degree to which b^ is expressed and thereby determines the relationship between A and c^/c.^ (Eq. 1). At low 0 values the relationship between A and c^/c.^ is nega- tive, at high 0 the relationship is positive, and at 0 = 0.32, A is constant at 4.4%o regardless of Ci/c.^. Equation 1 also predicts that for any given Cj/c^, an increase in 0 results in an increase in A. Given these relationships, varia- tion of A values in G4 plants can provide an indication of bundle sheath leakiness and its relationship to environmental stresses. To date, much work investigating variation of A in C4 plants has come from either labora- toiy gas exchange studies (Evans et al. 1986, Bowman et al. 1989, Henderson et al. 1992) or theoretical models (Peisker 1982, Farquhar 1983, Peisker and Henderson 1992). There is little direct information on environmental stresses that influence A in natural popula- tions of C4 plants (except see Walker and Sinclair 1992). Here we report on changes in A values for the G4 species Atriplex confertifo- lia found along a natural salinity gradient in Utah. The purpose of this study was to deter- mine if A values changed in relation to soil salinity under field conditions, and if these changes corresponded to variation in 0 values. Two previous laboratory studies have shown that higher soil salinity does increase A values in G4 plants and that this change is a result of greater 0 (Bowman et al. 1989, Meinzer et al. 1994). For A. confertifolia, we hypothesized that the same trend would be found over a transect of naturally increasing soil salinity. Methods Study Sites Four study sites of increasing salinity were chosen along a south-to-north transect in the northern end of Skull Valley (Tooele Gounty, UT) flanking the western slope of the Stans- buiy Mountain Range. The four sites range in elevation fiom 1366 m to 1286 m (Fig. 1). Site 1 (1366 m) is dominated by sagebiiish {Atiemisia tridentata) with low densities oi Atriplex con- fertifolia, Jiiniperus osteosperma, and Tetrady- mia spinosa. Weedy grasses and annual species of the Ghenopodiaceae are also found within disturbed areas of this and all other sites. Greasewood {Sarcohatus verrniculatus) is the dominant species at sites 2 (1317 m) and 3 (1294 m) with A. confertifi)lia co-occurring in low frequency. Site 4 (1286 m), along the mar- gins of the salt flats, is a heterogeneous site with a mixed community of salt-tolerant species. S. vermiculatits is the dominant species with moderate densities of Allenrolfea occidentalis, Atriplex gardneri, A. confertifolia, Chnjsothaninus viscidiflorus, Kochia ameri- cana, and Suaeda torreyana. Weather data for this transect are taken from the Grantsville weather station (Grants- ville, Tooele County, UT, 1307 m) located 17.3 km E and 8.2 km S from the center of our study transect. Leaf and Soil Samples Leaves of Atriplex confertifolia and soil samples were collected from each of the four transect sites in October 1991 and 1992, with the help of the 1991 and 1992 Plant Ecology classes from the Universit>' of Utah. Recently matured leaves of A. confertifolia were collect- ed from five to eight individuals per site in 1991 and three per site in 1992. Leaf samples were oven-dried (70 °G, 7 d), ground with mortar and pestle, and analyzed for carbon isotopic composition (Windy Ike, Delta S mass ratio spectrometer, Finnigan-MAT, San Jose, GA) relative to the Pee Dee Belemnite stan- dard. Analyses were done at the Stable Isotope Ratio Facility for Environmental Research (SIRFER, University of Utah, Salt Lake Gity, UT). Garbon isotope ratio values (6) were transformed to discrimination (A) values ^ = (Sa-5pMl + Sp) (2) where 5p is the measured carbon isotope ratio of the plant, and 5^ is the carbon isotope ratio of GO2 in the atmosphere (-.008 or -8%c; Farquhar et al. 1989). The standard per mil {%c) notation is used throughout for ease of presentation, and the overall, long-term error 1995] Carbon Isotope Discrimination in Atriplex 137 10 15 20 25 30 Transect distance from site 1 (km) 40 Fig. 1. Study transect in cross section. Shown is the topography over the transect and locahties of each study site based on the appro.ximate hnear distance from site 1. associated with carbon isotope determination is±0.11%o. Soil samples were collected from two depths (15-20 cm and 40-60 cm) in two to six excava- tion pits at each site. Approximately 200 g of freshly extracted soil from each hole and depth was placed immediately into soil canisters, sealed, and kept cool until analysis in the labo- ratory. In the lab one subsample per canister was removed for salinity' analyses. The remain- ing soil was used for gravimetric water content detennination based on the difference between soil fresh (wet) weight and dry weight (i.e., water content) relative to the soil dry weight. Soils were dried at 70°C for 7 d. In 1991 the soil salinity analysis was based on electrical conductivity (EC) of an aqueous solution extracted from a 1:2 soihdeionized water mixture, and in 1992 from a 1:5 soil: deionized water mixture. There was no evi- dence that the 1:2 mixture was ion saturated; thus, to standardize these ratios, the ECs of samples using a 1:2 solution were extrapolated to EC based on a 1:5 ratio assuming a linear dilution relationship. Tests confirmed that this extrapolation was valid even for EC values higher than those found in actual field samples. Although a more standard procedure for salinity determination is the "soil paste " method, the 1:5 ratio method we used is rec- ommended as a simpler technique to deter- mine relative salinity contents (Rhoades 1982) and is suitable for the purposes of this study (i.e., standardized comparison of relative salin- ities among sites). Additionally, the ECs of 1:5 ratio extracts are highly correlated with soil paste ECs for soils within and near our tran- sect (D. G. Williams unpublished data). Electrical conductivity is reported in ^mhos cm~l (1 jUmhos cm"^ = 0.1 mS m"^ = 0.502 mM NaCl), and the data were log transformed for statistical analyses. Interannual compar- isons of means for each soil trait were done by t tests, and correlations between soil trait and plant carbon isotope discrimination means were determined by Pearson product-moment correlation. Results Transect Characterization Salinity increased across the gradient in both the 1991 and 1992 samples; electrical conductivity increased by two orders of mag- nitude over the entire transect (Table 1). Site 1 was the least saline, and salinity progressively increased toward the highly saline site 4. There were few differences between years in soil electrical conductivity. Significant dif- ferences were found at only two sites and at only one depth per site. Furthermore, sites gave opposite results: soils of site 3 at the 15-20-cm depth had greater conductivity in 1991 than 1992 {t = 4.33, P < .01), and soils from site 1 at the 40-60-cm depth had higher conductivity in 1992 than in 1991 {t = 4.60, P < .01). Cravimetric water content also increased over the transect from site 1 to site 4 (Table 1). Soil water content was somewhat greater in 138 Great Basin Naturalist [Volume 55 Table 1. Soil propfrtie.s at two tleptli.s for site.s 1—4 along tlic .study transect (» = number of pits; one sample for each depth per pit). Soil water content was measured as gravimetric water content, and electrical conductivity is of an aque- ous extract from 1:5 soihwater mi.xture (extrapolated for 1991 from 1:2 ratio; see te.xt). Electrical El ectrical Soi 1 water Soi 1 water conductivity com ductivity coni tent (%) content (%) (/Ainhos/cm) (jun ihos/cm) @ 15-20 cm @40-60 cm @ 15-20 cm @40-60 cm Mean SE n Mean SE 11 Mean SE n Mean SE 11 October 1991 Sitel 4.66 0.300 4 5.34 0.234 4 89 15.7 4 70 3.2 4 Site 2 4.15 0.687 4 7.23 0.360 4 91 7.6 4 324 81.1 4 Site 3 11.79 1.446 4 17.24 0.892 4 2309 114.1 4 2066 657.7 4 Site 4 24.84 7.578 6 .39,41 7.841 6 3596 587.6 6 3382 530.7 6 October 1992 Sitel 2.89 0.454 2 3.81 0.402 3 84 6.3 3 93 3.7 3 Site 2 4.79 0.226 2 5.56 0.499 3 144 29.2 3 324 111.0 3 Site 3 2.46 0.270 2 10.26 3.672 3 546 459.3 3 984 858.5 3 Site 4 10.66 0.950 2 NA — — 1640 1440.0 2 3250 350.0 2 1991 than in 1992, but significant differences at both depths were found only at site 1 (15-20-cm depth, t = 3.34, P < .05; 40-60-cm depth, t = 3.52, P < .05). Rainfall over the 10- wk period prior to sampling in 1991 was much greater than that of 1992 (82.5 mm vs. 18.8 mm), which likely accounts for the trend of greater water content in the soils during the 1991 sample period. Carbon Isotope Discrimination Along the transect the carbon isotope dis- crimination for Atriplex confertifolia ranged from a low of 4.74 ± 0.96%^ at site 1 in 1992, to a high of 6.55 ± 0.1 l%c at site 3 in 1991 (Fig. 2). This range of nearly 2%c is high for C4 plants (Farquhar et al. 1989). The mean A value was always greater than 4.4%c, and for only a single sample was the individual shrub value less than 4.4%c. These high A values indicate that the mean 0 values were always greater than 0.32 (Eq. 1). With respect to the environmental parame- ters examined along the transect, mean leaf A was not significantly correlated with water content during any obsei-vation, but was posi- tively correlated with log EC (Fig. 2). Inclusion of the notably low A value of site 4 in 1991 resulted in a nonsignificant, positive trend (but when excluded, A was significantly correlated with log EC in 1991 at the deeper soil depth, R = 1.0, P < .01). In 1992 there was a highly significant, positive correlation ol A and log EC for both the shallow soils {R = .978, P < .05) and deeper soil depths (R = .999, P < .001) (Fig. 2). Discussion Variation in carbon isotope discrimination values of C4 plants is, in part, dependent upon the proportion of CO2 that is initially fixed by PEPC and ultimately diffuses out of the bun- dle sheath cells without being refixed (i.e., the leakiness, 0). Leakiness might be influenced by environmental stresses, such as salinity (Bowman et al. 1989, Meinzer et al. 1994), because such stresses could disrupt mem- brane properties or the biochemical coordina- tion between the C4 and C3 cycles operating in the mesophyll and bundle sheath cells, respectively (Peisker and Henderson 1992). The other component influencing variation of A in C4 plants is Cj/c^. Figure 3 illustrates how the relationship between A and c^/c^ depends upon the value of 0 (from Eq. 1), and provides a model for how changes in 0 and c^/c.^ can account for the changes in A values we observed. We found that A values of A. confertifolia increased by 2%c in concordance with increas- ing salinity (Fig 2). These A values were always greater than 4.4%o; therefore the 0 val- ues must be greater than 0.32 (cf Fig. 3). A 2%c increase in A values, at 0 > 0.32, cannot be explained solely by changes in c^/c.^^ given the typical range of c^/c.^^ values for C4 plants under ambient conditions (0.20-0.40; Pearcy and Ehleringer 1984). To do so would require either extreme leakiness values (0 > 0.6) or an increase of c\/c.^ with increasing salinity since A and Cj/c.j are positively related when 0 > 0.32. Leakiness values greater than 0.6 have 1995] Carbon Isotope Discrimination in Atriplex 139 O ■■E © o o X 6.5 Soil depth = 15 -20 cm T • i\ 1 If H -L + C- 5.5 " T • to -^ - o 4.5 3.5 - 1 Soil depth = 40-60 cm i-in 6.5 T -^ ^•l o ; 0—\ -L ^ -r • 5.5 - l-o^ -^ - o 4.5 ~ -a c 1 10 100 1000 10000 Log Soil Electrical Conductivity (^imhos cm"'') Fig. 2. Relationship between carbon isotope discrimina- tion (A) of Atriplex conferfifolia leaves and log electrical condnctivit>' (log EC) of soil at Kvo depths, 15-20 cm and 40-60 cm, for sites 1-4 along the transect. Closed symbols (•) are study site means for 1991, and open s\'mbols (O) are those for 1992. Error bars are ± ISE. never been reported, and the latter explana- tion is unlikely since salt stress typically decreases or does not change c^/c.^ (Long and Baker 1986, Flanagan and Jefferies 1988). A simpler explanation for the change in A values is that 0 increases with higher salinity. A 2%c increase based on changes in 0 values can be easily accommodated within the limits of Cj/c^ found for C4 plants (Fig. 3). Thus, changes in A values for A. conferfifolia are more likely due to an increase of 0 associated with a change in salinity; consequently, the presence of a significant relationship between A values and EC (Fig. 2). The trend of increasing A values with increasing salinity held in all but one site in the two-year study (site 4 in 1991). This devia- 9.0 8.0 7.0 6.0 < 4.0 3.0 2.0 1.0 0 ' //' -^' /^ 0 = 0.6/ / /• / / y -^ 0 = 0.5 y ^ /y ^ -^ ^^.-'--^^ / y^" ^ ^^^^ -""^ 0 = 0.4 / y /> ^^---^ , ■ yy /V ^^^-^ ^^>K^ ^„— — "^ ■ — — - J€^ --^ir^—" - ^^ ■ ^o^^>^ "--- — ____ ;:>^^ ~~~~-- -- 0 = 0.3' X \ ^^ ~^ ~~ " -^^^^ \\ "^ ^^"-~-^,^^^ . \ \ ■^ \0 = 0.2^ \ \ "^ 0 = 0.1 \ \ ^,.^ N 1 0.2 0.4 0.6 0.8 1 C/Ca Fig. 3. Model for the relationship between carbon iso- tope discrimination (A) and Cj/c^ (ratio of intercellular to ambient COo) based on Equation 1 and for 0 values rang- ing from 0.1 to 0.6. Dashed and solid lines represent the range of A values for each 0 value depicted, based on a high leaf temperature (34°C) where ^4 = -4.8%c (solid line) and a lower leaf temperature (25 °C) where h^ — -5.77cc (dashed line). tion could be due simply to the high degree of edaphic variability at site 4; this location had the greatest topographic variability, highest species diversity, and greatest overall variance for soil conductivity and water content (Table 1). Site 4 was also extremely wet in 1991 (near 40% water content at 40-60-cm depth), which may have diluted the salinity of these soils, thereby reducing the salinity experienced by the plants. Without a more detailed study, how- ever, this deviation remains unexplained. Previous studies have found contrasting patterns of the relationship between A and salinity. In a laboratory study with 11 C4 species, Henderson et al. (1992) found that 0 values were invariable and low, remaining at 0 ~0.21, thereby resulting in a negative rela- tionship between A and c^/Cg^ (Fig. 3). The small variation they observed in A values was attributed to changes in Cj/c^ values. However, in an earlier study with the C4 monocots Zea mays and Andropogon glomeratus. Bowman et al. (1989) found that A values of salt-stressed plants were more dramatically influenced by changes in c^/c.^ than were control plants. The increase of A values with salinity was ex- plained by a changing relationship between A 140 Great Basin Naturalist [Volume 55 and Cj/c.j due to increasing 0 values as the water status of salt-stressed plants declined through the day (Bowman et al. 1989). Recently, Mcinzer et al. (1994) also obseived that increasing salinit)' resulted in increases of A values. Using two sugarcane cultivars, they showed that change in A value could be ascribed to greater 0 values as salinity in- creased, and that variability of Cj/c., had much less impact on the increase of A values. In contrast, Walker and Sinclair (1992) reported that A values of two Australian C4 Atriplex species decreased at sites with increased salin- ity. The A values of these Australian Atriplex leaves were greater than 4.4%o, which could have been achieved only with a bundle sheath leakage greater than 0.32 (Fig. 3). Since the relationship between A and c^/c..^ is positive at 0 > 0.32 (Fig. 3), the Walker and Sinclair data suggest that salinity affected a decrease of Cj/c.^ and, therefore, a decrease of A. Our findings of a positive correlation be- tween A values oi Atriplex confertifolia and salinity are in contrast to findings of Walker and Sinclair (1992). Our observations, like those of Bowman et al. (1989) and Meinzer et al. (1994), suggest that changes in leaf carbon iso- tope discrimination result from an increased bundle sheath leakage when plants are exposed to a salinity stress. The mechanism of change in 0 values is likely to be associated with phys- ical changes in the bundle sheath permeability to CO2 (or to HC03~) and/or biochemical changes in the coupling of Rubisco and PEPC activity. Such biochemical changes due to salinity have been previously found. Guy and Reid (1986) have shown that salinity may reduce Rubisco activity in C3 plants without a concomitant decrease in PEPC activity. Increased salinity (NaCl) has also been shown to increase PEPC activity in some C4 halo- phytes (Shomer-Ilan et al. 1985). Any such increase in the activities of C4 carboxylation enzymes relative to those of C3 carboxylation enzymes in C4 plants should increase 0 values (Peisker and Henderson 1992). Thus, under natural conditions it appears that salinity could increase A values of A. confertifolia by influencing an increase in 0 values. The relationship between salt stress and 0 of C4 plants may be species specific or even population specific and may account for dis- crepancies among different studies of A values in C4 plants. For example, there is high vari- ability among pre\'ious studies of carbon iso- tope discrimination in Atriplex confertifolia; mean A values range from 4.4%c (Marino et al. 1992) to 6.9%c (Troughton et al. 1974). Yet, each of these observations is consistent with the notion that 0 values exceed 0.32 and are therefore high compared to nonhalophytic C4 species (Henderson et al. 1992). In the present study we have shown that salinity may be one factor that significantly influences variation of A values in C4 plants, most likely through an effect on bundle sheath leakiness. While variation in A values of C4 plants may provide new insights into plant- salinity dynamics along environmental gradi- ents, results also suggest that caution is neces- saiy when using A values of C4 plants to inter- pret historical changes in atmospheric CO2 concentrations and ^'^C values, as has been proposed by Marino et al. (1992). Acknowledgments We thank University of Utah students in 1991 and 1992 Plant Ecology classes for assis- tance in sample collection, Craig Cook for assistance in carbon isotope analyses, and Dr. David Williams for salinity analyses compar- isons. Dr. Williams and two anonymous reviewers also provided helpful comments on a previous version of this manuscript. Literature Cited Bowman, W. D., K. T. Hubick, S. von Caemmerer, and G. D. Farquhar. 1989. Short-term changes in leaf carbon isotope discrimination in salt- and water- stressed C4 grasses. Plant Physiologv 90: 162-166. Ehleringer, J., and R. W. Pearcy. 1983. Variation in quantum yield for CO2 uptake among C3 and C4 plants. Plant Physiolog\' 73; 555-559. Ehleringer, J. R., A. E. Hall, and G. D. Farquhar. 1993. Stable isotopes and plant carbon-water rela- tions. Academic Press, San Diego, CA. 555 pp. Evans, J. R., T. D. Sharkey, J. A. Berry, and G. D. Farquhar. 1986. Carbon isotope discrimination measured concurrentK' with gas exchange to investi- gate CO2 diffusion in leaves of higher plants. Australian Journal of Plant Physiology 13: 281-292. Farquhar, G. D. 1983. On the nature of carbon isotope discrimination in C4 species. Australian Journal of Plant Physiology 10: 205-226. Farquhar, G. D., J. R. Ehleringer, and K. T. Hubick. 1989. Carbon isotope discrimination and photosyn- thesis. Annual Review of Plant Physiology and Molecular Biology 40: 503-537. Flanagan, L. B., and R. L. Jefferies. 1988. Stomatal limitation of photos>'nthesis and reduced growth of 1995] Carbon Isotope Discrimination in Atriplex 141 the halophyte, Phmtiifio moritima L., at higli salinitv. Plant, Cell and Environment 11: 239-245. Guy, R. D„ and D. M. Reid. 1986. Photosynthesis and the influence of COq enrichment on 5^-^C values in a C3 halophyte. Plant, Cell and Environment 9: 65-72. Henderson, S. A., S. von Cafmmerer, and G. D. Farquhar. 1992. Short-term measurements of car- bon isotope discrimination in several C4 species. Australian Journal of Plant Physiology 19: 263-285. Jenkins, C. L. D., R. T. Furbank, and M. D. Hatch. 1989. Mechanism of C4 photosynthesis. A model describing the inorganic carbon pool in bundle sheath cells. Plant Physiology 91: 1372-1381. Long, S. E, and N. R. Baker. 1986. Saline terrestrial environments. Pages 63-102 in N. R. Baker and S. E Long, editors. Photosynthesis in contrasting envi- ronments. Elsevier Scientific Publishers, New York, NY. Marino, B. D., M. B. McElroy, R. J. Salawitch, and W. G. Spaulding. 1992. Glacial-to-interglacial varia- tions in the carbon isotopic composition of atmos- pheric COo. Nature 357: 461-466. Meinzer, F C, Z. Plaut, and N. Z. Saliendra. 1994. Carbon isotope discrimination, gas exchange, and growth of sugarcane cultivars under salinity. Plant Physiology 104: 521-526. MooK, W. G., J. C. Bommerson, and W. H. Staverman. 1974. Carbon isotope fractionation between dis- solved bicarbonate and gaseous carbon dio.xide. Earth and Planetary Science Letters 22: 169-176. O'Leary, M. H. 1988. Carbon isotopes in photosynthesis. BioScience 38: 325-336. Pearcy, R. W. , and J. Ehleringer. 1984. Comparative ecophysiology of C3 and C4 plants. Plant, Cell and Environment 7: 1-13. Peisker, M. 1982. The effect of CO2 leakage from bundle sheath cells on carbon isotope discrimination in C4 plants. Photosynthetica 16: 53.3-541. Peisker, M., and S. A. Henderson. 1992. Carbon: terres- trial C4 plants. Plant, Cell and Environment 15: 987-1004. Rhoades, J. D. 1982. Soluble salts. Pages 167-179 in Methods of soil analysis, part 2. Chemical and micro- biological properties. ASA-SSSA, Madison, WL Shomer-Ilan, a., D. Moualem-Beno, and Y. Waisel. 1985. Effects of NaCl on the properties of phospho- enolpyiaivate carboxylase from Suaeda nwnoica and Chloris gcnjana. Physiologia Plantaruin 65: 72-78. Troughton. J. H., E V Wells, and H. A. Mooney. 1974. Photosynthetic mechanisms and paleoecology from carbon isotope ratios in ancient specimens of C4 and CAM plants. Science 185: 610-612. Walker, C. D., and R. Sinclair. 1992. Soil salinity is cor- related with a decline in I'^C discrimination in leaves oi Atriplex species. Australian Journal of Ecology 17: 83-88. Received 20 May 1994 Accepted 16 August 1994 Great Basin Naturalist 55(2), © 1995, pp. 142-150 DEMOGRAPHY OF ASTRAGALUS SCAPHOIDES AND EFFECTS OF HERBI\ ORY ON POPULATION GROWTH Peter Lesica^ Abstract. — Losses in feeinulit\ due to predispersal lierbivon' can lie large; however, the effects of this loss on long- term population viabilit\' ha\'e rarely been investigated. I conducted a demographic study of Astragalus scaphoides (Fabaceae), a long-lived perennial endemic to east central Idaho and adjacent Montana, b\- following mapped individu- als at two sites from 1986 to 1993. Astragalus scaphoides suffers losses of predispersal fecundity averaging nearly 50% from insect seed predatioii and inflorescence predation by insects and livestock. Cattle reduced fecundit) by 0-85%. Nonetheless, estimates from matri.x projection models indicate that both sample populations had positive growth in most years. Elasticity analyses revealed that population growth occurred in spite of relatively small contributions by recniitment compared to growth and survival of nonreproductive plants. Results suggest that populations of this long- lived perennial depend little on reproduction and recruitment for growth and can persist in association with seasonal- rotation livestock grazing. Key words: demography, Iierbivory, livestock grazing, predation, matrix projection models, elasticity analysis. Astragalus, rare plant. The importance of herbivory in determining plant population dynamics and composition of vegetation has long been debated (Ehrlich and Birch 1967, Slobodkin et al. 1967, Belsky 1986). A great deal of evidence suggests a neg- ative impact of herbivory on the host plant (Harper 1977, Crawley 1983, Dirzo 1984); however, researchers have recently presented evidence for positive interactions (McNaughton 1986, Paige and Whitham 1987). A plant s life history plays an important role in determining the effects of herbivory. Loss of reproductive output from seed predators can be disastrous for an annual or biennial but may have little effect on a long-lived perenni- al. Furthermore, effects of herbivory will depend on the age or stage (e.g., seeds, adults) at which it occurs (Dirzo 1984). Most studies have focused on the effects of herbivores on particular components of fitness over relatively short time spans. This is unfortunate because it is the long-term effect on population growth that determines the importance of herbivory to population viability. Few studies have inte- grated the effects of herbivory on population dynamics and growth (Harper 1977; but see Louda 1982, 1983). Predation, particularly by exotic species, has often been cited as a threat to endangered plant populations (Greig-Smith and Sagar 1981, Parsons and Browne 1982, Willoughby 1987, Norton 1991, Pavlik et al. 1993). Negative im- pacts of herbivores were shown, but a causal link to declining population size has rarely been demonstrated. Astragalus scaphoides (Jones) Rydb. is endemic to a small area of east central Idaho and adjacent Montana (Barneby 1964). It was formerly a candidate for listing as a threatened or endangered species by the U.S. Fish and Wildlife Service (Category 3C; USDI-FWS 1993) and is currently listed as sensitive in Idaho (Moseley and Groves 1990) and Montana (Lesica and Shelly 1991). Most populations of A. scaphoides occur on public lands subject to livestock grazing. High levels of inflorescence and seed predation have been observed in some populations (Lesica and Elliott 1987a). Here I report the results of an eight-year demographic study of A. scaphoides at two sites. The puipose of the study is to document levels of herbivory and to assess its importance to population growth using stage-based transi- tion matrix models and elasticity analysis (de Kroon et al. 1986, Caswell 1989).' ' Di\ ision ol Biolowcal Science, Universih,' of Mdiitana, Missoula, MT 59812, ;incl Conservation BiologN Research, 929 Locust, Missoula, MT 59802. 142 1995] Demography and Herbivory in Astragalus 143 Methods Species Studied Astragalus scaphoides is a caulescent peren- nial with a taproot surmounted by a branched caudex. Reproductive individuals are 20-50 cm high with a cluster of pinnately compound basal leaves and 3-10 leaves at intervals along the erect stem. The inflorescence is composed of 1—4 racemes arising from the axils of the upper leaves. Each raceme is composed of a naked peduncle, 5-15 cm long, surmounted by a tight cluster of 10-30 flowers that expands in fruit. Nonreproductive individuals generally have 1-4 basal leaves and may have a sterile stem less than 15 cm tall with 1-5 leaves. The branching caudices of reproduc- tive plants may bear up to four stems and more than a dozen racemes (Barneby 1964, Lesica unpublished data). Astragalus scaphoides generally flowers dur- ing the first three weeks of June. The most conspicuous form of herbivoiy of these plants is the removal of inflorescences during flower- ing. Inflorescence predation has two principal sources: insects and livestock. Ants (subfamily Formicinae) and moth larvae [Melacosoma spp., family Lasiocampidae) were observed remov- ing inflorescences at a site near Haynes Creek in Idaho. Peduncles below the flowers were girdled, and withered inflorescences were often still present near the base of the plant. Inflorescence predation by livestock also occurred but differed from insect predation in that peduncles were all removed at the same height, and severed inflorescences were not found below the plants. In either case the cluster of basal leaves was usually left intact. It was possible to assign primary responsibility for inflorescence herbivory at a site in a partic- ular year to either insects or ungulates based on the appearance of damaged plants and the presence or absence of droppings, hoof prints, or trampled vegetation. However, it was not possible to unambiguously assign each case of herbivory to one or the other source. Inflor- escence predation by insects was obsei^ved at both study sites in all years that inflorescences were produced, but ungulate predation was common only at Sheep Corral Gulch. Predispersal seed predation occurred at both sites in most years. Lai-vae were collected from developing legumes in 1986 and identi- fied as weevils, small beetles in the family Curculionidae. Weevil larvae feed on maturing seeds and leave the developing or mature legume by creating a small hole in the outer wall. Seed predation by weevil larvae was inferred from the presence of fecula and/or an exit hole in the legume. Study Sites The Sheep Corral Culch population occurs in southern Beaverhead County, MT, on a gen- tle south-facing slope at 1920'm (T8S R12W S16). Mean July and Januaiy temperatures at Diflon, 32 km NW and 275 m lower, are 19.0° and -6.6° C, respectively. Mean annual precipi- tation is 241 mm. Vegetation is dominated by Artemisia tridentata and Agropijron spicatum. Aster scopulorum and Phlox hoodii are com- mon forbs. Livestock were managed on a rest- rotation system by which grazing occurred in different seasons in most consecutive years. Evidence of heavy spring grazing by livestock was observed in 1989, 1990, and 1993. The Haynes Creek population is in central Lemhi County, ID, approximately 48 km W of Sheep Corral Gulch. It occurs on a moderate southeast-facing slope at 1555 m (T19N R23E S2). Mean July and January temperatures at Salmon, 24 km NW and 365 m lower, are 16.2° and -6.7° C, respectively. Mean annual precipitation is 252 mm. Vegetation is domi- nated by Artemisia tridentata, Agropijron spica- tum, and Bromus tectoruni. This site was not grazed by livestock before early July during the course of the study. Field Methods Two permanent monitoring transects were established at each of the study sites in early July 1986 following methods outlined in Lesica (1987). Transects were located subjectively to represent the populations and were read in early July because fruits were mature or near- ly so, but seed dispersal had not yet begun. At each site the transects were parallel to each other and the slope. Each transect consisted of 50 adjacent l-m^ mapping quadrats placed along the transect line. The position of each A. scaphoides plant encountered in the quadrats was mapped and classified for three traits: (1) size, (2) inflorescence production, and (3) fecundity using the following classification: 144 Great Basin Naturalist [Volume 55 (1) Size classes: D Dormant (no abovegroiuid parts ohserved) S Small nonreproclucti\'es (1-3 leaves) L Large nonreproductives (> 4 leaves) R Reproductive (2) Inflorescence production: A Inflorescence produced no Iruit P Inflorescence was removed due to predation I Inflorescence produced at least one mature fruit (3) FecunditN': total number of mature fruits When stems were removed below the point of inflorescence articulation, I made a consei-va- tive estimate of the number of inflorescences removed based on the size of the remaining plant. Evidence of livestock and native ungu- lates (e.g., droppings, hoof prints, trampled \ egetation) was noted along each transect and for the site as a whole. I found that some plants would go unde- tected for one to several years but reappear in subsequent years (Lesica and Steele 1994). These "dormant" plants may have produced small leaves that had senesced and disap- peared by early July; however, my observa- tions in May and June suggest that most of them produced no vegetation on the years in (luestion. The presence of dormant plants can be inferred by comparing transect maps from tlie full sequence of years. The proportion of dormant plants ranged from 1% to 23%, with a mean of 10% in 1987-1991. Plants have "'dis- appeared" for as many as five years before re- appearing. However, in 1986-1992 at the two sites, 71% of dormant plants reappeared after one year, and 88% reappeared after two years (Lesica and Steele 1994). As a result, ca 10% of the plants were undetected in the first and last years of the study, and ca 3% were imde- tected in the second and second from last years. Thus, I have chosen to eliminate the first and last years (1986, 1993) of the study from demographic analysis, recognizing that a small (ca 3%) error still remains in mortality and recruitment estimates in 1987 and 1992. On years when fruit production was ade- quate, I collected 50 randomly selected mature fruits from at least 25 plants. I opened the pods, counted intact seeds, and recorded evidence of insect predation. Data Analysis Stage-structined transition matrix projec- tion models summarize the way in which sur- vival, growth, and reproduction at various life- history stages interact to determine population growth (van Groenendael et al. 1988, Caswell 1989). Matrix projections assume fixed transi- tion probabilities between stages in a popula- tion through time (Lefkovitch 1965, Menges 1990). They also assume density-independent population growth and thus do not give an accurate projection of long-term population future. Nonetheless, they can be used to sum- marize short-term population dynamics (Caswell 1989). One-year transition probabili- ties were estimated as the number of plants in life-stage class i moving into class j over the course of one year divided by the number of plants in stage / at the beginning of the year. This method assumes that an individual's tran- sition depends only on its life-stage class at the beginning of the period and is indepen- dent of its transition the previous year. The equilibrium growth rate (A,) is the dominant eigenvalue of the transition matrix (Lefkovitch 1965, Caswell 1989). }i > 1.0 indicates popula- tion increase, while X < 1.0 indicates decrease. X integrates the effects of sui-vival, growth, and fecundity of the different life-his- toiy stages into a single parameter There are two ways in which a reproductive plant can undergo a transition: (1) the plant itself moves into a different class or stays the same and (2) the plant produces progeny in one or more classes. These two prol^abilities (Recniit, Repro) are presented separately in the matrices but must be added together to solve for X. Details on the construction and use of matrix popula- tion models can be found in Caswell (1989) and Menges (1990). X was calculated using RAMAS/stage (Person 1991). Elasticity measures the relative change in the value of X in response to changes in the value of a transition matrix element. Elasticity matrices allow comparison of the relative con- tributions of various life-history transitions to population growth and fitness (de Kroon et al. 1986). Elasticities sum to unity, and regions of the matrix may be summed to compare the im- portance of growth and sui-vival to recruitment 1995] Demography and Herbivory in Astragalus 145 (Caswell 1989). Elasticities for nonreproductive plants are sums from the small (S) and large (L) classes. Elasticities were calculated using RAMAS/stage (Ferson 1991). When the majority of seeds pass directly from production to germination in less than one year, seeds should not appear as a sepa- rate stage in matrix models (Caswell 1989, Silvertown et al. 1993). Seeds oi' Astragalus scaphoides germinate readily without stratifica- tion (Lesica and Elliott 1987b), suggesting that most seeds germinate the same year they are produced. Nonetheless, A. scaphoides may fonii a seed bank. Not including a seed bank in the matrix model may affect the value of X (Kalisz and McPeek 1992), especially when it is <1.0. However, it will have little effect on analyses based on elasticities (Silvertown et al. 1993). I calculated separate elasticities for reproduc- tive transitions and recruitment by dividing the reproductive -I- recruitment elasticities proportionately between the two components. Losses to predation were estimated from the number of inflorescences lost using the calculated means for seeds/fruit and fruits/ inflorescence. Cumulative fecundity losses were calculated by multiplying the propor- tions of inflorescences and seeds remaining after predation and subtracting from one. Results Population Growth The number of Astragalus scaphoides plants in the transects at both sites increased by about one-third between 1986 and 1993 (Fig. 1). Equilibrium population growth rate (k) was >1.0 at both sites over the course of the study and was >2.5 at Sheep Corral Gulch in 1988-89 and 1990-91. At no time during the study was X < 0.8 at either site (Table 1). Survivorship Between 40% and 50% of the Astragalus scaphoides plants observed at the start of the study in 1986 were still alive in 1993 (Fig. 2). Approximately 50% of the 1989 cohort (the first large cohort recruited during the study) survived for more than 3-4 years. Taken together these results indicate that A. scaphoides is a long-lived perennial, with ca 50% mortality occurring in the first 3-4 years, but a large proportion of plants living to be > 10 years. Predation Inflorescence predation attributable to ungu- lates was virtually absent from the Haynes Creek population. Droppings and hoof prints of cattle were the only signs of ungulates at Sheep Corral Gulch. Droppings occurred in 3-9% of the mapping quadrats during the study. Inflorescence predation by insects occurred at both sites in all years. A significant number of inflorescences were produced in six of eight years at Haynes Creek, and inflorescence predation accounted for fecundity losses of 14-50% over the course of the study (Fig. 3). Most of this herbivory was attributable to insect damage. At Sheep Corral Gulch reproductive plants were common in only four of eight years. Inflorescence predation resulted in fecundity losses of 19-90%, and the proportion of inflorescences lost to preda- tion was highest in 1989, 1990, and 1993, years in which predation was due mainly to livestock (Fig. 3). Seed predation occurred at both sites in nearly every year in which significant fruiting occurred (Fig. 3). Overall, loss of seeds to wee- vil predation ranged from 0 to 33% with a mean of 18%. Insect seed predation was gen- erally higher at Sheep Corral Gulch than at Haynes Creek (Fig. 3). Losses of fecundity due to the combined effects of inflorescence and seed predation were 19-90% in 1986-1993, with means of 250 225 200 00 D 175 Q. o 150 (U XI E 125 -z. 100 75 50 87 88 89 90 91 92 Year Fig. 1. Density of Astragalus scaphoides plants in the two sample populations, 1987-1992. 146 Great Basin Naturalist [Volume 55 Table 1. Stage-based transition matrices for Aslra^ahis scaphoides at two sites in 1987-1992. Four stages are recog- nized: dormant (D), small nonreproductive (S), large nonreproductive (L), and reproductive (R). The reproductive and recruitment (Re) columns must lie added together before solving for X, the dominant eigenvalue (see Methods). >;Lo„,-, r^^rval r:„lr.lT 1987-88 ' 1990-91 From From To D S L R Re To D S L R Re D .67 .18 .20 0 + 0 D .14 .06 0 0 +0 S .11 .55 .24 0 + 0 S .21 .23 .06 0 + 9.86 L .22 .06 .36 0 +0 L .50 .42 .26 .29 + 2.42 R 0 0 .03 1.0 + 0 R .14 .12 .57 1.0 + .14 X- 1.18 X = 2.69 1988-89 1991-92 From From To D S L R Re To D S L R Re D .23 .04 .02 0 + 0 D .70 .24 .21 .25 + 0 S .17 .27 .05 0 + 4.0 S .30 .27 .37 .33 + .20 L .43 .53 .45 0 + 7.0 L 0 0 .14 .22 + 0 R .17 .08 .43 1.00 + .25 R 0 0 0 0 +0 X = 2.51 X = 0.83 1989-90 From To D S L R Re D .80 .17 .14 .06 + 0 S .10 .73 .37 .22 + .91 L .10 .02 ..35 .56 + .03 R 0 X = 0.97 .02 .01 .16 + 0 Haynes Creek - 1987-88 1990-91 From From To D S L R Re To D S L R Re D .50 .03 .04 0 +0 D .21 .03 0 0 +0 S .40 .45 .04 0 + 5 S .21 .34 .03 0 + .95 L 0 .24 .37 0 + .16 L .50 .31 .21 .18 + .27 R .10 X = 1.88 .06 .52 .60 + .20 R .07 x= : 1.31 .19 .66 .64 + 0 1988-89 1991-92 From From To D S L R Re To D S L R Re D .57 .13 .04 0 +0 D .75 .03 .06 .05 + 0 S .14 .42 .24 .05 + 1.10 S 0 .44 .30 .15 + .30 L .14 .16 .28 .14+ .14 L .25 .08 .36 .53 + .03 R .14 X= 1.13 .04 .32 .67 + .05 R 0 X = ( J.83 0 .11 .20 +0 1989-90 From To D S L R Re D .64 .12 .05 0 +0 S .27 .38 .14 .04 + .23 L 0 .20 .48 .27 + .04 R .09 ( X = 0.97 3 .33 .58 + 0 1995] Demography and Herbivory in Astragalus 147 100 90 80 70 60 50 40 30 20 • Sheep Corral V Haynes Sheep Corral Gulch 56 87 88 89 90 91 92 93 Year Fig. 2. Depletion curves for the 1986 sample popula- tions of Astragalus scaphoides at the two study sites. 51% and 44% at Haynes Creek and Sheep Corral Gulch, respectively (Fig. 3). Elasticity Analysis Elasticity gives the proportional impor- tance of demographic transitions to population growth. Elasticity matrices for five years of transitions for the two study sites are given in Table 2. Elasticities were summed into four life-history transition categories: (1) recruit- ment and sui-vival and growth of (2) dormant, (3) nonreproductive, and (4) reproductive plants (Fig. 4). Growth and sui'vival of nonre- productives was consistently important at both sites, with mean elasticities of 42% and 36% at Haynes Creek and Sheep Corral Gulch, respectively. Survival of dormant plants was important in two years at Sheep Corral Gulch and one year at Haynes Creek, with mean elasticities of 19% and 29%. Survival of repro- ductives had mean elasticities of 23% and 20% for the two sites, and mean elasticities for recruitment were 16% and 17%. Discussion Loss of Astragalus scaphoides fecundity due to inflorescence and predispersal seed predation was high at both sites, ranging from ca 20% to >90%. Further losses in reproduc- tive output due to ants or rodents may have occurred following dispersal. Recixiitment was the least important stage transition in the life history of A. scaphoides during my study. 0.6 0.4 0.2 0.0 {//} Inflorescence P^ Seed ^H Combined /X , /x /x /x /x 89 91 Year 93 Haynes Creek 0.4 0.2 0.0 \//} Inflorescence ^ Seed ^H Combined Fig. 3. Proportion of Astragalus scaphoides reproduc- tive output lost to inflorescence predation, predispersal seed predation, and the combination of the two in those years when significant flowering occurred at two study sites. Numbers of inflorescences in samples are given above bars. accounting for an average of less than 17% of population growth at both sites. High levels of inflorescence and seed herbivoiy are undoubt- edly one of the main reasons for the low con- tribution of recruitment to X in this species. Nonetheless, both sample populations became larger during the study. Furthermore, popula- tion growth rate was >1.0 in four of five years at both sites and never <0.8. Growth and sur- vival of dormant and nonreproductive plants contributed >60% to population growth at both sites. These results suggest that popula- tions of A. scaphoides can persist and even grow larger in spite of heavy losses in repro- ductive output and low recruitment. Large reductions in fecundity due to herbi- vores have been documented for Astragahis species (Green and Palmbald 1975) as well as 148 Great Basin Naturalist [Volume 55 Table 2. Elasticities for Astrufialus scaphukh's stage transition matrices at two sites for 1987-1992. The left three columns (D, S, L) represent nonreproductive growth and survival. The reproductive (R) column represents growth and survival of reproductives. The recniitment column (Re) represents recruitment from seed. . - - ^nf**^r\ \ cwv'^' C^ \\\t^'\-t _ 1987 -88 ■ - - kjiiccrp \_>uii 1990-91 D S L R Re D S L R Re D .048 .022 ,015 0 0 D .001 .005 0 0 0 S .004 .032 .009 0 .024 S .001 .018 .002 0 .185 L .0.32 .015 .055 0 .077 L .002 .082 .023 .014 .118 R 0 0 1988 .101 -89 .568 0 R .003 .099 .215 1991-92 .204 .029 D S L R Re D S L R Re D .001 .003 .001 0 0 D .686 .126 0 0 0 S .001 .017 .002 0 .139 S .126 .061 0 0 0 L .002 .081 .049 0 .143 L 0 0 0 0 0 R .003 D .057 1989 S .222 -90 L .224 R .056 Re R 0 0 0 0 0 D .413 .074 .015 .001 0 S .048 .292 .037 .002 .011 L .042 .007 .030 .005 .001 R 0 .017 .002 .004 0 — Haynes Creek - - 1987 -88 1990-91 D S L R Re D S L R Re D .002 .003 .002 0 0 D .001 .005 0 0 0 S .001 .045 .002 0 .141 S .001 .058 .004 0 .160 L 0 .082 .058 0 .153 L .004 .080 .038 .046 .068 R .003 .059 1988 .231 -89 .164 .055 R .001 .079 .193 1991-92 .264 0 D S L R Re D S L R Re D .055 .048 .006 0 0 D .617 .020 .041 .006 0 S .007 .086 .020 .005 .113 S 0 .042 .030 .003 .005 L .015 .064 .044 .028 .028 L .068 .018 .082 .021 .001 R .031 D .034 1989 S .109 -90 L .287 R .021 Re R 0 0 .036 .011 0 D .081 .028 .013 0 0 S .023 .061 .024 .007 .039 L 0 .065 .170 .093 .014 R .018 0 .134 .230 0 many other plant.s (Janzen 1971, Hendrix 1988, Louda 1989). Louda (1982, 1983) has shown that seed predation can lead to lowered recruitment; however, reductions in seed out- put will not necessarily lead to reduced recruitment if germination safe sites are limit- ing (Harper 1977). Analysis of the matrix pro- jection models suggests that recruitment is not limiting population growth of A. scaphoides. Recruitment from seed is likely to be im- portant to population growth for short-lived species and is essential for semelparous ones. Furthermore, successful reproductive episodes are rare for some perennial species in rigorous environments (Jordan and Nobel 1979). Signi- ficant reductions in a single reproductive bout could greatly increase chances of population extirpation for these sorts of species. On the other hand, many populations of long-lived plants will have more stable populations whose persistence is more dependent on the growth and survival of established plants (Silvertown et al. 1993). Survivorship curves indicate that Astragalus scaphoides is a long- lived species, and elasticity analysis suggests that recruitment is indeed less important to population persistence than growth and sur- vival of nonreproductive plants. 1995] Demography and Herbivory in Astragalus 149 Sheep Corral Gulch 0.8 >- 0.6 CD 0.4 0.2 0.0 ^B Dormant I I Non-reproductive )^\\| Reproductive lAAJ Recruitment imJ 89 90 Year 92 1.0 0.8 Haynes Creek ^B Dormant I I Non-reproductive - L\N Reproductive [XX] Recruitment >^ 0.6 C 0.4 - 0.2 - 0.0 ^ KL^ 90 Year 92 Fig. 4. Elasticities summed into four life-history transi- tion categories (recruitment and sur\'ival and growth of dormant, nonreproductive, and reproductive plants) for Astragalus scaphoides at two stud\' sites, 1987-1992. Inflorescence predation of Astragalus sca- phoides was greatest in years when livestock were present. In 1993 inflorescence predation was greater than 85%, and A. scaphoides was grazed in preference to the highly palatable grass, Agropyron spicatum (P Lesica personal observation). These observations suggest that livestock could nearly eliminate reproductive output under high stocking rates and repeated heavy spring grazing if carried on over a long enough period of time. However, results of my study suggest that A. scaphoides populations can persist if predation is moderate, at least in some years. Rotation grazing systems in which spring grazing occurs only one in three years appear to be compatible with the long-term persistence of A. scaphoides populations. These results have implications for other long-lived perennials exposed to livestock pre- dation. Upper portions of plants are most accessible to livestock, and newer growth is generally selected by livestock (Arnold and Dudzinski 1978, Vcilentine 1990). Furthemiore, sugars, such as found in flower nectar, also increase palatability (Arnold and Dudzinski 1978, Valentine 1990). Thus, livestock often remove only the upper portions of broad- leaved plants. Predation that mainly affects fecundity is likely to endanger populations only when grazing removes most inflores- cences consistently for many years because population growth is not likely to be limited by recruitment. On the other hand, grazing that lowers growth and survival (e.g., high- density stocking during periods of growth) will have a much more detrimental effect on popu- lation viability. Acknowledgments I am grateful to Joe Elliott, Anne Garde, and Lou Hagener for help in the field. James Liebherr of the Comstock Museum, Ithaca, NY, and Will Lanier of the Entomology Research Lab, Bozeman, MT, identified insects. Kimball Harper and an anonymous reviewer gave helpful comments on the manu- script. Funding was provided by the Idaho and Montana Bureau of Land Management and the Montana Natural Heritage Program. Literature Cited Arnold, G. W, and M. L. Dudzinski. 1978. Ethology of free-ranging domestic animals. Elsevier, Amsterdam. 198 pp. BEL.SKY, A. J. 1986. Does herbivory benefit plants? A review of the evidence. American Naturalist 127: 870-892. Barneby, R. C. 1964. Atlas of North American Astragalus, parts 1 and 2. Memoirs of the New Ybrk Botanical Garden 13: 1-1188. Caswell, H. 1989. Matri.x population models. Sinauer Associates, Sunderland, MA. 328 pp CR.WLEY, M. J. 1983. Herbivoiy, the dynamics of animal- plant interactions. University of California Press, Berkeley. 437 pp. DiRZO, R. 1984. Herbivory: a phytocentric viewpoint. Pages 141-165 in R. Dirzo and J. Saruklian, editors. Perspectives on plant population ecology. Sinauer and Associates, Sunderland, MA. Ehrlich, F R., and L. C. Birch. 1967. The "balance of nature" and "population control." American Naturalist 101: 97-108. Ferson, S. 1991. RAMAS/stage. Generalized stage-based modeling for population dynamics. Applied Bio- mathematics, Setauket, NY. 150 Great Basin Natl iulist [Volume 55 Green, T. W., and 1. C. Palmualu. 1975. En'fcts of insect seed predators on Astruflfilus ciharius and Astragalus utahensis (Leguminosae). Ecology 56; 1435-1440. CREic-Svirm, J., and G. R. Sacar. 1981. Biological causes of rarity in Carlina vulgaris. Pages 389-399 in H. Synge, editor, Biological aspects of rare plant conser- vation. John Wiley and Sons, Chichester, England. VAN Groenendael, J. M., H. DE Kroon, and H. Caswell. 1988. Projection matrices in population biology. Trends in Ecolog\' and Evolution 3: 264-269. Harper, J. L. 1977. Population biology of" plants. Aca- demic Press, London. 892 pp. IIendri.X, S. D. 1988. Herbivow and its impact on plant reproduction. Pages 246-263 in J. Lovett-Doust and L. Lovett-Doust, editors. Plant reproductive ecology. O.xford University Press, New York, NY. Janzen, D. H. 1971. Seed predation. Annual Review of Ecolog\' and Systematics 2: 465^92. Jordan, R W., and P S. Nobel. 1979. Infrequent estab- lishment of seedlings of Agave deserti (Agavaceae) in the northwestern Sonoran Desert. American [ounial ofBotany 66: 1079-1084. IC\Llsz, S., and M. A. McPeek. 1992. Demography of an age-structured annual: resampled projection matri- ces, elasticity analyses, and seed bank effects. Ecologn' 73: 1082-1093. DE Kroon, H., A. Plaiser, J. \1. van Groenendael, and H. Caswell. 1986. Elasticity: the relative contribu- tion of demographic parameters to population growth rate. Ecology 67: 1427-1431. Lefkovitch, L. P 1965. The study of population growth in organisms grouped by stage. Biometrics 21: 1-18. IjESICA, P 1987. A technique for monitoring nonrhizoma- tous, perennial plant species in permanent i)elt tran- sects. Natural Areas Journal 7: 65-68. Lesica, P, and J. C. Elliott. 1987a. Distribution, age stiTicture, and predation of Bittenoot milk^ etch pop- ulations in Lemhi County, Idaho. Report submitted to the Bureau of Land Management, Boise, ID. . 1987b. 1987 monitoring study of Bittenoot milk- vetch populations in Lemhi Count\, Idaho. Report submitted to the Bureau of Land Management, Boise, ID. Lesica, P, and J. S. Shelly. 1991. Sensitive, threatened and endangered vascular plants of Montana. Montana Natural Heritage Program Occasional Publication 1, Helena. 88 pp. Lesica, E, and B. M. Steele. 1994. Prolonged dormancy in vascular plants and implications for monitoring studies. Natural Areas Journal 14: 209-212. Louda, S. M. 1982. Distribution ecology: variation in plant recruitment over a gradient in relation to insect seed predation. Ecological Monographs 52: 25^1. . 1983. Seed predation and seedling mortality in the recruitment of a shrub, Haplupappus venetus (Asteraceae), along a climatic gradient. Ecologv 62: 511-521. . 1989. Predation in the cKnamics of seed regener- ation. Pages 25-51 in M. A. Leek, V. T Parker, and R. L. Simpson, editors, Ecolog\' of soil seed banks. Academic Press, New York, NY. McNalghton, S. J. 1986. On plants and herbivores. American Naturalist 128: 765-770. Menges, E. S. 1990. Population viability analysis for an endangered plant. Consei-vation Biology 4: 52-62. MosELEY, R., and C. Gro\ es. 1990. Rare, threatened and endangered plants and animals of Idaho. Idaho Natural Heritage Program, Boise. 33 pp. Norton, D. A. 1991. Trilcpidca udumsii: an obituary for a species. Conservation Biology 5: 52-57. P.\IGE, K. N., .\ND T. G. Whitham. i987. Overcompensation in response to mammalian herbivoiy: the advantage of being eaten. American Naturalist 129: 407-416. Parsons, R. E, and J. H. Browne. 1982. Causes of plant species rarity in semi-arid southern Australia. Bio- logical Consei"vation 24: 183-192. Pavlik, B. M., N. Ferguson, and M. Nelson. 1993. Assessing limitations on the growth of endangered plant populations, II. Seed production and seed bank d\namics of Erysimum capitafum ssp. angustatum and Oenothera deltoides ssp. howellii. Biological Conservation 65: 267-278. Silvertown, J., M. Franco, I. Pisanty; and A. Mendoza. 1993. Comparative plant demography — relative importance of life-cycle components to the finite rate of increase in woody and herbaceous perenni- als. Journal of Ecology 81: 465-476. Slobodkin, L. B., F E. Smith, and N. G. Hairston. 1967. Regulation in terrestrial ecosystems and tlie implied balance of nature. American Naturalist 101: 109-124. USDI-FisH AND Wildlife Sermce. 1993. Endangered and threatened wildlife and plants; review of plant ta\a for listing as endangered or threatened species; notice of review. Federal Register 58: 51144-51190. Valentine, J. F 1990. Grazing management. Academic Press, San Diego, CA. 533 pp. Willoughby, J. W 1987. Effects of livestock grazing on two rare plant species in the Red Hills, Tuolumne Count>', CiJifomia. Pages 199-208 in T S. Elias, editor. Conservation and management of rare and endan- gered plants. California Native Plant Society, Sacramento. Received 1 April 1994 Accepted 7 September 1994 Great Basin Naturalist 55(2), © 1995, pp. 151-157 LAHONTAN SAGEBRUSH {ARTEMISIA ARBUSCULA SSR LONGICAULIS): ANEWTAXON Alma H. Winwardl and E. Durant McArthur^ Abstract. — A new subspecies of Artemisia arbuscula is described, A. arbuscida ssp. lonp^icaiilis Winward & McArthur, ssp. nov. This ta.\on is a landscape dominant in portions of northwestern Nevada and adjacent California and Oregon at elevations from 1050 to 2000 m on shallow or argillic (clayey) soils. It differs from A. arbuscula ssp. arbuscula in its long floral stalks and large leaves. Moiphological, chemical, ecological, and cytological data suggest that it is of hybrid origin. It is he.xaploid (6x). We hypothesize that 2.x A. arbuscula ssp. arbuscula and 4x A. tridentata ssp. wijomin- gensis are its parents. Key words: Nevada, taxonomy, chemotaxonomy, allopolyploid, hybrid, Tridentatae. In preliminaiy repoi^ts we (Winward et al. 1986, 1991) provided a brief description of a new taxon of Ai~temisia found in northwestern Nevada and adjacent California and Oregon. We suggested referencing it by the common name Lahontan sagebrush pending a formal description. This paper provides that formal description and details concerning its ta.xono- my, distribution, general ecology, and origin. Taxonomy The new taxon is a member of subgenus Tridentatae of Artemisia, the true sagebrushes (Beetle 1960, McArthur et al. 1981, Shultz 1986). We suggested (Winward et al. 1986) that this taxon may have originated as a hybrid with parental lines consisting of low and big sage- brush (A. arbuscula and A. tridentata). Because of its general morphology and ecology, we sug- gested that it be considered a subspecies of A. arbuscula. Furtlier studies indicate that this new taxon is in fact best treated as a subspecies of A. arbuscula. The type specimen of A. arbuscula came from a collection along the arid plains of the Lewis (now known as the Snake) River (Nuttall 1841). Subsequent workers have submerged the species as a subspecies of big sagebrush, A. tridentata ssp. arbuscula (Hall and Clements 1923), or, in contrast, recognized a number of races and subspecies within the species (Ward 1953, Beetle 1960). Ward proposed that black sagebrush (A. nova) was best treated as a sub- species of A. arbuscida, but Beetle (1960) restored it to Nelson's (1900) original species status. Beetle (1960) recognized two sub- species of A. arbuscula, arbuscula and ther- mopola. His treatment has been generally accepted (Winward and Tisdale 1977, McArthur et al. 1981, Shultz 1986), although Winward (1980) has observed an unusual variant of A. arbuscula in eastern Oregon that reaches a height of 1 m. He suggested that further taxo- nomic treatment of A. arbuscula would be appropriate. During the past few decades fieldworkers in western Nevada have obsei^ved a sagebrush that does not fit the existing Artemisia taxo- nomic keys. Brunner (1972) termed this sage- brush "wide-lobe" with the comment, "Dr. Beetle feels this may be an ecotype of A. triden- tata ssp. wijomingensis. I concur." Others have referred to it as "wonder sagebrush," "junk sagebrush," or "N" sagebrush (Winward et al. 1986). Accessions of two populations (Trough Springs, Humboldt County, NV, cultures Ul and U58 and Leonard Creek, Humboldt County, NV, culture U55) of this taxon were established in common gardens of the Forest Service's Shrubland Biology and Restoration Research Work Unit at several locations around central Utah; there they were treated as an ecotv'iDe of A. tridentata ssp. wijomingensis following Beetle and Brunner (Brunner 1972; 'Range and Watershed Management, Intennountain Region, Forest Service, U.S. Department of Agriculture, Ogden, UT 84401, ^Shrub Sciences Laboraton-, Intermountain Research Station. Forest Senice, U.S. Department of Agriculture, Pro\'0, I'T 84606. 151 152 Great Basin Naturalist [Volume 55 e.g., McArthur and Pliimnier 1978, Welch and McArthur 1979, 1981, 1986, McArthur et al. 1981, 1985, McArduir and Welch 1982, Welch et al. 1986, 1987). The new taxon is a landscape dominant over much of its range (Winward et al. 1986), and both domestic and wild animals feed e.xtensivcK on it (Bmnner 1972, Welch and McArthur 1986, Winward et al. 1986, Welch et al. 1987). Description Artemisia arhuscula ssp. lonfiicaidis Win- ward & McArthur ssp. nov. Similis A. arhuscula ssp. arhuscula sed ramis floralibus nuilto lon- gioribus et foliis magnioribus differt (Similar to A. arhuscula ssp. arhuscula except flower stalks are nuich longer and leaves are larger). The longer flower stalks and larger leaves also differentiate ssp. longicaulis from ssp. thennopola, which differs from ssp. arhuscula and longicaulis by having deeply trifid leaves (Beetle 1959). We chose the common name Lahontan sage- brush because the old shorelines of Pleisto- cene Lake Lahontan are one of the centers of its current distribution and may have provided the ecological setting for the taxon's origin and development (Winward et al. 1986, 1991). Type: Toulon, Pershing County, Nevada, USA, 1053 m, S. C. Sanderson and E. D. McArthur 1593, 21 August 1986. Holotype: BRY. Isotypes: OGDF, RENO, SSLR and UTC. Other specimens examined: • Nevada, Douglas Co., Topaz Lake, Sanderson & McArthur 1594, (SSLP four sheets); • Nevada, Humboldt Co., Golconda, Plummers.n., 1985, (SSLP); • Nevada, Humboldt Co., Leonard Creek, Plummer & McArthur, s.n., 3 October 1975, culture U55, (SSLP); • Nevada, Humboldt Co., Trough Springs, Jackson Mountains, Plummer, Brunner, & McArthur, s.n., 3 October 1975, culture Ul, (SSLP); • Nevada, Humboldt Co., Trough Springs, Jackson Mountains, McArthur 1532, culture Ul, (SSLP); • Nevada, Humboldt Co., Trout Creek Basin, Jackson Mountains, McArthur 1501, (SSLR two sheets); • Nevada, Lyon Co., Dayton, Sanderson & McArthur 1595, (SSLI^ two sheets); • Nevada, Pershing Co., 6.4 km west of Toulon, McArthur & McArthur 1683, (SSLP two sheets); • Nevada, Washoe Co., Mustang, McArthur & McArthur 1684, (SSLP); • Oregon, Lake Co., 32 km east of Adell, Sanderson & McArthur 1590, (SSLP); • Oregon, Malheur Co., near McDermitt, Nevada, Winward, s.n. 31 October 1986, (OGDF; two sheets, SSLP). Distribution and Ecology Artemisia arhuscula ssp. longicaulis occurs on several hundred thousand hectares in northwestern Nevada and in adjacent areas of California and Oregon at elevations from about 1050 to 2000 m (Fig. 1). It often occurs in pure stands. It may also share dominance with other sagebrush taxa such as big sage- brush (A. tridentata ssp. tridentata and wyo- mingensis), low sagebrush (A. arhuscula ssp. arhuscula), and black sagebrush (A. nova). At lower elevations it is interspersed with salt desert shrub species such as shadscale {Atriplex confertifolia), Bailey greasewood {Sar- cohatus hadeyi). Mormon tea {Ephedra spp.), budsage {Artonisia spinescens), Shockleys desert thorn {Lycium shockleyi), and horse- brush {Tetradytnia spp.). Except for Artemisia, our taxonomy follows Welsh et al. (1993) and Mozingo (1987). The most common grass understoiy species at upper-elevation Lahontan sagebrush sites is bluebunch wheatgrass {Elymus spicatus). At lower elevations Thurber and desert needlegrasses {Stipa thurheriana and S. speciosa), and Indian ricegrass {Stipa hymenoides), bottlebrush squirreltail {Elymus elymoides), and Sandberg bluegrass {Poa secunda) are more common. Areas supporting A. arhuscula ssp. longicaulis receive between 175 and 350 mm of precipitation annually with most as wdnter precipitation. The frost-free season ranges from 90 to 110 days. Lahontan sagebrush grows most commonly on Aridisols, but at upper elevations it also occurs on MoUisols. Soil Conservation Service, U.S. De- paitment of Agricultin-e, personnel have located A. arhuscula ssp. longicaulis on at least 17 soil series. Generally, these soils have low available water-holding capacities and a shallow depth to an argillic horizon and/or bedrock. These soils are similar to those of low sagebrush (A. arhuscula ssp. arhuscula) communities 1995] Lahontan Sagebrush, A New Taxon 153 Susanville Austin NEVADA Fig 1. Extent of the known distribution o{ Artemisia arhuscula ssp. longicaulis. 154 Great Basin Naturalist [Volume 55 (Fosberg and Hironaka 1964, Zaniora and Tueller 1973, G. K. Brackley and C. A. Plumnier personal commnnication). General distributions of the three sub- species of A. arhuscula are as follows: ssp. arhuscula, western Wyoming and eastern Utah to eastern Washington and northeastern California; ssp. thennopola, western Wyoming and adjacent Idaho and northern Utah to northern Nevada and eastern Oregon; ssp. J()u12 inches DBH were defoliated. Stems exhibiting top-kill increased proportionately with per- cent defoliation. Four percent of subalpine fir stems over 5 inches DBH were killed by tus- sock moth. In the >12-inches diameter class, none of 7.3 Douglas-fir per acre were visibly defoliat- ed (Table I). Among subalpine fir in this class, 3% of 65.1 per acre were defoliator killed. Twenty-eight percent survived defoliation, while 69% were not visibly defoliated. Western balsam bark beede [Dryocoetes con- fiisiis Swaine) killed 4.9 subalpine fir stems per acre. These trees were attacked in 1991, coin- ciding with the peak of the tussock moth out- break. Table 1. Trees per acre condition sunimaiy of subalpine fir and Douglas-fir following a Douglas-fir tussock moth out- break. Blind Hollow, Wasatch-Cache National Forest, July 1993. Summary calculated from 10 variable/fi.xed plot pairs. SAF = subalpine fir, DF — Douglas-fir. Undamaged SAF DF Defoliation class Diameter Light Moderate SAF DF Heav)' VePi' SAF heav)- DF Mortality' class SAF DF SAF DF SAF DF 0-4.9" 5-8.9" 9-11.9" 12" -H 30.0 39.0 25.9 40.0 0.0 0.0 2.9 7.3 120.0 0.0 16.0 0.0 7.0 0.0 13.4 0.0 30.0 18.5 6.0 4.8 0.0 0.0 0.0 0.0 0.0 6.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.2 0.0 0.0 0.0 0.0 0.0 0.0 60.0 0.0 6.0 0.0 0.0 0.0 2.1 0.0 160 Great Basin Natuiulist [Volume 55 The 1992 survey found an average ol 3.1 pupae and 0.5 egg masses per three-branch samples. The 1993 survey found no current life stages on any sample tree, and no life stages were visible in the area. Baxter Sawmill Prior to the outbreak, composition for all live trees greater than 5 inches DBH was 65% subalpine fir, 25% aspen, and 10% Douglas-fir. Total live basal area was 176.1 sq ft/ac at the onset of the outbreak. Live basal area in 1993 was 112.8 sq ft/ac. Site elevations range from 7400 to 7900 ft. Aspect is south, southwest, west, and northwest on slopes vai^ving from 10 to 30%. Subalpine fir seedlings and saplings had considerable defoliator damage. More than 250 seedlings and saplings per acre, or 55% of stocking in this size class, died (Table 2). Most surviving seedlings and saplings were only lightly defoliated. Forty-nine percent of sub- alpine fir stems 5.0-11.9 inches DBH were killed by tussock moth. Trees with top-kill increased proportionately with percent defoli- ation. Only 3% of subalpine fir stems in the lightly defoliated category experienced top- kill, compared to 92% of surviving trees in the heavily and very heavily defoliated classes. In the >12-inches diameter class, Douglas- fir had 10% of 22.6 trees per acre defoliator killed. Fift\'-seven percent were not defoliated, with another 33% defoliated but surviving (Table 2). Among 38.5 subalpine fir per acre in this size class, 7% were defoliator killed and 77% were defoliated but sui-vived. Western balsam bark beetle has also been active at Baxter Sawmill, killing 38.2 subalpine fir per acre, mostly in 1990 or 1991. Annosus root disease {Heterobasidion anuosiiin [Fn] Bref ) was found on 4.6 subalpine fir per acre. The 1992 survey found an average of 4.8 pupae and 1.2 egg masses per three branches sampled. No cmrent life stages were found in 1993 on the plots or in the area. Additionally, no tussock moths were caught in pheromone traps placed in the Baxter Sawmill area in 1993. Amazon Hollow Prior to the outbreak, composition of all live trees greater than 5 inches DBH was 73% subalpine fir, 24% aspen, 2% Douglas-fir, and 1% lodgepole pine. Total live basal area was 125.5 sq ft/ac at the onset of the outbreak. Live basal area in 1993 was 72.2 sq ft/ac. Site elevations range from 7500 to 7800 ft. Aspect is east on slopes vaiying from 10 to 25%. One-hundred subalpine fir seedlings and saplings per acre, or 10% of stocking in that class, were killed (Table 3). Mortality in the three size classes greater than 5 inches DBH i-anged from 50 to 62%. Top-kill was common for all defoliation intensities. Of the sui'viving defoliated subalpine fir (>5 inches DBH), 60% had top-kill, including 63% of stems clas- sified as lightly defoliated. In the >12-inches size class, 28% of 4.3 Douglas-fir per acre were defoliator killed with another 16% defoliated but surviving (Table 3). Among 29.7 subalpine fir per acre in that class, 50% were defoliator killed and another 31% were defoliated but sui^vived. Western balsam bark beetle killed 2.6 sub- alpine fir per acre. Annosus root disease was found on 4.2 trees per acre. The 1992 sin-vey found 2.0 pupae and 0.6 egg masses per three branch samples. The 1993 sui'vey failed to detect any current life stages. Sample Tree Summaiy Two-hundred ninety-one host sample trees were rated for defoliation and monitored for Table 2. Trees per acre condition suniman' ot subalpine tir and Douglas-fir following a Douglas-fir tussock moth out- break, Baxter Sawmill, Wasatch-Cache National Forest, July 1993. Summaiy calculated from 13 \ariable/fixed plot pairs. SAF = subalpine fir, DF = Douglas-fir Undamaged SAF DF Defoli ation class Morta SAF Diameter Light SAF DF Moderate Me avy Veiy heavy SAF DF lit)' class SAF DF SAF DF DF 0-4.9" 23.1 0.0 13S.5 0.0 23.1 0.0 23.1 0.0 0.0 0.0 2,53.5 0.0 5-8.9" 12.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 29.3 0.0 9-11.9" 2.5 0.0 7.8 0.0 2.5 0.0 2.1 0.0 0.0 0.0 25.0 0.0 12" -h 3.4 12.S 17.9 6,5 2.7 1.0 0.0 0.0 1.3 0.0 2.7 2.2 1995] Tussock Moth on Subalpine Fir 161 Table 3. Trees per acre condition suninian of subalpine tir and Douglas-fir following a Douglas-fir tussock moth out- break, Amazon Hollow, Wasatch-Cache National Forest, July 1993. Summai-y calculated from 12 variable/fixed plot pairs. SAF = subalpine fir, DF — Douglas-fir iameter IISS Undamaged SAF DF Defol iation class Morta SAF D Lii ^ht Mod. erate Heavy SAF DF Veiy heavy SAF DF lity cl; SAF DF SAF DF DF 0-4.9" 500.0 0.0 275.0 0.0 75.0 0.0 0.0 0.0 25.0 0.0 100.0 25.0 5- -8.9" 6.8 0.0 13.4 0.0 4.3 0.0 0.0 0.0 0.0 0.0 40.4 0.0 9- -11.9" 5.0 0.0 9.5 0.0 3.1 0.0 0.0 0.0 0.0 0.0 23.0 0,0 U '." + 5.7 2.4 5.5 0.4 0.7 0.3 1,2 0.0 1.8 0.0 14.8 1.3 sur\i\'al (Tables 4, 5). Defoliator-caused mor- tality was found to increase with the degree of defoliation. In the very heavily defoliated class, 94% of subalpine firs and 100% of Douglas -firs were killed. None of the sample trees in the lightly defoliated class were killed. Incidence of top-kill also increased with degree of defoli- ation, although trees in the heavily and very heavily defoliated classes were more likely to succumb than exhibit top-kill. This parallels other tussock moth study results, where degree and incidence of top-kill is related to severity of defoliation (Wickman 1978). Sui-viving defoliated trees began to recover by 1993 (Tables 4, 5). Average defoliation rat- ing for lightly defoliated subalpine fir in 1992 was 7.7%. In 1993 the same trees had an aver- age rating of 3.9% with no visible defoliation of that year's needles. The other defoliation classes for subalpine fir and Douglas -fir had similar recoveries. Some of the most dramatic recoveries, however, can be partially attrib- uted to the most heavily defoliated trees of their respective classes succumbing and there- fore not being rated in 1993. Discussion Although Douglas-fir tussock moth had been previously captured in pheromone traps in Utah, the Wasatch-Cache outbreaks are the first to be documented in the state (Tunnock et al. 1985). More significantly, a literature review revealed the Wasatch-Cache outbreaks to be unique in that subalpine fir is apparently the preferred host type. Balch's (1930, 1932) stud- ies are the only that list subalpine fir as a pri- mary host. More recent literature indicates sub- alpine fir to be secondaiy to Douglas -fir, white fir, or grand fir (Wickman et al. 1981, Johnson and Lyon 1988). At the Wasatch-Cache out- breaks, subalpine fir appears to be preferred over Douglas-fir. All three study sites are in close proximity to stands where Douglas -fir is the primary overstory component. These Doug- las-fir stands experienced little or no visible defoliation. This contrasts to Balch's Jarbidge, NV, site where subalpine fir, limber pine, and quaking aspen "form practically the whole of the forest" (Balch 1932). Another exception to the tussock moth's pref- erence for Douglas-fir, white fir, or grand fir has been observed in urban areas along the Colorado Front Range. In these cases blue spruce {Picea piingens Engelm.) has been the preferred host over white fir and Douglas-fir (D. Leatherman-, personal communication). In Colorado's native forests Douglas-fir is the principal host. The defoliation pattern seen on the Wasatch- Cache National Forest outbreaks differed great- ly from previously recorded patterns, such as in Oregon's Blue Mountains. Wickman (1978) recommends estimating defoliation "according to the percent of crown totally defoliated from the top down. " That technique was abandoned for this study because most needle loss was distributed evenly throughout the crown rather than concentrated at the top. Application of Wickman's method would have misrepresent- ed many trees with significant defoliation by having them rated at < 10% defoliation. In other words, the Wasatch-Cache National Forest outbreaks did not fit the "top down" defoliation pattern observed in other outbreaks (J. Weatherby^, personal communication). This study indicates that subalpine fir may be locally more susceptible to tussock moth mortality than either grand fir or Douglas-fir. Despite the difference in percent defoliation "Entomologist, Colorado State Forest Service, ■^Entomologist, USDA Forest Service, Forest Pest Management. Boise, Idaho. 162 Great Basin Naturalist [Volume 55 Table 4. Condition of subalpine fir sample trees within tussock moth nionitorinij plots at Blind Hollow. Baxter Sawmill, and Amazon Hollow, Wasalch-Caehe National Forest, July 1993. Average 1992 Average 1993 Class limits defoliation defoliation- Total no. Top- -kill Mortality Defoliation class^ (% defoliation) (% defoliation) (% defoliation) of trees # % # ^/c Undamaged 0.0 0.0 0.0 51 0 0 0 0 Light 1-24 7.7 3.9 71 / 10 0 0 Moderate 25-74 39.5 36.3 31 14 45 3 10 Heav"\ 75-89 77.5 77.5 9 2 22 5 55 Very heavy >9() 91.3 72.5 63 2 3 59 94 'Trees assigned defoliation class based on 1992 deluliation ratings. ^Surviving trees from 1992 defoliation class. T.VBLE 5. Condition of Douglas-fir sample trees within tussock moth monitoring plots at Blind Hollow, Baxter Sawmill, and Amazon Hollow, Wasatch-Cache National Forest, July 1993. Average 1992 defoliation Average 1993 defoliation- Total no. Top-kill Mortality Defoliation classl (% defoliation) ('7c defoliation) (% defoliation) of trees # % # 9f Undamaged 0.0 0.0 0.0 17 0 0 0 0 Light 1-24 6.5 2.0 20 0 0 0 0 Moderate 25-74 40.0 21.7 3 0 0 0 0 Heavs' 75-89 80.0 65.0 2 0 0 1 50 Very heav\ >90 95.0 — 3 0 0 3 100 'Trees assigned defoliation class based on 1992 defoliation ratings. ^Surviving trees from 1992 defoliation class. estimation techniques, the Wasatch-Cache results can be compared to those of Wickman (1978). At the 90% defohation level, Wickman found 24% grand fir mortality and 30% Douglas - fir mortality (90% defoliation in Wickman's study means complete defoliation in the top 90% of the live crown). At the Wasatch-Cache outbreaks, 57% of subalpine fir defoliated at 90% were killed (90% defoliation using the methodology of this study means that 90% of the estimated total needle complement was consumed). At the 99% defoliation level, Wick- man found that grand fir died at 53% and Douglas-fir at 46%. This compares to 96% mortality on Wasatch-Cache subalpine fir rated at 95% defoliation. Within the infested stud\' areas, the degree of damage varies greatly from one plot to the next. One plot at Amazon Hollow had all host type defoliator killed, while a plot 100 m dis- tant was only lightly defoliated. Although the very heavily defoliated areas are restricted in size (usually less than 5 ac), the amount of mor- tality in these pockets is substantial. An area not sampled, at Ba.xter Sawmill due to salvage logging operations, included over 20 ac where virtually all host t\'pe was killed. Many of these areas are bounded by stands of similar compo- sition and density that were only lightly defoli- ated. In a study of five case histories in Oregon and California, Wickman et al. (1973) found almost one-half of tree mortality occurring in patches coinciding with high moth population centers. Douglas-fir tussock moth outbreaks typical- ly span two to four years. Moth populations develop rapidly and then abruptly subside after only one to two years of outbreak popula- tions (Wickman et al. 1981). The Wasatch- Cache outbreaks have followed this pattern. Moderate to heavy defoliation at Baxter Sawmill was first detected from aerial suney in 1990; defoliation was very heavy in 1991. In 1992 moth activity dramatically declined, and in 1993 no life stages were discovered by either visual inspection or pheromone trapping. While it is be>'ond this stud\ s scope to iden- tify causal agents that initiated the Wasatch- Cache outbreaks, it should be noted that a prolonged drought coincided with the infesta- tion. Most damage occurred on drier sites, such as ridge tops and southerly facing slopes. 1995] Tussock Moth on Subalpine Fir 163 This corresponds to patterns seen in other outbreaks (Bergstrom 1980). The affected trees were apparently drought stressed at the time of defohation. The sudden moth popula- tion collapse mimics that of other outbreaks where a nuclear polyhedrosis virus appears to be the major mortality factor (Wickman et al. 1973). Conclusion Although uncommon, Douglas-fir tussock moth can cause considerable damage to sub- alpine fir. While damage in the three study areas was variable, pockets of heavy defoliation had substantial subalpine fir mortality. Larger- diameter trees are apparently less susceptible to mortality except in these pockets where vir- tually all host tv'pe was killed. Although a minor component in the heavily defoliated areas, local- ly Douglas-fir appears to be less-preferred host material. All study areas are in close prox- imity to Douglas-fir stands that exhibited little or no tussock moth activity. Western balsam bark beetle and annosus root disease con- tributed to subalpine fir mortality, though visi- bly minor relative to defoliator impacts. While forecasting losses in volume would be difficult based on this study, the fate of individual trees can be reasonably predicted given degree of defoliation. Acknowledgments I am grateful for the many people who helped with this project. David Leatherman (Colorado State Forest Sendee), Julie Weather- by, Steve Munson, and John Anhold (all Forest Pest Management, Intermountain Region) pro- vided critical review of the manuscript. Alan Dymerski, John Guyon, Dawn Hansen, John Anhold, Valerie DeBlander (all Forest Pest Management, Intermountain Region), Jill Ansted, Craig Yanase, and Lisa Robinson (all Utah Department of Agriculture) assisted with data collection. Julie Weatherby and John Anhold provided input for the sui-vey design. Dawn Hansen, Cindy Hampton, John Cuyon, and Bent Olsen (all Forest Pest Management, Intermountain Region) helped with data pro- cessing, table preparation, and editing. Irene Voit (Intermountain Research Station) assisted with the literature search. Literature Cited Balch, R. E. 1930. The fir tussock moth reveals abihty to cause serious damage. Forest Worker 6(2): 17-18. . 19.32. The fir tussock moth (Hemerocmnpa psetido- tsiigata McD.). Journal of Economic Entomology 25; 1143-1148. Bergstrom, D. 1980. New lessons from old tussock moth outbreaks. USDA Forest Service, Pacific Northwest Research Station. 4 pp. Berryman, a. a. 1988. Dynamics of forest insect popula- tions. Plenum Press, New York, NY. BousFiELD, W, R. Eder, and D. Bennett. 1985. User's guide and documentation for insect and disease damage survey (INDIDS). Rl-85-19. USDA Forest Service, Northern Region, Missoula, MT. Johnson, W. T, and H. H. Lyon. 1988. Insects that feed on trees and shrubs. 2nd edition. Cornell University Press, Ithaca, NY. Ollieu, M. 1978. Detection of Douglas-fir tussock moth in the Intermountain Region using baited sticky traps. USDA Forest Service, Intermountain Region, Ogden, UT. 7 pp. TuNNOCK, S., M. Ollieu, .\nd R. W Thier. 1985. Histoiy of Douglas-fir tussock moth and related suppression efforts in the Intermountain and northern Rocky Mountain regions — 1927 through 1984. USDA Forest Service, Report 8.5-13. Intermountain and Northern Regions, Missoula, MT. Weatherby, J. C, K. A. Knapp, B. R. G.\rdner, J. Roberts, and P Mocettlnt. 1992. A biological evaluation of the Douglas-fir tussock moth outbreak in southern Idaho, 1991. USDA Forest Sei-vice, Report R4-92- 01. Intermountain Region, Ogden, UT. WiCKNL^N, B. E. 1978. How to estimate defoliation and predict tree damage. USDA Agriculture Handbook 550. Wickman, B. E., R. R. Mason, and C. G. Thompson. 1973. Major outbreaks of the Douglas-fir tussock moth in Oregon and California. USDA Forest Service. General Technical Report PNW-5. Wickman, B. E., R. R. Mason, and G. C. Trostle. 1981. Douglas-fir tussock moth. USDA Forest Service. Forest Insect and Disease Leaflet 86. Received 21 April 1994 Accepted 14 November 1994 Great Basin Naturalist 55(2), © 1995, pp. 164-168 SEASONAL NUTRIENT CYCLING IN POTAMOGETON PECTINATUS OF THE LOWER PROVO RIVER C. Mel Lytle' and Bruce N. Sniitlii-2 Abstiuct. — A common submersed aquatic plant of Great Basin wetland and riverine systems, Potamogeton pectina- tiis L. (sago pondweed) is a key waterfowl food. Nutritional qualities of submersed aquatics in the Great Basin are little understood. The puipose of this study was to determine the seasonal element cycling and nutritional qualities of P. pcctiiialus drupelet, leaf and root tissues from the lower Provo River. Leaf tissue protein was 27% (dry weight) in July, hut declined to 15% by December Diiipelet protein content was 9% in July and 6.5% in October. Lignocellulose in leaf tissue was lowest in July at 34% and increased as the season progressed. Percent fat was highest in leaf tissue at 12% in |ulv Sugars were highest in P pectinatus leaf tissues in December and July. Calcium and magnesium concentrations increased in P pectinatus tissues over the entire season. Leaf tissue zinc was 329 ppm (diy weight) in October Leaf iron concentration was highest in September at 1184 ppm, while root tissue iron was 7166 ppm. Manganese content in leaf tissue peaked in October at 4990 ppm. Copper concentrations in leaves and roots were variable. High protein in leaf tis- sue would benefit local nesting and brooding waterfowl populations that feed on this aquatic. Trace metal concentrations in leaf and root tissues, fi-om possible anthropogenic activities, appear veiy high during fall migratoiy months. Metal bioaccumulation by this species in other Great Basin wetlands and possible metal toxicity in waterfowl warrant ftuther study. Key words: sago pondweed. Potamogeton pectinatus, nutritional qualities, trace element cycling, metal bioaccumulation, waterfowl. A common submersed aquatic plant of the Great Basin, Potamogeton pectinatus is a key priman' producer in fresh and sahne wetlands (Kantrud 1990). Waterfowl feed on all plant parts including drupelet, leaf, and root tissues (Anderson and Ohmart 1988, Korschgen et al. 1988). Sherwood (1960) noted that whisding swans {Olor columhianus) fed heavily on tubers and drupelets during fall migration in the Bear River Migratory Bird Refuge and Ogden Bay Refuge. Other waterfowl species — Canada geese {Branta canadensis), mallards {Anas platy- rhynchos), pintails {Anas acuta), gadwalls {Anas strepera), canvasbacks {Aytha vallisneria), and redheads {Aytha americana) — also fed on P. pectinatus leaf and root tissues. Localized inter- mountain trumpeter swan {Cygnus buccinator) populations are also largely dependent on sub- mersed aquatic plants as food (Anderson et al. 1986, Henson and Cooper 1993). Little is known concerning nutrient dynam- ics and seasonal element cycling of P. pectina- tus from Great Basin wetlands (Kadlec and Smith 1989). Consequently, how this aquatic species may affect waterfowl nutrition is poorly understood. Most assumptions concerning body condition and nutritional requirements are based on studies from other areas of North America. Yet, energy and sustenance required by waterfowl species that frecjuent the Great Basin are largely provided by resident aquatic plants. Of these, P. pectinatus, Ruppia mariti- ma L. (widgeon grass), Scirpus mahtimus L. (alkali bulrush), Scirpus pungens L. (Olney three-square), Scirpus acutus L. (hardstem bulrush), and Zannichellia palustris L. (horned pondweed) are common plant species man- aged in national refuges and waterfowl man- agement areas. Potamogeton pectinatus is con- sidered the most important of these species for diving and dabbling ducks (Kadlec and Smith 1989). The purpose of this study was to determine the seasonal element concentra- tions and nutritional qualities of P. pectinatus from a local Great Basin river drainage. Methods Plant harvests were conducted monthly from three locations within the lower Provo River drainage from July 1991 to December 1991: (1) just below Deer Creek dam (40°24'N, 'Department of Botany and Bange Science. 401 WIDE, Brinliarii VounK I iii\ersit>. Provo. UT 84602. ^Author to whom correspondence should be addressed. 164 1995] Seasonal Nutrient Cycling in P. pectinatus 165 Table 1. A range of measured water column and sedi- ment characteristics, pH, and electrical conducti\'it\' (EC) from the lower Provo River drainage. Water Claritv clear-opaque Velocity (m/sec) 0-0.4 Depth (cm) 5-60 Temperature (°C) 3-14 pH 7.4 EC (/j,mhos/cm'^) 42.5 Sediment >120 3-12 6.9 1570 Table 2. Mean exchangeable Fe and Mn from lower Provo River drainage sediments (ppm dr>' weight ± S.E., n > 3). Means sharing the same letter are not significantly different [P < .05). Depth (cm) 0-7 7-15 15-22 22-30 Fe 61.6 ± 0.7a 56.5 ± 1.4a 61.3 ± 1.7a 57.1 ± 0.8a Mn 19.2 ± 1.2a 11.4 ± 0.8b 9.4 ± 2.1b 12.0 ± 1.7b lll°3rw, elev. 1603 m), (2) near the Sundance turnoff (40° 22'N, lir34'W, elev. 1560 m), (3) =200 m from the mouth of the Provo River near Utah Lake (40°14'N, 111° 44 'W, elev. 1347 m). Water column and sedhiient charac- teristics measured in the lower Provo River are found in Table 1. Sediment conditions ranged from stony with gravelly patches to silty-clay mud. Stands of P. pectinatus were most abundant on muddy sediments. Whole plants (leaf and root tissues) of P. pec- tinatus were sampled in replicate from each location. Drupelets, shoot (stems and leaves) tissues, and belowground (root, rhizomes, and turions) tissues were separated from plant litter and sediments. Invertebrates were removed from samples when rinsed in warm water (38 °C). Cleaned samples were rinsed in deionized water and dried in a forced-air oven at 70 °C. Plant, sediment, and water samples were analyzed at Brigham Young University, Department of Agronomy and Horticulture, Plant and Soil Analysis Laboratoiy Diy plant tissue samples were weighed and ground in a Wiley Mill to pass a 40-mesh screen, and 0.25- g samples were digested in Folin-Wu tubes with 5 ml of concentrated HNO3. Samples were left covered for 16 h before digestion in an aluminum block for 1 h at 100 °C. Three ml of 70% HCIO4 was added, and samples were refluxed at 200 °C until the solution cleared (approx. 30 min). Samples were then brought to 50-ml volume with deionized water (Orson et al. 1992). Element contents were detected by direct aspiration into a Perkin-Elmer Model 5000 Atomic Absorption Spectrophotometer. All blanks and standards were run with the same procedures. Percent total nitrogen and phosphorus were determined using a Kjeldahl digestion followed by analysis with an ALP- KEM rapid-flow analyzer. Sediment (0-30 cm) and water (1000 ml) samples were obtained from the same loca- tions and intervals as plant samples. Sediments were air-dried and extracted for exchangeable iron (Fe) and manganese (Mn) with diethylenetriaminepenta-acetic acid (DTPA) and detected by atomic absorption spectroscopy. Water samples were analyzed for pH, electrical conductivity (;Umhos/cm^), and available Fe and Mn with an Orion Micro- processor Ion-analyzer/901 pH meter, a wheat- stone bridge, and by atomic absorption spec- troscopy. Mean concentrations and standard errors (S.E.) were determined for each plant, sedi- ment, and water sample. To determine if sig- nificant variation in plant tissue nutrient and element concentrations existed between the different months, we used analysis of variance (ANOVA) where month was considered the fixed effect and sample site the experimental unit in a repeated measures design. If signifi- cance (P < .05) was found, Tukey's multiple comparison procedures were used to separate means. Results and Discussion Available Fe and Mn concentrations in water samples were 0.06 ± 0.01 and 0.001 ppm. Sediment exchangeable Fe and Mn contents were found between the normal soil range of 5-65 ppm. Yet, under anoxic conditions that are common in sediments, Fe and Mn may be- come more available for root uptake (Spencer and Brewer 1971, Tisdale et al. 1985; Table 2). Significant differences in sediment exchange- able Mn were found between surface sedi- ments (0-7 cm) and the rest of the sampled profile (Table 2). Element concentrations and forage quali- ties were determined for P. pectinatus tissues from July to December. Leaf and root tissue dry matter, as a percentage of fresh weight, remained constant at 6-7%, with the highest 166 Great Basin Naturalist [Volume 55 dry matter content observed in October. Throughout the season, P. pectinatus element and forage composition varied with growth stage. Significant variation in leaf tissue pro- tein was found (F = 21.69; d.f = 4,14; P < .001) between July, September, October, and December (Table 3). Drupelet protein content was higher in July than in October. In all months sampled, leaf tissue piotein was high- er than drupelet protein. Percent protein in leaf tissue was higher than values reported in other studies (Linn et al. 1975, Kantrud 1990). Acid detergent fiber (ADF) analysis revealed that leaf tissue was lowest in lignocellulose (fiber) in July, but significant differences (F = 3.03; d.f = 4,14; P = .07) in fiber content were not observed as the year progressed. Linn et al. (1975) found P. pectinatus leaf fiber content of 33% that is similar to values obsei-ved in this study. Increased fiber content would decrease the overall forage quality of leaf tissue. Significant variation did exist (F = 177.40; d.f = 4,14; P < .001) in leaf tissue fat content and was highest in July. Total non- structural carbohydrates (sugars) in leaf tissues were highest in December and differed from all other months (F = 42.19; d.f = 4,14; P < .001). By October, drupelet fat and sugar con- tent were both higher than values found in July. Percent nitrogen (N) and phosphorus (P) in leaf tissue reached peak concentrations in July but were significantly lower by December (F = 23.37; d.f = 4,14; P < .001) (F = 79.30; d.f = 4,14; P < .001; Tible 4). Veniiaak et al. (1983) stated that P. pectinatus played an important role in P cycling in aquatic systems. Cultured P. pectinatus grown in water relatively high in phosphate (PO4-P) (0.3 ppm) bioaccumidated p32 jq 4738 times the amount found in the water column. Nitrogen and P content in P. pectinatus can be well above that required for plant growth; this would indicate luxuiy con- sumption of these elements (Jupp and Spencer 1977, Ho 1979, Madsen 1986). Significant concentrations of calcium (Ca) and magnesium (Mg) accumulated (F = 29.12; d.f = 4,14; P < .001) (F = 278.71; d.f = 4,14; P < .001) in leaf tissue between Jul\' and December. This may indicate abiotic deposition, though no vis- ible encrustation on exterior leaf or stem sur- faces was obsei-ved. Hutchinson (1975) report- ed that P. pectinatus leaves were higher in Ca, Fe, K, Mg, Na, and several micronutrients than other aquatic plants. Yet, no mention of time sampled was given for these mineral concen- trations. Therefore, no knowledge of seasonal accumulation was determined. Potassium (K) content was highest in September and differed significantly from percent K content in July (F Table 3. Mean piotein, fiber, fat, and sugar content in P. pectinatus drupelet and leaf tissue over fi\'e months. Forage quality constituents expressed as % diy weight ± S.E., n > 3. Means sharing the same letter are not significantly differ- ent (P < .05). Monti 1 Tissue Protein ADF^' Fat TNC'^ July Leaf 27.4 ± 0.3a .34.2 ± 0.9a 12.2 ± 0.1a 8.3 ± 0.1a Aug. Leaf 24.9 ± 0.3ab 35.6 ± 2.8a 6.5 ± 0.2b 8.1 ± 0.4a Sept. Leaf 2L4±0.3b 39.7 ± 0.4a 6.8 ± 0.2b 7.9 ± 0.2a Oct. Leaf 20.3 ±L4b .37.9 ± 1.3a 7.1 ± 1.1b 8.6 ± 0.1a Dec. Leaf 15.1 ± 0.3c 38.1 ± 0.5a 5.9 ± 1.1c 11.0±0.1b July Dnipelet 9.0 ± 0.5 33.4 ±0.6 6.1 ±0.7 12.0 ± 0.3 Oct. Drupelet 6.5 + 0.8 36.3 ±1.3 7.4 ±0.8 16.3 ±1.2 "Acid detergent fiber (ADF), a measure of percent lignocellulose or fiber ''Total nonstructural carbohydrate (TNC), a measure of sugars Table 4. Mean mineral element concentration in P. pectinatus leaf tissue over five months. Element content expressed as % dr\' weight ± S.E., n > 3. Means sharing the same letter are not significantly different {P < .05). Month Tissue N P K Ca Mg S (% j,-^- wt.) 1,3 ± 0.1a July Leaf 4,4 ± 0.7a 0.6 ± 0.1a 1.9 ± 0.2a 0.3 ± ().04a L2±0.1a Aug. Leaf 2.8 ± 1.0a 0.5 ± 0.1a 3.5 ± 0.1b 1.2 ± 0.1a 0.5 ± 0.01b Sept. Leaf 3.4 ± 0.1a 0.5 ± 0.2a 3.7 ± 0.1b 1.4±0.04ab 0.6 ± 0.02c 1.8 ± 0.2b Oct. Leaf 3.3 ± 0.2a 0.5 ± 0.1a 3.1 ± 0.1b 1.7 ± 0.1b 0.6 ± 0.01c Dec. Leaf 2.4 ± 0,1b 0.2 ± 0.1b 3.0 ± 0.3b 2.3 ± 0.1c 0.7 ± 0.02c 0.6 ± 0.1c 1995] Seasonal Nutrient Cycling in P. pectinatus 167 Table 5. Mean trace element concentration in P. pectinatus leaf tissue over five months. Element content expressed as ppni cli"y weight ± S.E., n > 3. Means sharing the same letter are not significantly different (P < .05). Month Zn Mn Cu July Leaf Aug. Leaf Sept. Leaf Oct. Leaf Dec. Leaf (ppni diy wt.) 213 ± 14a 633 ± 67a 122 ± 6a 21±4a 185 ± 10a 1097 ± 58b 1744 ± 101b 10±lb 211 ± la 1184 ± 75b 3861 ± 117c 10 ± lb 329 ± 4b 963 ± 73b 4990 ± 48d ll±Ob 295 ± 13h 1038 ± 63b 21.30 ± 65b 8±0b = 26.40; d.f. = 4,14; P < .001). Percent sulfur decreased between July and December (F = 13.41; d.f. = 2,10; P = .03; Table 4). Zinc (Zn) concentration in leaf tissue was significandy higher (F = 36.56; d.f = 4,14; P < .001) in October and December than in all other months (Table 5). Mean Fe content was higher in August leaf tissue than in July (F = 12.59; d.f = 4,14; P = .001), after which Fe content remained fairly constant throughout the remainder of the sample period. Leaf tissue Mn content increased through the season and was highest in October (F = 587.38; d.f = 4,14; P < .001; Table 5). Dudkin et al. (1976) found that P. pectinatus, growing in polluted coastal waters of the Black Sea, accumulated Mn to 0.5% (dry weight). This Mn concentra- tion corresponds to values found in this study. Yet, Mn concentrations in water and sediment from the lower Frovo River appear normal. Copper (Cu) in leaf tissue varied significantly (F = 44.48; d.f = 4,14; P < .001), with high concentrations in July followed by lows in August through December (Table 5). Root tissues (root, rhizomes, and turions) of P. pectinatus were not separated for analysis. Mean root tissue forage qualities, compared to leaf tissues, were lower in percent protein but higher in fat content (Table 6). Phosphorus was the only mineral element with a concen- tration higher in root tissues than in leaf tissues. Mineral (N, P K, Ca, and Mg) contents of root tissues in this study were similar to contents found in other studies (Kollman and Wali 1976, Van Vierssen 1982). High trace metal concentrations were also found in root tissues. Like leaf tissues, mean Fe and Mn concentra- tions in root tissues appear inordinately high. Conclusions and Future Research Seasonal variation did exist in forage quali- ties and nutrient concentrations in P. pectina- tus. Protein content in leaf tissue was highest in the summer months when P. pectinatus was growing rapidly. By fall and early winter, protein content decreased but was still higher than concentrations found in drupelets. Apparently, protein content in P. pectinatus leaf tissue from the lower Provo River was higher than concentrations reported elsewhere. High pro- tein content in leaves and stems in the sum- mer months would greatly benefit nesting and brooding waterfowl that feed on this aquatic species. Drupelet fat and sugar content was higher than that for leaf or root tissues in October. This would tend to confirm why dru- pelets are so eagerly sought after by staging and migrating waterfowl. Trace metal (Fe and Mn) contents in leaf and root tissues accumu- lated over the season and were very high by fall. However, water and sediment concentra- tions appear normal. It should be determined whether the trace metal concentrations observed are of natural or anthropogenic ori- gin. Future research should develop a greater understanding of heavy metal accumulation in this and other key Great Basin aquatic plant species. Table 6. Forage quality, mineral and trace element con- centration of P. pectinatus root tissue (root, rhizome, and turions) averaged over five months. Forage quality con- stituents and mineral content expressed as % and ppm dry weight! S.E., n > 3. Protein ADF Fat TNC^ (% dn' wt.) 13.0 ±1.0 ndl' ' 10.9 + 3.0 11.9 ±1.3 N P K Ca Mg (% dry wt.) 2.1 ±0.2 0.4 ±0.1 2.9 ±0.2 1.5 ±0.2 0.3 ±0.1 Zn Fe Mn Cu (ppm di-y wt.) 167 ± 25 7166 ± 1438 ' 2051 ± 570 14.8 ± 3.3 ^Total nonstructural carbohydrate (TNC), a measure of sugars "Not detemiined 168 Cheat Basin Natur.\list [Volume 55 Acknowledgments Funding and materials foi- this stud>' were provided by the Department of Botany and Range Science at Brigham Young University and the Utah Chapter of the Wildlife Society. Literature Cited Anderson, B. W, and R. D. Ohmart. 1988. Stmcture of the winter duck commiinit\' on the Lower Colorado River: patterns and processes. Waterfowl in winter University- of Minnesota Press, Minneapolis. 624 pp. Anderson, D. R., R. C. Herron, and B. Reiswk;. 1986. Estimates of annual survival rates of tnmipeter swans handed 1949-82 near Red Rock Lakes National Wild- life Reflige, Montana. Journal of Wildlife Management 50:218-221. DuDKiN, M. S., I. V. Areshidze, and G. D. Lukina. 1976. Chemical composition of seaweed in the coastal waters of the Black Sea Ukrainian-SSR USSR. Rastitel' nye Resursy 12: 133-137. Henson, R, and J. A. Cooper. 1993. Trumpeter swan incuhation in areas of differing food quality. Journal of Wildlife Management 57: 709-716. Ho, Y. B. 1979. Inorganic mineral nutrient level studies in Potamogeton pectbuitus L. and EnteromoiyJw prolifera in Forfar Loch, Scodand. Hydrohiologia 62; 7-15. Hutchinson, G. E. 1975. A treatise on limnolog>'. Volume III: Linmological botany. John Wiley & Sons, New York, NY, 660 pp. JUPP, B. P, and D. H. Spencer. 1977. Limitations on macro- phytes in a eutrophic lake. Loch Leven. I. Effects of phytoplankton. Journal of Ecology 65: 175-186. Kadlec, J. A., and L. M. Smith. 1989. The Great Basin marshes. In: L. M. Smith, R. L. Pedersen, and R. M. Kaminski, editors. Habitat management for migrating and wintering waterfowl in North America. Te.xas Tech University Press, Lubbock. 651 pp. Kantrud, H. a. 1990. Sago pondweed (Potamogeton pectinatus L.): a literature review. U.S. Fish Wildlife Service Resource Publication 176. 89 pp. Kollman, a. L., and M. K. Wall 1976. Intraseasonal variations in environmental and productixity rela- tions of Potamagcton pectinatus communities. Archiv fuer Hydrobiologie, Supplementband 50: 439-472. KoRSCHGEN, C. E., L. S. Georce, .\nd W. L. Green. 1988. Feeding ecology of canvasbacks staging on Pool 7 of the Upper Mississippi River Waterfowl in winter. University of Minnesota Press, Minneapolis. 624 pp. Linn, J. G., E. j. Staba, R. D. Goorich, J. C. Meiske, and D. E. Otterby. 1975. Nutritive value of dried or ensiled aquatic plants. I. Chemical composition. Journal of Animal Science 41: 601-609. Madsen, J. D. 1986. The production and physiological ecolog}' of the submerged acjuatic macroph\ te com- munity in Badfish Creek, Wisconsin. Unpublished doctoral dissertation. University of Wisconsin, Madison. 449 pp. Orson, R. A., R. L. Simpson, and R. E. Good. 1992. A mechanism for the accumulation and retention of heavy metals in tidal freshwater marshes of the LJpper Delaware River Estuaw. Estuarine, Coastal, and Shelf Science 34: 171-186.' Sherwood, G. A. 1960. The whistling swan in the West with particular reference to Great Salt Lake Valley, Utah. Condor 62: 370-377. Spencer, D. W, and P. G. Brewer. 1971. Vertical advec- tion, diffusion and redox potentials as controls on the distribution of manganese and other trace metals dissolved in water of the Black Sea. Journal of Geo- physical Research 76: 5877-5892. Tisdale, S. L., W. L. Nelson, and J. D. Beaton. 1985. Soil fertility and fertilizers. 4th edition. Macmillan Publishing, New York, NY 753 pp. Van Vierssen, W. 1982. The ecology of communities dominated by ZannicJieUia ta.xa in western Europe. III. Chemical ecology. Aquatic Botany 14: 259-294. Vermaak, J. E, J. H. Swanepoel, and H. J. Schoonbee. 1983. The phosphorus cycle in Germiston Lake with special reference to the role of Potamogeton pectinatus L. Pages 317-321 in ProceecUngs of the international symposium on aquatic macrophytes, Nijmegan, Netherlands. Received 31 May 1994 Accepted 3 Jamianj 1995 Great Basin Naturalist 55(2), © 1995, pp. 169-173 FACTORS INFLUENCING FISH ASSEMBLAGES OF A HIGH-ELEVATION DESERT STREAM SYSTEM IN WYOMING Bernard Carter^ and Wayne A. Hubert^ Abstract. — Seven fish species were found in the Bitter Creek drainage of southwest Wyoming, but only speckled dace (RJiinichthijs oscuhis), flannelmouth sucker [Catostomus latipinnis), and mountain sucker (Catostomtis phityrhynchus) were indigenous. No relationships were foimd between fish standing stocks and habitat features, but species richness was related to elevation and stream width. No fish were found above an elevation of 2192 m. Only the most downstream study reach had more than three species present. Two indigenous species, speckled dace and moimtain sucker, and a nonnative species, fathead minnow (Pimephales promelas), were predominant fishes in the drainage. These three species withstand intermittent stream flows that are common in the drainage. Key words: fish, streams, desert, Wyoming, habitat, distribution. Fish communities in streams become more complex as habitat diversity increases along the length of a stream. Variation in fish com- munity' stnicture within a stream system can fol- low patterns of zonation or addition. Specific fish communities can be associated with zones defined by water temperature or geomoi-pho- logic features, or community complexity can increase with progression downstream by addition of species (Moyle and Nichols 1973, Guillory 1982, McNeely 1986, Hughes and Gammon 1987, Platania 1991, Rahel and Hubert 1991). However, such patterns may differ in arid drainages of the western United States with depauperate ichthyofauna (Cross 1985). Little is known about the fish communities in high-desert stream systems in southwestern Wyoming. Annual precipitation over most of these drainages is < 16 cm, with much of it as snow in headwater areas during late winter and thunderstorms during late summer Discharge is highest during spring runoff, and streams frequently become intermittent during sum- mer and winter. Because these systems in Wyoming are at high elevations (>1800 m above mean sea level), water temperatures are cool compared with other desert streams. The climate in these areas typically consists of diy, moderately warm summers with long, cold winters. The puipose of this study was to (1) describe fish species present in a high-desert stream system in southwestern Wyoming and (2) determine the factors that influence fish abun- dance and community structure within the drainage. Study Area The study was conducted in an intermittent drainage. Bitter Creek, a tributary to the Green River in the Red Desert of southwest Wyo- ming (Fig. 1). The study area consists of Bitter Creek and four tributaries — Little Bitter, Salt Wells, Bean Springs, and Gap creeks. Frequent- ly, no measurable surface flow occurs in Bitter Creek at Bitter Creek, WY, during midsum- mer and midwinter (flow data available in the Water Resources Data System at the Wyoming Water Resources Center, University of Wyo- ming, Laramie). Bitter Creek at Salt Wells, WY, generally has no measurable surface flow from July to February. Salt Wells Creek has more persistent flows near its mouth, but records of no measurable flows occur in mid- summer and midwinter When no measurable flow occurs in these streams, isolated pools of standing water can be found in the stream channels. Elevation of the study area ranges from 1800 to 2400 m. Streams in the Bitter Creek drainage typi- cally are downcut by at least 1.5 m, with steep clay banks having no vegetation. Riparian vege- tation consists of grasses and sagebrush {Artem- isia spp.); upland vegetation is primarily the latter ^Wyoming Cooperative Fish and Wildlife Research Unit, University of Wyoming, Laramie. \V\' 82071-.3166. The unit is jointK- supported by the University of Wyoming, Wyoming Game and Fish Department, and National Biological Survey. 169 170 Great Basin Naturalist [Volume 55 Salt Wells, Wyoming Rock Springs, Wyoming ?>^' iixef G<' ,e^ Bitter Creek, Wyoming CD 03 O WY0MIN6 \ Research location 0^- o 3r 0^' >#> 1 03 5> ) 95^1 20 km Fig. 1. Map of the Bitter Creek drainage, \\T, showing the location of the 16 study readies. Baxter and Simon (1970) reported four fi.sh species in collections at two sites in Bitter Creek. Speckled dace {Rhinichthijs osculus), fathead minnow {Pimephales promelas), and mountain sucker {Catostomus platijrhtjnchus) were reported from a site al:)out 10 km upstream from the mouth. Bluehead sucker {Catostomus discobolus) was the only species reported from a site near Salt Wells. Methods Sixteen 100-m-long study reaches were selected to represent variation in stream size and habitat in the drainage during summer 1993. Wetted width, mean depth, and sub- strate were determined across transects at 10- m intervals. Dominant substrate at each tran- sect was visually determined following Bain et al. (1985): sand-silt (<2 mm diameter), gravel (2-16 mm), pebble (17-64 mm), cobble (65-256 mm), and boulder (>256 mm). Water velocity was determined within each reach using the dye flow mediod (Binns 1982). Stream discharge at time of sampling was computed from width, depth, and velocity. Alkalinity, hardness, and pH were mea- sured at the time of sampling. Alkalinity and hardness were determined with field test kits (Hach Model A1-36DT), pH with an electron- ic meter Mean elevation and channel slope at each study reach were estimated from 7.5- minute topographic maps. Fish were sampled in each 100-m reach by electrofishing. Small-mesh (6.4-mm) block nets were placed at each end, and two or three elec- trofishing passes were made over the entire reach. Three-pass depletion estimates of species abundance were made in most reaches. Two- pass depletion estimates were used when >80% offish captured by the first two passes were captured during the first pass. Fish abundance was computed using the Zippin method (Platts et al. 1983). All fish were 1995] Desert Fishes 171 weighed to enable computation of standing stock estimates. Standing stocks of individual species, total standing stock of all species, and number of species in a reach were evaluated for their relation to nine habitat variables using simple- linear and multiple-regression analyses. Independent variables were included in regression models if they were significant at P < .05. We further limited inclusion of de- pendent variables in multiple-regression mod- els to ones that were not correlated at P < .05. Computations were performed using Statistix 4.0 (Analytical Software 1992). Results Seven fish species were collected: speckled dace, fathead minnow, Utah chub {Gila atrarici), Bonneville redside shiner {Richardsoniiis balteatus hydrophlox), mountain sucker, white sucker {Cotostomus commersoni), and flannel- mouth sucker (C. latipinnis). Abundance varied substantially among study reaches (Table 1). Mean total standing stock of all species was 3.0 g/m'^ and ranged from 0 to 21.3 g/ni'^. No fish were found in the four reaches above 2192 m. Habitat features varied among the 16 study reaches (Table 2). Flow was measurable at all reaches. Stream width, water velocity, and dis- charge increased downstream. Sand-silt sub- strate occurred over >90% of almost all study reaches. Alkalinity, pH, and hardness also increased downstream. No significant relations were found between any of the nine habitat variables and standing stocks of individual species or total standing stock of all species. However, there were significant relations between the number of species and four habitat variables: NS = 20.88 - 0.0091 E (F = .0003, R^ = .61), NS = 0.13 + 0.812 W (F = .0010, R2 = .52), NS = 3.40 - 11.008 V (F = .029, fi2 = .33), and NS = 0.57 + 31.245 F (F = .022. R^ = .32), where NS = number of species, E = eleva- tion in meters, W = mean wetted width in meters, V = water velocity in meters per sec- ond, and F = flow in cubic meters per second. The best two-variable model was NS = 14.36 - 0.0065 E + 0.53 VV (F < .0001, R'- = .80). As study reaches declined in elevation and as width, water velocity, and discharge increased, the number of species increased. Because the most downstream reach on Bitter Creek had twice as many species as any other reach and flow at the reach was en- hanced by discharge from a sewage treatment plant, we assessed relations with the omission of that reach. Again, no relationships were found between any of the habitat variables and standing stocks of fish, but the number of species (NS) was significantly related to eleva- tion (E) and water velocit)' (V): NS = 15.95 - 0.0068 E (F = .0014, R- = .55), and NS = 3.00-10.11 V (F = .0018, R^ = .51). Among the 15 study reaches with a maximum of three species present, the number of species increased with decline in elevation and water velocit}'. Discussion Of the seven fish species in the Bitter Creek drainage, only three — speckled dace, flannel- mouth sucker, and mountain sucker — are indigenous (Baxter and Simon 1970). Absence of fish above 2192 m is probably due to a cli- mate that is too cold for warmwater fishes. Additionally, no trout occur naturally or have become naturalized in the watershed. The number of species increased with pro- gression from headwater to downstream reaches (Table 1). With the exception of the most downstream reach on Bitter Creek, no more than three species — specked dace, mountain sucker, and fathead minnow — were found in any of the study reaches. The high- elevation reaches with fish tended to have predominantly or exclusively speckled dace and mountain sucker. Much of the increase in species richness with downstream progression was due to the most downstream reach on Bitter Creek where six species were found (Table 1). Four of six species were not natives — fathead min- now, white sucker, Utah chub, and Bonneville redside shiner Mountain sucker was not found in this reach, but it was common throughout most of the Bitter Creek drainage. While this study reach was lowest in elevation among the 16 study reaches, it also was downstream from the outfall of the wastewater treatment facility 172 Great Basin Naturalist [Volume 55 O) m ^ d t- c£> lO »o 5 CO o) — O' — ^ d d d — : p — ; q — < o o -— I 00 00 in t~- d ^ CO d 00 05 t^ Ol f>] -H oj CO r-^ p in — ; q oi t>i oq o6 oq oq oi q 00 t-; CO oi oi ^ d> ^ ys O^ oq oq — < o o ^ i J, a 1- 5 u s X 1000) inity g/L X 10 Iness o X bC ^ ~ 3 3 ■| a 0; !Ll j;"! s § ^ > S3 E^ S ,-* 0; OJ ""'^ V -^^ — rt K cii S S ^ 5 u w ^ X 1995] Desert Fishes 173 for Rock Springs, WY, and was only 13 km upstream from the confluence of Bitter Creek and the Green River. The more permanent flows due to the wastewater treatment facihty may have enabled fish not adapted to inter- mittent flows to persist in this reach. Also, the relatively short distance to the Green River may enable upstream migration of fish to this reach, contributing to higher species diversity. Repeated invasion of nonnative species from downstream reservoirs maintains species diversit)' in the Virgin River, UT (Cross 1985). Also, human disturbances have been found to create environmental conditions favorable to nonnative fish in California (Moyle and Nichols 1973, 1974). Therefore, enhanced flows due to the wastewater treatment facility and invasion of nonnative species from the Green River probably contribute to the diversity of fish in the downstream portion of Bitter Creek. During summer 1993, flowing water occurred at all study reaches when they were sampled. Precipitation in spring and summer 1993 was substantially greater than normal, enabling measurable surface flows to persist during summer. However, study reaches up- stream from the outfall of the Rock Springs, WY, wastewater treatment facility are frequent- ly intermittent during late summer. Fathead minnow has been described previously as a species associated with intermittent streams (Baxter and Simon 1970, Pflieger 1975). Our observations indicate that two indigenous species — speckled dace and mountain suck- er— and one introduced species — fathead minnow — can survive in the frequently inter- mittent streams. Consequently, these three fishes are the only species occuning over most of the Bitter Creek drainage, but fathead min- nows tend to be limited to lower elevations than the two native species. It is not known how the invasion by fathead minnow may affect the native speckled dace and mountain sucker in this desert stream system. Acknowledgments We thank M. Fowden, K. Johnson, W. Wengert, and R. Wiley, Wyoming Game and Fish Department, for their assistance, and H. Li, T. Patton, F Rahel, R. Wiley, and an anony- mous referee for review of the manuscript. The project was supported by the Wyoming Game and Fish Department. Literature Cited Analytical Software. 1992. Statistix Version 4.0 user's manual. Analytical Software, St. Paul, MN. Baln, M. B., J. T. Finn, and H. E. Booke. 198.5. Quantifying stieam substrate for habitat analysis stud- ies. North American Journal of Fisheries Manage- ment .5: 499-500. Baxter, G. T, and J. R. Simon. 1970. Wyoming fishes. Bulletin 4. Wyoming Game and Fish Department, Cheyenne. Binns, N. a. 1982. Habitat quality inde.x procedures manu- al. Wyoming Game and Fish Department, Cheyenne. Cross, J. N. 1985. Distribution offish in the Virgin River, a tributary of the lower Colorado River Environ- mental Biology of Fishes 12: 13-21. GuiLLORV, V. 1982. Longitudinal gradients of fishes in Thompson Creek, Louisiana. Southwestern Naturalist 27: 107-115. Hughes, R. M., and J. R. Gammon. 1987. Longitudinal changes in fish assemblages and water qualit>' in the Willamette River, Oregon. Transactions of the American Fisheries Society 116: 196-209. McNeeley, D. L. 1986. Longitudinal patterns in fish assemblages of an Ozark stream. Southwestern Naturalist 31: 375-380. Moyle, R B., and R. D. Nichols. 1973. Ecology of some native and introduced fishes of the Sierra Nevada foothills in central California. Copeia 1973: 478-490. . 1974. Decline of native fish fauna of the Sierra Nevada foothills, central California. American Mid- land Naturalist 92: 72-83. Pflieger, W. L. 1975. The fishes of Missouri. Missouri Department of Conser\'ation, Columbia. Platania, S. R 1991. Fishes of the Rio Chama and upper Flio Grande, New Mexico, widi preliminan' comments on their longitudinal distribution. Southwestern Naturalist 36: 186-193. Platts, W. W., W. F Megahan, and G. W Minshall. 1983. Methods for evaluating stream, riparian, and biotic conditions. LISDA Forest Service General Technical Report INT-138. 70 pp. R\hel, F J., AND W. A. Hubert 1991. Fish assemblages and habitat gradients in a Rocky Mountain-Great Plains stream: biotic zonation and additixe patterns of community change. Transactions of the American Fisheries Society 120: 319-332. Received 21 April 1994 Accepted 3 October 1994 Great Basin Nahiralist 55(2), © 1995. pp. 174-176 SPECIATION BY ANEUPLOIDY AND POLYPLOIDY IN MIMULUS (SCROPHULARIACEAE)^ Robert K. Vickcn', Jr.2 Key wards: Miimiliis. speciation. cvoluthni, aiwuploidy, polyploidy. Speciation by aneuploid and polyploid changes in chromosome numbers is so common in flowering plants as to be almost a character- istic of the angiosperms. Elegant examples of these patterns of evolution are exhibited by monkey flowers of the genus Mimiihis (Scro- phulariaceae). The genus Mimulus contains some 150 species occurring in western North and South America with a few outlying species in eastern North America, Japan, Vietnam, the Himalayas, New Zealand, Australia, and South Africa. The center of diversity is California, with a second- ary center in Chile. Some species are annuals of deserts, grasslands, or forests; some are biennials of marshy places; some are herbaceous peren- nials from springs, streamsides, or lake-shore habitats; and others are woody shrubs of the dry California chapanal. The species fonii clus- ters reflecting these various life forms. There are 8-10 such clusters commonly recognized as sections of the genus Mimuhis (Crant 1924, Fennell 1951, Chuang and Heckard personal communication). Chromosome numbers of over 50 species (Table 1), that is, approximately one-third of the Mimulus species, have been ascertained by Vickery and his co-workers (Vickeiy 1978, Vickery, Chu et al. 1981, Vicken, Simpson et al. 1981, Vickery et al. 1982, 1985,' 1986, 1990, un- published) and by Chuang and Heckard (per- sonal communication). Chromosome numbers reveal intriguing patterns of evolution by aneu- ploidy and polyploidy. First, let us consider the base chromosome numbers of the eight main sections of the genus. Section Munulastrutn has a ])ase num- ber of X = 7; Eunanus and Erytliranthe have base numbers of x = 8; Paradanthus 8, 9, 10; Eumimulus 8, 11, 12; Oenoe 9; DipJacus 10; and Simiolus 14, 15, 16, 30. Base numbers of the sections suggest extensive evolution by both aneuploidy and poKploidv'. For the genus as a whole, the base number appears to be x = 8, inasmuch as the other plausible base num- ber, X = 7, is found only in one, apparently de- rived, desert species, M. moJuiven.sis Lemmon (Table 1). Next, let us consider the chromosome num- bers by individual species. All species counted thus far are the same in each of several sections, specifically, in Mitnulastrum, Erythranthe, Oenoe, and Diplacus. The other sections are polymorphic for their species' chromosome numbers and frequently exhibit speciation by aneuploidy and/or polyploidy, often in com- plex combinations. For example, the various species of section Eumimulus exhibit n = 8, 11, and 12; species of section Eunanus exhibit n = 8, 10, and 16; species of section Paradan- thus exhibit n = 8, 9, 16, 17, 18, and 30; and species of section Simiolus exhibit n = 13, 14, 15, 16, 24, 28, 30, 31, 32, 46, and 48 (Table 1). Section Simiolus, which shows by far the most speciation by aneuploidy and/or poly- ploidy of all sections of the genus, consists of six species groups, that is, complexes of related species and varieties. First is the M. guttatus complex, centered in California; it has as its base number x = 14, with aneuploid forms at n = 13 and n = 15 (Table 1), as well as tetra- ploid forms with n = 28. Second is the alpine (western United States) M. tdingii complex with its base number of x = 14 and aneuploid forms at n = 15, /] = 16, and an unusual pol> - ploid form at n = 24. The third species group is the M. dentilobus complex of southwestern United States and northwestern Mexico with its base number of x = 16 and an aneuploid form at n = 15. Fourth is the M. luteus complex 'a talk presented 4 September 1993 as part of the symposium, "Plant Evolution. ^Biology Department, University of Utali, Salt Lake City, UT 84112 USA. lit tlie National Institute ol't^enetics, Mishima. Japan. 174 1995] Notes 175 T\BLE 1. Chromosome mmibers in tlie genus Miimihis li\' sections (counts by Chuang and Heckard and 1)\' Vicken and co-workers; see text for references). Taxon n = Mimiilastrum Gray (.v = 7) M. mohavensis Lemnion 7 Eumiimilus Gray {x = 8, 11, 12) M. alatus Aiton 11 M. gracilis R. Br 8 M. ringens L. 8, 12 Eunanus Gray (.r = 8) M. bolanderi Gray 8 M. layneae (Greene) Jepson 8 M. brevipes Bentham 8 M. cusickii (Greene) Piper 8 M. nanus Hook. & Arn. 8 M. torreyi Gray 10 M. biglorii Gra\ 16 Paradantlms Grant (.v = 8, 9, 10) M. bicolor Haitweg ex Bentham 8 M. filicaulis Watson 8 M. breweri (Greene) Coville 16 M. floribundus Douglas 16 M. moschattis Douglas 16 M. laiidens (Gray) Greene 16 M. arenarius Grant 16 M. primidoides Rydb. 9, 17, 18 M. repens R. Br 10 M. nepalensis Bentham 16, .30 Enjtbrandw Greene {x = 8) M. cardinally Douglas 8 M. eastwoodiae Rydb. 8 M. lewisii Pursh 8 M. nelsonii Grant 8 M. ntpestrh Greene 8 M. verbenaceits Greene 8 Oenoe Gray (x = 9) M. picttis (Cunan) Gray 9 M. tricolor Lindl. 9 M. pygmaeus Grant 9 (or 10?) M. pilosellus Greene 9 Diplacus Gray (.t = 10) M. aridiis (Abrams) Grant 10 M. aurantiacus Curt. 10 M. calycinus Eastw. 10 M. clevelandii Brandg. 10 M. fasiculatus (Pennell) McMinn 10 M. longiflorus (Nutt.) Grant 10 M. puniceus (Nutt.) Steud. 10 Simiohts Greene (.t = 14, 1.5, 16) M. gtttatus Fischer e.\ DC. 14, 1.5, 28 M. laciniatus Gra\' 14 M. nasutus Greene 13, 14 M. glaucescens Greene 14 M. platycalyx Pennell 15 M. tilingii Regel 14, 15, 24, 28 M. gernnipanis Weber 16 M. dentilobus Rols. & Fern. 15, 16 M. wiensii Vicker>' 16 M. glabratus HBK 1.5, .30, 31 Af. andicolus HBK 46 M. pilosiuscidus HBK 46 M. extemiis (Skotts.) Skotts 46 M. luteus L. 30, 31, .32 A/, cupreus Dombrain 31 Undescribed n. sp #A 16 n. sp #B 32 n. sp #C .32, 48 ± 1-4 from the central and southern Andes of South America. Its base number is x = 30, but there are n = 31 and n = 32 forms as well. Fifth, there is the M. glabratus complex that ranges from Canada to Patagonia. Its varieties in cen- tral North America exhibit the base number of the complex, x = 15. In the Rio Grande drainage we find tetraploids with n = 30. From northern Mexico to southern Colombia we find the aneuploid tetraploid n = 31 vari- eties of the complex. From Ecuador south to southern Argentina and including the Juan Fernandez Islands off the coast of Chile, we find the aneuploid hexaploid species and vari- eties with n = 46 chromosomes. Apparently, each change in chromosome number facilitat- ed an adaptive radiation further south. Last is the M. wiensii complex of the mountains of western Mexico with its base number of .t = 16 and three apparent new species that are morphologically distinct and reproductively isolated (Vickeiy et al. unpublished). One has n = 16 chromosomes, one has n = 32 chromo- somes, and the third has two forms — one with n = 32 chromosomes and the other with n = 48 ± 1—4 chromosomes (incipient aneuploidy?). How does speciation by aneuploidy and polyploidy occur? We carefully examined meiosis in M. glabratus var. utahensis and M. glabra- tus var. fremontii, two of the widespread diploid varieties of the M. glabratus complex, and their intervarietal F^ hybrids. First, of 1317 cells examined in diakinesis or meta- phase of first meiosis (Tai and Vickery 1970, 1972), 1090 exhibited regular 15 bivalent chromosomes. Another 23 cells, or 1.7%, had aneuploid numbers of chromosome pairs rang- ing from only 6 to as many as 13, plus 4-18 univalents. These cells presumably could pro- duce aneuploid gametes, at least in some cases. A sizeable minority, 204 cells, exhibited 14 II and 2 I, or 13 II and 1 IV, or complement frac- tionation with its uneven groupings of chro- mosomes. These cells might produce aneu- ploid gametes also. Second, of 782 additional cells observed in Anaphase I, 294 (37.5%) exliibited unequal disjunction, laggard chromo- somes, or chromatin bridges. These cells also could result in aneuploid gametes. Some 47 of these abnormalities occurred in M. glabratus wax. fremontii, only 18 occurred in M. glabratus var utahensis, but most, 229, occuiTcd in die in- tervarietal hybrids. Thus, varieties differ in their potential for producing aneuploid gametes. 176 Cheat Basin Naturalist [Volume 55 and intcrvarictal hybrids are particular!)' prone to do so. This suggests to me that natur- al hybridization probably plays a significant role in evolution in monkey flowers. Finding occasional plants in various populations with aneuploid chromosome numbers indicates that aneuploid gametes not only are produced, but actually function. Third, of 95 cells exam- ined in Anaphase II, 22 were polyploid and could presumably lead to polyploid gametes. Thus, we see significant numbers of the veiy cytological abnormalities in the basic diploid varieties that could lead to evolution by aneu- ploidy and polyploidy, that is, to the veiy pat- terns of evolution that we actually see in the M. glabratus complex. Literature Cited Grant, A. L. 1924. A monograph of the genus Miinulus. Annals of the Missouri Botanical Garden 11; 99-389. Pennell, F. W. 19.51. Mimuhis. Pages 688-731 in L. Abrams, Illustrated flora of the Pacific States. Volume 3. Stanford University Press, Stanford, CA. Tai, W., and R. K. Vickery, Jr. 1970. Cytogenetic rela- tionships of key diploid members of the Mimuhis glahratus complex (Scrophulariaceae). Evolution 24: 670-679. . 1972. Unusual cytological patterns in microsporo- genesis and pollen development of evolutionaiy sig- nificance in the Mimulu.s glabratu.s complex (Scrophu- lariaceae). American Journal of Botany .59; 488-493. Vickery, R. K., Jr. 1978. Case studies in the evolution of species complexes in Mimuhis. Evolutionarv' Biolog\ 11; 404-506. Vickery, R. K., Jr., Y. E. Chu, K. Fine.man, and S. Pt'RC:ELL. 1981. Chromosome number reports on the Scrophulariaceae in lOPB Chromosome number reports LXX presented by Askell Love. Taxon 30: 68. Vickery, R. K., Jr., M. Simpson, and M. Nellestein. 198 1. Chromosome number reports on the Scrophu- lariaceae in lOPB Chromosome number reports LXX presented by Askel Love. Taxon 30: 68-69. Vickery, R. K., Jr., S. A. Werner and E. D. MacArthur. 1982. Chromosome number reports on the Scrophu- lariaceae in lOPB Chromosome number reports LXXV presented by Askell Love. Taxon 31; 360. Vickery, R. K., Jr., S. A. Werner , D. R. Phillips, and S. R. Pack. 1985. Chromosome counts in section Siiniohis of the genus Mimuhis. X. The M. gkibnifiis complex. Madrono 32: 91-94. Vickery, R. K., Jr., B. Y. Kang, T K. Mac, S. R. Pack, and D. A. Phillips. 1986. Chromosome counts in Mimuhis sect. Enjthranthc (Scrophulariaceae). III. Madrono .33; 264-270. Vickery, R. K., Jr., E R^^hmen, S. R. Pack, and T. Mac. 1990. Chromosome coimts in section Simiohis of the genus Mimuhis (Scrophulariareae). XI. M. ghibratus complex (cont.). Madrono 37; 141-144. Received 6 J uhj 1994 Accepted 24 September 1994 Great Basin Naturalist 55(2), © 1995, pp. 177-180 SPECIATION IN MIMULUS, OR, CAN A SIMPLE FLOWER COLOR MUTANT LEAD TO SPECIES DIVERGENCE?^ Robert K. Vickeiy Jr.- Key words: Mimulus, Eiythranthe, speciation, reproductive isolation, flower color mutations, pollinators, bumble- bees, hummingbirds. The general pattern of speciation in nature has been clear for a long time — the diver- gence of portions of a population, usually small (Levin 1993), usually in geographic isolation (Mayr 1976), and the accumulation of genetic changes by selection and/or genetic drift (Crow and Kimura 1970) that produce reproductive isolation and normally character divergence as well. The critical step is reproductive isolation, and yet that step — except for polyploid forma- tion which in itself is not always effective (DeWet 1980) — has rarely been observed actually happening in nature. A promising group in which to study speciation events in progress is section Erythranthe of the genus Mimidus (Vickeiy 1978). The six species of monkey flowers compris- ing section Erijthranthe are moisture-requiring, herbaceous perennials 1-10 dm in height, with variously shaped, opposite leaves and bilabiate flowers that have four stamens, one style with a bilobed sensitive stigma, and five corolla lobes that range in color from orange to red — rarely yellow — and from lavender-pink to magenta-pink — rarely white. See Grant (1924) for further details. When considered species by species, corollas of M. cardinalis Douglas vary from orange to red — rarely yellow — and are sharply and fully reflexed, hummingbird- pollinated flowers. Corollas of M. verbenaceus Greene are partially reflexed; that is, the upper two corolla lobes are reflexed, whereas the lower three are gently recurved. Flowers are orange -red to red — rarely yellow — and also are hummingbird-pollinated. Corollas of M. nelsonii Grant are partially re-flexed also and have orange-red to red flowers, which are longer than those of M. verbenaceus (6-7 cm versus 4-5 cm). Corollas of M. eastwoodiae Rydberg and M. rupestris Greene, the two cliff-dwelling species, are partially reflexed, red, and typically hummingbird-pollinated also. And last, flowers of the Rocky Mountain variety of M. lewisii Fursh are magenta-pink with all five corolla lobes gently recurved rather than reflexed, thus forming a bee-land- ing platform; flowers of the Sierra Nevada variet\' of Al lewisii are lavender-pink — rarely white — with corolla lobes thrust foi"ward. Both varieties of M. lewisii are bee-pollinated. Mimulus lewisii flowers and those of M. east- woodiae and M. rupestris produce only modest amounts of nectar, whereas the other species produce abundant nectar (Table 1). Thus, the species differ markedly in flower shape, flower color, nectar production, and, consequently, in pollinators sei'vicing the flowers. In the forma- tion of the six species, evolution appears to have responded to selection imposed by polli- nator preferences and ecological opportunities. The result is that members of the complex have radiated into a wide variety of different habitats and niches. A bright yellow-flowered mutant has appeared on the scene in this setting of polli- nator-driven, ecologically opportunistic evolu- tion. In two populations of M. cardinalis bright yellow-flowered morphs have become well established. One population is in the Siskiyou Mountains of Oregon, which is the northern limit of the range of M. cardinalis (Grant 1924). The other population is on Cedros Island, Baja California, and is at the southern limit of the species range. As Mayr (1976) suggests, new forms often evolve from isolated populations such as these on the periphery of a species 'The opening talk in the s\niposiuni, "Mechanisms of Speciation in Higher Plants," given 1 September 1993 at the XV International Botanical Congress, Yokohama, Japan. ^Biology Department, University of Utah, Salt Lake City, UT 84112 USA. 177 178 Great Basin Natur/VLIst [Volume 55 Table 1. Nectar production in the species of section Enjthranthe measured at OSOO h in the wild (Vickery- and Sutherland 1994). Averages are based on 20 or more mea- surements from a population representative of each species or variet>'. Species \olinne in Ml Cf sut^ar M. canlinalis 3.9 11.5 M. vcrbenaceus 6.7 5.8 M. nipestris 0.9 19.0 M. eastwoodiae 1.5 13.7 M. nelsonii 18.3 19.2 M. lewisii Roclcv Mountains 0.5 0.5 Sierra Ne\ada 0.7 11.3 range adjacent to new ecological opportuni- ties. A bright yellow-flowered morph of M. verhenaceus has appeared also and become well established in a population growing in an isolated spring area, Vasey's Paradise, at the bottom of the Grand Canyon of the Colorado River, AZ, that species' northwesteiTi limit. Flower colors in section Enjthranthe are due to various combinations of six anthocyanin pigments — three pelargonidins (apricot-pink) and three cyanidins (lavender-pink) — and at least one carotene pigment (Pollock et al. 1967). The lavender to magenta flowers of M. lewisii are due to various combinations of the pelar- gonidin and cyanidin anthocyanin pigments without die yellow carotene. Flowers of the red- flowered species have all or most of the six anthocyanin pigments plus the carotene pig- ment. Red color results from a visual blend of pink pigments and yellow pigment. Yellow- flowered plants have a pair of recessive genes at one locus that suppresses anthocyanin pro- duction (pink pigments), leaving just the yel- low carotene pigment showing. So, a sin^h' mutation, when homozygous, changes flower coh)rfrom red to yellow. If the change from red to yellow flowers leads to a change in pollinators, for example, from hummingbiids to bumblebees or hawk- moths, then the first major step in reproduc- tive isolation has been established by a single gene change (when homozygous)! Once repro- ductive isolation has been established by color differences, presumably selection would fine- tune it, e.g., by favoring more tubular flowers for hummingbird-pollinated flowers and by favoring a landing platform morphology and nectar guides for bee-pollinated flowers. Are pollinators required for seed set in Mimulus cardinalis or do the flowers self-polli- nate? To test these hvo questions, I used the fact that M. cardinalis flowers are borne in pairs. I grew plants of red- and of yellow-flowered M. cardinalis from Cedros Island in the green- house of the Biology Department, University of Utah. The greenhouse is free of pollinators. I carefulK hand-pollinated one flower of each of ten pairs of red flowers and of ten pairs of \'ellow flowers. The hand-pollinated flowers of both the red-flowered and \'ellow-flowered plants set moderate numbers of seeds per cap- sule (50-150), while the unpollinated flowers set no seeds at all. This finding corroborates my earlier observations on the Cedros Island M. cardinalis (Vickery 1990) that flowers do not self-pollinate and that pollinators are required for seed set. Are the rewards for pollinators the same in yellow flowers as in red? That is, do yellow flowers and red flowers produce equal volumes of nectar with the same concentrations of sug- ars? Red flowers of the Cedros Island M. car- dinalis produced an average (based on flowers from 30 greenhouse-grown plants) volume of 9.5 fx\ of nectar with 18.2% sugar Yellow flow- ers produced an average (based on measure- ments of flowers from 40 greenhouse-grown plants) of 10.9 (jl\ of nectar with 23.0% sugar. There is so much variation that these values are not significantly different. Finally the key c|uestion, do pollinators show a preference for red or yellow flowers? To study this question, I placed 24 red-flowered and 24 yellow-flowered plants in a random arrange- ment in a meadow in the Red Butte Can\'on Natural Area in the Wasatch Mountains be- hind the University of Utah and observed pol- linators that visited this experimental popula- tion. Pollinators that came were humming- birds and bumblebees, with rare visits from flies, but no hawkmoths or honey bees. Pollinators were observed for three 50-min periods on each of 5 d. On 28 July 1987 there were 55 hummingbird visits to the 39 red flowers present and 20 visits to the 35 yellow flowers. Chi-square = 14.379, p <' .001, which indicates a significant preference for red flowers. That da\' there were 10 bumble- bee visits to red flowers and 12 to yellow flow- ers. Chi-square = 0.1818, no significant pref- erence. On 31 July there were 176 humming- bird visits to the 42 red flowers in bloom that day in the population and 40 \'isits to the 21 yellow flowers. Chi-square = 70.246, p < .001, 1995] Notes 179 which indicates a significant preference for red. That day there were six l:)umhlebee visits to red and one to yellow. There were too few bumblebee visits for a meaningful .v- value to be calculated. The same pattern of three observation periods was continued on 2-4 August, but once again there were too few pol- linator visits to obtain meaningful .v^ values. Apparently, most hummingbirds had migrated south and there were few bumblebees all sea- son that year. On die first day of the experiment when the plants had just been placed in the meadow all pollinators would be naive for both red- and \ellow-flowered M. cordinalis plants inasmuch as Red Butte Canyon is hundreds of miles from the nearest M. cardinalis popula- tion in northern Arizona. Therefore, the highly significant preference for red appears to be real and not the result of learned behavior. Apparently, hummingbirds strongly preferred the red flowers but also visited the yellow flow- ers to some extent. The few bumblebee visits did not suggest a preference. Results show that the change in flower color from red to yellow did affect the frequencies of pollinator visits, but not in an all-or-none way that would immediately establish repro- ductive isolation. However, the change would probably be enough to initiate partial, incipi- ent reproductive isolation. Would M. verbenaceus with its normal red morph and mutant yellow morph produce the same reactions in pollinators? The flowers of M. verbenaceus differ from those of M. cardinalis in that only the upper two corolla lobes are reflexed, whereas all five of those of M. cardi- nalis are reflexed. Both species sometimes have wild populations with orange-red flowers instead of the typical red flowers. For the M. verbenaceus experiment, plants of red-flowered and yellow-flowered individuals from Vasey's Paradise in the Grand Canyon plus plants of an orange-red-flowered popula- tion from Yecora, Sonora, Mexico, were placed on a lawn by clumps of native Gambel oak at the mouth of Parley's Canyon, Salt Lake City, UT This location had an abundance of pollina- tors in contrast to the paucity of pollinators in the Red Butte Canyon meadow used previ- ously. The test population was observed for 15 periods of 1 h each at different times of day from 26 July through 8 August 1988. On aver- age, there were 73 red flowers, 87 orange flow- ers, and 136 vellow flowers (see Vicken' 1990 for daily details of numbers and chi-square calculations). On average, bumblebees visited them 24, 56, and 128 times, respectively; and hummingbirds 43, 98, and 52 times, respective- ly (Vickery 1990). Bumblebees significantly eschewed red and orange flowers and prefer- entially visited yellow flowers. Hummingbirds significantly preferred orange, visited red flowers in proportion to their fi-equency in the population, and significantly eschewed yellow flowers. Results for M. verbenaceus are much clearer than those for M. cardinalis. There is a definite preference for yellow by bumblebees and a clear avoidance of yellow by humming- birds. Thus, this color change has lead to sig- nificant, partial isolation between the normal orange- and red-flowered morphs and the yel- low-flowered mutant morph under the condi- tions of this experiment. Would M. cardinalis react like M. verbe- naceus in the better experimental locality at the mouth of Parley's Canyon? To probe this question, I added red-, orange-, and yellow- flowered morphs of M. cardinalis to the M. verbenaceus red-, orange-, and yellow-flow- ered moiphs of the previous experiment. The new experiment was run 8-17 August 1988, with the population being observed for 15 periods of 1 h each at different times of day. On average there were 61 red, 57 orange, and 22 yellow flowers of M. cardinalis (see Vickeiy 1990 for day-to-day numbers and chi-square calculations). On average, bumblebees visited them 28, 30, and 29 times, respectively, and hummingbirds 59, 60, and 6 times, respectively. Bumblebees eschewed red and orange flowers and significantly preferred yellow flowers despite their low numbers in the population. Hummingbirds significantly eschewed yellow flowers and preferentially visited orange flow- ers. M. verbenaceus plants were run again at this time with M. cardinalis plants and exhibit- ed the same attractiveness or lack of attrac- tiveness to the pollinators as before. The pres- ence of M. cardinalis flowers did not alter pol- linator response to M. verbenaceus flowers. The color shift from red (or orange) to yellow leads to marked, partial reproductive isolation in M. verbenaceus as well as in M. cardinalis. How effective is the partial reproductive isolation? To test this, I placed 198 plants of M. verbenaceus — one-sixth yellow-flowered and five-sixdis red-flowered to simulate a popu- lation with a well-established mutant — in four 180 Great Basin Natuiulist [Volume 55 experimental areas: the experimental garden on the University of Utah campus. Red Butte Canyon Natural Area, the mouth of Parley's Canyon, and at Silver Fork, Big Cottonwood Canyon, Salt Lake County, UT. 1 harvested seeds of each plant and planted seeds han'est- ed from 20 yellow-flowered plants and grew them to flowering. If pollinators were visiting the flowers at random, then they should pick up and cany pollen from red flowers five times more often than pollen from yellow flowers. Pollen loads and resulting seed sets were well below the 500-1500 seeds per capsule that may occur in M. verbenaceus. So, results were not skewed by saturation of the stigma. Also, assuming all else to be neutral such as relative growth rates of yellow- and red-pollen tubes, speed of flowering of red- and yellow-flow- ered plants, randomness of placement of red- and yellow-flowered plants, and sample size of red- and yellow-flowered plants, then the expected five-to-one visitation rate should hold. Inasmuch as red is genetically dominant to yellow, then five-sixths of the seedlings should be red-flowered and one-sixth yellow- flowered; that is, of the 214 seedlings grown, 178 should be red-flowered and 36 yellow- flowered. In fact, there were 86 red-flowered seedlings and 128 yellow-flowered seedlings. The ratio is 2 red to 3 yellow flowers, which is far from the expected ratio of 5 red flowers to 1 yellow flower. This suggests considerable pollinator faithfulness to one color or the other. However, in addition to pollinator faith- fulness there could be self-pollination. Mimulus cardinalis does not self-pollinate but M. verbenaceus does at the average rate of 10 seeds per capsule. Average normal seed set is 110 seeds per capsule. Therefore self-pollina- tion would account for 9% of the yellow-flow- ered seedlings; i.e., 9% of the 214 seedlings, or 19 seedlings, would be expected to be yellow- flowered as a result of self-pollination. Of the remaining 195 seedlings, five-sixths, or 162, would be expected to be red, and one-sixth, or 33, would be expected to be yellow. Therefore, I should expect to observe 162 red-flowered seedlings and 52, i.e., 33 + 19 (the results of self-pollination), yellow-flowered seedlings instead of the 86 red-flowered and 128 yellow- flowered seedlings actually obsened. This is a highly significant difference (.r^ = 146.730, p < .0001) and greatly strengthens the point of pollinator faithfulness. Clearly, pollinator preference for yellow and faithfulness to yel- low are having a large effect, though not an all-or-none effect. We are seeing strong incipi- ent reproductive isolation due to color change. In different areas with different conditions and different guilds of pollinators the effect might be less or might be stronger, even lead- ing eventually to effective reproductive isola- tion and speciation. Acknowledgments I appreciate the financial support of the U.S. National Science Foundation, Grant BSR- 8306997. I thank Dr. Stephen Sutherland for nectar measurements and for carrying out the Red Butte Canyon experiment on M. cardinalis. Literature Cited Crow, J. E, and M. Kimura. 1970. An introduction to population genetics theory. Harper & Row, New York, NY. 591 pp. DeWet, J. M. J. 1980. Origins of polyploids. Pages 3-15 in W. H. Lewis, ed.. Polyploidy. Plenum Press, O.xford. Grant, A. L. 1924. A monograph of the genus Mimulus. Annals of the Missouri Botanical Gardens 11; 99-389. Levin, D. A. 1993. Local speciation in plants: the rule not the exception. Systematic Botany 18: 197-208. Mayr, E. 1976. Evolution and the diversity of life. Harvard University Press, Cambridge, MA. 721 pp. Pollock, H. G., R. K. Vickery, Jr., and K. G. Wilson. 1967. Flavonoid pigments in Mimulus cardinalis and its related species. I. Anthocyanins. American Joumiil of Botany 54: 695-701. Vickery, R. K., Jr. 1978. Case studies in the e\ olution of species complexes in Mimulus. Evolutionai-y Biology 11:404-506 . 1990. Pollination experiments in the Mimulus car- (linalis-M. Icwisii complex. Great Basin Naturalist 50: 153-159. Vickery, R. K., Jr. and D. Sutheri^vnd. 1994. Variance and replenishment of nectar in wild and greenhouse populations of Mimulus. Great Basin Naturalist 54: 212-227. Received 6 July 1994 Accepted 27 September 1994 Great Basin Naturalist 55(2), © 1995, pp. 181-182 FALL LAMB PRODUCTION BY A CALIFORNIA BIGHORN SHEEP Matthew McCoyl, Wcilt Bodie^, and EiRoy Taylor^ Key words: paiiiirition. California bighorn sheep, Ovis canadensis, Idaho. Parturition is timed to maximize sui^vival of offspring (Thompson and Turner 1982). Parturition occurring outside an optimum time period lowers reproductive fitness and, there- fore, should be selected against. Timing of parturition in bighorn sheep {Ovis canadensis) has been related to resource abundance (Geist 1974, Bunnell 1982, Thompson and Turner 1982, Risenhoover and Bailey 1988) and climat- ic conditions (Stewart 1982). Parturition varies by latitude between subspecies (Thompson and Turner 1982) and by elevation within sub- species (Risenhoover and Bailey 1988). Peak- lambing periods occur in March for desert bighorn sheep (O. c. nelsoni; Hanson 1960, Sandoval 1980, Witham 1983), May for Cali- fornia bighorn sheep (O. c. californiana; Jones 1950), and early June for Rocky Mountain bighorn sheep (O. c. canadensis; Bunnell 1982, Thompson and Turner 1982). Unusual lambing periods such as January for desert bighorn (Russo 1956, Welles and Welles 1961) and July for Rocky Mountain bighorn (Stewart 1982) were attributed to extremes in climatic condi- tions and elevations. Vegetation in the Big Jacks Creek drainage, Owyhee County, ID, is dominated by sage- brush {Artemisia spp.), bluebunch wheatgrass {Pseudoroegneria spicata), and Sandberg blue- grass {Poa sandbergii). Climatic conditions are characterized by warm, dry summers and cool winters. Total precipitation from November 1988 through March 1989 was equal to the 10- year average; however, most precipitation occurred as rain in November and March. Ten ewes, one ewe lamb, and three ram lambs from Chilcotin, BC, and two rams from East Fork Owyhee River, ID, were transplanted to Big Jacks Creek during February and March 1988. Four ewes were fitted with radio-collars and periodically located to monitor their movements and status. Three radio-collared ewes were observed with lambs in May 1988. A fourth collared ewe (34) was obsei-ved with a lamb (A) less than two weeks old (based on size and behavior) on 26 October 1988. An average gestation period of 174 days for bighorn sheep (Shackleton et al. 1984) indicated conception occurred about 25 April 1988. Ewe 34 and lamb A were located monthly through March 1989. During 1987-1990, onset of parturition occurred from 11 April to 3 May, and mating activity was observed between October and December in an adjacent drainage. The birth of lamb A occurred approximately six months out of cycle. Ewe 34 was observed 4 January 1990 with a lamb (B) that appeared to have been born during the normal lambing period (April-June 1989). Lamb B was conceived be- tween October and December 1988 while ewe 34 was nursing lamb A. Ewe 34 may not have bred in 1987, or stress related to transplanting may have caused her to abort. Stress can affect any aspect of repro- duction (DeForge 1976). Contact with rams during March and April 1988 may have caused ewe 34 to come into estrus. Presence of males has been found to induce estrus in female merino sheep (Watson and Radford 1960) and feral goats (Coblentz 1980). Recurrent estrus was observed in a cow elk {Cervus elaphiis) that was associated with bulls but not bred during previous estrus periods (Morrison 1960). Lamb survival has been related to forage quality (Wehausen et al. 1987, Festa-Bianchet ^Idaho Department of Fish and Game, 3101 S. Powerline Road, Nampa, ID 83686. Present address: Bureau of Land Management, 3948 Development Avenue, Boise, ID 83705. 2ldaho Department of Fish and Game, 3101 S. Powerhne Road, Nampa, ID 83686. •^Bureau of Land Management, 3948 Development Avenue, Boise, ID 83705. 181 182 Gkeat Basin Natuk.\list [Volume 5.' 1988a), precipitation patterns (as they affect plant growth; Douglas and Leslie 1986), popu- lation density (Douglas and Leslie 1986), and mother's age (Festa-Iiianchet 1988a). Cheatgrass hrome {Broiniis tectormn) seedlings were avail- able in November, and Sandberg bluegrass greenup was observed in Januaiy South-facing slopes were generally free of snow soon after storms. Cattle grazing occurred in riparian areas and on plateaus adjacent to drainages, areas that received limited use by bighorns during summer and lambing periods. Bighoin and mule deer {Odocuileiis hemionus) popula- tions were at low densities. Competition for forage was probably not a limiting factor. Festa-Bianchet (1988b) reported that lambs born to ewes four to nine years old had signifi- cantly higher survival rates than those born to two- to three-year-old ewes. Ewe 34 was esti- mated to be five years old in 1988. Mild cli- matic conditions, availability of green forage during weaning, limited competition for for- age, and probable previous lambing experi- ence allowed ewe 34 to raise an out-of-season lamb and survive concurrent fall/winter lacta- tion and gestation periods. This observation suggests that under favorable conditions bighorn sheep may be able to successfully reproduce outside generally observed repro- ductive periods. Literature Cited Bunnell, E L. 1982. The lambing period of mountain sheep: synthesis, hypotheses, and tests. Canadian Journal of Zoology 560: 1-14. CoBLENTZ, B. E. 1980. A unique ungulate breeding pat- tern. Journal of Wildlife Management 44: 929-933. Deforce, J. R. 1976. Stress: Is it limiting bighorn? Desert Bighorn Council Transactions 19: 30-31. Douglas, C. L., and D. M. Leslie. 1986. Influence of weather and density on lamb sur\i\al of desert mountain sheep. Journal of Wildlife Management 50: 153-156. Eesta-Bianchet, M. 19S8a. Nursing behavior of bighorn sheep: correlates of ewe age, parasitism, lamb age, birthdate and sex. Animal Behavior 36:1445-1454. . 1988b. Age-specific reproduction of bighorn ewes in Aliierta, Canada, Journal of Mammalogy 69: 157-160. Geist, V. 1974. On the relationship of ecology and behav- ior in the evolution of ungulates. Pages 235-246 in \. Geist and E Walthers, editors. The behavior of imgu- lates and its relation to management. International Union Consenation Nature Publication. Hanson, G. 1960. Lamb survival on the Desert Game Range. Desert Bigliom Council Transactions 4: 60-61. Jones, E L. 19.50. A survey of the Sierra Nevada bighorn. Pages 29-76 in Sierra Club Bulletin 1950. .MoRRLSON, J. A. 1960. Characteristics of estrus in captive elk. Behaviour 16: 84-92. Risenhoover, K. L., AND J. A. Bailey. 1988. Growth- rates and birthing period of bighorn sheep in low- elevation environments in Colorado. Journal of Mammalogy 69: 592-597. RUSSO, J. P 1956. The desert bighorn sheep in Arizona. Arizona Game and Fish Department, Wildlife Bulletin 1. Sando\'AL, a. V 1980. Preferred habitat of desert bighorn sheep in the San Andres Mountains, New Mexico. Unpublished thesis, Colorado State University, Fort Collins. 282 pp. SlIACKLETON, D. M., R. G. PETERSON, J. HA'iAVOOD, AND A. Bottrell. 1984. Gestation period in Ovis canaden- sis. Journal of Mammalogy 65: ■337-.338. Stewart, S. T. 1982. Late parturition in bighorn sheep. Journal of Mammalogx' 63: 154-1.55. Thompson, R. W, .\nd J. C. Turner. 1982. Temporal geo- graphic variation in the lambing season of bighorn sheep. Canadian Journal of Zoology 60: 1781-1793. W.^TSON, R. H., and H. M. R\DF()RD. 1960. Influence of rams on the onset of oestrous in merino ewes in the spring. Australian Joimial of Agricultural Research 2: 65-71. Wehausen, J. D., V. C. Bleicii, B. Bloxg, and T. L. RUSSI. 1987. Recruitment dynamics in a southern California mountain sheep population. Joinnal of Wildlife Management 51: 86-98. Welles, R. E., and F B. Welles. 1961. The bighorn of Death Valley. Fauna of the National Parks of the United States, Fauna Series 6. 242 pp. WiTHAM, J. B. 1983. Desert bighorn sheep in southwest- ern Arizona. Unpublished dissertation, Colorado State University, Fort Collins. 93 pp. Received 22 November 1993 Accepted 20 June 1994 Great Basin Naturalist 55(2), © 1995, pp. 183-187 AGE, GROWTH, AND REPRODUCTION OF LEATHERSIDE CHUB {GILA COPEI) Jerald B. Johnsonl-^, Mark C. Belk'i, and Dennis K. Shiozawa^ Key words: Gila copei, leathersicle chub, life history, reproduction, age. growth. The leatherside chub {Gila copei) is a small cyprinid native to eastern and southern areas of the Bonneville Basin of Utah, Idaho, and Wyoming, to Wood River, Idaho, and to regions of the Snake River, Idaho and Wyoming, above Shoshone Falls (Baxter and Simon 1970, Simpson and Wallace 1982, Sigler and Sigler 1987). Gila copei is currently listed as a can- didate for federal protection under the Endangered Species Act. Conservation efforts for G. copei would benefit from accurate life histoiy data, yet the life history of G. copei is not well known. This species was thought to live less than five years (Sigler and Sigler 1987). Based on bright col- oration of males and abdominal distension in females, Sigler and Miller (1963) concluded G. copei spawns between lune and August. Using similar coloration criteria on males from Sulphur Creek, WY, Baxter and Simon (1970) suggested breeding occurred in late summer; Simon (1951) found females distended with eggs in early August. Other than these few obser\'ations, no data on age, growth, or repro- duction are available. We present data on age, growth, and repro- duction of leatherside chub from central Utah; these data were generated as a first step to understanding and protecting this potentially threatened, endemic fish species. Study Site Age and growth data were obtained from 36 G. copei collected from Thistle and Main creeks, both tributaries to larger rivers that flow into Utah Lake. Thistle Creek, a tributaiy to Spanish Fork River, was sampled in May, September, and October 1993 {n = 25) (39°52'N, lir32'W) at an elevation of approx- imately 1700 m. Main Creek flows directly into Deer Creek Reservoir (an impoundment of Provo River) and was sampled in Inly 1993 (n =11) 500 m upstream from the reservoir (40°24'N, lir28'W) at an elevation of 1650 m. Chubs used for determining reproductive pat- terns (below) were collected from the latter site in 1978-79. Creeks at both locations flow slowly at low gradient through meadows. The Main Creek site is downstream from beaver dams; tall grasses and small trees grow along banks. Collections for both creeks were made from vegetated pools separated by shallow riffles; stream substrate is silt, gravel, and boulders. Materials and Methods Because Gila copei is a species of special concern, our permit was limited to 40 speci- mens, and care was taken to collect the entire size range. Following collection by elec- troshocking, fish were placed on ice and trans- ported to Brigham Young University (BYU) where they were stored frozen. Individuals were then thawed, rinsed in water, blotted diy, and weighed (0.001 g) on a Denver Instmment XD-1200D® electronic balance; standard length (SL) was determined (0.01 mm) using Fowler Ultra Cal III® electronic calipers. Ages were determined by grinding otoliths (lapilus) to a thin section and counting opaque bands under a Leica dissecting microscope (40X). Opaque bands were validated as annuli using a marginal increment analysis; because juvenile (ages 1-2) and adult fish (ages 3-8) demonstrated distinct growth rates, they were evaluated separately. Identification of annuli was facilitated by generating digitized images of otoliths on a video monitor using a Hitachi 'Department of Zoolog>'. Brigham Young University', Provo, UT 84602. ^Address correspondence to this author 183 184 Great Basin Natufl^list [Volume 55 CCTV® camera fitted to a Heerliru^g Wild® dissecting microscope. Annual growth incre- ments along the longest axis of the otolith were then measured (0.001 mm) using video image analysis software (Mocha release 1.0, Jandel Scientific; Rundel 1993), which reduces measurement errors introduced when reading otoliths directK under a microscope (McGowan et al. 1987). Size at age was back-calculated from otolith measurements using a modified Fraser-Lee formula (Campana 1990): L, = L„ + (L,, - LJ(R, - R„)/(R, - R„), where L,. is estimated SL at age x, L^. is length at capture, R^ is otolith radius at age x, and R^. is otolith radius at capture. L^ is estimated length at swim-up (estimated at 4 mm from data on Gila atraria; Varley and Livesay 1976), and R„ is otolith radius at swim-up (estimated from otoliths at 0.01 mm). There was no significant difference in back- calculated lengths at age I between Main and Thistle creek chubs (Main Creek, n = 11; Thistle Creek, n = 25; T = 1.96, d.f = 34, P = .06) . Numbers of age II (n = 2) and age III (n = 3) fish from Main Creek precluded statis- tical comparisons; however, back-calculated lengths at age II and age III for Main Creek fish were within the range of comparably aged fish from Thisde Creek. Hence, growth data for the two populations were combined. An age- growth curve was generated for the combined samples by averaging back-calculated sizes at age. Leatherside chub collected in 1993 were sexed by dissection and examination of gonads; individuals lacking mature gonads were classi- fied as juveniles. Immature testes were trans- lucent and threadlike, while mature testes were opaque (white or pinkish) and firm. Reproductive states of ovaries were deter- mined according to criteria in Holden and Berry (1983); immature ovaries were small, translucent, and lacked yolked ova; mature ovaries were larger and contained both imma- ture ova and firm, yolked ova. Reproductive data were obtained from a collection of 176 adult leatherside chubs archived in the Monte L. Bean Museum at BYU (#5592-5619, 5629-5686, 5688-5775). Monthly collections from Main Creek (August 1978 to September 1979) were made using minnow traps, hand nets, and electrofishing gear, and preserved in formalin. Daily temper- ature was recorded from September 1978 to July 1979. Standard length was measured (mm), and presei-ved wet mass (0.01 g) was recorded, for each specimen. Gonads from all {n = 176) individuals >50 mm SL were removed and weighed (0.001 g). No fish <50 mm SL had en- larged gonads. A gonadosomatic index (GSI) was generated for each fish using the follow- ing formula (Andreasen and Barnes 1975): GSI = (gonad weight / body weight) X 100. Mean monthly GSI values were used to deter- mine onset and duration of spawning. Ova counts were made on nine fish collected in May 1979. The relationship between number of ova present and SL was evaluated by linear regression. Results Opaque bands on leatherside chub otoliths appear to be valid annuli as demonstrated by an increase in the marginal growth increment throughout the growing season for both adult and juvenile fish (Fig. 1). U.iiO - ■ = Ages 1-2 E^ E 0.20 - • = Ages 3-8 \ /* c 1 r^ 0) r E 0.15 1 ^ r - o / r - 0.10 - CD / J - C / CD i f TO 0 05 « / ^ 0.00 - » { - 1 1 1 1 1 10 11 Month Fig. 1. Mean inarginal increment widths (±2 S.E.) mea- sured from otoliths in Gila copei (n = 36). Immature age classes (1-2) and mature age classes (3-8) plotted sepa- rately. 1995] Notes 185 Ages of 36 G. copei collected in 1993 ranged from one to eight years, with SL of 38-110 mm (Table 1). Chubs grew rapidly to -58 mm SL at about age II (Fig. 2, Table 1). From age II on, annual growth was slower and fairly uniform. Mean GSI values for males and females (Fig. 3a) were highest for both sexes in spring with maxima in May (female GSI = 12.3, male GSI = 2.7). Increasing water tem- peratures from Januaiy through May (Fig. 3b) were associated with increased GSI values for both sexes. Average water temperature in May, corresponding to GSI maxima, was 9.4° G. Fecundity (as measured by ovimi counts) in- creased with SL for females collected in May 1979 and ranged from 938 in a 67-mm-SL, 5.9-g female to 2573 in a 92-mm-SL, 14.6-g female. Average count for leatherside chubs collected in May 1979 was 1813. Significant correlations existed between SL and fecundity (R2 = .82, P < .05, n = 9) and weight and fecundity (fi2 = .72, P < 0.05, n = 9). Discussion A maximal age of eight years in our sample of G. copei indicates a life span much longer than previously thought (Sigler and Sigler 1987). Longevity in G. copei may be a life his- toiy trait that has evolved in response to living Table 1. Capture and back-calculated standai^d lengths (SL) of Gila copei from Thistle and Main creeks, central Utah. N SLat capture Mean back-calculated SL at annulus Age Mean Range 1 2 3 4 5 6 7 8 1 S 44 38-49 32 2 9 76 65-85 41 65 3 2 87 71-104 42 68 82 4 1 85 — 51 65 77 83 5 1 97 — 38 53 70 87 96 6 7 92 83-110 35 52 63 73 82 89 7 7 94 88-105 36 54 66 73 80 86 91 8 1 96 — 31 46 54 62 67 76 86 93 Overall means 37 58 67 74 81 87 90 93 100 r c CO ■o c Oi Fig. 2. Mean back-calculated standard lengths at age (±2 S.E.) for Gila copei {ii indicates estimated age at first reproduction. 36) in central Utah. Shaded block 186 GiucAT Basin Naturalist [Volume 55 a 18 1 I I r A S 0 N D J F M A M Month J J A JASONDJ FMAMJ JA Month Fig. 3a. Mean gonadosomatic indices (±2 S.E.) for male and female Gihi copei (n = 176); b, mean monthly temper- atures (±2 S.E.) from August 1978 to July 1979 in Main Creek, Wasatch County, UT. in an environment where annual precipitation and stream flow vaiy considerably. Successful chub reproduction and recruitment may be uncertain in any given year. An extended life span would increase the likelihood that appro- priate environmental conditions for reproduc- tive success would be met at some time in an individual's life; thus, longevity may be a "bet- hedging" strategy (Stearns 1976) for living in unpredictable conditions. The growth pattern of G. copei is typical of other fishes in which rapid juvenile growth decreases at the onset of sexual maturity as finite energy resources are allocated to both growth and reproduction (Roff 1984). The inflection point in the growth curve (Fig. 1), coupled with the facts that the smallest fish with dexeloped gonads collected in 1993 was 65.2 mm SL and no fish in the museum collec- tion <50 mm SL had enlarged gonads, sug- gests that first reproduction in G. copei occurs at age II. High OS I in May followed by decreased GSI in June and minimal values in July and August (Fig. 3a) indicates that peak spawning occurred in May, with some activity possibly extending into early June. Gila copei appar- entl)' follows a pattern of reproduction common to various cyprinids living in temperate climates (Munro et al. 1990). This pattern is character- ized by the onset of spawning in late spring followed by a period of gonadal recrudescence and inactivity; size of gonads begins to increase in autumn and continues through winter, with final maturation occurring in early spring. If temperature influences the onset of spawning, differences in temperature (as a function of latitude) between Main Creek (this study) and southwestern Wyoming (Simon 1951) could explain the discrepancy between onset of chub spawning at these locations (May vs. August). A more detailed investigation of G. copei will be required to resolve questions of differences in reproductive and life histoiy characteristics among populations. Literature Cited Andreasen, J. D., .WD J. R. B.\rnes. 1975. Reproductive life histoiy oi Catostomu.s aniens and C. discobolus in the Weber River, Utah. Copeia 1975; 643-648. Baxter, G. T, and J. R. Si.vion. 1970. Wyoming fishes. Wyoming Game and Fish Department, Cheyenne. 168 pp. Camfana, S. E. 1990. How reliable are growth back-calcu- lations based on otoliths? Canadian Journal of Fish- eries and Aquatic Science 47; 2219-2227. HOLDEN, M. A., AND C. R. Berry. 1983. Vitellogenesis in the Utah chub {Gila atraria) and its use in evaluating reproduction in a transferred population. Encyclia 60: 32-42. McGowAN, W E, E. D. Prince, and D. W Lee. 1987. An ine.vpensive microcomputer-based system for making rapid and precise counts and measurements of zona- tion in \'ideo displa\ed skeletal structures in fish. Pages 385-395 in R. C. Summerfelt and G. E. Hall, editors. Age and growth of fish. Iowa State Univer- sity Press, Ames. Munro, A. D., A. R Scott, and T J. Lam. 1990. Repro- ductive seasonality in teleosts; environmental influ- ences. CRC Press, Inc., Boca Raton, FL. 254 pp. Roff, D. A. 1984. The evolution of life histoiy parameters in teleosts. Canadian Journal of Fisheries and Aquatic Science 41: 989-1000. Rundel, R. 1993. Mocha image analysis software: user's manual. Jandel Scientific, San Raliiel, CA. 189 pp. 1995] Notes 187 SiGLER, W. E, AND R. R. MiLLER. 1963. Fishes of Utah. Utah Department of Fish and Game, Sah Lake Cit\'. 203 pp. SiGLER, W. F, AND J. W. SiGLER. 1987. Fishes of the Great Basin: a natural histor>'. University of Nevada Press, Reno. 425 pp. Simon, J. R. 1951. Wyoming fishes. Wyoming Game and Fish Department, Cheyenne. 129 pp. Simpson, J. C., and R. L. W.all.\ce. 1982. Fishes of Idaho. University Press of Idaho, Moscow. 238 pp. Stearns, S. C. 1976. Life history tactics: a review of the ideas. Quarterly Review of Biology 51: 3—47. Varley, J. D., AND J. C. Ll\ESAY. 1976. Utah ecology and life history of the Utah chub, Gila atraria, in Flaming Gorge reservoir, Utah-Wyoming. Utah Division of Wildlife Resources, Publication 76-16, Salt Lake Cit\'. 29 pp. Received 15 June 1994 Accepted 7 September 1994 Great Basin Naturalist 55(2), © 1995, pp. 188-191 CONSUMPTION OF A TOXIC PLANT {ZIGADENUS PANICULATUS) BY MULE DEER William S. Longland'-^ and Charlie Clements^ Key words: death eamas. Zitiadfiius, iiinle deer, poisonous plants, cocrolution. The abundance of green vegetation in nature can yield false impressions of the availability of food resources to herbivores because many plants have evolved anti-herbivore defenses. Defensive mechanisms commonly include incorporation of distasteful or toxic secondary chemical compounds into plant tissues. Effects of different compounds on consumers range from mild (unpalatable) to severe (illness or death fi-om poisoning). Herbivores have conse- quently evolved a host of means for coping with defensive compounds, resulting in an evolu- tionary arms race between plants and herbi- vores (Freeland and Janzen 1974). Although evidence of plant/herbivore coevolution can be found for herbivores ranging from phyto- phagous insects to mega-vertebrates, we con- centrate specifically on mule deer {Odocoileus hemoniiis) feeding on toxic plants. Because domestic grazing animals lack a coevolutionaiy history with the plant commu- nities in which they forage, they are often affected by toxic secondary compounds to a greater degree than native herbivores. This has significant economic impact on the range livestock industry due to direct losses, such as death, reduced fecundity, or reduced weight gain, and to indirect costs of minimizing such losses (Nielsen et al. 1988, James et al. 1992). Historical familiarity with local plant assem- blages has provided herbivores foraging in their native ranges with two advantages over introduced domestic counterparts (Freeland and Janzen 1974, Laycock 1978, Laycock et al. 1988). First, native mammals often avoid eating toxic plant species that are eaten by domestic grazers. For example, toxic plants eaten by livestock, such as azalea {Azalea spp.) and lark- spur {Delphinium spp.), are avoided by mule deer even when these plants are abundant (Dixon 1934). Second, in most cases of native ungulates eating a plant species that is toxic to domestic animals, the plant does not produce noticeable toxic effects in the fomier, indicat- ing that native herbivores may possess detoxi- fication mechanisms for some plant toxins (Lay- cock 1978). Thus, deer consume without adverse effects a variety of plants poisonous to live- stock (Stoddart and Rasmussen 1945, Dean and Winward 1974). Reciprocal examples in which native plants are toxic to native herbi- vores, but benign to domestic animals, are lacking in the literature. Herein we report on four years of obsei-va- tions of an eastern Sierra Nevada mule deer herd feeding on substantial quantities of foot- hill death camas {Zigadenus paniculatus), a lilia- ceous bulb plant that is toxic to domestic sheep, cattle, and horses (Fleming et al. 1921, Kings- buiy 1964, James et al. 1980, Panter et al. 1987). The genus Zigadenus includes several species, all containing toxic steroidal alkaloids (James et al. 1980). Death camas emerges earlier than most plants, making it particularly hazardous for spring grazing of livestock (Panter and James 1989). These plants have been variously described as "the most important poisonous plants in the western U.S." (Kingsbury 1964) and "the most dangerous poisonous plants in North America" (Clarke and Clarke 1975). Foothill death camas has been described as one of the more toxic Zigadenus species (Kingsbuiy 1964, James et ah 1980). Our study site is located at T20N, R18E, S36 just west of Reno, NV, on an alluvial fan at the southern base of Peavine Mountain. Woody vegetation is dominated b\' basin big sage- brush {Aiiemisia thdentata tridentata) and bit- terbrush {Purshia tridentata). Death camas emerges at this site in mid- March, flowers in lUSDA, Agricultural Research Service, 920 Valle\ Koad, Reno, NV 89512. ^Address correspondence to this author. 188 1995] Notes 189 April, and remains green into May. A herd of mule deer, usually numbering 20-25 animals, has foraged extensively in this area from October to May since we began making obser- vations in fall 1988. We first noticed deer consuming death camas on 28 March 1989 (before plants flow- ered) and confirmed this with additional obser- vations in all subsequent years. Examination of death camas foliage immediately after deer left the foraging patches consistently revealed fresh herbivore damage. We found that deer herbivoiy left a characteristic leaf damage pat- tern, with most or all leaves of a foraged plant cleanly bitten off perpendicular to their long axes. In addition to direct obsei^vations of deer consuming death camas, fresh deer pellet groups were found in patches of plants ex- hibiting this characteristic damage pattern during all five springs (1989-1993). During observation periods we found no evidence of deer exhibiting toxic effects from death camas consumption, and neither we nor personnel from the Nevada Department of Wildlife (which surveys deer in the area by air) have found any fresh deer carcasses in the vicinity. Each year from 1990 through 1993 we walked 10-12 permanently located, parallel transects and categorized all death camas plants seen as either eaten or uneaten by deer Transects were 500 m long, 20 m wide (i.e., we generally saw all plants occurring < 10 m from the transect lines), and spaced 30 m apart. Usually, deer removed only the distal 2-5 cm of leaves, but on several occasions we found plants eaten to within 2 cm of ground level. Plants were considered eaten regardless of the amount of leaf removed. We tested these data for temporal differences in frequency of death camas consumption by comparing numbers of eaten versus uneaten plants among the four years of the study using a G-test of indepen- dence. We similarly tested for spatial effects on consumption by comparing eaten versus uneaten plant counts among individual tran- sect lines within years. There are at least two potential explana- tions for the partial consumption of leaves that we noted. Perhaps ends of leaves are less toxic than leaf bases, and deer preferentially con- sume less-toxic plant parts. Kingsbuiy (1964) suggests that death camas bulbs are the most toxic part of the plants, and a gradient of decreasing toxicity could occur from bulbs to ends of leaves. Alternately, deer may occasion- ally sample plants in their environment (Free- land and Janzen 1974), and removal of short leaf segments may represent cautious sam- pling of a plant deer find undesirable. The lat- ter possibility (sampling) seems less likely than the former (selectivity) because we have observed individual deer feeding on several death camas plants consecutively. Furthermore, total numbers of plants consumed on our tran- sects were several orders of magnitude greater than the number of deer foraging in the study area, and it seems unlikely that deer would have to sample repeatedly so many plants to discover they are undesirable. We found significant annual variation in the frequency of death camas consumption, rang- ing from 3.8% to 18.9% of total plants counted showing evidence of deer herbivory (G = 232.8, df = 3, F < .0001; Table 1). Maximum and minimum percentages of plants eaten (Table 1) illustrate that frequency of herbivory also varied spatially; in each of the four years we sampled there was significant variation among transects in numbers of plants eaten (F < .001 for all years). While the minorit)' of plants in the local death camas population were eaten, the values in Table 1 also represent a surprisingly high frequency of herbivory on a plant species with such a notorious reputation. The relatively low proportions of damaged plants indicate that deer may be selective for particular death camas plants. This is support- ed by the fact that deer generally ate only a few non-neighboring plants from large patches of death camas; rarely did the majority of plants within a patch show evidence of her- bivory. The apparently selective use of indi- vidual death camas plants, significant tempo- ral and spatial variation in death camas use, and infrequent extensive herbivory on small patches of plants could be due to variation among plants or patches in toxicity or to dif- fering availabilities of superior foods leading to variation in the use of toxic foods. Our observations suggest that death camas is more palatable to deer than to domestic cattle or sheep. Domestic animals must be force-fed death camas in captivity experiments (Fleming 1918, Fleming et al. 1921, Panter et al. 1987) and must be stressed or left with few alterna- tive foods in nature before they consume it (Panter et al. 1987). Mule deer at our study site, however, occur at a low density and consume 190 Great Basin Naturalist [Volume 55 Table 1. Numbers and percentages of foothill death canuis plants consiinucl in nuile deer along 500-ni transects, 1990-1993, at Peavine Mountain (Washoe County, NV). Xuniher of Number of plants" Pla nts eaten per transect {%) Yeai- transects Total f':aten M axinunii M inimuin XtS.D. 1990 12 2646 501 29.0 7.3 18.6 ± 10.6 1991 12 2726 259 44.7 2.6 16.4 ±13.4 1992 10 3073 118 32.6 1.6 8.3 ± 9.6 1993 10 3799 202 15.4 2.5 8.0 ± 4.0 ■'huliiilcs ccinihini-d data from all tiansccts death camas each spring akliougli alternative plants are available. Because bitterness is a general property of alkaloids (Laycock 1978), death camas is quite bitter Most herbivores apparently find bitterness distasteful (Laycock 1978); howexer, bitterbrush {Purshia tridentata), which is named for its bitterness, is a pre- ferred browse plant of mule deer. Although bitterbrush is also consumed by domestic ungulates, it is not highly preferred by them, perhaps because bitterness is a greater feed- ing deterrent to domestic animals than to deer Native herbivores have been observed con- suming a variety of plant species known to be toxic to domestic herbivores (Laycock 1978), including an anecdotal report of mule deer in Utah consuming death camas and several other toxic plants (Stoddart and Rasmussen 1945). Recent work stimulates the interesting possi- bility that herbivores consume specific toxic plants to rid themselves of gut parasites (Barbosa et al. 1991, Gauld and Gaston 1992). However, this hypothesis only addresses why toxic plants are consumed lather than wh\' the consumers are physiologically able to tolerate the toxins. Although we can only speculate about reasons mule deer are less affected by death camas toxicity than domestic ruminants, a likely explanation is that deer possess rumen microflora that have acquired the ability through natural selection to detoxify this plant (Freeland and Janzen 1974, Laycock 1978). Such selection is perhaps to be expected for native ruminants because the microflora com- mimity has seen prolonged exposine to native toxic plants. It is certainly possible, however, that deer are able to detoxify death camas by some other mechanism. For example, since deer are browsers, their diets include large amounts of tannins (Cooper and Owen-Smith 1985, Bobbins et al. 1987) that may precipitate the alkaloids in death camas into a harmless tannate (Freeland and Janzen 1974). Because even limited past exposure of a herbivore to a particular toxin can result in reduced toxic effects, selection for detoxifying rumen microflora may also account for intra- specific variation in toxicity among individuals of a domestic species. Such individual \'aria- tion in susceptibility to death camas toxicity has been reported in force -feeding experi- ments with domestic sheep (Fleming et al. 1921, Kingsbun' 1964). Perhaps it is possible to utilize this indixidual variation in selectively breeding for reduced vulnerability to particu- lar toxins. Currently, most domestic grazing animals are products of artificial selection for productivity, rather than for resistance to envi- ronmental challenges. Another avenue for applied research con- cerns the possibility of ameliorating effects of toxic plants through the transfer of rumen innocula from animals resistant to specific tox- ins to those that are susceptible. Jones (1985) reported that transfer of rumen cultures from goats that were resistant to poisoning by Leii- caena leucocephala to susceptible goats and steers eliminated adverse effects of Leucaena consumption in the previously susceptible ani- mals. This example suggests that even inter- specific transfer of rumen fluids may effective- ly reduce toxic effects in some cases. Deer herbivory we witnessed on Peavine Mountain may affect the demography of the local death camas population. Defoliation exper- iments indicate that death camas probably suf- fers reduced reproductive output after her- bivory (Tepedino 1982, Knapp 1986). While plants adapted to herbi\'or\' ma\' compensate for loss of biomass by allocating additional energy to growth and/or reproduction, highly toxic species instead employ an evolutionary strategy of defense against herbivory and thus may not exhibit compensation (Gates 1975, Laycock 1978). When such defenses are cir- cumvented bv herbivores with detoxification 1995] Notes 191 mechanisms, toxic plants should experience reduced fitness. Acknowledgments We thank Dr. Jeanne Chambers, Dr. Kip Panter, and two anonymous reviewers for thoughtful reviews of the manuscript. This paper is a contribution of the USDA, Agricul- tural Research Senice, Consei-vation Biology of Rangelands Unit, Reno, NV Literature Cited Barbosa, E, E Gross, and J. Kemper. 1991. Influence of plant allelochemicals on the tobacco hornworn and its parasitoid, Cotesia congregata. Ecology 72: 1567-157,5. Gates, R. G. 1975. The interface between slugs and wild ginger; some evolutionary aspects. Ecology 56: .391-400. Glarke, E. G. G., and M. L. Glarke. 1975. Veterinary toxicologv'. Macmillan Eublishing, New York, NY. GooPER, S. M., AND N. Owen-Smith. 1985. Gondensed tannins deter feeding by browsing ruminants in a South African savanna. Oecologia 67: 142-146. Dean, R. E., and A. H. Winward. 1974. An investigation into the possibility of tansy ragwort poisoning of blacktailed deer Journal of Wildlife Disease 10: 166-169. Dl.xON, J. S. 19.34. A study of the life history and food habits of mule deer in California, part 2. Food habits. Galifornia Fish and Game 20: 31.5-354. Fleming, G. E. 1918. Range plants poisonous to sheep and cattle in Nevada. Nevada Agricultural E.xperi- ment Station Bulletin 95. Fleming, G. E., N. F Eeterson, M. R. Miller, and L. H. Wright. 1921. Death camas. Flants poisonous to sheep cattle. Nevada Agricultural Experiment Station Bulletin 101. Freeland, W J., AND D. H. Janzen. 1974. Strategies in herbivory by mammals: the role of plant secondaiy compounds. American Naturalist 108: 269-289. Gauld, I. D., AND K. J. Gaston. 1992. Flant allelochemi- cals, tritrophic interactions and the anomalous diver- sity of tropical parasitoids: the "nasty' host hypothe- sis. Oikos 65: 3.53-357. James, L. F, R. E Keeler, A. E. Johnson, M. G. Williams, E. H. Gronin, and J. D. Olsen. 1980. Flants poison- ous to livestock in the western states. USDA-SEA Agricultural Information Bulletin 415. James, L. F, D. B. Nielsen, and K. E. Fanter. 1992. Impact of poisonous plants on the livestock industiy Journal of Range Management 45: 3-8. Jones, R. J. 1985. Leucaena toxicity and the ruminal degradation of mimosine. Fages 111-119 in A. A. Seawright, M. E Hegarty, L. E James, and R. F. Keeler, editors, Elant toxicology. Eroceedings of the Australia-U.S. Eoisonous Elant Symposium, Brisbane, Queensland Department of Erimary Indirstries, Yeerongpilly. Kingsbury, J. M. 1964. Eoisonous plants of the United States and Ganada. Erentice-Hall, Englewood Gliffs, NJ. Knapp, A. K. 1986. Ecophysiolog\' of Zigadenti.s luittallii, a toxic spring ephemeral in a warm season grassland. Oecologia 71: 69-74. Laycock, W. a. 1978. Goevolution of poisonous plants and large herbivores on rangelands. Journal of Range Management 31: 335-342. Laycock, W. A., J. A. Young, and D. N. Ueckert 1988. Ecological status of poisonous plants on rangelands. Eages 27-42 in L. F James, M. H. Ralphs, and D. B. Nielsen, editors, The ecology and economic impact of poisonous plants on livestock production. Westv'iew Eress, Boulder, GO. Nielsen, D. B., N. R. Rimbey, and L. F James. 1988. Economic considerations of poisonous plants on live- stock. Eages .5-15 in L. E James, M. H. Ralphs, and D. B. Nielsen, editors. The ecology and economic impact of poisonous plants on livestock production. Westview Eress, Boulder, GO. Eanter, K. E., and L. F James. 1989. Death camas — early grazing can be hazardous. Rangelands 11: 147-149. E\NTER, K. E., M. H. Ralphs, R. A. Smart, and B. Duelke. 1987. Death camas poisoning in sheep: a case report. Veterinaiy and Human Toxicology 29: 4.5-48. Bobbins, G. T, S. Mole, A. E. Hagerman, and T. A. Hanley. 1987. Role of tannins in defending plants against njminants: reduction in dn.' matter digestibil- ity? Ecology 68: 1606-1615. Stoddart, L. a., and D. I. Rasmussen. 1945. Deer man- agement and range livestock production. Utah Agri- cultural Experiment Station Circular 121. 17 pp. Tepedino, V J. 1982. Effects of defoliation on reproduction of a toxic range plant, Zigademis paniculatiis. Great Basin Naturalist 42: .524-.528. Received 12 October 1993 Accepted 30 August 1994 C;reat Basin Naturalist 55(2), © 1995, pp. 192 USE OF AN UNUSUAL FOOD SOURCE BY ROCK WRENS (TROGLODYTIDAE) PolK K. Piiillipsl and Allen F. Sanhoni^ Key nonls: Rock Wrens, foixl source, forci'^in'^i,. Salpinctes obsoletiis, Tro'Jodiithhte, jecdiiinhchavior On 12 Jul>' 1993 we obsei-ved an interesting I'xchange between an achilt Rock Wren [Salpinctes obsoletiis) and two juveniles. While at Toroweap Point on the north rim of the errand Canyon we observed an adult wren accompanied by two juveniles near our vehi- cle. All three birds walked beneath the vehicle by the rear wheel, but the adult moved imme- diately to the front end whereupon it hopped onto the front bumper and began to inspect the grill. The adult fr)und and ate an insect that had been trapped in the grillwork. While stand- ing on the bumper, the adult began to vocalize after consuming the insect. The juveniles appeared to show a positive phonotactic response to these calls, stopped foraging under the rear of the vehicle, and moved to the front. After the juveniles arrived at the front of the vehicle, the adult continued collecting insects from the grill. The adult ate none of these insects but mereh' held them in its beak while walking back and forth across the bumper The adult continued to vocalize, periodically paus- ing to face the juveniles. Then it continued for- aging in the grill. It appeared to us that the adult was showing the insects to the young. Neither of the young birds joined the adult on the bumper, however, and within a few min- utes the adult and juveniles flew off, not to return that afternoon. As far as we have been able to determine, this sort of acquired or derived behavior has not been reported previ- ously for Rock Wrens nor for any member of the famiU' Troglodytidae. Other obsen'crs have noticed birds taking advantage of unusual food sources, such as the opening of milk bottles (Fisher and Hinde 1949). There is generally a question, however, as to whether the behavior was bv chance or learned. One possible explanation for our obsei'vations is that the parent was tutoring the offspring about the availability of food in xehicle grill- work. Tutoring and obsei^vational learning have been documented in laboratory experiments in blackbirds (Mason et al. 1984), tits (Sheriy and Galef 1984, 1990), and pigeons (Palameta and Lefebvre 1985), and have also been document- ed in the wild in other birds using usual food sources (Schaadt and Rymon 1982). We have no way of knowing whether the adult we observed was attempting to teach what we believe were its offspring about an unusual food source. We hope this observation will stimulate further study of feeding in fledgling birds with the possibility of discoveries in social learning. Literature Cited Fisher, J., and R. A. Hindi:. 1949. The opening of milk bottles by birds. British Birds 42; 347-357. Mason, J. R., A. H. Arzt, and R. F. Reidinger. 1984. Comparative assessment of food preferences and aversions acquired by blackbirds via observational learning. Auk 101: 796-803. Palameta, B., and L. Lefebvre. 1985. The social trans- mission of a food finding technique in pigeons: What is learned? Animal Behaviour 33: 892-896. Schaadt, C. E, and L. M. Rymon. 1982. Innate fishing behavior of ospreys. Raptor Research 16: 61-62. Sherry, D. F, and B. G. Galef, Jr. 1984. Cultural trans- mission without imitation: milk bottle opening by birds. Animal Behaviour 32: 937-938. . 1990. Social learning without imitation: more about milk bottle opening by birds. Animal Behaviour 40: 987-990. Received 13 June 1994 Accepted 16 November 1994 'Biology Dc|)aitiMciit, Miami-Dack- Coii 2Sl'Ii()()I of Natural and Hcaltli Scienc-ef inimity CIol , Barr\' LInl c-nc North Campus, 11:380 N.W. 27th Avenue, Miami, FL .331(17-349.5. .(•rsity, 11.3()() N E, Second Avenue, Miami Shores, FL 33Kil-669.'5. Address reprint requests to this 192 INFORMATION FOR AUTHORS The Great Basin Naturalist welcomes previously unpublished manuscripts pertaining to the biologi- cal natural history of western North America. Preference will be given to concise manuscripts of up to 12,000 words. Simple species lists are dis- couraged. 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 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, 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. Consuk 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 Naturalist for changes and refer to the CBE Stijle Manual, 5th edition (Council of Biology Editors, One Illinois Center, Suite 200, 111 East Wacker Drive, Chicago, IL 60601-4298 USA, phonE: (312) 616-0800, FAX: (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, appendices, 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 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 fonn 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 BASIN NATURALIST voi 55 no 2 ap ii 1995 CONTENTS Articles Diets ot >'oiing Colorado sqiiavvfish and otlier small fish in backwaters of the Green River, Colorado and Utah Robert T. Mnth and Darrel E. Snyder 95 Invertebrate fauna of wastewater ponds in southeastern Idaho Karen L. Cieminski and Lester D. Flake 1 05 Growth and reproduction in an alpine cushion plant: Astragalus kentrophyta var implexus Wayne R. Owen 117 Calileuctra, a new genus, and two new species of stoneflies from California (Plecoptera: Leuctridae) W. D. Shepard and R. W. Baumann 1 24 Carbon isotope discrimination in the C4 shrub Atriplex confertifolia along a salinity gradient Darren R. Sandquist and James R. Ehleringer 1 35 Demography of Astragalus scaphoides and effects of herbivory on population growth Peter Lesica 1 42 Lahontan sagebrush {Artemisia arhuscula ssp. longicaulis): a new taxon Alma H. Winward and E. Durant McArthur 151 Douglas-fir tussock moth {Orgyia pseudotsugata McDunnough) on subalpine fir in northern Utah E. Matthew Hansen 1 58 Seasonal nutrient cycling in Potamogeton pectinatus of the lower Provo River . . C. Mel Lytle and Bruce N. Smith 1 64 Factors influencing fish assemblages of a high-elevation desert stream system in Wyoming Bernard Carter and Wayne A. Hubert 1 69 Notes Speciation by aneuploidy and polyploidy in Mimulus (Scrophulariaceae) Robert K. Vickery, Jr. 1 74 Speciation in Mimulus, or, Can a simple flower color mutant lead to species divergence? Robert K. Vickery, Jr 177 Fall lamb production by a California bighorn sheep Matthew McCoy, Walt Bodie, and ElRoy Taylor 181 Age, growth, and reproduction of leatherside chub {Gila copei) Jerald B. Johnson, Mark C. Belk, and Dennis K. Shiozawa 1 83 Consumption of a toxic plant {Zigadenus panicidatus) by mule deer William S. Longland and Charlie Clements 1 88 Use of an unusual food source by Rock Wrens (Troglodytidae) Polly K. Phillips and Allen F Sanborn 1 92 6»r>^ H E GREAT BASIN NATURALIST VOLUME 55 NO 3 — JULY 1995 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, Bojce, 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 Biologv; John CanoU University Department of Environmental Resource Sciences University Heights, OH 44118 University of Nevada-Reno, 1000 Valley Road Boris C. Kondratieff R^"«' ^V 89512 Department of Entomology, Colorado State Rorert 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. All are at Brigham Young University. Ex Officio Editorial Board members include Steven L. Taylor, College of Biologv' and Agriculture; H. Duane Smith, 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 1995 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 scholarlv 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 © 1995 b\- Brigham Young University ISSN 0017-3614 Official publication date: 21 July 1995 7-95 750 15072 The Great Basin Naturalist Published at Provo, Utah, by Brigham Young University ISSN 0017-3614 VoLU M E 55 3 1 J u LY 1 995 N o. 3 Great Basin Naturalist 55(3), © 1995, pp. 193-200 BENTHIC COMMUNITY STRUCTURE IN TWO ADJACENT STREAMS IN YELLOWSTONE NATIONAL PARK FIVE YEARS AFTER THE 1988 WILDFIRES G. Wayne MinshalU, Christopher T. Robinson^ Todd V Royerl, and Samuel R. Rushforth^ Abstract. — Ph\sical characteristics, benthic macroin\'ertebrates, and periph>'ton assemblages in two adjacent head- water streams in Yellowstone National Park were evaluated five years after the 19SS wildfires. The catchment of one stream was burned by wildfire (burned stream) while the other catchment was unburned (unbumed stream). Physical measures revealed channel alteration in the burned stream relative to the unbin-ned stream. Periphyton biomass was lower in the burned than the unburned stream (29.2 vs. 50.5 g/m^ AFDM, respectively), further demonstrating the unstable physical conditions of that system. Kendall's coefficient of concordance (an index of similarity) between diatom assemblages was 0.22, indicating distinct assemblage composition between streams. Navicida pennitis Hust. was the most abundant diatom in the burned stream while Hannaea arcus (Ehr.) Patr was dominant in the unbumed stream. Macroinvertebrate taxa richness, density, and biomass were all greater in the unburned stream, although Chironomidae was the most abundant taxon in both streams. Results suggest the removal of ten-estrial/riparian vegetation by wildfire can directly influence stream benthic assemblages by altering the inherent disturbance regime of the physical habitat templet. Key words: wildfire, streams, disturbance, inacroinvertebrates, diatoms, benthic habitat, Yellowstone National Park. Physical disturbance, acting at various spa- and Minshall 1992, Robinson et al. 1994, Mihuc tial and temporal scales, often is the predomi- et al. in press, Robinson and Minshall in press). nant factor structuring stream benthic com- In lotic ecosystems, physical disturbance also munities (Minshall 1988, Resh et al. 1988). may constrain the estabhshment of biotic con- Further, physical disturbances may be viewed trols, such as competition and predation, on in a hierarchical framework, with the effects of benthic community structure (McAuliffe 1984, small-scale disturbances altered (intensified or Minshall and Petersen 1985, Resh et al. 1988). mediated) by large-scale disturbance events Wildfire burned extensive portions of the (sensu O'Neill et al. 1986). Wildfire, as a large- Greater Yellowstone Ecosystem during the scale disturbance, directly influences stream summer of 1988. Over 32% of the streams in biotic structure and function by affecting the Yellowstone National Park (YNP) were affected physical habitat of stream ecosystems (Minshall to varying degrees by wildfires (Minshall et al. et al. 1989, Minshall and Brock 1991, Richards 1989, Minshall and Brock 1991, Robinson et al. 'Streuni Ecolog\ Center. Department of Biological Sciences, Idaho State Uni\ersit>'. Pocatello, ID 8.3209. ^Department of Botan\' and Range Science, Brigham Young University. Pro\o, UT 84602. 193 194 Great Basin Naturalist [Xbliinie 55 1994). Minshall and Brock (1991) sumniarized the immediate effects of the fires on YNP's stream ecosystems and h\'pothesized on the mid-term (10-25 yr) and long-term (50-300 yr) effects. They suggested that most adverse short-term effects on streams resulted from increased sediment load and channel erosion caused by increased overland runoff following precipitation events and snowmelt. The in- tensity and frequency of short-term effects were hxpothesized to decrease by year 5 as riparian conditions improve (see Richards and Minshall 1992). In general, mid- and long- term effects on streams, including recovery to prefire conditions, should coirespond to vegeta- tive regrowth in burned catchments (Minshall at al. 1989, Minshall and Brock 1991). The present study compared physical char- acteristics and benthic community structure in two streams five years after the 1988 wildfires. The streams are adjacent second-order (after Strahler 1952) tributaries of the South Fork Cache Creek. The catchment of one stream was bunied during the 1988 wildfires, while the catchment of the other was essentially un- burned. The spatial arrangement of these streams (adjacent basins) provided a treatment/ reference situation where confounding factors of climate and geology are minimized when comparing differences among the study streams. However, the study lacks true repli- cation of the burned and unburned treatments (sensu Hurlbert 1984) and must be viewed as a simple comparison study. Phenomenological studies and/or two stream comparisons are common in stream ecology (e.g., Wallace et al. 1986, Robinson et al. 1993, Scarsbrook and Townsend 1993) and are capable of providing valuable insights (Townsend 1989). The present study can be viewed as a natural "experiment" with observed differences between the two streams attributed to the effects of wildfire. In that context, the study provides insights on general patterns of lotic ecosystem recoveiy to an unpredictable, large-scale disturbance (Townsend 1989, Lamberti et al. 1991). Methods The study streams, located in the northeast comer of YNI^ were surveyed on 19 July 1993. One stream had over 80% of its catchment burned during the 1988 Yellowstone wildfire (hereafter, burned stream; 110°0r30"W, 44°50'00"N), while less than 10% of the catch- ment of the other stream was bunied (hereafter, unburned stream; 110°01'00"\V, 44°49'30"N). Climate of the area is typical of the northeiTi Rocky Mountains, with precipitation primarily occurring as snow during the winter months. Both streams drain catchments primariK' veg- etated (prior to the fire in the buiTied stream) by coniferous forests of lodgepole pine {Pinus conturta) and Engelmann spruce {Picea engel- mannii). Riparian vegetation consisted of wil- low {Salix), rose {Rosa), and alder {Alniis). Sui^veys were conducted approximately 0.5 km above the confluence of the two streams. Physical characteristics were measured in each stream at five cross-sectional transects, each sit- uated approximately 50 m apart. Measurements made at each transect included stream width at baseflow, stream width at bankfull discharge, and stream cross-sectional profile (for calcula- tion of width:depth ratios). Discharge was cal- culated in each stream at the most suitable transect following the methods of Platts et al. (1983). In addition to measurements at each transect, 100 randomly selected rock substrata along a 100-m length of stream (located within the outermost cross-sectional transects) were measured for size (length of the longest axis) and percent embeddedness. Embeddedness was defined as the percent coverage of the rock (three-dimensional surface) by fine sedi- ments. Large boulders that protruded through the water surface were not used in substratum size measurements. Water depth and near-bed water velocity also were recorded at each of the 100 random locations. Near-bed water velocity was measured with a small Ott C-1 current meter approximately 2 cm above each substratum. One periphyton sample was collected from a suitable (flat-surfaced, medium-sized) rock substratum at each cross-sectional transect using a method described in Robinson and Minshall (1986). Samples were frozen in the field in a Taylor-Wharton 3DS dry shipper charged with liquid nitrogen and returned to the laboratory for processing. In the laboratoiy, samples were extracted in 10 ml of methanol for 24 h (Holm-Hansen and Riemann 1978). One 3-ml subsample was then removed from each sample and analyzed for chlorophyll a using a Gilford Instruments (Model 2600) spec- trophotometer. The remaining periphyton material from each sample was used for algal 1995] Wildfire and Benthic Communities 195 biomass determination, expressed as grams ash-free dry mass (AFDM) per m^. The mater- ial was dried at 50 °C for 24 h, weighed on a Sauter balance (Model AR 1014), ashed at 550 °C for a minimum of 3 h, rehydrated, redried at 50 °C, then cooled to ambient tem- perature in a desiccator and reweighed. The difference in weights equaled the AFDM of the sample. Diatom samples were collected in each stream, after Robinson and Rushforth (1987), from three to five rock substrata representing the predominant habitat type (typically riffles). Samples were composited, preserved with 5% formalin, and returned to the laboratory. The composite sample was boiled in concentrated nitric acid, rinsed, mounted in Naphrax moun- tant, and examined under lOOOX oil immersion using a Zeis RA microscope with Nomarski optics (St. Clair and Rushforth 1976). Relative abundances of diatom taxa were determined by counting a minimum of 1000 diatom valves from each stream. Diatoms were analyzed in terms of species richness, Simpson's index, and Kendall's coefficient of concordance (an index of similarity using all taxa with a relative abundance >1%). Other algal groups such as Chlorophyta (green algae) and Cyanobacteria (blue-green algae) were not abundant at the time of sampling and thus were not consid- ered in the present study. One benthic sample was collected from a riffle/run habitat (pools were rare and not sam- pled) near each transect and analyzed for macroinvertebrates and benthic organic mat- ter (BOM). Samples were collected using a Surber sampler (250 /xm mesh), preserved with 5% formalin, and returned to the labora- tory. Woody debris >5 cm in length that was collected in the benthic samples was rinsed of invertebrates and removed from the samples. In the laboratory, macroinvertebrates were hand-sorted from the benthic detritus with the aid of a 3X dissecting microscope, identified to the lowest feasible level (usually genus), enu- merated, dried at 50°C for a minimum of 48 h, then cooled to ambient temperature in a des- iccator and weighed. Dry weights, in milli- grams, were determined on a Cahn (Model 25) electrobalance. The benthic detritus from each sample was used for BOM determination. The quantity of BOM, expressed as g AFDM/ m^, was determined as described above for periphyton. Macroinvertebrates were analyzed in terms of density (no./m-), biomass (mg/m^), taxa richness, Simpson's index, and relative abundances. Chi-squared analysis was used to test for statistical differences in median substratum size between the two streams (Zar 1984). Independent sample t tests were used to com- pare the other characteristics for differences between the two streams. Prior to the ^-test analysis all data were log (x -I- 1) transformed, except substratum embeddedness and the rela- tive abundance of invertebrate taxa (both per- centage measures), which were arcsine (square root [x]) transformed (Zar 1984). Tabular re- sults are presented as untransformed means and standard deviations. All statistical analyses were performed on SYSTAT (Wilkinson 1990). Results Baseflow discharge was equal in the two streams (0.2 m'^/s), reflecting the similar catch- ment size of the burned (22 km^) and unbumed (26 km^) streams. Mean baseflow width, near- bed water velocity, and BOM were not signifi- cantly different between the two streams [P > .05). Substratum embeddedness was signifi- cantly greater in the burned stream (P = .01), although the difference between mean values was not large (burned = 62.9, unburned = 52.8). It is not known whether this statistical difference was biologically meaningful or sim- ply a reflection of the large sample size {n = 100). Water depth at baseflow (F < .01) was lower and stream width at bankfull discharge greater (P = .03) in the burned stream than the un- burned stream. Although not statistically sig- nificant (P = .06), the ratio of stream width: depth was greater in the burned than the un- burned stream (216 and 91, respectively). The general appearance of the two streams was dis- tinctly different (Fig. 1; Minshall personal ob- servation). Large, woody debris and streamside riparian vegetation, which provide bank and channel stability, were noticeably absent in the burned stream. Mean substratum size was not significantly different between the two streams (P > .05) in 1993, possibly because large boulders were not recorded in the measurements (see Fig. 1). We collected additional data on substratum size in August 1994 and included large boulders in the measurements. Further, substrata within 196 Ghkat Basin Naturalist [Volume 55 ^-f. _"** ■j^ ' Fig. I. HepresLMitatixe photographs of the huined (upper) and iinl)urned (lower) streams Fixe >ears after the 1988 wildfire. Note absence of large, woody debris and streaniside riparian xegetation in the l)inned stream. 1995] Wildfire and Benthic Communities 197 the bankfull channel were measured in 1994, in contrast to measures being recorded only within the baseflow channel in 1993. The 1994 results showed that mean substratum size was significantly larger in the unburned than in the burned stream (F < .01). A comparison of median substratum size showed similar results to that of mean substra- tum size. Median substratum size was not dif- ferent between the two streams when mea- surements excluded large boulders and were confined to the baseflow channel (P > .05). However, when measurements included large boulders and encompassed the bankfull chan- nel, the difference in median size was signifi- cant (F < .01). Whether large boulders were present in the burned stream prior to the wild- fire has yet to be detemiined. However, in other streams influenced by intensive wildfire, large boulders were obsei-ved to be buried by inor- ganic debris (primarily gravel and fine sedi- ments) within five years following wildfire (Minshall personal obsei'vation). The burned stream contained less periphy- ton chlorophyll a (F = .06) and AFDM (F < .01) than did the unburned stream (Table 1). Diatom species richness was greater in the burned (34 taxa) than in the unburned stream (27 taxa; Table 2). Simpson's index was lower for the burned than the unburned stream (0.12 and 0.42, respectively). Kendall's coefficient of concordance for the two diatom communities was 0.22, suggesting distinct assemblage com- position among sites. For example, Navicida pennitis Hust. was the most abundant species in the burned stream, constituting 24.7% of the assemblage, while Hannaea orciis (Ehr.) Patr. comprised 63.1% of the assemblage in the unburned stream (Table 2). Mean macroinvertebrate density and bio- mass were lower in the burned than unburned stream (Table 3), but the differences were not significant (F > .05). For example, mean den- sity' in the burned stream was 9960 individuals/ m^, while the unburned stream had 16,950 indi- viduals/m^, and mean biomass (dry weight) was 1960 and 3200 mg/m- in the burned and unburned streams, respectively. Taxa richness and Simpson's index both were reduced in the burned stream, although the difference was significant only for Simpson's index (F = .04) (Table 3). The burned stream contained a mean of 15 taxa per benthic sample compared to a mean of 20 taxa for the unburned stream. The mean Simpson's index was 0.57 for the burned stream and 0.73 for the unburned stream. Chironomidae was the most abundant taxon in both streams (Table 4), although their relative abundance was significantly greater (P = .03) in the unburned stream. There were no statistical differences (F > .05) in relative abundances of other taxa common to both streams (Hydracarina, Simuliidae, Baetis bi- caudatiis, Cinygnnilo, and Zopada columhiana). Discussion Alterations of the surrounding terrestrial landscape by major unpredictable disturbances such as hurricanes, volcanic eruptions, or wild- fire directly influence streams draining the Table 1. Means (SD) and P values for physical characteristics measured in the study streams. Burned Unbi P \'alue Baseflow width (m) Near-bed velocity (cm/s) BOM(g/m2) Embeddedness (%) Baseflow depth (cm) Bankfull width (m) Bankfull width:depth ratio Periphyton chl a (mg/m-) Periphyton AFDM (g/m^) Mean substratum size (cm) Mean substratum size (cm) Median substratum size (cm) Median substratum size (cm 5.9 (2.6) 10.3 (0.1) 1.5 (0.8) 62.9 (28.5) 16.9 (11.0) 35.0 (11.9) 216 (101) 8.9 (5.5) 29.2 (3.7) 14.4 (10.5) 15.5 (15.5) 12.0 11.0 4.5 (1.8) .46 10.7 (0.1) .71 2.5 (1.2) .20 52.8 (30.0) .01 24.3 (12.5) <.01 16.6 (1.4) .03 91 (21) .06 32.1 (19.5) .06 .50.5 (9.3) <.01 15.0 (14.2) .26 27.6 (27.4) <.01 11.0 .72 17.0 <.01 Baseflow channel, large boulders excluded. Bankfull channel, large boulders included. 198 Great Basin Natuhaijst [Volume 55 Table 2. Community measures and relative aliuudanees (%) for the diatom assemblage of each study stream. Tahi>K 3. Means (SD) and P values of macroinvertehrate eomnnmity measures for the study streams. Burned Unburned Species richness 34 27 Simpson s index (C) 0.12 0.42 Navictila pcnniti.s I hist. 24.7 5.1 Nitzsch id dissipatii (Kuetz.) Grun. 17.3 4.2 Achnanthes lanceolata (Breb.) Grun. 9.8 1.7 Nitzschia paleacea Grun. 7.6 3.1 Navicuhi (iriensis Hust. 4.8 2.4 Hannaca arciis (Ehr.) Patr. 2.1 63.1 Burned Uiihunic'tl P Mill affected watersheds. For example, the Mt. St. Helens eruption of 1980 dramatically changed drainage patterns and river networks, elimi- nated terrestrial vegetation, and caused major debris flows that scoured stream channels (Wilzbach et al. 1983, Hawkins 1988). However, high spatial variation in the intensity of these major disturbances may occur, causing tempo- ral differences in recovery patterns (Yount and Niemi 1990). In catchments of YNP the rela- tive area burned ranged from <10% to >90% (Minshall and Brock 1991). Further, the degree of alteration of stream habitat was highly correlated with percent of catchment burned (Robinson and Minshall in press). In the present study significant differences were observed in the benthic habitat of the two streams. The width:depth ratio of the burned stream was greater than that of the unburned stream. Anderson (1992) also observed in- creased widthidepth ratios following major disturbances in streams of the Cascade Moun- tains. With large boulders included in the mea- surements, the unburned stream exhibited sig- nificantly greater substratum size. Gurtz and Wallace (1984) demonstrated that large sub- strata could mediate the effects of large-scale disturbances by providing stable habitat for benthic organisms. At the time of sampling, the burned stream did not contain the larger- sized substrata found in the unburned stream. It is probable that the larger substrata in the burned stream were buried by inorganic sedi- ments following the wildfire, as has been observed in other YNP streams (Minshall per- sonal observation). Thus, one effect of the wild- fire appeared to be alteration of the substrata in such a manner as to make the benthic habitat more susceptible to future disturbances (e.g., Gurtz and Wxllace 1984). Densit\ (no./ni^) 9963 (47.30) 16,948 (899Si .31 Biomass (mg/m^) 1956(1056) 3198(1274) .19 Ta.xa richness 15.0 (3.7) 20.4 (2.6) ..58 Simpsons indf.xfC) 0..57 (0.09) 0.73(0.10) .04 Lamberti et al. (1991) found that faunal densities and macroinvertehrate species rich- ness had recovered within one year following a major debris flow in an Oregon stream. In central Idaho, however, streams disturbed by wildfire and unburned reference streams showed little similarity in macroinvertehrate assemblages, even after five years of recovery (Richards and Minshall 1992). Similarly, in the present study the influence of wildfire was still apparent after five years. Macroinvertehrate community structure was not similar between the two systems, despite their close proximity to each other (0.5 km). Most researchers agree that recovery of the benthic community will correspond to recovery of the surrounding landscape (Steinman and Lamberti 1988, Minshall et al. 1989, Lamberti et al. 1991, Minshall and Brock 1991, Anderson 1992, Richards and Minshall 1992, but see Hawkins 1988). Primary producers (lotic algae) may recover sooner than consumers (macroinvertebrates and fish) because of their much shorter life cycles, and subsequenfl}' they may influence recoveiy of the higher trophic levels (Steinman and Mclntire 1990). In the present study, peri- phyton biomass (as AFDM) in the unburned stream was 1.7X greater than in the burned stream, implying a present lack of recoven' by primary producers in the burned system. Macroinvertehrate taxa richness also was greater in the unburned stream than in the burned stream. How functional or structural recovery of macroinvertebrates is related to algal recoveiy following wildfire has yet to be determined, but provides an interesting and important avenue for future research. Algae have shorter life cycles and reduced mobility relative to macroinvertebrates, and possibly the two groups respond differently to large- scale disturbances. 1995] Wildfire and Benthic Communities 199 Table 4. Mean (SD) density and relative abundance of the 10 most abundant niacroin\ertebrate taxa from each stream. These taxa constituted >909'f of the assemblage in their respective streams. Taxa Density Relative (no./m-) abimdance (%) - - Burned Chironomidae 5437 (2218) 59.7 1 (19.2) Simuhidae 2737 (4583) 18.2 1 (24.1) Baetis bicaitdatits 576 (346) 8.1 (6.3) Hydracarina 148 (232) 1.5 (2.3) Epeoriis olbertae 95 (123) 1.0 (1.2) Zapada coluinbiana 90 (72) 1.0 (0.7) Epcorits longimaniis 80 (125) 0.5 (0.7) Ciiujgmula 75 (71) 0.8 (0.7) Eluthrogena 37 (52) 0.3 (0.3) Ameletus cooki 22 (15) - Unbumed 0.3 (0.2) Chironomidae 14,676 (8289) 84.7 (6.7) H\dracarina 372 (253) 2.1 (1.6) Cinygmula 314 (187) 1.7 (0.8) Zapada cohiiubiana 310 (215) 1.6 (0.6) Simuliidae 307 (181) 3.8 (5.3) Baetis bicaitdatus 125 (60) 1.4 (1.7) Dninella coloradensis 75 (64) 0.4 (0.5) Wniacophila angelita 73 (55) 0.4 (0.2) Rliyacoplula tiicida 52 (29) 0.3 (0.1) Kogotiis 52 (35) 0.3 (0.2) After five years of recovery, the channel of the burned stream still appeared unstable as indicated by different diatom assemblages between the two streams. For example, the small, adnate diatom Naviciila pennitis Hust. was predominant in the burned stream but was found in relatively low abundance in the unburned stream. N. pennitis was predomi- nant in other YNP streams influenced by the 1988 wildfires, and it has been suggested that a diatom community with an abundance of N. pennitis is indicative of more physically dis- turbed stream environments (Robinson et al. 1994). Further, Robinson et al. (1994) showed diatom recoveiy among 14 streams in Yellow- stone was inversely related to degree of dis- turbance by wildfire. Similarly, Steinman and Lamberti (1988) found little recovery, after six years, in the composition of algal communities in intensively disturbed streams of Mt. St. Helens. In summary, benthic community re- covery patterns appeared to be related to the recovery of stream physical habitat which, five years after the 1988 wildfires, still displayed evidence of instability. When examining the recovery of benthic communities following large-scale disturbance, one must remain aware of the connections between the terrestrial land- scape, lotic habitat, and benthic organisms. Acknowledgments We thank Vincent Archer, Michael Bray, Justin Gill, and especially Scott Relyea for assistance in the field. Cecily Nelson, Jason Nelson, Mark Overfield, and Jeffrey Varricchione assisted in the laboratory. Suggestions from Dr. Richard Hauer and two anonymous reviewers greatly improved the manuscript. The research was partially sup- ported by Grant No. 725 from the Faculty Research Committee, Idaho State University. Additional funding was provided through Yellowstone Ecosystem Studies (Dr. Robert Crabtree, Director) and the Department of Botany and Range Sciences, Brigham Young University. Literature Cited Anderson, N. H. 1992. Influence of disturbance on insect communities in Pacific Northwest streams. Hydro- biologia 248: 79-92. GuRTZ, M. E., AND J. B. Wallace. 1984. Substrate-medi- ated response of stream invertebrates to distm-bance. Ecologv' 65: 1556-1569. Hawkins, C. P 1988. Effects of watershed vegetation and disturbance on invertebrate community structure in western Cascade streams: implications for stream ecosystem theoiy Verhandlungen der Internationale Vereinigung fiir Theoretische und Angewandte Limnologie 23: 1167-1173. Holm-Hansen, O., and B. Riemann. 1973. Chlorophyll a determination: improvements in methodology. Oikos 30: 438-447. HURLBERT, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecological Mono- graphs 54: 187-211. La.mberti, G. a., S. V Gregory, L. R. Ashkenas, R. C. Wildman, and K. M. S. Moore. 1991. Stream eco- system recoveiy following a catastrophic debris flow. Canadian Journal of Fisheries and Aquatic Sciences 48: 196-208. McAuLlFFE, J. R. 1984. Competition for space, distur- bance, and the stnicture of a benthic stream commu- nity. Ecolog>' 65: 894-908. MlHUC, T., G. W. iviiNSHALL, AND C. T. RoBlNSON. In press. Responses of benthic macroinvertebrate populations in Cache Creek Yellowstone National Park to the 1988 wildfire. In: D. G. Despain and P Schullery, editors. The ecological implications of fire in Greater Yellowstone: 2nd biennial conference on the Greater Yellowstone Ecosystem. National Park Service, Yellowstone National Park. MiNSHALL, G. W. 1988. Stream ecosystem theoiy: a global perspective. Journal of the North American Bentho- logical Societv' 7: 263-288. 200 Great Basin Natur.\list [Volume 55 MlNSHALL, G. VV., AND J. T. BROt:K. 1991. Ohscrvfd and anticipated effects of forest fire on Yellowstone stream ecosystems. Pages 123-135 in R. B. Keitcr and M. S. Boyce, editors. The Greater Yellowstone Ecosystem: redefining Americas wilderness heritage. Yale University Press, New Haven, CT. MlNSHALL, G. W, AND R. C. Peteksen. 1985. Towards a theory of macroinvertebrate community structure in stream ecosystems. Archives fiir Hxdrohiologia 104; 49-76. MlNSHALL, G. W., J. T Brock, and J. D. Vahley. 1989. Wild- fire and Yellowstone's stream ecosystems. BioScience 39; 707-715. O'Neill, R. V, D. L. DeAngelis, J. B. Waide, .\nd T. H. E Allen. 1986. A hierarchical concept of ecosystems. Princeton University Press, Princeton, NJ. 257 pp. I'lATis. W. S., W. E Megahan, and G. W. Minshall. 1983. Methods for evaluating stream, riparian, and hiotic conditions. USDA Intermountain Forest and Range E.\periment Station, Ogden, UT. General Technical Report I NT- 138. 70 pp. Resh, V. H., et al. 1988. The role of distiubance in stream ecology. Journal of the North American Benthological Societx' 7: 433^55. Richards, C., and G. W. Mlnshall. 1992. Spatial and temporal trends in stream macroinvertebrate com- munities; the influence of catchment disturbance. HydrobioIogia241; 173-184. Robinson, C. T, and G. W. Minshall. 1986. Effects of disturbance fi^equency on stream benthic communi- ty structure in relation to canopy cover and season. Journal of the North American Benthological Society 5: 237-248. _. In press. Physical and chemical responses of streams in Yellowstone following the 1988 wildfire. In D. G. Despain and P Schullery, editors. The eco- logical implications of fire in Greater Yellowstone; 2nd biennial conference on the Greater Yellowstone Ecosystem. National Park Service, Yellowstone National Park. R0BIN.SON, C. T, AND S. R. Rlshforth. 1987. Effects of physical disturbance and canopy cover on attached diatom community structure in an Idaho stream. Hydrobiologia 154;' 49-59. Robinson, G. T, G. W. Minshall, and L. Van Every. 1993. Seasonal trends and colonization patterns of macro- invertebrate assemblages in two streams with con- trasting How regimes. Great Basin Naturalist 53; 321-331. Robinson, G. T, S. R. iirsiiEOHTii, and G. VV. Minshall. 1994. Diatom assemblages of streams influenced by wildfire. Journal of Phycology 30; 209-216. ScARSBROOK, M. R., AND G. L. TowNSEND. 1993. Stream community structure in relation to spatial and tem- poral variation; a habitat templet study of two con- trasting New Zealand streams. Freshwater Biology 29; 395-410. St. Clair, L. L., and S. R. Rushforth. 1976. The diatom flora of the Goshen Warm Springs Ponds and Wet Meadows, Goshen, Utah, U.S.A. Nova Hedwigia 24; 353-125. Steinman, a. D., and G. A. Lamberti. 1988. Lotic algal communities in the Mt. St. Helens region six years following the eruption. Journal of Phycology 24; 482-489. Steinman, A. D., and G. D. McIntire. 1990. Recovery of lotic periphyton communities. Environmental Management 14; 589-604. Strahler, a. N. 1952. Hypsometric (area-altitude) analy- sis of erosional topography. Geological Society of America Bulletin 63; 11 17-1 142. Townsend, C. R. 1989. The patch dynamics concept of stream community ecology. Journal of the North American Benthological Society 8; 36-50. Wall.\ce, J. B., D. S. Vogel, and't. E Cuffney. 1986. Recovery of a headwater stream fi-om an insecticide- induced community disturbance. Journal of the North American Benthological Society 5; 115-126. Wilkinson, L. 1990. SYSTAT; the system for statistics. SYSTAT, Inc., Evanston, IL. 677 pp. Wilzbach, M. a., T. H. Dudley, and J. D. Hall. 1983. Recovery patterns in stream communities impacted by the Mt. St. Helens eruption. Water Resources Research Institute, Oregon State University, Gorvallis. Report No. WRRI-83. 33 pp. YOUNT, J. D., and G. J. NiEMl. 1990. Recovery of lotic communities and ecosystems from disturbance — a narrative review of case studies. Environmental Management 14; 547-569. Zar, J. H. 1984. Biostatistical analysis. Prentice-Hall Inc., Englewood Cliffs, NJ. 718 pp. Received 16 November 1994 Accepted 2 March 1995 Great Basin Naturalist 55(3), © 1995, pp. 201-212 EFFECTS OF BROWSING BY NATIVE UNGULATES ON THE SHRUBS IN BIG SAGEBRUSH COMMUNITIES IN YELLOWSTONE NATIONAL PARK Francis J. Singer^ and Roy A. Renkin^ Abstract. — The effects of elk {Cerviis elaphus), pronghorn (Antilocapra americuna), and mule deer (Odocoileus liemionus) browsing on shi-ubs in big sagebrush {Artemisia tridentata) communities were monitored over a 31-year peri- od in Yellowstone National Park. Ungulates were restricting Wyoming big sagebrush (spp. wyomingensis) heights, size, and recruitment on the lower-elevation stratum only, while no such suppression was observed on the high-elevation stratum. Parallel increases in mountain big sagebmsh (spp. vaseijana) densities and cover occuired over the study period on both browsed and unbrowsed sites at the higher-elevation stratum, although big sagebrush, green rabbitbrush (Chnjsothamnus viscidiflorus), and horsebnish (Tetradymia canescens) were slightly taller and crown sizes were slightlv larger on unbrowsed than browsed sites. Wyoming big sagebrush utilization (percent leader use) was eight times higher (x = 87 ± 7.2% by pronghorns, mule deer, and elk) on the low-elevation winter range stratum (the Boundary Line Area [BLA] portion of the winter range), while mostly mountain big sagebrush with leader use averaged only 11 ± 4.1% (near- ly all by elk) on the high-elevation range stratum. In addition, annual aboveground biomass production of big sagebrush did not differ between browsed and unbrowsed study sites on the high-elevation stratum of the winter range. Population turnover was higher on browsed big sagebrush at the high-elevation plots; seedling germination and survival rates were higher on browsed plots versus unbrowsed plots. No difference was observed in percent dieback of big sagebmsh adult plants between browsed and unbrowsed plots at the higher stratum. Browsing did not influence the number of leaves or seedstalks per plant (P > .05), but leaves averaged 45% longer and seedstalks 42% longer on browsed big sagebmsh. Ungulate browsing, however, apparently suppressed production, germination, and sur\'ival of Wyoming big sagebmsh on the low-elevation stratum. Numbers of Wyoming big sagebmsh declined 43% and cover declined 29%, 1957-1990, on browsed sites on the BLA. Annual biomass production on browsed sites at the low-elevation stratum was only 6-35% that of unbrowsed sites, and big sagebrush recruitment was less on browsed sites. Percent leader use of big sagebrush did not differ between the period of ungulate reductions, 1962-1969, and the 19S0s on the lower stratum (.v = 87% leader use), but utilization was less on higher portions of the winter range during the period of elk reductions (x = 2%) than during the 1980s following cessation of elk controls (.v = 11%). Key words: big sagebrush browsing, noiihcrn Yellowstone elk, pronghorn. mule deer Cen'us elaphus. Native populations of elk {Cervus elaphus), models suggest 8-15% fewer elk and 10-25% bison {Bison bison), and pronghorn {Antilocapra fewer bison would occupy the system if wolves americana) were artificially reduced in Yellow- were recovered (Carton et al. 1990, Boyce 1993, stone National Park (YNP), particularly from Mack and Singer 1993). 1942 through 1967 (Meagher 1973, Houston Early workers expressed concern about 1982). Reductions were terminated in 1967 apparent overbrowsing and declines in big sage- when an experimental management program brush {Artemisia tridentata) due to possible of natural regulation was initiated (Cole 1971, overabundant populations of elk and prong- Houston 1976, 1982). Elk and other ungulate horn. As early as the 1930s, Rush (1932) and numbers tripled after cessation of controls, and Cahalane (1943) reported losses of big sage- concerns were expressed over high ungulate brush over lower-elevation areas of the north- densities (Chase 1986, Kay 1991). Appropriate ern winter range. Rush (1932) reported that numbers of ungulates for the park are unknown less-palatable rabbitbrushes {Chnjsothamnus since no similar control area exists where wolves spp.) were increasing. KJttams (1950) concluded {Canis hipus) are present and where ungulate that big sagebrush numbers were declining at migrations are completely unrestricted by both lower and higher elevations of the north- humans (Cayot et al. 1979, Peek 1980). Ungulate em winter range. He felt that physical distur- densities are likely slightly above natural con- bances of big sagebrush by elk during cold ditions, in that three independent computer periods (shattering and trampling) and an ^Division of Research, Box 168, Mammoth Hot Springs, Yellowstone National Park, WT 82190, and Colorado State Universit\; Fort Colhns, CO 80523. ^Resources Management Division, Box 168, Yellowstone National Park, WT 82190. 201 202 Great Basin Naturalist [Volume 55 absence of l)ig sagebrush reprocliielioii con- tributed to the dechne. Dechnes in big sage- brush at the lower-elevation Boundary Line Area (BLA) were attributed by Kittanis (1950) to excessive lexcls of browsing b\' pronghorn. Park management established a goal to reduce the pronghorn herd by 50% (Kittams 1959); by 1969 pronghorn numbers were artificially re- duced from 600-800 to less than 200 dn-ough a combination of artificial reductions and severe winters (Barmore 1980). Houston (1982) provided alternative inter- pretations concerning big sagebrush. He re- ported increases in big sageliiaish numbers over all the northern winter range except the BLA near Gardiner, MX where numbers declined. Houston (1982) compared photos taken during the 1860s to photos retaken in the 1970s. He attributed the increase in big sagebrush at higher elevations to fire suppression and the decline in big sagebrush in the BLA to a return to more natural conditions following the removal of intense grazing by livestock in the early 1930s when the area was added to the park. In 1986 the U.S. Congress directed the National Park Sei^vice (NPS) to conduct a study to evaluate whether native ungulates were overgrazing the northern winter range (Congressional Record 1986). Our objectives were to document trends in big sagebrush abundance on a series of per- manently marked plots from 1958 to 1990. Height, canopy size, twig lengths, and annual production of shrubs were compared between browsed and unbrowsed sites. Study Area Shrub sampling was conducted on un- browsed (exclosed) and paired browsed sites at eight ungulate exclosures erected in 1958 and 1962 on Yellowstone's northern winter range. The eight exclosures, 2 ha in size, were locat- ed on gently rolling upland steppe ridge and the intervening swale habitats (Fig. 1). We divided the study area into a low-elevation stratum (the BLA of Houston 1982) with two exclosures, and a much larger, high-elevation stratum (n = 6 exclosures), based on large dif- ferences in ungulate species, elevation, snow- pack, precipitation, and big sagebrush sub- species (Fig. 1). Underlying soils are typic calciborolls, aridic haploborolls, and aridic calciborolls (Lane 1990). Precipitation averages 30 cm at the low- elevation exclosures and 55 cm at the high- stratum exclosures (Houston 1982, Despain 1991). The northern winter range is approxi- mate!)' 1100 km- and is located along the upper Yellow stone River drainage. Elevations range from 1500 m at the low-stratum exclo- sures to about 2200 m at the highest exclosure. The northern winter range is lower, warmer, and drier than the remaining higher plateaus of YNP (Houston 1982). As a result, 80% of the ungulates in the park during winter are found on the northern winter range (Singer 1991). Dominant shrubs at all the study sites include two subspecies of big sagebrush at the high-elevation stratum, nearly all mountain big sagebrush (A. t. vaseijana) with some basin big sagebnish {Aiicinisia tridentata tridentata), while Wyoming big sagebrush (A. t. wyotningensis) occurs in the low-stratum study sites. The big sagebrush subspecies vary markedly in their site requirements, growth, and preferences by ungulates (Beetle 1960, Welch et al. 1981, Beede and Johnson 1982, McArthur and Welch 1982). Rubber rabbitbrush {Chnjsothamnus noiiseosiis), green rabbitbrush {Chn/sothcnnuiis viscidiflorus), and horsebrush {Tetradymia canescens) are found at all study sites. Grayia spinosa and Athplex canescens occur at the low-stratum study sites (Houston 1982). Dominant grasses are bluebunch wheatgrass {Pseiidoroegneria spicato). Idaho fescue {Festuca idahoensis), junegrass {Koeleria pyramidata), bluegrasses {Poa compresses P- sandbergii, or P. pratensis), and thick-spike wheatgrass (A. dasystachyum; Houston 1982, Wambolt et al. 1987, Despain 1991). Pronghorn and mule deer occupy only the low-elevation stratum of approximatcK' 52 km^ within Yellowstone National Park (Barmore 1980, Houston 1982, Singer 1991). About one- half of the pronghorn population also sum- mers in the low-elevation stratum. The lower- elevation (about 1500 m) exclosures are locat- ed in typical, nearly snow-free, rolling xeric shrub and mixed grassland habitats. Elk occu- p\' both strata and the entire winter range of 810-1000 km- (Houston 1982, Singer 1991). Only elk and bison winter near the high-eleva- tion stratum exclosure sites. Winter snow depths near the higher exclosures (1639-2200 m) are typically 0.4-0.6 m, which are excessive for pronghorn and deer. The abrupt elevation rise for Mt. Everts separates the high and low strata. 1995] Ungulate Browsing in Yellowstone Park 203 YELLOWSTONE NATIONAL PARK Scale I 1 5 km Ungulate • Exclosures Fig. L Map of the northern Yellowstone ungulate winter range and the high-elevation stratum (six exclosures) and low-elevation stratum (two exclosures) big sagebrush study sites. Pronghorn, mule deer, and the Wyoming subspecies of big sagebrush were found only at the low-elevation stratum. The bison winter range expanded from about 130 km^ in the 1960s in the higher stratum to about 460 km^ in the late 1980s during a peri- od of bison population and range expansion (Meagher 1989, Singer and Norland 1995). Periodic bison use of the low-elevation stratum occurred following the population expansion in the late 1980s (Meagher 1989), but vegeta- tion measures reported here are nearly all prior to any bison use of the low-elevation stratum. Elk numbered about 8000 in 1958-1962 when initial monitoring of the sagebrush belt transects began. Elk were subsequently further reduced by artificial controls to less than 5000 in 1967 (Houston 1982). After cessation of con- trols, elk steadily increased, with counts rang- ing from 16,000 to 19,000 from 1982 to 1989 (Singer et al. 1989, Singer 1991). Bison were also artificially controlled until 1967. After cessa- tion of controls, bison on the northern range increased from less than 100 in 1967 to 850 by 1988 (Houston 1982, Meagher 1989). Prong- horn were artificially reduced from 600-800 to <200 (Barmore 1980), and pronghorn num- bers remained <200 until about 1981. During the 1980s — apparently due to milder win- ters— pronghorn increased to about 600 (Singer 1991). Mule deer counts increased from 1000 in 1985 to 2300 in 1988 over the entire deer winter range, the majority of which lies north of the park boundary (Singer 1991). Conversely, mule deer counts just within the park bound- aries declined from 230 in the 1960s (Barmore 1980) to about 100 in 1988, in spite of the overall herd increase. Methods Ungulate Densities and Diets Average ungulate densities near the exclo- sures were based on actual aerial counts made from fixed-wing aircraft as described in 204 Great Basin Naturalist [Volume 55 Barmore (1980), Houston (1982), Meagher (1989), Singer (1991), and Singer and Norland (1995). Densities are uncorrected for visibilitv bias (Samuel et al. 1988) and therefore repre- sent minimum axerage densities for the study periods — undoubtedly some animals were missed on the counts (Singer et al. 1989). Diets of all fom- ungulates found near the study sites on the northern winter range were estimated for each of three winters, December-March 1985-1988, from microhistological analysis of fecal samples (Washington State University, Wildlife Habitat Laboratory, Pullman). Each sample was a composite of 5 g of fresh dung material from 6-12 dung piles. Aggregate average percentages are reported for signifi- cant species and plant groups. To avoid confu- sion between similar species, fresh samples were collected for groups of animals immedi- ately after the groups had vacated an area. Bighorn sheep (Ovis canadensis) use steeper tenain on the northern range, and moose {Alces dices) are found at higher elevations; neither species was observed near the study sites. Shrub Utilization Rates Winter ungulate herbivory rates were sam- pled on the browsed transects in late winter- spring before leaf emergence (usuallv late April) in 1963-1969, 1987, 1989, and 1990. Percent twig utilization was obtained from counts of all browsed and unbrowsed twigs on each shrub located in the transect. Diameters at basal point and browsing point were measured on 20 ran- dom shoots on every fifth browsed shrub of each species, and bite sizes were estimated fol- lowing Pitt and Schwab (1990). Trends in Big Sagebrush, 1958-1990 Five exclosures were erected in 1957 and three more in 1962 (n = 8 total). The exclosures were placed in sites representative of mixed big sagebrush/bunchgrass communities. Paired belt transects (each 1.5 m x 30.5 m = 46.5 m^) were permanently located inside and outside eight of the exclosures (one per exclosure treat- ment) in big sagebrush communities (Canfield 1941, Parker 1954). Each matched pair of transects was as nearly comparable as possible in terms of slope, aspect, elevation, shrub species, and shrub cover (Barmore 1980, Houston 1982); nevertheless, differences might have occurred. Sampling of transects occurred at the date of exclosure, which should reveal any initial site differences. The transect for exclosure was selected arbitrarily. Heights and species of all shrubs found on the belt tran- sects were recorded in 1958, 1962, 1967, 1974, 1981, 1986, and 1990. Numbers of indi- vidual shrubs and any shrub seedlings were tallied. Aerial cover of all shrubs was mapped on graph paper, and shrub cover was later esti- mated using a grid (Barmore 1980, Houston 1982). Detailed Site Comparisons in 1986 and 1987 Shruli belt transects were not replicated at a site {n = 1 transect per treatment per loca- tion, 13 transects total) and were useful pri- marily for long-term trends and assessment of pretreatment conditions (Parker 1954). In 1986 and 1987 more intensive and better replicated measurements (n = 15 plots per treatment) were gathered; 15 circular plots, each 1.7 m in radius (9.3 m^), were randomly located in big sagebrush stands both inside and outside six e.xclosures. The tallest height, widest diameter, and perpendicular diameter were recorded for each shrub within each plot. The number of totally dead shrubs was recorded. The percent- age of dead material on partialK' li\ e shnibs was estimated. At eveiy fifth shrub of each species, lengths and diameters of 10 randomly sampled twigs were measured, and eveiy vegetative twig and reproductive stalk were counted. A mini- mum of 100 twigs of each species fi'om each site was collected, dried, and weighed. Canopy area for each individual shrub was estimated following Peek (1970) from the formula for the area of an ellipse: area \\)cnd^ where d^ = largest diameter and d^ = its per- pendicular diameter. All plots and long-term transects were located more than 25 m from exclosure fences to avoid the effects of snow- drifts or ungulate trails along the fences. Mean shrub height, largest crowii area, shoot numbers and lengths, and total shrub cover were compared using a two-way ANOVA, with browsing and exclosure location as treatments. The six exclosures should l)e considered repli- cations of one treatment (browsing) with 5 d.f used to test for differences among the treat- ment and replications. Replications at a location 1995] Ungulate Browsing in Yellowstone Park 205 included the 15 random plots in each treat- ment (15 X 2 X 8 locations = 240 plots of 9.3 m^). Nonparametric procedures were used for percent twig utilization comparisons between the 1960s and 1980s and for other data that were nonnormal or with unequal variances (F- max tests; Sokal and Rohlf 1981). Frequency distributions of shrubs in browsed and un- browsed plots were compared using the Kolomogorov-Smirnov test (Zar 1974). All dif- ferences discussed are significant at the P < .05 level unless othei^wise indicated. Aboveground biomass production of shrubs was estimated from the numbers of shrubs per plot times the average number of reproductive and vegetative shoots per plant times the aver- age dry weight of shoots. Regression equations for dry weight of shoots (independent vari- able) were calculated following MacCracken and Viereck (1990) from diameters at base (de- pendent variable) and length (dependent vari- able) of shoots. Separate regression equations were calculated for reproductive and vegetative shoots of big sagebrush, green rabbitbrush, and rubber rabbitbiaish. The regressions on diy weight were applied to the sample of all twig diameters and lengths to estimate average twig biomass. Results Ungulate Densities and Diets Ungulate densities approximately doubled during the study period on the low-elevation stratum (Table 1), whereas they approximately tripled on the high-elevation stratum during the same period (Table 1). Pronghoni consumed 81% shrubs in their diet, followed by mule deer 50%, elk 8%, and bison 1% (Table 2). Pronghorn diets were 49% big sagebrush, mule deer consumed 23%, and elk diets were only 4%. The higher combined ungulate den- sities and the presence of pronghorn and mule deer, both of which eat more big sagebrush, suggest that ungulate herbivoiy on big sage- brush will be greater on the low-elevation stratum study sites. Since pronghorn consume 12x more big sagebrush in their diets than elk and 2x more than mule deer, and since prong- horns also spend summers in the low-elevation stratum, we suspect pronghorn were the most important herbivore on big sagebrush on the low-elevation stratum. Table 1. Average minimum densities of elk, pronghorn, mule deer, and bison near the lower- and higher-ele\'ation exclosures on the northern winter range of Yellowstone National Park. Reported densities are based upon actual counts from fixed-wing aircraft (Houston 1982, Meagher 1989, Singer 1991) and are uncorrected for visibility bias. Ungulate density (no. /km-) Lower Higher Ungulate e.xclosures e.xclosures 1965-1968 Elk 6 6 Pronghorn 3 0 Mule deer 4 0 Bison 0 1 Total 13 7 1985-1988 Elk 16-19 16-19 Pronghorn 7-10 0 Mule deer 2 0 Bison tr 2 Total 25-31 18-21 Shrub Utilization Rates Big sagebrush utilization rates were consis- tently high (87%) and did not differ between 1963-1969 and 1985-1988 at the low-eleva- tion stratum study sites dominated by the more palatable (to pronghorn) Wyoming big sage- brush (Table 3, Mann-Whitney U tests, P > .05). Pronghorn and elk reductions during 1962-1967 apparently did not result in any decrease in percent leader use of Wyoming big sagebrush on the low stratum. Green rabbit- brush was also used heavily at the low-stratum sites where deer and pronghorn occuned (Table 3). Utilization rates of big sagebrush at the higher sites dominated by mostly mountain big sagebrush, however, increased about sixfold after ungulates increased threefold (Table 3, P < .05). Use of green rabbitbrush did not in- crease significantly at the high stratum during this period of ungulate increase. Percent leader use of big sagebrush at the lower-elevation sites averaged 87%, but leader use averaged only 11% at the higher sites. Bite sizes averaged 73% of vegetative shoots and 83% of repro- ductive shoots {n = 180 measured diameters of browsed shoots and 540 unbrowsed vegeta- tive and reproductive shoots). Consumption of annual aboveground biomass of big sagebrush by ungulates averaged about 68% at the low- elevation stratum sites and 9% at higher-ele- vation stratum sites. 206 Great Basin Naturalist [Volume 55 Table 2. Mean percent ofslirnl);, in winter diets ol lour uny;ulates on Yellowstone's northern winter range, 198.5-1988, determined by microhistological analysis of feces (x ± SE). Ungulate (no. aggregate Big Hahhit- Eiirc )tUl Fri nged Total samples) sagebnish hrus h" I ana \ta s; :ige shnibh ;'' X SE .V SE X SE X SE X SE Elk (28) 3.8 3.1 l.(i 2.7 0.4 1.0 1.0 2.0 7,8 4.1 Bison (25) 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 1.3 1.6 Mule deer (21) 23.2 15.1 7 2 4.(i 0.7 1.0 17.9 10.7 49.7 20.9 Pronghorn (20) 48.7 18.0 5.3 3.5 5.8 6.0 18.5 13.7 80.5 15.V ^Ral)l)itl)nisli (iiclmlt's Chnjsuthatnnm numcostt.s and C. lisklijldni.s ''Total slinilis also includes Poptilus spp.. Salix spp.. and Atripli-i spp Big Sagebrush Trends in Densities, Heights, and Cover, 1958-1990 Big sagebrush on belt transects in the lower stratum differed at the time of exclosure in 1958-1962. Densities were similar, but average heights were 50% and cover about 60% those values on transects selected for exclosure (Table 4). Big sagebrush densities, heights, and cover, however, were similar between browsed and unbrowsed transects at the time of exclosure on the high-elevation stratum. Apparently, ungulates were suppressing Wyoming big sagebrush on the low-elevation stratimi during the study period. Wyoming big sagebrush densities decreased 43% and big sagebrush cover decreased 29% on the low- stratum browsed site over the 31 -year period (Table 4). Density and cover of Wyoming big sagebrush increased dramaticalK' (350% and 830%, respectively) in the unbrowsed sites of the low sti-atum. Big sagebrush individuals were taller on unbrowsed sites (Table 4). Herbivory effects were less on the high- elevation stratum study sites, and all trends were similar for browsed and unbrowsed sites. Mountain and basin big sagebrush density declined, and canopy cover increased on both browsed and unbrowsed belt transects, 1958-1990 (Table 4). Wyoming and basin big sagebrush density declined 39%, but cover increased 39% on browsed sites over the 31 years. Heights of big sagebrush increased on both browsed and unbrowsed sites, but more on unbrowsed sites (Table 4). Detailed Site Comparisons of Densities, Cover, and Biomass Production in 1986 and 1987 Densities of big sagebrush (F = 50.9), total canopy cover of big sagebrush (F = 8.1), indi- vidual shrub crown area (F = 22.5), and heights of big sagebrush (F = 79.8, P < .05) differed between a much larger sample of browsed (n = 15) and unbrowsed plots {n = 15 per loca- tion, n = 180 total) sampled in 1986 and 1987. In each case, however, location was also signif- icant, and the interaction between location and browsing was significant. For example, sage- bmsh individuals were 59% taller on unbrowsed plots at six e.xclosure sites, but at the Blacktail exclosures sagebrush plants were taller on browsed plots. Heights of big sagebrush, green rabbitbrush, and horsebrush increased with elevation in both treatments. As a conse- quence of this exploratory analysis and signifi- cant interactions with location, our division of plots into a high and low strata appeared justi- fied, and we anaK'zed data from the lower and higher study sites separately in all subsequent analyses. Big sagebrush individuals were shorter and crowns smaller in browsed versus unbrowsed T.-\BLE 3. Percent of twigs browsed in big sagebrush communities on ^'ellowstone's northern range. Total un- gulate numbers increased twofold at the lower e.xclosures and threefold at the higher e.xclosures between 1963-1969 and 198.5-1988. The same transects of 46.5 m- each {n = 5) were sampled both periods; only these five browsed tran- sects were sampled 1963-1969. Location 196.3- -1969 198.5- -1988 Shmb species (»i = transects) X SE X SE Low elevation {n =2) Big sagebrush 88.0 4.2 86.8 7.2 Green rabbitbrush-' 70.1 10.5 Spiny hopsage-' 14.8 4.3 High elevations {n - 3) Big sagebrush 1.9 0.8 11.6 3.5* Green rabbitbrush 6.7 3.6 8.9 2.9 Horsebrush 46.6 11.4 'Only big sagebrush utilization was sampled 1963-1969, and green rabbitbrush at only the higher exclosures. *P < .05, according to Mann-Whitney U tests. 1995] Ungulate Browsing in Yellowstone Park 207 Table 4. Changes in densih', heights, and canopy cover of individnal big sagebnish shrubs between lime of exclosure placement in 1958 and 1990 on permanently marked 46.5 m^ shioib transects, Yellowstone's northern winter range. Heights Canopy cover Density of shrubs (< :'m) (m2/46.5 m-) 1958-1962 1990 1958-1962 1990 1 1958-1962 1990 Treatment X SE T SE X SE X SE X SE X SE Lower exclosures" Browsed 21 "■ 12 2 0.7 0.3 0.5 0.1 19 4 16 4 Unbrowsed 23 15 103 27 1.8 0.6 16.7 0.9 28 7 50 9 Higher exclosures'' Browsed 67 IS 41 1.7 1.9 0.9 5.9 1.6 12 2 42 7 Unbrowsed 72 34 37.S 6.2 1.6 0.2 8.6 0.9 10 2 82 9 *Big sagebrush subspecies in these transects, mostly A. t. wyomingensis, are apparent!) high!) palatable to pronghorns. ''Big sagebrush subspecies include mostly A. t. vaseyana. exclosure sites at low elevations (F = 29.8, 14.3, respectively), but there was no difference in heights or crowoi sizes due to browsing at the high-elevation sites (F > .05, Table 5). Horsebrush was shorter and crowns were smaller on browsed and unbrowsed exclosure sites at the higher elexations only (F = 14.5, 4.6, Table 5). Common rabbitbrush was short- er on browsed plots at the lower elevations, but it was taller on browsed plots at the higher- elevation exclosure sites (Table 5). Density of Wyoming big sagebrush was less on browsed versus unbrowsed plots at the lower exclo- sures (F = 14.7), but there was no effect of browsing at the higher exclosures (Table 6). No difference in the number of dead big sage- brush individuals was observed between browsed and unbrowsed plots at either eleva- tion category (F > .05). Twenty-two times more seedlings of the year were obserx^ed on browsed than unbrowsed plots at higher ele- vations (F = 2.7, Table 6). Big sagebrush contributed 82-99% of annual aboveground shrub production in these shrub communities. Browsing did not consistently influence the production of big sagebrush or green rabbitbrush at higher exclosure sites, but browsed rubber rabbitbrush produced less biomass at higher-elevation sites (Table 6). Botii Wyoming big sagebrush and rubber rabbit- brush produced much less aboveground bio- mass on browsed sites on the low study sites (Table 6). There was no influence from browsing on the number of vegetative or reproductive shoots per shrub for big sagebrush or green rabbit- brush. Reproductive shoots averaged 42% longer (Friedman test, Xr^ = 38, n = 6 locations, F < .05), and vegetative shoots averaged 45% longer on browsed versus unbrowsed big sagebiTish (Friedman test, %j.- = 42, n = 6 loca- tions, F < .05, Table 7). There was no effect of browsing on length of reproductive shoots of green rabbitbrush (F > .05). Discussion Other studies indicate mountain big sage- brush is preferred and eaten at a higher rate by mule deer and elk than Wyoming big sage- brush, while basin big sagebrush is the least preferred (Sheehy and Winward 1981, Welch et al. 1981, Fersonius et al. 1987). Our obsen^a- tions initially appear in constrast widi diis gener- alization; we observed 70% more winter utili- zation on Wyoming big sagebrush than moun- tain big sagebrush. Too few basin big sage- brush occurred on the study sites to draw any conclusions. Our data do not constitute a pala- tability test, however, in that mountain and Wyoming subspecies did not occur at the same study sites. We suspect pronghorn were the primary herbivore on Wyoming big sagebrush in lower study sites; pronghorn find the Wyo- ming subspecies highly palatable (Beetle 1960, Beetle and Johnson 1982), and that subspecies was more available to all ungulates due to shallow snows and more winds in the low stra- tum. Ungulate preference for big sagebrush subspecies also varies between locales (Welch et al. 1981, McArthur and Welch 1982); for example, Dietz and Nagy (1976) found Wyoming big sagebiiish was prefen^ed by mule deer in Colorado. Mountain and basin big sagebrush seedling germination, establishment, and survival were apparently enhanced by browsing and ungu- late grazing (possibly due to secondary effects 208 Great Basin Naturalist [Volume 55 Table 5. Iiulixidual slinil) crown and liciiilits of shnihs in lirowscd aiitl iiiihrowsed (protected) sites on Yellowstone's northern elk winter range. Samples were drawn Ironi in — ISO) plots oi 9.3 ni- eaeli located randoniK in browsed and nnhrovvsed sites in 1986 and 1987. ( jown area (cm-) Heigh ts (cm) Shrill) sjiecies Unbrowsed Br ■( )w sed Unl browsed Bi owsed Location J SE X SE X SE X SE Big sagebrush' Lower (Artrwy) 678 85 347 66* 50 2 37 4* Higher 798 90 524 51 79 3 71 2 Horscbrush Lower 45 7 37 13 15 1 11 2 Higher 575 222 71 93* 63 10 27 3* Common rabl)i thrush Lower 287 109 278 85 78 8 43 3* Higher 196 42 881 590* 59 5 50 11 Green rabbitbrush Lower 76 13 104 55 36 11 28 3 Higher 742 70 392 42 79 3 53 2* *Significant difference behveen grazed and control means using ANOVA, P < .05. "Big sagebrush subspecies included lower exclosures — A. t. wijomingensis only; higher exclosures — nii.xed piiinihiliniis nf A. I tridriitiita ; nearK' all A. t. vaseyanu. id -A. (- vasvifana. but Table 6. Estimated annual production (g/m-) of the most common shrubs in browsed and unbrowsed big sagebrush communities at six e.xclosures on Yellowstone's northern winter range (n — 15 plots each in both browsed and unbrowsed treatments at each site). Wyoming big sagebrush is found only at the lower-stratimi exclosures. and mi.xed populations of nearly all mountain with some basin big sagebiiish at the higher exclosures. Lower elev ations Higher elevations Unbrowsed Browsed Unbrowsed Br owsed Exclosure locatioTi X SE X SE X SE X SE Estimated biomass (g/m2) Big sagebrush 18 1.9 73.6 72.6 Green rabbitbrush 0.1 1.5 3.1 5.8 Common rabbitbrush 0.6 0.3 5.9 4.1 No. big sagebrusli indi\'iduals/9.3 m- No. alive 16 2 2 1* 13 2 15 2 No. dead 1.3 0.4 0.6 0.5 3.3 0.8 5.9 1.9 No. seedlings 0.8 0.3 0.2 0.1 0.2 0.1 4.4 1,4* *P < .0.5 according to t tests. No tests were conducted on biomass since it was estimated from a product of no. of plants X average no. of shoots X average weight of shoots. Tests were conducted on each of those parameters separately, however (see text and Table 7), suggesting statistically significant differences at the lower elevations. such as reductions of herbaceous vegetation competition and ungulate hoof action) at the higher winter range, but the opposite trend was observed on lower sites. McArthur et al. (1988) also observ'ed more big sagebrush seed- lings on a site browsed by mule deer in winter than on an imbrowsed site. The physical act of ungulate grazing, with its accompanying hoof action, greater soil disturbance, more bare ground, and less standing dead vegetation and Utter, may provide conditions more suitable to big sagebrush gemiination. Big sagebrush indi- viduals are smaller on browsed sites, which may also benefit establishment and survival of seedlings due to reduced competition for light, soil moisture, and other resources. Ungulate herbivory suppressed big sage- brush on die lower-elevation sites, where almost no recruitment of Wyoming big sagebrush occuiTcd on browsed sites; apparenth' few seed- lings survive the intense browsing. Wyoming big sagebrush reproduces more successfully 1995] Ungulate Browsing in Yellowstone Park 209 Table 7. Numbers and lengths of reproductive and vegetative stalks on shrubs in browsed and unbrowsed plots in big sagebrush communities on Yellowstone's northern range (n = no. shnibs). No. 1 reproductive flowers/shrub Length (cm) of flower st, alks Unbrowsed Brow sed Ui ibrowsed Browsed Species Location X SE a" SE X SE X SE Big sagebrush Lower (Wyoming spp. Higher (basin and only) 15 6 13 10 4 0.2 8 0.6** mountain spp.) Green rabbitbrush 21 7 17 6 14 0.7 18 0.6** Higher 10 3 16 10 11 0.4 12 0.4 No 1. vegetative shoots/shn lb Length (cm) i 2 0.1 of vegetative : 3 shoots Big sagebrush Lower 99 23 88 18 0.4 Higher Green rabbitbrush 83 31 59 12 5 0.5 7 0.5** Higher 3.3 7 52 20 6 0.2 8 0.4** *P < .0.5. **P < .01. Differences between numbers in browsed and unbrowsed plots were tested vrith f tests and lengths v\ith Mann-\\'hitne\ U tests. than the other subspecies on xeric sites (Welch and Jacobson 1988), and the xeric, sodic clay soils of the low stratum are clearly more suit- able to Wyoming big sagebrush. At tlie high lev- els of ungulate herbivory we observed (rough- ly 68% biomass removal), the Wyoming sub- species is presently suppressed by ungulates. The ability of Wyoming big sagebrush to recover from herbivory is less than for moun- tain and basin big sagebrush. Wyoming big sagebrush is shorter (individuals often do not exceed 0.3 m), seedling growth rates are lower, and current annual growth is less than for the other two subspecies (McArthur and Welch 1982, Booth et al. 1990). The approximately 66% decline in numbers of mule deer using the lower stratum within the park over the past two decades may be due to the localized Wyo- ming big sagebrush decline. Pronghom did not decline in the lower stratum during the same period, but pronghom, unlike mule deer, were artificially reduced well below carrying capac- ity levels during the 1960s (Houston 1982), and they may still be recovering from the reductions. Increases in height and cover of big sage- brush are reported after protection from ungu- lates. Robertson et al. (1970) reported big sage- brush cover increased 76% after 30 years of protection from browsing, although mean heights declined 12%. Heights and crov^ni sizes were similar, but live cover by big sagebrush was greater on unbrowsed sites on a mule deer winter range, primarily due to a greater die- back of browsed big sagebrush (McArthur et al. 1988). Average crown dieback was 64% in the browsed area and 17% in the unbrowsed area (McArthur et al. 1988). Mule deer use was heavy (370 deer-use days/ha), and dieback of big sagebrush occurred after two successive win- ters of heavy snowfall (McArthur et al. 1988). Browsing by native ungulates stimulated seedstalks and leaves of big sagebrush and leaves of green rabbitbrush on the study sites. Stagnation of shrubs occurred inside big game exclosures after only two years of exclosure — nonuse of big sagebrush resulted in an aver- age 36% reduction in biomass production over clipped plants (Tueller and Tower 1979). Numbers of sprouts of green rabbitbrush were similarly increased by clipping (30% herbage removal), and new growth was longer, leaves were larger, and leaves remained green for one month longer (Willard and McKell 1978). On the other hand, browsing of more than 80% of the leaders of mountain big sagebrush by mule deer resulted in a reduction of 50-93% in total number of seedstalks per plant and a reduction of 0-53% in length of seedstalks (Wagstaff and Welch 1991). Grazed grasses on the northern Yellowstone winter range have higher protein levels (Coughenour 1991), and grazing stimulates aboveground growth of grasses (Frank and McNaughton 1993). Increased vigor in new growth of browsed shrubs on the Yellowstone northern winter 210 Great Basin Naturalist [Volume 55 range is consistent with these observations of" grasses, and shrub vigor may be the result of increased rates of nutrient cycling due to ungulate defecation and urination (McNaugliton 1979). In addition, plant competition is reduced and water availability may be increased on browsed sites on the northern winter range due to smaller crown sizes and fewer transpir- ing tissues for indixidual shrubs. Historic mean duration between fires was 25 years on the Yellowstone northern winter range, but due to fewer fire starts and active fire suppression, no significant burning of grasslands occurred between 1870 and 1988 (Houston 1973, Romme and Despain 1989). Big sagebrush communities had not yet achieved climax postfire state on the northern winter range as indicated by increases in heights and cover of both browsed and unbrowsed big sagebrush individuals between 1958 and 1990. Browsed big sagebnish communities on higlier- elevation ranges were replacing themselves; many successful seedlings and small individuals were observed on browsed versus unbrowsed sites. Lomasson (1948) observed almost no reproduction for 40 years in a stand of big sagebiTish, but then reproduction increased as the original population began dying. Average life span of big sagebrush is 53-72 years, and in a mature, undisturbed stand, most big sage- brush individuals are in the 55-59-year age class (Roughton 1972). Sagebrush recovery following fire varies from a few years to 30 years depending upon environmental condi- tions for reestablishment (Sneva 1972, Harniss and Murray 1973). If most big sagebrush com- munities we studied on the northern winter range last burned in the 1840-1890 period (Houston 1973), then most big sagebrush pop- ulations should have approached senescence and population turnover at the time of the 1986-87 sampling. Ungulate herbivory levels on the lower study stratum restricted growth, establish- ment, and survival of big sagebrush at the time of this investigation, although browsed big sagebrush communities were stable or in- creasing at the higher elevation. Suppression of growth and reproduction of plants by in- creasing native ungulates can result in a new, altered plant-ungulate equilibrium (Sinclair 1977, Caughley 1981). If unnatural (human- caused) concentrations of ungulates cause plant alterations, the situation is not accept- able under NFS policy (U.S. Department of the Interior 1988). Houston (1982) concluded densities of ungulates in the BLA were unnat- ural and artificially high due to animal avoid- ance of hunting outside the park. If so, some form of ungidate management — control, en- couragement of migrations — is justified on the BLA. Elk and pronghorn reductions in the 1960s, however, did not reduce percent leader use or improve the declining status of big sagebrush in the BLA. Either effective ungu- late densities remained the same near the big sagebrush study sites, the ungulate reductions did not go on long enough, or high preference for the Wyoming subspecies by pronghorn maintained high levels of herbivoiy in the area during the control period. We caution that we were unable to calculate appropriate or recommended ungulate herbi- voiy levels or ungulate densities for the north- ern winter range. Our data included two dichot- omous periods in ungulate management. The first period of our study, 1958-1968, was clear- ly a period of ungulate underpopulation dur- ing which time elk, bison, and pronghorn were controlled far below ecological carrying capacity (ECC) densities (Barmore 1980, Houston 1982, Boyce 1993, Mack and Singer 1993, Singer and Norland 1995). The second period of our investigations, 1986-1988, likely was a time of ungulate densities in excess of natural conditions, at least for elk and bison. This statement is not based upon any compar- isons to conti^ol conditions (no similar ecosystem exists with wolves and nondisrupted migra- tions for a comparison), but upon the conclu- sions of Houston (1982) that elk concentra- tions were unnaturally high in the low-eleva- tion BLA stratum, and computer predictions that elk and bison would number 8-25% less following wolf restoration (Garton et al. 1990, Boyce 1993, Mack and Singer 1993). Pronghorn densities in relation to EGG are unknown — one author feels coyotes {Caiiis kitrans) are suppressing pronghorn on the northern Yellow- stone winter range and that, following wolf restoration, coyotes will decline and prong- horn will further increase (Berger 1991). Wolf restoration occurred on the study area during tlie winter of 1994-95, providing an opportunity to test the effects of wolves upon ungulate- plant interactions in the Yellowstone ecosys- tem (Gook 1993). 1995] Ungulate Browsing in Yellowstone Park 211 Acknowledgments The research was funded by the U.S. Depart- ment of the Interior, National Park Service, Natural Preservation Program, Washington D.C., and Yellowstone National Park. The authors acknowledge J. Varley and R. Barbee for administrative support, and D. Frank, W. Wiens, J. Whipple, G. Kittel, M. Hennen, J. Meek, and M. Harter for field assistance. D. Swift, M. Coughenour, A. Beetle, E. Durant McAithur, and J. Whipple reviewed the manuscript. Literature Cited Barmore, W, J. 1980. Population characteristics, distribu- tion and habitat relationships of six ungulates in northern Yellowstone Park. Unpublished report, Yellowstone National Park files. Mammoth, \VY. Beetle, A. A. 1960. A study of sagebrush: the section Tridentata of Artemisia. University of Wyoming, Agricultural E.xperiment Station Bulletin .368. 68 pp. Beetle, A. A., and K. L. Johnson. 1982. Sagebrush in Wyoming. Wyoming Agricultural Experiment Station Bulletin. 68 pp. Berger, J. 1991. Greater Yellowstone's ungulates; myths and realities. Consei^vation Biology 5: 353-363. Booth, G. D., B. L. Welch, and T. L. C. Jacobson. 1990. Seedling growth rates of 3 subspecies of big sage- brush. Journal of Range Management 43: 432-436. Boyce, M. S. 1993. Predicting the consequences of wolf recovery to ungulates in Yellowstone National Park. Pages 234-269 in R. S. Cook, editor. Ecological issues on reintroducing wolves into Yellowstone National Park. U.S. National Park Service Science Monograph 22. Denver, CO. Cahalane, V. H. 1943. Elk management and herd regula- tion— Yellowstone National Park. Transactions of the North American W'ildlife Conference 8; 95-101. Canfield, R. 1941. Application of the line interception method in sampling range vegetation. Journal of Forestry 39: 388-394. Caughley, G. 1981. Overpopulation. Pages 7-19 in P A. Jewell, S. H. Holt, and D. Hart, editors, Problems in management of locally abundant wild animals. Academic Press, New York. 361 pp. Cayot, L. J., J. Prukop, and D. R. Smith. 1979. Zootic cli- max vegetation and natural regulation. Wildlife Society Bulletin 7; 162-169. Chase, A. 1986. Playing God in Yellowstone. Atlantic Monthly Press, Boston. 446 pp. Cole, G. F 1971. An ecological rationale for the natural or artificial regulation of ungulates. Transactions of the North American Wildlife Conference 36: 417-425. Congressional Record. 1986. Senate S. 12613. U.S. House of Representatives. Cook, R. S., editor. 1993. Ecological issues on reintro- ducing wolves into Yellowstone National Park. National Park Service Science Monograph 22. Denver, CO. Coughenour, M. B. 1991. Biomass and nitiogen responses to grazing of upland steppe on Yellowstone's north- ern winter range. Journal of Applied Ecology 28: 71-82. Despain, D. G. 1991. Yellowstone vegetation, conse- quences of environment and historv' in a natural set- ting. Roberts Rinehart Publishing, Boulder, CO. 239 pp. DiETZ, D. R. and J. G. Nagy. 1976. Mule deer nutrition and plant utilization. Pages 71-78 in C. W. Workman and J. B. Low, editors. Mule deer decline in the West — a symposium. Utah State University, College of Natural Resources, Logan. 134 pp. Frank, D., and S. J. McN,\ughton. 1993. Interactive ecology of plants, large mammalian herbivores and drought in Yellowstone National Park. Unpublished doctoral dissertation, Syracuse University, Syracuse, NY 150 pp. Carton, E. O., R. L. Crabtree, B. B. Ackerman, and C. Wright. 1990. The potential impact of a reintro- duced wolf population on the northern Yellowstone elk herd. Pages 3-59 in Yellowstone National Park, U.S. Fish and Wildlife Service, University of Wyo- ming, University of Idaho, Interagency Study Team, University of Minnesota Cooperative Park Studies Unit, editors. Wolves for Yellowstone? Report to the United States Congress. Volume II. Research and analysis. Harniss, R. O., and R. B. Murray 1973. Thirty years of vegetation change following burning in sagebrush- grassland range. Journal of Range Management 26: 322-325. Houston, D. B. 1973. Wildfires in northern Yellowstone National Park. Ecology .54: 1111-1117. . 1976. Research on ungulates in northern Yellow- stone National Park. Pages 11-27 in Research in the parks. Transactions of the National Park Symposium, National Park Sei-vice Symposium No. 1. . 1982. The northern Yellowstone elk: ecology and management. Macmillan Publishing Co., Inc., New York. 474 pp. Kay, C. E. 1991. Yellowstone's northern elk herd: a critical evaluation of the "natural regulation" paradigm. Un- published doctoral dissertation, Utah State University, Logan. 490 pp. Kittams, W. H. 1950. Sagebrush on the lower Yellowstone range as an indicator of wildlife stocking. Yellow- stone National Park files. Mammoth, WY. 14 pp. . 1959. Future of the Yellowstone wapiti. Naturalist 10: 30-39. Lane, J. 1990. Characterization and comparison of soils inside and outside of grazing e.xclosures on Yellow- stone National Park's northern winter range. Un- published master's thesis, Montana State University, Bozeman. Lomasson, T. 1948. Succession in sagebrush. Journal of Range Management 1: 19-21. MacCracken, J. G., AND L. a. Viereck. 1990. Browse regrowth and use by moose after fire in interior Alaska. Northwest Science 64: 11-18. Mack, J. A., and F J. Singer. 1993. Using Pop-II models to predict effects of wolf predation and hunter har- vests on elk, mule deer, and moose on the northern range. Pages 49-74 in R. S. Cook, editor. Ecological issues on reintroducing wolves into Yellowstone National Park. National Park Service Science Mono- graph 22. Denver, CO. McArthur, E. D., and B. L. Welch. 1982. Growth rate differences among big sagebrush (Artemisia tridentata) accessions and subspecies. Journal of Range Manage- ment 35: 396-401. 212 Cheat Basin Natufl\list [Volume 55 McArthur, E. D., a. C. Blauer, and S. C. Sanderson. 1988. Mule deer induced mortaliU' of mountain hi^ sagebrush. Journal of Range Management 41; 114-117. McN.AUGHTON, S. J. 1979. Grazing as an optimi/.ation pro- cess: grass-ungulate relationships in the Serengeti. American Naturalist 113; 691-703. MEAf;iiER, M. M. 1973. The bison of Yellowstone National Park. National Park Service Science Monograph Series 1. 161 pp. . 1989. Range expansion by bison of Yellowstone National Park. Journal of Mammalogy 70; 670-675. Parker, K. W. 1954. A method for measuring trend in range condition on National Forest ranges with sup- plemental information for measurement of vigor, composition, and browse. USDA Forest Service Report. 37 pp. Peek, J. M. 1970. Relation of canopy area and volume to production of three woody species. Ecology 51; 1098-1101. . 1980. Natural regulation of ungulates. Wildlife Society Bulletin 8: 217-227. Personius, T. L., C. L. Wambolt, J. R. Stephens, and R. G. Kelsey. 1987. Crude terpenoid influence on mule deer preference for sagebrush browse. Journal of Range Management 40: 84—88. Pitt, M. D., and F. E. Schwab. 1990. Assessment of a nondestructixe method for estimating browse use. Journal of Wildlife Management 54: 175-179. Robertson, J. H., D. L. Neal, R. McAdams, and R T. TUELLER. 1970. Changes in crested wheatgrass ranges under different grazing treatments. Journal of Range Management 23; 27-.34. ROMME, W. H., and D. G. Despain. 1989. Historical per- spective on the Yellowstone fires of 1988. BioScience 39; 695-699. ROUGHTON, R. D. 1972. Shrub age structure on a mule deer winter range in Colorado. Ecology' 53; 615-625. Rush, W. M. 1932. Northern Yellowstone elk study. Mon- tana Fish and Game Commission, Helena. 131 pp. Samuel, M. D., E. O. Carton, M. W. Schlegel, and R. G. Carson. 1988. Visibility bias during aerial sun'eys of elk in north-central Idaho. Journal of Wildlife Man- agement 51; 622-630. Sheehy, D. R, and a. H. Winward. 1981. Relative palata- l:)ility of seven Artemisia taxa to mule deer and sheep. Journal of Range Management 34: .397-399. Singer, E J. 1991. The ungulate prey base for wolves in Yellowstone National Park. Pages .323-348 in R. B. Keiter and M. S. Boyce, editors. The Greater Yellow- stone Ecosystem: redefining America's wilderness heritage. Yale University' Press, New Haven, CT. Singer, F J., and J. Norland. 1995. Niche relationships within a guild of ungulates following release from artificial controls, Yellowstone National Park, Wyo- ming. Canadian Journal of Zoology 72; In press. Singer, F J., W. Sghreier, J. Oppenheim, and E. O. Carton. 1989. Drought, fires, and large mammals. Bioscience 39; 716-722. Sinclair, A. R. E. 1977. The African buffalo; a study of resource limitation of populations. University of Chicago Press, Chicago, IL. 355 pp. Sneva, F a. 1972. Grazing return following sagebrush control in eastern Oregon. Journal of Range Manage- ment 25; 174-178. SOKAL, R. R., AND E J. RoHLE 1981. Biometry. W. H. Freeman, San Francisco, CA. TuELLER, R J., AND J. D. TowER. 1979. Vegetation stagna- tion in three-phase big game exclosures. Journal of Range Management 32; 258-263. U.S. Department of Interior, National Park Service. 1988. Management policies. U.S. Government Print- ing Office, Washington, DC. Wagstaff, F J., AND B. L. Welch. 1991. Seedstalk pro- duction of mountain big sagebiiish enhanced througli short-term protection from heavy browsing. Journal of Range Management 44: 72-74. Wambolt, C. L., R. G. Kelsey, T. L. Personius, K. D. Striby, a. F McNeal, and K. M. H.wstad. 1987. Preference and digestibility of three big sagebrush subspecies and black sagebnish as related to crude terpenoid chemistry. Pages 71-73 in E D. Provenza, J. T. Flinders, and E. D. McArthur, compilers. Sym- posium on plant-herbivore interactions. USDA Forest Service, General Technical Report INT-222. Ogden, UT. Welch, B. L., and T L. C. Jacobson. 1988. Root growth of Artemisia tndentata. Joum;d of Range Management 41;. 3.32-334 Welch, B. L., E. D. McArthur, and J. N. Davis. 1981. Differential preferences of wintering mule deer for accessions of big sagebrush and black sagebrush. Journal of Range Management 34: 409-411. Will.\rd, E. E., and C. M. McKell. 1978. Response of shrubs to simulated browsing. Journal of Wildlife Management 42: 514-519. Zar, J. H. 1974. Biostatistical analysis. Prentice-Hall Inc., Englewood Cliffs, NJ. 620 pp. Received 13 September 1994 Accepted 19 January 1995 Great Basin Naturalist 55(3), © 1995, pp. 213-224 SOFT SEDIMENT BENTHIC MACRO INVERTED RATE COMMUNITIES OF THE GREEN RIVER AT THE OURAY NATIONAL WILDLIFE REFUGE, UINTAH COUNTY, UTAH Eric R. Wolzl and Dennis K. Shiozawa^-^ Abstract. — Benthic macroinvertebrates from four habitat types (river channel, ephemeral side channel, river back- water, and seasonally inundated wetland) were e.xamined from the Green River at the Ouray National Wildlife Refuge, Uintah County, UT, June-August 1991. Four major taxa (Nematoda, Oligochaeta, Diptera: Ceratopogonidae, and Chironomidae) were quantified. Cluster analysis of densities showed that habitat types with comparable flow conditions were the most similar. Highest to lowest overall benthic invertebrate densities of the four habitats were as follows: ephemeral side channel, river backwater, seasonally inundated wetland, and river channel. Nematodes were the most abundant taxon in all habitat t\'pes and sample dates e.xcept the August sample of the river channel and river backwater and the July sample of the seasonally inundated wetland. Key words: benthic macroinvertebrates, Nematoda, Oligochaeta, Ceratopogonidae, Chironomidae, river benthos, wetland, benthos. Green Riven In 1962 Flaming Gorge Dam was completed on the Green River in northeastern Utah. This, in addition to dikes constnicted along the river's course and the introduction of nonnative fishes, has altered natural conditions such that many native fishes have reached the brink of extinc- tion and are now listed as endangered species. Grabowski and Hiebert (1989) studied the Green River below Flaming Gorge Dam and noted the importance of backwaters as nursery habitats to introduced and native fishes. They found the most important food items to be ben- thic macroinvertebrates, predominantly chiro- nomid larvae. Their investigation was confined to two habitats: the main channel and river backwaters. We also studied benthic commu- nities of the river channel and back-water habi- tats and two additional habitats — seasonally inundated wetlands and ephemeral side chan- nels. No published information exists about the community structure of benthic macro- invertebrates in these latter two habitat types. Benthic invertebrates of large rivers are poorly known. Difficulty in sampling, the amount of time needed to process samples, identification of specimens after collection, and heterogeneity of habitats make study diffi- cult and often expensive. Studies of riverine systems have utilized divergent methodologies. Some studies randomly sample an entire river cross section and do not attempt to quantify dif- ferent river habitat types (Grzybkowska 1989, Grzybkowska et al. 1990, Munn and Brusven 1991). Other studies have been directed toward specific river habitats such as riffles (Rader and Ward 1988, Morgan et al. 1991), floodplains (Gladden and Smock 1990), or tailwaters of re- servoirs (Swink and Novotny 1985). Relatively few have simultaneously studied multiple habitat types in a single river system (Beckett et al. 1983, Grabowski and Hiebert 1989). Our purpose was to determine densities and community assemblages of the major ben- thic macroinvertebrates in four Green River habitats: river channel, ephemeral side channel, river backwater, and seasonally inundated wet- land. Benthic samples were taken from lune through August 1991, in the Green River at the Ouray National Wildlife Refiige, Uintali County, UT USA. Study Sites The Green River originates in Wyoming and flows south through eastern Utah to its conflu- ence with the Colorado River (Fig. 1). It adds more volume to the Colorado River system than any other tributary. In eastern Utah, at river km 404, the Green River enters the Ouray National 'Chadwick & Associates, Inc., Littleton, CO 80120. ^Department of Zoology, Brigham Young University', Prove, UT 84602 USA. ■'Author to whom correspondence should be addressed. 213 214 Great Basin Naturalist [Volume 55 Wildlife Refuge. This seetion of the river has the lowest gradient of the entire Green River system. Riparian vegetation consists of willow and tamari.x with occasional cottonwoods. We collected monthly samples in the Ouray National Wildlife Refuge (see also Fig. 2). In addition to benthic samples, water chemistry was determined for each habitat type on each sample date (Table 1). Salinity and conductivity were recorded with a YSI meter (Yellowstone Instruments); turbidity was measured with a nephelometer; and hardness, pH, and alkalini- ty were determined with a Hach Kit (Hach Chemical Corporation). Water chemistry was recorded at three locations per sample area on each sample date. At each site, a min-max ther- mometer was placed near the benthos-water interface at the time of sampling and left for 10 days. Substrate composition was estimated visually. River Channel The river channel was sampled approximate- ly 1.3 km north of the United States Fish and Wildlife Service (USFWS) hatchery on the Ouray National Wildlife Refuge. Sampling was adjacent to a sand bar that decreased water tmbulence and prevented shifting sands. Water chemistry values were relatively stable. Turbid- ity was substantially higher during the August sample. Substrate consisted mostly of sand with little silt and detritus. Water levels were too high during June (peak flow) to allow sampling. Ephemeral Side Channel During high flows the Green River will occupy various smaller channels that are diy during low-flow intervals. We have named such habitats "ephemeral side channels." The ephemeral side channel studied was approxi- mately 2.75 km south of the USFWS hatchery. For most of the year water levels in the main channel were below the level of the ephemeral side channel. However, during peak flow, water filtered through a wooded area and gathered into the channel, which was 10 m wide and 500 m long. As the river level dropped, flow slowed and eventually stopped. Because the side channel dried up shortly after the July sample, no August sample was taken. Most notable of the water chemistry measurements was the increase of salinity and alkalinity when comparing June to July. Water temperature also deviated more during July. Substrate con- sisted mostly of firm silt and detritus with little sand. Sediment deposition contributed little to the site during our study. SEASONALLY I^aINDATED WETLAND 100 Kilometers Fig. 1. Regional map showing the location of the Oura\' National Wildlife Refuge. Fig. 2. Local map of the Ouray National Wildlife Refuge, Uintah Count\; UT, showing the location of sam- pling sites. 1995] Macroinvertebrates of the Green River 215 Table 1. Mean ± standard deviation water cheniistiy temperature in °C, salinity in percent, condiicti\it\ CaCOg). values from Green River sample sites, June-August 1991 (n — 3, in /xmhos, turliidity in NTUs, hardness and alkalinity in ppm Habitat type Date M in. /max temp. pll Salinity Conductivity Turbidity Hardness Alkalinity River channel 7/15 * 8. 14 ±.09 .04 ± .0 753 ± 6 183 ±318 411±0 183 ± 10 8/12 20.5/26.5 8.48 ±.10 .04 ± .01 718 ± 8 402 ±41 320 ± 20 205 ±17 Ephemeral side channel 6/3 20.5/30.5 9.0 ±0 .03 ± .06 326 ± 10 57 ±6 183 ± 20 171 ±0 7/1 16/30.5 9.14 ±.16 .12 ±.03 445 ± 5 127 ±21 228 ± 10 240 ± 17 River back\\ater 7/10 20.5/29.5 7.98 ± .23 .01 ±.01 523 ± 23 57 ±9 228 ± 10 183 ± 20 8/8 19/26.5 8.59 ±.12 .03 ± .0 730 ±111 45 ±11 268 ± 40 228 ± 26 Seasonallv inundated we tland 6/10 19.5/26.5 9.0 ±0 ,02 ± .0 314 ±8 52 ±8 154 ±0 143 ± 10 7/12 22/32 8.37 ±.11 .02 ± .01 446 ± 20 36 ±8 205 ±0 223 ±0 8/15 22/29.5 8.93 ±.1 .01 ± .0 345 ± 13 195 ± 17 171 ±17 154 ± 0 *Theniioineter lost River Backwater River backwaters are submerged during high flows and do not emerge as distinct entities until the river drops. For this reason the river backwater was not sampled during peak flow (June). The river backwater we sampled, located just upstream of the river channel site described above, was approximately 10 m wide X 50 m long and 1.3 m deep. Turbidity, alkalinity, and pH were highest during the August sample. Substrate consisted mostly of loose silt and detritus with virtually no sand. Silt and detritus were constantly being deposited during the study period. Seasonally Inundated Wetland This site, commonly called "Old Charlie's Wash," is a shallow floodplain wetland man- aged by the USFWS for waterfowl and is located approximately 4.3 km south of the USFWS hatchery. As the river rises in the spring, water enters Old Charlie's Wash and, at peak flow, retaining structures are put in place to create a 43-ha pond and to prevent the impounded water from receding as rapidly as the river. By early fall the water in Old Charlie's Wash is nearly depleted by seepage and evaporation. Turbidity increased dramati- cally during the August sample, and conduc- tivity, hardness, and alkalinity peaked during the July sample. Substrate consisted of firm silt, detritus, and sand. Methods Sampling Samples were collected during the summer of 1991 (Tables 2-5). Initial sampling of the ephemeral side channel and seasonally inun- dated wetland occurred just after river flow peaked in early June, but samples for the river channel and backwater habitats were not col- lected because the water level was too high. All four habitats were sampled during July and all but the ephemeral side channel during August. Fifty core samples were taken along a 30-m transect at each site. Each sample was collect- ed with a clear aciylic tube, 450 mm long x 47 mm in diameter (Shiozawa 1985), which was pushed into the substrate to a depth of 60-80 mm. Sediment from each sample was preserved in 5% formalin with rose bengal stain added to aid in sample sorting. Sample Processing In the laboratory we washed each sample to separate organisms from sediments using the following procedure. First, the formalin was drained and replaced with tap water. The sample was then gently stirred to resuspend the sediments and poured into a plastic tray (36.5 cm X 31.5 cm X 6 cm) through which a small volume of warm water flowed. The out- flowing water, laden with small sand and clay particles, detritus, and benthic invertebrates, was filtered through a 63-^tm screen. Larger 216 Great Basin Naturalist [Volume 55 Table 2. Densities of bentliic iiivertehrates (#/ni-) from tlic Green River, river channel hal)itat, Onray National Wildlife Reftige, Ouray, UT 15 July 1991 12 August 1991 # of samples # ot samples Taxon Density/m2 (95% C.L.) processed Density/m2 (95% C.L.) processed Nematoda 24,881 (13,107-47,302) 6 2421 (2063-2840) 5 Oligochaeta 3426 (2565-4570) 18 11,182 (7497-16,678) 5 Insecta Ceratopogonidae 3608 (2731-4767) 27 13,026 (9316-18,215) 5 Chironomidae 4150 (2798-6155) 5 3516 (2454-5037) 30 Earlv instars 1037 3016 Chirottomus 346 0 Cijphomella 0 58 Lenziella 576 0 Paramerina 115 0 Paratendipes 0 96 Polijpedihim 1844 269 Procladiiis 115 0 Psectrocladius 115 0 StempellineUa 0 58 Tanijtarsus 0 19 T.\BLE 3. Densities of benthic invertebrates {#/m~) from the Green River, ephemeral side channel habitat, Ouray National Wildlife Refuge, Oura>, UT 3 June 1991 1 July 1991 # of samples # of samples Taxon Density/m2 (95%^ C.L.) processed Densit y/m2(95%C.L.) processed Nematoda 261,680 (88,934-769,968) 5 302,603 (215,886-424,154) 5 Oligochaeta 2728 (2096-3546) 15 12,796 (10,681-15,329) 5 Insecta Ceratopogonidae 0 30 0 5 Chironomidae 2325 (1843-2927) 30 8185 (6385-10,491) 5 Earlv instars 979 2075 Chironomus 1134 3112 Cryptochironomus 0 115 Cryptotendipes 19 461 Lenziella 96 1383 Polypedilum 19 692 Procladius 0 346 Tanypus 19 0 Tanytarsits 58 0 .sediment particles (sands and structural clays) that remained in the plastic tray were periodi- cally examined for specimens. If none were found, the sediments were discarded. Material collected on the screen was stored in 70% ETOH. Samples sorted were randomly chosen from the 50 samples taken at each site and date. Each sample was placed in glass petri dishes (from one to six dishes depending on the amount of material) and sorted under a dissecting micro- scope (see Tables 1-4 for number of samples processed). Four major taxa (Nematoda, Oligo- chaeta, Ceratopogonidae, and Chironomidae) were counted. Only Chironomidae were iden- tified to the generic level. Miscellaneous taxa were also recorded but were not quantified (see Table 5). The number of samples sorted fiom each site and sampling date was determined as follows: 5 of the 50 samples were randomly selected and the four major taxa were counted. Because of their contagious distribution (determined by calculating variance to mean ratios), numbers of individuals of each taxon were then log transformed (x + 1). The variance and mean 1995] Macroinvertebrates of the Green River 217 Table 4. Densities of benthic invertebrates (#/ni-) from the Green River, river backwater habitat, Ouray National Wildhfe Refuge, Ouray, UT 10 July 1991 8 August 1991 # of sam pies # of samples Taxon Density/m- (95% C.L.) process ied Density/rn- (95% C.L.) processed Nematoda 54,872 (24,350-123,650) 5 134,183 (94,656-190,542) 5 Ohgochaeta 26,642 (14,622-48,495) 9 164,731 (101,881-266,728) 5 Inseeta Ceratopogonidae 96 (90-107) 30 461 (385-552) 30 Chironoinidae 31,125 (15,356-63,089) 5 22,863 (12,139-13,136) 6 Earlv instars 8877 7301 Chironomus 7032 6340 Lenziella 346 1249 Polypedilum 14,179 5860 Prochidiits 461 1345 Psectrodadius 115 0 Tamjtarsus 115 769 Table 5. Densities of benthic invertebrates (#/m2) from the Green River, seasonally inundated wetland habitat, Ouray National Wildlife Reftige, Ouray, UT 10 June 1991 12 July 1991 15 August 1991 Densitv/m^ # of samples Densit>7m- # of samples Density/m- # of samples Ta.\on (95% C.L.) processed (95% C.L.) processed (95% C.L.) ] Drocessed Nematoda 7133 (4534-11,266) 8 80,694 (38,595-168,713) 5 88,533 (83,125-94,784) 5 Oligochaeta 4573 (3402-6141) 30 87,150 (39,242-193,547) 10 22,249 (11,930-41,494) 5 Inseeta Ceratopogonidae 0 30 0 14 2478 (1941-3165) 20 Chironomidae 903 (895-915) 30 23,055 ; (13,707-38,780) 14 3977 (2816-5617) 10 Earlv instars 96 8769 2479 Ablabesmijia 0 124 0 Chironomus 154 41 576 Cricotopus 19 453 0 Cnjptochironomus 134 206 0 Cryptotendipes 58 947 346 Glyptotendipes 58 988 0 Lenziella 115 1112 0 Microtendipes 0 1029 0 Paratamjtarsiis 231 6505 58 Polypedilum 19 2388 173 Procladius 0 124 58 Psectrocladius 0 41 0 Tanypus 0 124 173 Tamjtarsus 0 206 115 Zavrelia 19 0 0 were used in the following formula to estimate the number of samples to process (Elliot 1977): N = S2 where N = number of samples to process, S = variance, d = level of accuracy desired for the sample (in this case 0.1), and x = the mean. For our samples d was chosen to be 0.1, for an accuracy within 10% of the mean. If, after five samples were processed, N was <5 for a spe- cific taxonomic group, no more samples were processed for that group. Those taxa for which N was >5 were counted in an additional sam- ple. The mean and variance for taxa not elimi- nated were again calculated using the addi- tional sample value(s) and above formula. This 218 Great Basin Naturalist [Volume 55 process continued until N was less than the number of samples already processed for the taxon. Because of time and financial constraints, we never picked more than 30 samples for any specific habitat and sample date. All sorted samples were preserved in 70% ETOH. Chironomids were removed from 70% ETOM and placed in distilled water for 10-15 min prior to clearing. Individual specimens were placed in hot (-80 °C) 10% KOH (Cranston 1982) for 5-15 min to clear (larger specimens lequired more time to clear). After clearing, specimens were transferred to distilled water for at least 5 min. Each specimen was then placed in glycerine on a microscope slide for identification. Only late instars were identifi- able. Representative specimens of each genus encountered were permanently mounted. Specimens were classified to the generic level using keys by Mason (1968), Wiederholm (1983), and Merritt and Cummins (1984). Data Analysis Average densities (#/m") and 95% confi- dence limits for each of the four main taxa and each genus of Chironomidae were calculated for each sample site and date. Because density distributions were contagious, 95% confidence intervals were calculated for each of the four main taxa using a logarithmic transformation suggested by Elliot (1977; Tables 2-5). These values were then applied to the arithmetic mean (Shiozawa and Barnes 1977). Confidence inter- vals were not calculated for each genus in the Chironomidae because densities of some genera were too low. Cluster analysis was performed using the statistical package NTSYS-pc (Rohlf 1992). Several dissimilarit>' measures, including Bray- Curtis, Canberra's, and Renkonen s, were used to generate distance matrices. A comparison of each of these matrices to the original data showed that the Bray-Curtis measure (Bray and Curtis 1957) provided the best "fit " of the cluster analysis to the data. Average linkage clustering of the Bray-Curtis distances, based on the mean number of individuals/m^ of each species between habitat types and sample dates, was done with the unweighted pair-group method using arithmetic averages (UPGMA; Krebs 1989). Results Invertebrates Nematodes occurred in eveiy sample pro- cessed and were most abundant in the July sample of the ephemeral side channel habitat (302,603/m-) and least abundant in the river channel August sample (2421/m-; Tables 2-5). They comprised the majority of benthic inver- tebrates in all habitats and sample dates except Table 6. Functional group (Merritt and Cummins 19S4) and habitat association of Chironomidae genera from the Green River, Ouray National Wildlife Refuge, Ouray, UT. Fimctional group Taxon Collectors Predators Shredders Unknown Habitat association* Ahlahcsmijia X SIW Chir(»u)intis X RC,ESC,RB,SIW CladotaiHjtarsiis X RC,ESC,RB,SIW Criartopus X X SIW Cnjptnchironomus X ESCSIW ('njptotendipes X ESCSIW nr. CijphomeUa X RC C.lyptolcndipes X X SIW Microtcndipes X SIW Paraiiichna X RC ParatcDiiitarsiis X SIW Paratendipi's X RC Polijpcdilum X X X RC,ESC,RB,SIW Pwcladius X X RC,ESC,RB,SIW Pscctroclddius X X RC,RB,SIW nr Steinpellinella X RC Tamjpus X X ESC.SIW Tanijtarsus X RC,ESC,RB,SIW Zavrelia X SIW lumifl. ESC = ephemeral side channel, RB = ri\ (. r l)ack\\atf r, SIW = seasonully iiiinitlalrd \ 1995] Macroinvertebrates of the Green River 219 the August river channel and river backwater habitats and the July wetland sample. Oligochaetes were present in all habitat types and on all sample dates. Densities ranged from a low of 2728/m^ in the June ephemeral side channel sample to a high of 164,73 l/m^ in the July river backwater sample (Tables 2-5). The lowest abundance of Ceratopogonids was observed in the July river backwater sam- ple (96/m^). Their density was 136X greater in the river channel August sample (13,026/m^; Tables 2-5). Ceratopogonids were absent from both June and July samples of the seasonally inundated wetland and the ephemeral side channel. Ninteen chironomid genera were collected during this study. Fourteen genera were found in the July seasonally inundated wetland sam- ples, and five genera occurred in the August river channel and river backwater samples. Seven genera occurred in only one habitat or on only one date. Si.x genera were found in the seasonally inundated wetland habitat only, and four occuned onK' in the river channel. No chi- ronomid genus was unique to the ephemeral side channel or the river backwater. The genus Polypedihim was collected in all habitat types and on all sample dates. Total chironomid densi- ties were least (903/m-) in the June sample of the seasonally inundated wetland and greatest (31,125/m^) in the July river backwater sample (Tables 2-5). Unidentifiable early instars were collected in all habitat types and in all sample periods and comprised 86% of the river chan- nel sample in August. The most common func- tional group category of the Green River chi- ronomids was collectors followed by predators and shredders. Specific functional group and Green River habitat association for each genus are presented in Table 6. Other insects found in the samples are list- ed in Table 7. Density estimates would not be valid for these taxa because of their ability to avoid the core sampler. Cluster Analysis The UPGMA cluster analysis of the benthic invertebrate communities in each habitat type and sample date indicated that sites with similar flow conditions tended to cluster together (Fig. 3). A matrix comparison of original distances calculated using the Bray-Curtis coefficient with distances implied from the dendrogram is presented in Figure 4. Correlation between the two was high {R = .907), implying that the dendrogram is an accurate representation of Table 7. Other insects encountered in the Green Ri\ or ecos> stem, June-August 1991. River Ephemeral Riv er Seasonally inundated channel side channel backwater wetland Taxon July August June July July August June Juh' August Coleoptera Hydrophihdae (larvae) X Diptera Chironomidae (pupae) X X X X X Empididae (larvae) X X Simuliidae (lan'ae) X Ephemeroptera Baetidae Baetis (nymph) X X X X X Callibaetis (nymph) X Caenidae Caenis (nymph) X X X Tricorythidae Tricorythodes (nymph) X Hemiptera Corixidae X X X Odonata Coenagrionidae Ischnura (nymph) X X Gomphidae (nymph) X Plecoptera Perlodidae (nymph) Isoperla X 220 Great Basin Natur.\list [Volume 55 1.00 I 0.75 Bray-Curtis Distance 0.50 I 0.25 •0.00 Seasonal Wetland -June River Channel -July River Channel - August Seasonal Wetland -July River Backwater - August Seasonal Wetland - August River Backwater -July Side Channel - June Side Channel -July Fig. 3. UPGMA cluster analysis of Green River habitat t)pes located in the Ouray National Wildlife Refuge. the original Bray-Curtis distances. Ephemeral side channel samples show the greatest simi- larity (least distance), and wetland and back- water sites are more similar to one another Discussion Nematoda The importance of free-living nematodes in aquatic systems has not been extensively stud- ied. Aquatic nematodes are known to be micro- botrophic, predaceous, and/or parasitic during one or more of their life stages (Poinar 1991). Due to the scarcity of adequate keys and their small size, nematodes are seldom listed beyond the phylum designation in most studies and may not even be quantified. In studies of aqua- tic systems where nematodes are quantified, highest densities have been found in lakes. Strayer (1985) and Nalepa and Quigley (1983) reported that nematodes comprised 60% and 80%, respectively, of all benthic metazoans in Mirror Lake, NH, and in Lake Michigan with means of 680,000/m2 (Minor Lake) and 260,000/ m2 (Lake Michigan). In contrast. Palmer (1990) in Goose Creek and Gladden and Smock (1990) on the floodplain of Colliers Creek reported that nematodes comprised a much smaller percentage (6% of total invertebrates) and occurred at diminished densities (1000-15,000/ m^ and 1746/m2, respectively) in lotic systems. In our study nematode density estimates from the seasonally inundated wetland June sample (7133/m2) and the July and August river channel samples (24,881/m- and 2421/m^, respectively) are comparable to densities pre- viously reported from lotic systems (Gladden and Smock 1990, Palmer 1990). Density esti- mates for all other sites and dates (54,872- 302,603/m^) are more similar to densities in lentic habitats (see above). Greater densities are achieved in the more stable benthic envi- ronments provided by calmer waters and finer sediment particle size. In their study of White Clay Creek, Bott and Kaplan (1989) found that nematode densities were greater in silt than in sand. In our study the highest densities are also associated with a low sand content in the substratum. Low densities reported for the June sample of the seasonally inundated wet- land site reflect the relatively short time that water had been on the sample site. Of the four major invertebrate groups collected in this study, nematodes accounted for 8% of the individuals in the river channel August sample and 98% in the June ephemeral side channel. Nematodes accounted for 67.7% of all organisms observed. Palmer (1990), using a 3.3-cm-dia. core and 44-yam mesh, reported that nematodes constituted only 4-15% of the Goose Creek community, with a mean of 9%. Her data are similar to our river channel values. High nema- tode densities and their high percentage of the total invertebrates that we report from the ephemeral side channel, river backwater, and seasonally inundated wetland are unusual and should be compared to samples taken at similar locations in this and other large rivers using comparable methods. Oligochaeta Freshwater oligochaetes are a well-studied and diverse group found in every type of estu- arine and freshwater habitat. They feed mostly on bacteria living in soft sediments (Brinkliurst and Gelder 1991). The amount and quality of 1995] Macroinvertebrates of the Green River 221 i.uu- • • •• 0.75- i • B V) Q • «•• • • • • Oh • 0.25- n nn- • T 0.00 0.25 0.50 Original Distance .75 1.00 Fig. 4. Comparison of original dissimilarih' matrix and implied matrix from the dendrogram. organic matter found in the sediment are pri- mary factors determining which species will be present in a particular area (Brinkhurst and Cook 1974). We identified our specimens only to class level. Oligochaete densities in nonpol- luted lakes are lower than those in organically polluted waters. Densities in Mirror Lake ranged fi'om .30,000 to 33,000/m2 (Strayer 1985). Jonasson and Thorhauge (1976) reported oligo- chaete densities in Lake Esrom, Denmark, of 6000-12,000/m2. Brinkhurst and Cook (1974) found that densities of the three most common tubificids in the more polluted areas of Toronto Harbor ranged from 51,000 to 197,000/m2. Oligochaete densities in nonpolluted lotic sys- tems tend to be lower. Grzybkowska and Witczak (1990) report oligochaete densities in the lower Grabia River, Poland, ranging from 110 to 900/m2, and Palmer (1990) reports den- sities from 5000 to 15,000/m2 in Goose Creek, VA. Densities from polluted lotic systems can approach 200,000/m2 (Koehn and Frank 1980). Oligochaete densities in the seasonally in- undated wetland June sample (87,150/m2) and river backwater August sample (164,731/m2) are comparable to values observed in polluted systems described above. Densities from both ephemeral side channel samples (2728 m^ and 12,796/m2) and both river channel samples (3426/m2 and ll,182/m2) are comparable to those in Goose Creek (Palmer 1990). In general, oligochaete densities in our study were higher in habitats with the least amount of water flow (seasonally inundated wetland and river back- water habitat types). Terrestrial vegetation invades wetlands during dry periods, and when the water returns the following spring, decaying vegetation forms a rich food base. Backwater habitats retain fine particles, including detri- tus, being transported by the river; as summer progresses, this creates an enriched food base. These factors are the likely reason for the con- vergence oligochaete densities in these two hab- itats with those in organically polluted systems. Ceratopogonidae The study of ceratopogonids has mainly centered on adults because of their economic importance (Davies and Walker 1974). Larvae inhabit a variety of habitats including tree holes, leafiDacks, and pitcher plants, but are usu- ally most numerous in shallow areas of streams, lakes, and ponds (Bowen 1983). Aquatic forms are mostly predaceous (Merritt and Cummins 1984), but several species are known to consume algae and plant debris (Kwan and Morrison 1974). Corkum (1990) investigated streams associ- ated with different land-use types in south- western Ontario and found densities of 50/m2 in "forested" sites, 480/m2 in "mixed" sites, and 5300/m2 in "farmland" sites. Adamek and Sukop (1992) found maximum densities of only 222 Great Basin Natur.\list [Volume 55 1/m^ on over-flooded meadows in Czechoslo- \akia. In Lake Norman, NC, Bowen (1983) reported a mean lar\'al ceratopogonid density of767/m2. Ceratopogonid densities reached a peak in tiie August river channel sample (13,026/ m^) — much higher than any reported in the litera- ture above. In their study of the Green River, Grabowski and Hiebert (1989) did not report densities, but did conclude that ceratopogo- nids were more abundant in river channel samples than in backwaters. Our study supports this conclusion. Average densities for the river channel July and August samples were 3608/m- and 13,026/m2, respectively, compared to 96/m- and 461/m2 for the backwater July and August samples. Ceratopogonid lai^vae were complete- ly absent from the ephemeral side channel as well as the June and July seasonally inundated wetland samples. Chironomidae Chironomidae are typically the most abun- dant macroinvertebrates in lentic (Strayer 1985) and lotic (Grzybkowska and Witczak 1990) sys- tems. Studies of relatively small geographical areas have reported impressive numbers of species. For instance, Douglas and Murray (1980) found 142 species in Killarney Valley, Ireland. High diversity of chironomids makes them important as indicators of environmental condition (Wingard and Olive 1989). They are also abundant and provide an important food source for fish (Brown et al. 1980, Winkel and Davids 1987, Grabowski and Hiebert 1989), waterfowl (Titmus and Baddock 1980), and other migratoiy birds (Bowman 1980). We identified 19 chironomid genera from our sites within the Green River ecosystem. Other investigations of lotic systems have yield- ed similar numbers — 12 genera in the upper Tuscarawas River, OH (W^ingard and Olive 1989), 24 genera in the River Frome, England (Finder 1980), 25 genera in the Mississippi River (Beckett et al. 1983), and 36 genera in Juday Creek, IN (Berg and Hellenthal 1991). Grabowski and Hiebert (1989) studied the Green River in the same general area consid- ered in our study and also identified 19 genera. However, only seven of the genera reported by the latter authors were found in our study: Chiron(»mis, Cricotopus, Cryptochirunouius, Polijpediliiin, Procladius, Tamjpus, and Tanytarsus. Densities of chironomids in aquatic sys- tems can van substantialK. In a study of Lake Vissavesi, Finland, Paasivirta and Koskenniemi (1980) reported densities of 64/m^ in a coarse debris habitat and 2997/m- in a moss-grown site. Jonasson and Lindegaard (1979) reported 59,000/m2 from Lake Myvatn, Iceland. Vari- ability in lotic systems has also been docu- mented. Finder (1980) reported densities from a low of 48/m2 to 6273/m- in a chalk stream in England, and Grzybkowska (1989) found 10,664/m^ in the River Grabia, Foland. While no distinct trends e.xist when comparing chiro- nomid densities in lentic and lotic SNstems, den- sities are influenced by sediment size (Faasivirta and Koskenniemi 1980, Beckett et al. 1983). Chironomid densities from the July and August river channel samples were 4148/m2 and 3516/m2, respectively. River backwater samples were 31,125/m2 and 22,864/m2 for the same times. Grabowski and Hiebert (1989) reported maximum chironomid densities in the same area of the Green River of less than lOO/m^ for the river channel and 2800/m2 for river back'waters — substantially less than our estimates. It is possible that annual differences in seasonal discharge, area of the sampling device, and later sampling period all contrib- uted to this discrepancy. However, because of significant differences in mesh size (63-/u,m ours, 600-/xm Grabowski and Hiebert's), data of Grabowski and Hiebert and ours cannot be considered equivalent. It is worth noting that mesh sizes larger than 100 [xm have been shown to negatively bias density estimates (Strayer 1985). Community Similarity Cluster analysis of the data showed that, in general, habitat t\'pes clustered together inde- pendent of sample date, suggesting that the different habitat types studied in the Green River are distinct. Beckett et al. (1983), for ex- ample, studied five habitats in the Mississippi River and also found them to remain composi- tionally distinct regardless of flow and sample date. Distril)ution and abundance of benthic macroinvertebrates characteristic of these habitat types have been attributed to flow con- ditions and sediment size in our study. Since flow conditions are the major determinant of particle size, flow conditions are likely the determining factor. This conclusion has also 1995] Macroinvertebrates of the Green River 223 been reached by other investigators (Beckett et al. 1983, Statzner and Higler 1986). Grabowski and Hiebert (1989) conchided that benthic macroinvertebrate densities in backwaters of the Green River were higlier than those of the river channeh Our data suggest that the seasonally inundated wetland and ephemeral side channel are also valuable habi- tats and have the potential to contribute sub- stantial biomass to the Green River system. Oligochaete and chironomid densities report- ed in our study are comparable to other lotic systems (Koehn and Frank 1980, Finder 1980, Grzybkowska 1989, Grzybkowska and Witczak 1990, Palmer 1990). High densities of nema- todes and ceratopogonids imply that these groups may be very important in the overall energetics of the Green River system. Both should be studied more intensely. The overall dynamics of these communities is undoubtedly associated with seasonal changes in flow as well as year-to-year variability in annual discharge. This study, while describing a backwater, river site, side channel, and floodplain wetland over a short time interval, does not allow a full assessment of either annual or spatial variabil- ity. It is clear that some sort of successional colonization of various habitats occurs; for instance, floodplain wetlands are maximum in extent during highest spring-early summer flows, but their faunal development lags peak flooding. Back-waters do not exist during high flows, but as floodplains diminish with reced- ing water levels, back-water habitats develop. Again their faunal assemblages tend to lag be- hind the emergence of recognized back-waters. While we documented what appears to be seasonal succession within habitat type, such changes should not be assumed the norm. Until a detailed study is undertaken for the Green River or Colorado River system with replicate habitats over at least a full year period, our observations must be considered tentative. Further, annual discharge can vary tremen- dously from year to year, depending upon fac- tors such as drought cycles and their link with El Nino dynamics in the Pacific. Thus, what is seen in one year may not be representative of all years. Such factors introduce additional variables that should be considered when attempting to understand the dynamics of the benthos of the Green River. Literature Cited Adamek, Z., and I. SUKOP, 1992. Invertebrate communi- ties of fomier soiitliem Moravian floodplains (Czecho- slovakia) and impacts of regulation. Regulated Rivers: Research and Management 7: 181-192. Beckett, D. C, C. R. Bingham, and L. G. Sanders. 1983. Benthic macroinvertebrates of selected habitats of the lower Mississippi River. Journal of Freshwater Ecology 2: 247-261. Berg, M. B., and R. A. Hellenthal. 1991. Secondaiy pro- duction of Chironomidae (Diptera) in a north tem- perate stream. Freshwater Biology 25: 497-50.5. BOTT, T. L., and L. a. K\pl.\n. 1989. Densities of benthic protozoa and nematodes in a piedmont stream. Journal of the North American Benthological Society 8: 187-196. Bowen, T. W. 1983. Production of the predaceous midge tribes Sphaeromiini and Palpomyiini (Diptera; Ceratopogonidae) in Lake Norman, Nortli Carolina. Hydrobiologia 99: 81-87. Bowman, C. M. T. 1980. Emergence of chironomids from Rosterne Mere, England. Pages 291-295 in D. A. Murray, editor, Chironomidae: ecolog>', systematics, cytology, and physiology. Proceedings of the 7th International Symposium on Chironomidae, Dublin, Ireland. Brw; J. R., AND J. T. Curtis. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecological Monographs 27: 325-349. Brinkhurst, R. O., AND D. G. Cook. 1974. Aquatic earth- worms (Annelida: Oligochaeta). Pages 143-156 in C. W. Hart, Jr, and S. L. H. Fuller, editors. Pollution ecology of freshwater invertebrates. Academic Press, New York. Brinkhurst, R. O., and S. R. Gelder. 1991. Annelida: Oligochaeta and Branchiobdellidi. Pages 401-433 in J. H. Thoip and A. P Covich, editors, Ecolog\ and classification of North American freshwater inverte- brates. Acedemic Press Inc., San Diego, CA. Brown, A. E., R. A. Oldham, and A. Warlow. 1980. Chironomid larvae and pupae in the diet of brown trout {Salmo tnitta) and rainbow trout {Salmo gaird- neri) in Rutland Water Leicestershire. Pages 323-329 in D. A. Munay, editor Chironomidae: ecology; sys- tematics, cytology, and physiology. Proceedings of the 7th International Symposium on Chironomidae, Dublin, Ireland. CORKUM, L. D. 1990. Intrabiome distributional patterns of lotic macroinvertebrate assemblages. Canadian Journal of Fisheries and Aquatic Sciences 47: 2147-2157. Cranston, R S. 1982. A key to the Umae of the British Orthocladiinae (Chironomidae). Freshwater Biological Association. Scientific Publication 45. Davies, E G., and a. R. Walker. 1974. The isolation of ephemeral fever virus from cattle and Culiciodes midges in Kenya. Veterinary Record 95: 63-64. Douglas, D. J., and D. A. Murray. 1980. A checklist of Chironomidae (Diptera) of the Killamey Valley catch- ment area Ireland. Pages 12.3-129 in D. A. Murray editor, Chironomidae: ecology, systematics, cytologx; and ph\'siolog>-. Proceedings of the 7th International Symposium on Chironomidae, Dublin, Ireland. Elliot, J. M. 1977. Some methods for the statistical anal\- sis of samples of benthic invertebrates. Freshwater Biological Association, Scientific Publication 25. 224 Great Basin Naturalist [Volume 55 Gladden, J. E., and L. A. SM(X:k. 1990. Macroinvcrte- brate distribution and production on the floodplains of two lowland headwater streams. Freshwater Biology 24: 533-545. CiRABOWSKi, S. J., AND S. D. HiKBKivi. 1989. Some aspects of trophic interactions in selected backwaters and the main channel of the Green River, Utah, 1987-1988. Bureau of Reclamation, Environmental Sciences Section, Denver, GO. GkzyhK{)\\sk:\, M. 1989. Production estimates of the dom- inant ta.xa of Chironomidae (Diptera) in the modi- fied, Hi\er W'idawka, and the natural. River Grabia, central Poland. H>dn)biologia 179; 245-259. Grzybkow.ska, M., and J. WiTCZAK. 1990. Distribution and production of Ghironomidae (Diptera) in the lower course of the Grabia River (central Poland). Freshwater Biology 24: 519-531. Gr7.vbk{)\vsic\, M., J. Hejduk, and P Zielinski. 1990. Seasonal dynamics and production of Chironomidae in a large lowland river upstream and downstream from a new reservoir in central Poland. Archiv fiir Hydrobiologie 119: 439-455. JoNASSON, R M., and C. Lindegaard. 1979. Zoobenthos and its contribution to the metabolism of shallow- lakes. Ergebnisse der Limnologie (supplement to Archiv fiir Hydrobiologie) 13: 162-180. JONASSON, R M., AND E Thorhauge. 1976. Population dynamics of Potamothrix hainmoniensis in the pro- fundal of Lake Esrom with special reference to envi- ronmental and competitive factors. Oikos 27: 19.3-203. KOEHN, T, .\ND G. Fr^^nk. 1980. Effect of thermal pollution on the chironomid fauna in an urban channel. Pages 187-194 in D. A. Murray, editor, Ghironomidae; ecology, systematics, cytology, and physiology. Pro- ceedings of the 7th International Symposium on Ghironomidae, Dublin, Ireland. Krebs, G. K. 1989. Ecological methodology. Harper Gollins Publishing Go., New York. KwAN, \V. E., AND E O. Morrison. 1974. A summan of published information for field and laboratoiy stud- ies of biting midges, Cidicoides species (Diptera; Geratopogonidae). Annals of the Entomological Society of Quebec 19: 127-137. M.vsoN, W. T, Jr. 1968. An introduction to the identifica- tion of chironomid larvae. Federal Water Pollution Gontrol Administration, U.S. Department of the Interior, Cincinnati, OH. Merritt, R. W, AND K. W. Gummins, editors. 1984. An introduction to the aquatic insects of North America. Kendall/Hunt Publishing Go., Dubuque, lA. Morgan, R. R, II, R. E. Jacobsen, S. B. Weisberg, L. A. MgDowell, and H. T. Wilson. 1991. Effects of flow alteration on benthic macroinvertebrate com- miMiities below the Brighton hydroelectric dam. Journal of Freshwater Ecology 6: 419-429. MuNN, M. D., and M. a Brusven. 1991. Benthic macro- invertebrate communities in nonregulated and regu- lated waters of the Glearwater River, Idaho, U.S.A. Regulated Rivers: Research and Management 6; 1-11. Nalepa, T. F, and M. a. Quigley. 1983. Abundance and biomass of the meiobenthos in nearshore Lake Michigan with comparisons to the macrobenthos. Journal of Great Lakes Research 9: 530-547. Palmer, M. A. 1990. Temporal and spatial dynamics of meiofauna within the hyporheic zone of Goose Greek, Virginia. Journal of the North American Benthologi- cal Society 9: 17-25. P.\ASiviRTA, L., AND E. KosKENNlEMl. 1980. The Ghiro- nomidae (Diptera) in two polyhumic reservoirs in western Finland. Pages 23.3-238 in D. A. Murray, editor, Ghironomidae: ecolog}-, systematics, cytology, and physiolog). Proceedings of the 7th International Symposium on Ghironomidae, Dublin, Ireland. PiNDER, L. G. V. 1980. Spatial distribution of Ghironomidae in an English chalk stream. Pages 15.3-161 in D. A. Murra\', editor, Ghironomidae; ecology, systematics, cytology, and physiology. Proceedings of the 7th International Symposium on Ghironomidae, Dublin, Ireland. POINAR, G. O., Jr. 1991. Nematoda and Nematomoqiha. Pages 249-283 in J. H. Thoip and A. P Govich, editors, Ecology and classification of North American fi-esh- water invertebrates. Academic Press Inc., San Diego, GA. Rader, R. B., AND J. V. Ward. 1988. Influence of regulation on environmental conditions and the macroinverte- brate community in the upper Colorado River. Regu- lated Rivers: Research and IVlanagement 2; 597-618. ROHLF, E J. 1992. NTSYS-pc. Numerical ta.xonomy and multivariate analysis system. Version 1.70. Depart- ment of Ecology and Evolution, State University of New York, Stony Brook. Shiozawa, D. K. 1985. The seasonal community structure and drift of microcmstaceans in Valley Creek, Minne- sota. Canadian Journal of Zoology 64: 16.55-1664. Shiozawa, D. K., and J. R. Barnes. 1977. The microdis- tribution and population trends of larval Tanypus stellatu.s Cociuillett and Clunmomits frommeri Atchley and Martin (Diptera: Chironomidae) in LUah Lake, Utah. Ecolog>- .58; 610-618. Statzner, B., and B. Higler. 1986. Stream hydraulics as a major determinant of benthic invertebrate zonation patterns. Freshwater Biolog\' 16: 127-139. STR.AYER, D. 1985. The benthic micrometazoans of Mirror Lake, New Hampshire. Archiv fiir Hydrobiologie Supplement 72: 287-426. SwiNK, W D., and J. E NovoTNY. 1985. Invertebrate colo- nization rates in the tailwater of a Kentucky flood- control resei-\'oir. Journal of Freshwater Ecology 3: 27-34. TiTMUS, G., and R. M. Baddogk. 1980. Production and emergence of chironomids in a wet gravel pit. Pages 299-305 Hi D. A. Murray, editor, Ghironomidae: ecol- ogy, systematics, cytologv', and physiolog\'. Proceedings of die 7di International Symposiimi on Ghironomidae, Dublin, Ireland. WiEDERHOLM, T, EDITOR. 1983. Ghironomidae of the Holarctic region: keys and diagnoses. Part 1, Larvae. Borgstroms Tryckeri AB, Motala. WiNGARD, G. J., and J. H. Olive. 1989. Larval Ghirono- midae (Diptera) of the upper Tuscarawas River of northeastern Ohio, U.S.A. Journal of Freshwater Ecology 5; 93-102. VViNKEL, E. K., and T. Davids. 1987. Cyprinid fish and water mite reducing chironomid populations. Entomo- logica Scandinavica Supplementum 29: 26.5-267. Received 24 January 1994 Accepted 14 November 1994 Great Basin Naturalist 55(3), © 1995, pp. 225-236 ALPINE VASCULAR FLORA OF THE TUSHAR MOUNTAINS, UTAH Alan C. Tliyel Abstract. — The Tushar Mountains of southwestern Utah rise to a maximum elevation of 3709 m, with tiniherline and krummholz reaching maximum elevations of 3438 m and 3566 m, respectively. Voucher specimens were collected from the alpine region during eight field seasons to inventor^' this largely unknown alpine flora. Listed are 171 vascular plant species from 102 genera and 34 families that occur in eight types of plant communities within an alpine area of about 19.3 km-. The seven largest families are Asteraceae (29 species), Poaceae (20), Brassicaceae (13), Rosaceae (12), Cyperaceae (11), Car}ophyllaceae (10), and Fabaceae (8). Thirteen species are restricted to the alpine area. The perenni- al herb growth form accounts for 86.4% of the flora, .5.9% of the species are shrubs, and the remaining species are annu- als to short-lived perennials. Bedrock in the alpine region is entirely of Tertiary igneous origin. Vegetation cover and species richness are highest on an andesite ash-How tuff and latite flow and lowest on hydrothermally altered inter- caldera rhyolites and tuffs. Forty-four species (26.0% of the indigenous flora) also occur in the Arctic, and 13 species are at a southern margin of distribution. Eight taxa (4.7% of the flora) are local or regional endemics. The majority' of the alpine species appear to have migrated to the range by way of the contiguous mountain system to the north; statistical comparison with neighboring alpine floras shows the flora to be most similar to the floras of the Wasatch Mountains, Uinta Mountains, and Teton Range, with Sorensen's similarity indices of 52.8, 50.2, and 48.8% respectively. Key words: Utah, Colorado Plateau, Tushar Mountains, alpine vascular flora, alpine vegetation, plant geography. The Tushar Mountains, located in south- western Utah in the High Plateaus section of the Colorado Plateau at the eastern margin of the Great Basin (Fig. 1), reach a maximum ele- vation of 3709 m at the summit of Delano Peak. This elevation is surpassed within the state only by peaks in the Uinta Mountains and La Sal Mountains. A diverse alpine environment and flora occur on the 11 peaks that rise above the elevation of timberline. The alpine area is isolated. Though minor patches of alpine vege- tation occur on the Fish Lake Platetiu 66 km to the northeast and Markagunt Plateau 81 km to the south, the nearest extensive alpine area occurs in the Wasatch Mountains (Mount Nebo) 157 km to the north. The purposes of this paper are to document this isolated alpine flora, briefly describe the alpine plant communities, and determine possible migrational pathways to the Tushars by means of statistical and qual- itative comparisons with neighboring alpine areas. Study Area The Tushar Mountains have a length of 60 km and a width of 36 km at the widest point. Vertical relief exceeds 2000 m, with a low eleva- tion of 1695 m at the confluence of Clear Creek and the Sevier River. The range is located within an area of large-volume Tertiary (Oligo- cene to Miocene) volcanic activity known as the Marysvale volcanic field and is composed mostly of volcanic rocks (Cunningham and Steven 1979). Structurally, the range consists of a plateau-like, north-trending, up-faulted block bordered by structural valleys formed from down-faulted blocks; the High Plateaus section is thus structurally transitional be- tween the Basin and Range Province and the Colorado Plateau Province (Hunt 1987). The major faulting that produced the current linear ranges of the High Plateaus occured between 8 million and 5 million years ago (Steven et al. 1984). Topography and soil development in the alpine area are strongly influenced by the two volcanic formations exposed near timberline and above (Fig. 2). The mostly plateau-like to domelike ridges in the southern and eastern portions of the alpine region (including Delano Peak) are composed of calc-alkaline basaltic andesite flows and tuffs of the Bullion Canyon Volcanics (Cunningham et al. 1983) on which two soil complexes consisting of mollic cry- oborolls, argic pachic cryoborolls, pachic 1465 North 300 \\est #22. Pro\o, UT 84601. 225 226 Great Basin Naturalist [Volume 55 38- Fig. 1. Map of Utah showing location of Tushar Mountains (in black) and other mountainous areas above 2286 m in elevation (generalized and adapted from Smouse and Gurgel [1981]). ciyohoroUs, litliic ciyoboroUs, and rock outcrops are recognized (U.S. Forest Sei-vice 1993). The more mountainous noilliem and western portions of the alpine region are composed of intercaldera siliceous alkali rhyolite lava flows, lava domes, and ash-flow tuffs of the Mount Belknap Volcanics which have been hydrother- mally altered in many places (Cunningham and Steven 1979) and which are resistant to weathering. The summit pyramids of Mount Belknap (3699 m) and Mount Baldy (3695 m) consist of steep talus slopes and cliffs; portions of these talus slopes lack a cliff at their head and thus appear by definition (Washburn 1979) to be block slopes formed from periglacial frost action. Soil development in this region is limited to areas too small to map, and all alpine exposures of this formation as mapped by Cunningham et al. (1983) are classified by soil scientists as a cirqueland-rubbleland-rock outcrop complex (U.S. Forest Sei-vice 1993). Pleistocene glaciers produced several well- defined cirques on the eastern side of the crest where glacial ice descended to a low elevation of about 2500 m (Callaghan 1973). Glaciation also occurred on the western side of the crest as evidenced by glacial striations in the Poison [TT| Mount Belknap Volcanics m Bullion Canyon Volcanics Fig. 2. Map of the central Tushar Mountains with out- lined portions approximating the areas above 3383 m in elevation. Location of igneous formations is generalized from Cunningham et al. (1983). Creek drainage. Periglacial patterned ground in the form of stone stripes, stone circles, and stone nets occurs on the main ridgecrest be- tween the elevations of 3432 and 3600 m. Climate of the Rock\' Mountain alpine zone is characterized by Billings (1988) to have a mean temperature of less than 10° C for the warmest month. Climatic data are unavailable for the alpine area in the Tushar Mountains. A weather station located in an Engelmann spiaice {Piceo engehnannii) community 3.6 km south of the alpine region at an elevation of 3136 m has a mean annual temperature of 1.7°C, the warmest months being June, July, and August with mean monthly temperatures of 9.2, 11.7, and 1().9°C, respectively (three years of rec- ords). Most of the mean annual precipitation of 909 mm (12 years of records) falls as snow. Rainfall from summer thunderstorms is high- est in August, with an average of 74 mm, while the months of June and July receive averages 1995] TusHAR Mountains Alpine Flora 227 of 46 and 45 mm, respectively (Soil Conserva- tion Sei-vice 1993). An alpine region is defined by Bliss (1985) to be tlie area above die climatic limit ot upright tree growth, although it may include patches of krummholz. The average elevation of tim- berline in the Tushar Mountains occurs at about 3383 m (11,100 ft), with a corresponding alpine area of about 19.3 km^ above this eleva- tion. This alpine area, located between 38°20'04" and 38°27'47" North latitude and 112°19'32" and 112°26'42" West longitude, extends from Signal Peak in the north to Lake Peak at the south over a distance of 14.5 km. Much of the alpine area, centered about 25 km ENE of the city of Beaver, is accessible by Forest Service road 123, which crosses the crest of the range at an elevation of 3505 m. Timberline coincides with the upper limit of continuous forest and reaches a maximum ele- vation of 3438 m on a minor ridgecrest on the western (windward) side of the range. Timber- line occurs as low as 3341 m on lower ridges and is edaphically depressed even lower on some talus slopes. Engelmann spruce and lim- ber pine {Pinus flexilis) are the only arboreal species found at timberline. Subalpine fir {Alnes lasiocarpa) and aspen {Popidus tremidoides) approach timberline with maximum known ele- vations of 3365 and 3292 m, respectively. The krummholz limit, consisting of Engelmann spruce, occurs at about 3566 m on the steep, south-facing slope of Mount Baldy; this slope is protected from prevailing winds by a ridge extending southwesterly from the summit. Methods Voucher specimens were collected from 1984 through 1993 from throughout the range in preparation of a checklist of the vascular plants of the Tushar Mountains. Collected specimens were deposited in the herbarium of Brigham Young University and a search was made in this herbarium for other pertinent specimens. The total known flora for the range consists of 971 taxa representing 924 species, 381 genera, and 89 families (Taye 1994). The alpine region was visited during die same period except for the years 1986 and 1989. Only those species found above local timberline are included in this study. Species nomenclature and life form classifi- cation follow Welsh et al. (1993). Sorensen's Index of Similarity (Mueller- Dombois and Ellenberg 1974) was used to compare the alpine flora with neighboring alpine floras to determine possible migrational path- ways to the Tushars. Differences in nomencla- ture among the floras were largely resolved with the references of Dorn (1988), Weber and Wittmann (1992), and Welsh et al. (1993). Subspecific taxa were not used in statistical comparisons. Alpine Vegetation Eight types of alpine plant communities were recognized based on qualitative obsewa- tions; future intensive study of the vegetation will likely expand this classification. As noted for alpine communities in the Uinta Mountains (Lewis 1970), boundaries between plant com- munities are usually diffused. Cushion plant. — Low-growing species dominate the windswept ridgecrests where soils are shallow and outcrops of bedrock and rub- ble formed in place by frost-heaving are com- mon. Dominant species include Carex ehj- noides, Cerastiiim beeringianum, Erigeron compositus, Festuca ovina, Geiim rossii. Phlox pidvinata, and SUene acaidis. Dry meadow. — The warmer and drier south- and west-facing slopes are characterized by a plant cover in which bare soil is generally pre- sent between individual plants; rock cover is frequently high and soil disturbance from pocket gophers is common. Spruce krummholz is common at lower elevations. Common species in this extensive community type include Achdiea mdlefolinm, Astragalus miser, Carex elynoides, CastUleja parvula, Cymopterus lem- monii, Elymus trachycaulus, Haplopappus macronema, Heleniiim hoopsii. Phlox pidvina- ta, Potentilla glandidosa, P. gracilis, Poa seciin- da, and Ribes montigenum. Geum rossii occurs occasionally in usually mesic microhabitats. Alpine populations of Gentiana parry i, Jiiniperas communis, Sambucus racemosa, Thalictrum fendleri, and Viola nuttallii occur onl\^ in this community type at low elevations. Mesic meadow. — Plant cover is generally higher on suitable (nontalus or bedrock) north- and east-facing slopes and near drainage bot- toms and is occasionally cai-pet-like where suf- ficient soil development has occurred. Alpine avens {Geum rossii) is perhaps the most com- mon species in this community. Komarkova 228 Great Basin Natukalist [Volume 55 (1979) found tliis species to he most abundant on slopes with deep soil profiles and moderately proloniied snow cover Other common species in this extensive community type include Arenaria obtusiloha, Artemisia scopulorum, Carex heteroneura, Cerastiinn heerinf^ianwn, Erigeron simplex, Luziila spicata, Pcdiculans parnji, PJilox puhimita, Poa reflexa, Pohj^^onwn bistortuides, Salix arctica, Saxifraga rhotn- boidea, and Silene acaidis. Wet meadow. — A few wet meadows occur adjacent to rivulets and below long-lasting snowdrifts. Common species include Geiim rossii. Polygonum bistortoides, and Primula parriji. Caltha leptosepala, Potentilla diversifo- lia, Seduin rhodanthum, and Stellaria longipes reach their upper ele\'ational limit of 3627 m in this community type on the north-facing slope of Delano Peak. Rivulet. — Alpine rivulets from melting snowfields are mostly transitoiy and occur only in the southern (Bullion Canyon Volcanics) por- tion of the alpine region. Cardamine cordifo- lia, Deschampsia cespitosa. Delphinium occi- dentale van barbeyi, and Mertensia arizonica occur at lower elevations while Caltha lep- tosepala, Pedicularis parryi, and Salix arctica occur more commonly at higher elevations. Polygonum bistortoides and Primula parryi are common throughout this community type. Two rivulets on the flanks of Delano Peak (to about 3444 m in elevation) persist throughout the summer; Epilobium halleanum, Juncus drummondii, Mimtdus tilingii. and Saxifraga odontoloma occur at their upper elevational limit at these continually moist and marginally alpine sites. Gravelly barren. — This is perhaps the most distinctive alpine community type in the Tushar Mountains. It occurs on saddles of ridge- crests and on man\' of the higher tributar\' drainages between Lake Peak and Mount Bel- knap where snow accumulations are long last- ing; plant cover is only 0 to about 20% on largely unaltered, gravelly, grayish parent mate- rial. Some of the relatively few species that occur here are Calyptridium umbellatum, Elymus scribneri, Ivesia gordonii. Phlox pul- vinata, Poletnonium viscosum, Senecio amplec- tens, and S. canus. The endemic Draba sobo- lifera frequently flowers in gravel at the edge of receding snowbanks. Isolated 'hanging' patches of Geum rossii turf are sometimes present, indicating possible recent erosion of surround- ing material. GravclK barrens usualK inter- grade into dry meadow or talus/scree commu- nities over relatively short distances. Bedrock. — Plant growth on exposures of bedrock is restricted to rock crevices and ledges where pockets of soil have accumulated. Species present include many of those present on sim- ilar aspects in surrounding dry and mesic meadow communities. A unique assemblage of species that occasionally occurs on north- facing exposures includes A)ien\isia scopulonmi, Carex heteroneura, Cystopteris jragilis, Erigeron compositus, Geum rossii, Oxyria digyna, Saxifraga caespitosa, S. debilis, and Stellaria longipes. Talus/scree. — Colluvial deposits are most common in the area composed of the Mount Belknap Volcanics and along the glaciated por- tions of the main ridgecrest. Arenaria nuttallii, Cerastium beeringianum, Erigeron compositus, and Polemonium viscosum are commonly pres- ent on all exposures while Artemisia scopulonmi, Geum rossii, and Primula parryi are more com- mon on more mesic north- and east-facing slopes within this community type. The Flora The alpine flora of the Tushar Mountains consists of 171 species from 102 genera and 34 families. The largest families are Asteraceae (29 species), Poaceae (20), Brassicaceae (13), Rosaceae (12), Cyperaceae (11), Caryoph\'lla- ceae (10), Fabaceae (8), Ranunculaceae (7), and Scrophulariaceae (7). The largest genera are Carex, Poa, and Potentilla with 11, 8, and 7 species, respectively, while Saxifraga and Senecio are each represented by five species. Bromus inermis and Taraxacum ojficinale are the only introduced species occurring above timberline. The species list is presented near the end of this paper Thirteen taxa appear to be restricted to the alpine area: Astragalus australis van glabrius- culus, Carex elynoides, C haydeniana, C. nar- dina, Claytonia inegarJiiza, Hyincnoxys grandi- flora. Lychnis apetala van kingii, Poa patter- sonii, Potentilla concinna, Salix arctica, Saxifraga caespitosa, Townsendia condensata, and Valeriana acutdoba. Three taxa (1.8% of the alpine flora) are en- demic to high elevations in the Tushar Moun- tains. Draba sobolifera and Senecio castoreus are most common in gravelly barren and talus/ 1995] TusHAR Mountains Alpine Flora 229 scree communities above timberline while Cirsium eatonii var. harrisonii is most common on subalpine talus/scree slopes. Other Utah endemics found in the alpine are Agoseris glauca var. cronquistii. Astragalus perianus, CastiUeja parvula var parviila, Gilia fridactyhL and Lesquerella icarclii. The perennial herb life form accounts for 86.4% of the indigenous alpine flora. This fig- ure includes 143 species of angiosperms (110 dicots and 33 monocots), one spikemoss, and two ferns. Ten species of shrubs (5.9% of the flora) are present (two of which are gymno- sperms). The remaining 13 indigenous taxa are considered to be annual or biennial to short- lived perennials. Only 1.8% of the flora {Cheno- podiiun atrovirens, Gentianella tenella, and Polygonum douglasii) is classified as strictly annual though Spira (1987) reports Gentianella tenella to be strictly biennial in the alpine of the White Mountains, CA. Perennial herbs increase in importance at higher elevations and comprise 94.6% of the species (53 of 56 taxa — with exceptions being Androsace scpten- trionalis, Draba crassifolia, and Salix arctica) known to occur in the area of 0.6 km^ above the elevation of 3596 m on Delano Peak. A similar life form composition is reported for the alpine flora of the Teton Range (Spence and Shaw 1981). Species richness and habitat diversity are greatest in the vicinity of Delano Peak be- cause of this peak's geologic substrate, glacial history, and elevation. Erosion of the Bullion Canyon Volcanics has produced a mostly plateau-like topography conducive to soil for- mation and associated meadow communities. The northern and eastern slopes of Delano Peak, though glaciated, are relatively gentle as compared to the cliff-like glacial headwalls present along much of the main ridgecrest; per- sistent snowdrifts (sometimes lasting through- out the summer), which are necessary for the growth of some alpine species (Billings 1978) and which provide moisture to lower eleva- tions, are thus able to form on these less-inso- lated, high-elevation, leeward slopes. All eight types of plant communities and a minimum of 101 species (59.8% of the indigenous alpine flora) are known to occur within a radius of 1.0 km of the summit within an area of 3.14 km^ (16.3% of the alpine area). The northern portion of the alpine region composed of the weathering-resistant Mount Belknap Volcanics is floristically poor despite the presence of the second and third highest peaks; no vascular plants were obsei'ved above the elevation of 3536 m on Mount Belknap. A depauperate alpine flora of about 65 species occurs on the ridgecrest cushion plant com- munities, block slopes, and in the talus/scree and gravelly barren communities and small patches of mesic meadow that occur on the ridges and flanks of these summits; Crypto- gramma crispa and Poa pattersonii apparently occur in the alpine only on this fonnation, how- ever Soil formation and plant growth on this substrate may be hindered by unfavorable nutrient availability as occurs locally in hydro- thermally altered, highly acidic exposures at the base of the range (Salisbuiy 1964). Plant Geography The Tushar Mountains are located on the western margin of a floristically similar high- land region known as the Southern Rocky Mountains. This area, which includes most of Colorado and parts of adjacent states, contains the greatest concentration of alpine tundra in the United States outside of Alaska (Weber 1965). One hundred fifty-five of the 169 indige- nous alpine species of the Tushar Mountains are also reported by Weber and Wittmann (1992) for the flora of Colorado. Statistical comparison with 14 neighboring alpine floras shows the Tushar alpine flora to be most similar to the adjacent northerly floras of the Wasatch and Uinta ranges of Utah and the Teton Range, Wyoming, with Sorensen's similarity indices of 52.8, 50.2, and 48.8%, respectively (Table 1). The relatively continu- ous "Teton-Wasatch-High Plateau mainland mountain system" (Harper et al. 1978), which is perhaps best illustrated as an elevated (2000 m and higher elevation) corridor in Figure 19 in Reveal (1979) over which direct migration of alpine species may have occurred during glacial times (Billings 1978) and which has previously been noted to be a migration route for Utah's boreal species (Harper et al. 1978, Welsh 1978, Reveal 1979, and Welsh 1993), has thus likely been a primary source area for development of the alpine flora of the Tushar Mountains. In particular, Calyptridium wnhel- latum, Cymopterus hendersonii, Synfhris pin- natifida, and Townsendia condensata appear to have migrated to the Tushars via this north-to- 230 Great Basin Natuiulist [Volume 55 Table I. Moristic siinilarit\ indices hi'twecii tlu' alpine flora oi tlic Tnshar Nlonntains, UT, and representative nei^li- boring alpine floras. The index of similarity used is that of Sorensen (Miieller-Donil)ois and Ellenherg 1974). Mainland area floras arc part of a relatively continuous mountain s\'stem such as the Teton-Wasatch-High Plateau system in contrast to the more isolated mountain floras of the Great Basin and portions of the Colorado Plateau (Harper et al. 1978). Number of Alpine-to-alpine indigenous dist ance from Percent Flora^ alpii lie species Tnshar Mts (km) similarity M.MNLA.ND ARKAS 1. Wasatch Mountains, UT 202 157 52.8 2. Uinta Mountains, UT 257 269 50.2 3. San Juan .Mountains, CO 250 410 44.9 4. Sawatch Range, CO 285 507 45.4 5. Teton Range, VVT 216 573 48.8 6. Indian Peaks area, CO 249 596 42.1 7. Pioneer Mountains, ID 130 600 36.8 S. Sangre de Cristo Mountains, NM 157 627 40.5 9. Beartooth Plateau, WT-MT 185 750 36.2 Mountain Islands 10. Henr\' Mountains, UT 47 136 32.4 11. Snake Range, NV 43 171 25.5 12. Deep Creek Mountains, UT 81 198 51.2 13. San Francisco Peaks, AZ 82 332 44.6 14. Ruby Mountains, NV 150 345 42.0 ^Alpine floras are from tlie following sources; (1) ,\rnow ct al. (1980) and \ouclier specimens from Allred (1975) and Collins (1980); (2) Lewis (1970), Goodrich and Neese (1986), and Goodrich (1994); (3) Webber et al. (1976) and Hartman and Rottman (1985); (4) Hartman and Rottman (1988); (5) Spence and Shaw (1981); (6) Komarkova (1979); (7) Moseley and Bematas (1992); (8) Baker (1983); (9) Johnson and Billings (1962); (10) Neese (1981); (11) Lewis (1973); (12) McMillan (1948); (13) Schaack (1983) and Schaack and Morefield (1985); (14) Loope (1969) and Lewis (1971). south route, inasmuch as they occur in west- ern Wyoming (Dorn 1988) but are unreported from Colorado (Weber and Wittmann 1992). A total of 158 of the indigenous Tushar alpine species are reported by Dorn (1988) for the flora of Wyoming. The alpine flora of the Tushar Mountains is more similar to that of the Teton Range, Wyoming, than to any of the compared Colorado alpine floras despite a greater distance of up to 160 km (Table 1). Pro.ximity along the same migrational pathway thus appears to be an important factor in floristic similarity. The apparent effectiveness of the Teton- Wasatch-High Plateau migration route is fur- ther illustrated by 13 boreal species occuning in the alpine of the Tushars which are apparently at a southern margin of distribution within the longitudes of Utah: Astrafialus ati.stralis, Carex nardina. Lychnis apetala, and Salix arctica are arctic species (Polunin 1959) not known to occur further south in Utah (Albee et al. 1988) or in adjacent Arizona (Lehr 1978). A total of 44 alpine species from the Tushars (26.0% of the indigenous alpine flora) are reported by Polunin (1959) as also occurring in the Arctic. Other alpine species at an apparent southern margin of distribution are Antcwiaria cdpinci var media. Arenaria mittalUi, Ccdijptridiwn um- beUatinih ClunnaerJwdos erccta, Claytonia >neg- orhiza, Hymenoxys grandiflora, Poa pattersonii, Saxifraga adscendens, and Townsendio con- demata. Nonalpine boreal species at a southern margin of distribution in this range include Arnica diversifolia. Aster engelmannii, Carex deweyana, C. hoodii, C. hizidina, Draba lance- olata, Hieracium graciJe, Leucopoa kingii, Microseris nutans, and Mitella pentandra. Though migration of high-ele\'ation species has occurred between the Colorado Rockies and the La Sal Mountains of southeastern Utah (Holmgren 1972, Welsh 1993), significant migra- tion of alpine species fijrther west to the Tushar Mountains has perhaps been limited by an area of relatively low elevation termed the "Colorado Plateau migrational barrier" by Hadley (1987). The isolated Hemy Mountains, located midway between the Tushar and La Sal ranges (Fig. 1), have a meager alpine flora of 47 species; absent there are common alpine species such as Geum rossii, Oxyria digyna. Polygonum bistortoides, and Silcne acaulis (Neese 1981). These and other alpine species may have been eliminated from the Henry Mountains by the warmer post-glacial hyp- sithermal chmate (Neese 1981), however, thus 1995] TusHAR Mountains Alpine Flora 231 masking the true effectiveness of the Colorado Plateau as a migrational barrier to high-eleva- tion species. Species richness, which is strongly corre- lated with area on mountains (Harper et al. 1978, Hadley 1987), also appears to affect flor- istic similarity as the Tushar alpine flora gen- erally has higher indices of similarity with the larger and generally more distant mainland floras (Table 1); Harper et al. (1978) note that the isolated mountain floras of the Inter- mountain West have fewer species per unit area than adjacent mainlands and also an uneven stocking of species as a result of greater ran- domness of colonization and/or extinction. The isolated alpine floras of the east central Great Basin to the west of the Tushars (Loope 1969), the San Francisco Peaks to the south (Moore 1965), and the Henry Mountains to the east (Neese 1981) are slightly to extremely depau- perate examples of the Rocky Mountains alpine flora. The Tushar alpine flora is also slightly de- pauperate in comparison with most other neighboring mainland area floras (Table 1); this is likely due to the limited alpine area (in comparison, the Uinta Mountains have an alpine area of about 1000 km^ [Lewis 1970]), scarcity of wet meadows and rivulets, and presence of the talus-forming Mount Belknap Formation. The smaller Tushar alpine flora may be a factor in the relatively low maximum similarity index of 52.8% with the Wasatch Mountains; Hartmann and Rottman (1988) report a similarity index range of 72.5-73.3% between the larger alpine floras in Colorado. The alpine flora and vegetation of the Tushar Mountains are remarkably diverse given the relatively small alpine area. Interesting, too, are the number of endemic taxa and species that reach a southern limit of distribution here. The wide-ranging alpine species Claytonia megarhiza, Poa pattersonii, and Saxifraga adscendens are disjunct here with other in- state distributions only in the Uinta and La Sal ranges, while Townsendia condensato occurs nowhere else in the state (Albee et al. 1988, Welsh et al. 1993). Alpine environments are in general fragile and easily susceptible to disturbance (Billings 1973). This fragility is locally compounded by poor soil-forming characteristics of some igne- ous members and by the questionable intro- duction of Rocky Mountain goats to the range in 1986. There is evidence these animals feed on the endemic CastiUeja parvula, and they endanger the species diversity of the alpine area by grazing at scarce alpine wet sites. Acknowledgments I am grateful to Dn Stanley L. Welsh, curator of the herbarium at Brigham Young University, for his assistance and encouragement with this study initiated as part of a graduate program. Kaye Thorne, assistant curator of the herbari- um, provided sustained aid in herbarium research. Dr. Wesley B. Niles gave helpful comments on an earlier version of the manu- script, and Dr. Kimball T Harper graciously loaned me pertinent references from his per- sonal library. Ray Wilson of the Soil Conser- vation Service, Salt Lake City, provided data on climate for the area, and Mike Smith of the U.S. Forest Service office in Richfield, UT, provided information on soils. Information on plant specimens from the Tushars was provid- ed by the following individuals: Linda Allen, assistant curator of the Intermountain Herb- arium at Utah State University; Dr. Patricia K. Holmgren, director of the herbarium at The New York Botanical Garden; Ann Kelsey, cura- torial assistant at the Garrett Herbarium, University of Utah; and Tim Ross, senior cura- torial assistant at the Rancho Santa Ana Botanic Garden. This checklist would be less complete without the efforts of early botanists such as Marcus E. Jones and Drs. Walter R Cottam, Bertrand F. Harrison, and Bassett Maguire as well as the more recent prolific collecting by Dr. Stanley L. Welsh, Dr. N. Duane Atwood, Mont E. Lewis, and Joel Tuhy Annotated List of Vascular Plants The following list of families, genera, and species is arranged alphabetically within the divisions of Cronquist et al. (1972). Community type(s) and maximum elevation noted for each taxon are based on field notes and herbarium specimen label information. The following abbreviations are used for community types: cushion plant (CP), dry meadow (DM), mesic meadow (MM), wet meadow (WM), rivulet (RI), gravelly barren (GB), bedrock (BR), and talus/scree (TS). Frequency of occurrence for most taxa is estimated using the following scale from Thorne (1967): rare, 1-3 collections or observation stations; infrequent, 4-7 sta- 232 Great Basin Naturalist [Volume 55 tions; frequent, 8-12 stations; connnon, 13 + stations. Life form is listed as a = annual, ab = annual or biennial, ap = annual to short- lived perennial, bp = biennial to perennial, p = perennial herb, and s = shrub. Speeies that also oeeur in the Arctic (Polunin 1959) are fol- lowed by an asterisk (*). I collected Botrychiuin lunaria, Junctis mertensianiis, Pediciilaris groenlandica, and Salix planifolia at a seep below local timber- line at an elevation of 3389 m, and Draba lanccolata has been collected at timberline (Welsh et al. 14015). These and other taxa may eventually be discovered from the alpine area. Erigeron humilis and Taraxacum ceratophorum have recently been reported for the Tushars (Cronquist 1994), but I have seen no speci- mens. Division Lycopodiophyta Selaginellaceae Selaginella watsonii Underw.; rock crevices in CP, DM. MM, BR. and TS to 3658 m; common; p. Division Polypodiophyta Polypodiaceae Crijptogramma crispa (L.) R. Br. var. acrostichoides (R. Br.) C. B. Clarke; TS to 3304 m; rare; p. Cystopteris fragilis (L.) Bernh.; rock crevices in DM, MM, HI, BR, and TS to 3.505 m; frequent; p.* Division Pinophyta Cupressaceae Juniperiis communis L. var. depressa Pursh; DM at 3444 m in slielter of houlder on south-facing slope; rare; s.* Pinaceae Picea engelmannii Parry; DM, MM, and TS to 3566 m; common; s. Division Magnoliophyta Class Magnoliopsida Apiaceae CymopteruH hendersonii (Coult. & Rose) Cronq.; CR 15 K, and TS to 3627 m; frequent; p. CymopteruH lemmonii (Coult. & Rose) Dorn [Pseudo- cymopterus montanus (Gray) Coult. & Rose]; CR DM, MM, RI, and TS to 3700 m; common; p. Asteraceae Achillea millefolium L. ssp. hmulosa (Nutt.) Piper; DM, MM, and RI to .3548 m; connnon; p.* Agoseris aurantiaca (Hook.) Greene var. purpurea (Gray) Cronq.; MM(?) to ca 3505 m; rare; p. Agoseris glauca (Pursh) Raf. var. cronquistii Welsh; DM to 3353 m; infretjuent; p. Agoseris glauca (Pursh) Raf. var. dasycephala (T. & G.) Jepson; (^P to 3414 m; rare; p. Anteiinaria alpina (L.) Gaertner var media (Greene) Jepson [A. media Greene]; CR MM, and BR to 3487 m; connnon; p. Antennaria microphylla Rydb.; DM to 3536 m; rare; p. Arnica mollis Hook.; BR/MM to 3444 ni; rare; p. Artemisia frigida VVilld.; CP and DM to 3505 m; rare; s.* Artemisia ludoviciana VVilld. var. incompta (Nutt.) Cronq.; (JR DM, MNL and RI to 3475 m; common; p. Artemisia scopulorum Gray; MM, WM, RI, BR, and TS to 3703 ni; common; p. Cirsium eatonii (Gray) Robins, var. harrisonii Welsh; TS to 3444 m; rare; p. Crepis nana Richards.; CR GB, and TS to 3475 m; fre- quent; p.* Erigeron compositus Pursh var. glahratus Macoun; CR DM, GB, BR, and TS to 3706 m; common; p.* Erigeron simplex Greene; MM to 3700 m; fiequent; p. Erigeron speciosits (Lindl.) DC. var. uintahensis (Cronq.) Welsh [£. uintahensis Cronq.]; DM to 3414 m; rare; p. Erigeron ursinus D. C. Eaton; CR MM, and RI to 3536 ni; common; p. Haplopappus dementis (Rydb.) Blake; MM and GB to 3578 m; common; p. Haplopappus macronema Gray; CR DM, GB, and TS to 3.536 ni; connnon; s. Helenium hoopesii Gray [Dugaldia hoopesii (Gray) Rydb.]; DM and MM to 3566 m; common; p. Hymenopappus filifolius Hook. var. nudipes (Maguire) Turner; DM and GB to 3561 m; infrecjuent; p. Hymenoxys grandiflora (T. & G.) Parker; "grassy tun- dra above timberline at 350.5 m; rare; p. The only record from the range is K. E Parker et al. 6354 at the Rancho Santa Ana Botanic Garden. Senecio amplectens Gray var. hohnii (Greene) Har- rington; MM, GB, and BR to 3700 m; common; p. Setiecio canus Hook.; DM and GB to 3609 m; com- mon; p. Senecio castoreus Welsh; CR GB, and TS to 3536 m; infrequent; p. Senecio eremophilus Richards, var. kingii (Rydb.) Greenman; DM and MM to 3536 m; infrequent; p. Senecio werneriaefolius (Gray) Gray; TS to 3505 m; frequent; p. Solidago multiradiata Ait.; DM, MM, GB, and BR to 3700 ni; common; p.* Solidago parryi (Gray) Greene [Haplopappus parryi Gray]; MM to 3505 m; infrequent; p. Taraxacum officinale Weber; DM and TS to 3536 m; infrequent; introduced p. Townsendia condensata D. C. Eaton; CR and GB at 3.50.5-3609 m; infrequent; p. Boraginaceae Mertensia arizonica Greene; DM, MM, and RI to 3505 m; common; p. Brassicaceae Arahis drummondii Gray; DM to 3414 m; infrequent; I,p. Arahis lemmonii Wats.; MM and BR to 3402 m; rare; p. Cardamine cordifolia Gray; RI to 3444 m; infrequent (localh common); p. 1995] TusHAR Mountains Alpine Flora 233 Descurainia richardsonii (Sweet) Schuiz van brevipes (Nutt.) Welsh & Reveal; RI and TS to 3475 m; infre- quent; ah. Draba aiirea Vahl; MM and BR to 3688 m; rare; p.* Draba crassifolia Graham; MM and RI to 3700 m; fre- quent; ap.* Draba sobolifera Rydb.; MM, GB, BR, and TS to 3688 111; common; p. Draba stenoloba Ledeb.; MM to 3505 m; rare; ap.* Erysimum asperum (Nutt.) DC.; DM to 3441 m; rare; bp. Lesquerella wardii Wats.; DM and GB to 3609 m; fre- quent; p. Physaria chambersii Rollins van chambersii; GB to 3414 ni; rare; p. Smelowskia cahjcina C. A. Mey. van americana (Regel & Herder) Drury & Rollins; CH DM, MM, BR, and TS to 3703 111; common; p.* Thlaspi montaniitn L. van montanum; Cf! MM, and TS to 3475 m; common; p. Caprifoliaceae Sambucus racemosa L. van microbotrys (Rydb.) Kearney & Peebles; DM and TS to 3444 m; infrequent; s. Car)'ophyllaceae Arenaria mittallii Pax; CP MM, GB, and TS to 3505 m; common; p. Arenaria obtiisiloba (Rydb.) Fern.; MM and WM to 3676 m; common; p.* Arenaria rubella (Wahl.) J. E. Sm.; CR DM, MM, and RI to 3688 m; frequent; p.* Cerastium beeringianum C. & S.; CR DM, MM, WM, BR, and TS to 3700 m; common; p.* Lychnis apetala L. van kingii (Wats.) Welsh [L. kingii Wats.]; CR DM, and MM at 3536-3688 m; frequent; p.* Lychnis drummondii (Hook.) Wats.; DM, MM, and BR to 3487 m; frequent; p. Sagina saginoides (L.) Britt.; MM and RI to 3414 m; rare; hp.* Silene acaulis L. van subacaulescens (F. Williams) Fern. & St. John; CR MM, WM, BR, and TS to 3676 m; common; p.* Stellaria longipes Goldie; DM, MM, WM, BR, and TS to 3627 111; common; p.* Stellaria umbellata Turcz.; MM, RI, and TS to 3615 m; frequent; p. Chenopodiaceae Chenopodiiim atrovirens Rydb.; DM in disturbed soil (pocket gophers?) at 3548 m; rare; a. Crassulaceae Sedum rhodanthum Gray; MM and WM to 3627 m; infrequent; p. Fabaceae Astragalus australis Fisch. van glabriuscultis (Hook.) Isely [A. aboriginum Richards.]; CF and GB at 3505-3609 m; infrequent; p.* Astragalus miser Dougl. van oblongifolius (Rydb.) Cronq.; DM, MM, and GB to 3706 m; common; p. Astragalus periamis Barneby; DM and GB to 3566 m; infrequent; p. Lupinus argenteus Pursh van rubricaulis (Greene) Welsh; DM to 3463 m; rare; p. Lupinus lepidus Dougl. van utahensis (Wats.) C. L. Hitchc. [L. caespitosus Nutt. van utahensis (Wats.) B. Cox]; DM, MM, and GB to 3572 m; fiequent; p. Oxytropis oreophila Gray van oreophila; CR DM, MM, and GB to 3706 ni; common; p. Oxytropis parryi Gray; DM and MM to 3633 m; infre- quent; p. Trifolium longipes Nutt. var. rusbyi (Greene) Harrington; MM to 3597 m; frequent; p. Gentianaceae Gentiana parryi Engelni.; DM to 3389 m; rare; p. Gentianella amarella (L.) Borner; DM and MM to 3535 m; fretiuent; ah.* Gentianella tenella (Rottb.) Borner; MM and WM to 3566 m; rare; a (h?).* Swertia radiata (Kellogg) Kuntze [Frasera speciosa Dougl.]; MM and TS to 3475 m; rare; p. Grossulariaceae Ribes cereum Dougl.; DM and BR to 3536 m; fre- quent; s. Ribes inerme Rydb.; DM (among rocks) and TS to 3438 m; rare; s. Ribes montigenum McClatchie; DM, MM, RI, and TS to 3627 m; common; s. Hydrophyllaceae Phacelia hastata Dougl.; DM in gravelly soil to 3444 m; rare; p. Phacelia sericea (Graham) Gray var. ciliosa Rydb.; DM to 3475 m; rare; p. Lamiaceae Monardella odoratissima Benth.; TS to 3475 m; rare; p. Linaceae Linum perenne L. ssp. lewisii (Pursh) Hulten; MM in gravelly soil at 3536 m; rare; p.* Onagraceae Epilobium angustifolium L.; TS to 3414 m; rare; p.* Epilobium halleaniun Hausskn.; RI to 3444 m; rare (locally common); p. Epilobium saximontanum Hausskn.; RI to 3487 m; rare (locally common); p. Polemoniaceae Gilia tridactyla Rydb.; CR and TS to 3414 m; rare; p. Phlox pulvinata (Wherry) Cronq.; CR DM, MM, GB, BR, and TS to 3706 in; common; p. Polemonium pulcherrimum Hook. var. delicatum (Rydb.) Cronq.; DM and MM to 3444 m; infrequent; p.* Polemonium viscosum Nutt.; DM, MM, GB, BR, and TS to 3633 m; common; p. Polygonaceae Eriogontim umbellatum Toit. var. porteri (Small) Stokes; DM, MM, and BR to 3566 m; frequent; p. Oxyria digyna (L.) Hill; MM, GB, BR, and TS to 3658 m; common; p.* 234 Grkat Basin Natufi\list [Volume 55 Polygonum histortoides Pursh; MM, \\M, and HI to 3676 111; commoii: p. Polygonum doiighisii Greene var. douglasii., DM to 3444 111; rare; a. Rumex salicifolius Weinm. ssp. triangulivalvis Danser; DM, MM, RI, and BR to 3499 in; frequent; p. Portulacaceae Calyptridium umhellatum (Torr.) Greene var. caudi- cifera Griiy, MM and i'Ai to 3536 ni; infre(iiient; ap. Claytonia megarhiza (Gray) Parry; BR and TS at 3475 to 3615 m; rare; p. Lewisia pygmaea (Gray) Robins.; MM and Rl to 3597 ni; frefjiient; p. Primulaceae Androsace septentrionalis L.; DM, MM, RI, and TS to 3700 m; eoinmon; ab.* Primula parryi Gray; MM, WM, RI, BR, and TS to 3658 m; common; p. Ranunculaceae Anemone multifida Poir.; CP and MM to 3487 m; rare; p. Aquilegia scopuJorum Tidestr.; TS to 3438 m; infre- quent; p. A.S noted in \\elsh et al. (1993), some specimens are completeK' transitional with A. caendea James. Caltha leptosepala DC. var. leptosepaloj MM, WM, and RI to 3627 m; frecjuent; p. Delphinium occidentale (Wats.) Wats. var. barbeyi (Huth) Welsh [D. barbeyi (Huth) Huth]; DM, RI, and TS to 3475 m; comnion; p. Ranunculus eschscholtzii Schlect.; TS to ca 3490 m; rare; p. Ranunculus inamoenus Greene; DM and RI to 3597 m; common; p. Thalictrum fendleri Engelm.; DM (in shelter n( Ril)es niontigciUDiij to 3414 m; rare; p. Rosaceae Chamaerhodos erecta Bunge var. parviflora (Nutt.) C. L. Hitchc.; CP and DM to 3505 m; rare; lip. Geum rossii (R. Br.) Sen var. turbinatum (Rydb.) C. L. Hitchc; CR DM, MM, WM, RI, GB, TS, and BR to 3700 m; common; p.* Ivesia gordonii (Hook.) T. & G.; DM and GB to 3609 m; infrequent; p. Potentilla concinna Richards, var. proxima (Rydb.) Welsh & Johnston; DM and TS at 3353 to 3536 in; infre- quent; p. Potentilla diversifolia Lehm. var. diversifolia; WM, RI, and TS to 3627 m; frecjuent; p. Potentilla glandulosa Lindl. var. intermedia (Rydb.) C. L. Hitchc; DM, MM, and TS to 3487 in; common; p. Potentilla gracilis Dougl. var. pulcherrima (Lehm.) Fern.; DM to 3463 m; frequent; p. Potentilla hippiana Lehm.; DM to 3414 m; p. Potentilla ovina Macoun var. decurrens (Wats.) Welsh & Johnston; (>P and DM to 3475 ni; infreciuent; p. Potentilla pensylvanica L. var. pensylvanica; CP DM, MM, and TS to 3700 ni; common; p.* Ritbus idaeus L. ssp. melanolasius (Dieck) Focke.; TS to 3414 in; rare; s. Sibbaldia procutnbens L.; MM, RI, and BR to 3627 m; common; p.* Salicaceae Salix arctica Pallas var. petraea Anderss.; MM, WM, and RI at 3444 to 3676 m; frecjuent (locally common); s.* Sa.xifiagaceae Heuchera rubescens Torr. var. rubescens; BR to 3444 m; rare; p. Saxifraga adscendens L. var. oregonensis (Raf.) Breitung; MM (ainony; rocks) to 3676 m; rare; p. Saxifraga caespitosa L. var. minima Blake; MM, WM, and BR at 3566 to 3676 in; infrequent; p.* Saxifraga debilis Engelm.; MM and BR to 3658 m; common; p. Saxifraga odontoloma Piper; RI to 3444 m; rare; p. Saxifraga rhomboidea Greene; DM, MM, WM, and RI to 3700 m; common; p. Scrophulariaceae Castilleja miniata Dougl.; DM to 3535 m; infrequent; p. Castilleja parvida Rydb. var. parvtda; DM and MM to 3688 111; common; p. Mimulus tilingii Regel; RI to 3414 m; rare; p. Pedicularis parryi Gray var. parryi; MM, WM, and RI to 3627 m; common; p. Penstemon whippleanus Gray; MM and BR to 3450 m; frecjuent; p. Synthris pinnatiftda Wats. var. laciniata Gray; DM, MM, WM, RI, and BR to 3627 m; common; p. Veronica ivormskjoldii R. & S.; MM and RI to 3487 m; rare; p.* Valerianaceae Valeriana acutiloba Rydb.; DM and MM at 3414 to 3567 m; infrecjuent; p. Valeriana edulis Nutt.; CR DM, and MM to 3599 m; infrecjuent; p. Valeriana occidentalis Heller; DM to 3353 m; rare; p. Violaceae Viola canadensis L.; BR and TS to 3444 m; rare; p. Viola nuttallii Pursh; DM to 3414 m; rare; p. Class Liliopsida Cyperaceae Carex albonigra Mack.; CR and MM to ca 3658 m; infrecjuent; p. Carex ebenea Rydb.; RI to 3444 m; rare; p. Carex egglestonii Mack.; DM to 3414 m; rare; p. Carex elynoides H. T. Holm; CR DM, MM, and TS at 3353 to 3706 m; common; p. Carex haydeniana Olney; MM, RI, GB, and BR at 3414 to 3566 in; common; p. Carex heteroneura W. Boott var. chalciolepis (H. T. Holm) F. Hermann; the intergrading var epapillosa F. Hermann also occurs in the range though perhaps not in the alpine; MM and BR to ca 3658 m; common; p. Carex tnicroptera Mack.; DM (?) to 3414 m; rare; p. Carex nardina Fries; MM at 3505 m; rare; p.* Carex nova Baile> ; unknown community at ca 3505 m; rare; p. Carex phaeocephala Piper; CR DM, MM, and GB to 3566 m; common; p. Carex rossii F. Boott; DM (?) to ca 3353 m; rare; p. 1995] TusHAR Mountains Alpine Flora 235 Juncaceae Jimcus drttmmondii E. Mey.; RI to 3444 ni; rare; p. Luzula spicata (L.) DC; MM, WM, and BR to 3627 m; coinnion; p.* Liliaceae Zigademis elegans Pursh; MM, WM, and RI to 3536 m; infrequent; p.* Poaceae Agrostis variabilis Rydb.; MM to 3383 m; rare; p. Bromus ciliatus L.; MM to 3414 ni; rare; p. Bromiis inermis Leysser; roadside adjacent to MM at 3487 m; rare; introduced p. Calamagrostis piirpiirascens R. Br.; TS to 3414 m; rare; p.* Deschampsia cespitosa (L.) Beauv.; MM and RI to 3499 ni; infrequent; p.* Elymus ehjmoides (Raf.) Swezey [Sitanion hystrix (Nutt.) J. G. Sm.]; DM to ca 3505 m; rare; p. Elymus scribneri (Vasey) Jones [Agropyron scribneri Vasey]; DM, GB, and TS to 3578 m; common; p. Elymus trachycaulus (Link) Gould [Agropyron tra- chycauhim (Link) Malte]; DM and NLM to 3566 m; com- mon; p.* Festuca ovina L. van brevifolia (R. Br.) Wats.; CR DM, MM. GB, BR, and TS to 3706 m; common; p.* Phleum alpimim L.; MM and RI to 3487 m; frequent; p.* Poa arctica R. Br.; CP, MM, WM, BR, and TS to 3700 m; frecjuent; p.* Poa fendleriana (Steudel) Vasey; DM to 3383 m; fie- quent; p. Poa glauca Vahl [P. glauca ssp. rupicola (Nash) W. A. Weber; P interior Rydb.]; CE DM, MM, GB, and TS to 3536 m; common; p.* Poa nervosa (Hook.) Vasey; TS to 3414 m; infrequent; p. Poa pattersonii Vasey; TS at 3505 m; rare; p. Poa pratensis L.; MM to 3444 m; rare?; possibly intro- duced p.* Poa reflexa Vasey & Scribn.; MM and RI to 3536 m; common; p. Poa secunda Presl [P. sandbergii Vasey]; CP DM, and TS to 3475 m; frequent; p. Stipa lettermanii Vasey; DM and MM to 3475 m; fre- quent; p. Trisetum spicatum (L.) Richter; CP, MM, BR, and TS to 3700 m; common; p.* Literature Cited Albee, B. J., L. M. Shultz, and S. Goodrich. 1988. Atlas of the vascular plants of Utah. Utah Museum of Natural Histon; Salt Lake Cit>'. 670 pp. Allred, K. W 1975. Timpanogos flora. Unpublished tlie- sis, Brigham Young University, Provo, UT. 178 pp. Arnow, L., B. Albee, and A. Wyckofe 1980. Flora of the central Wasatch Front, Utah. 2d edition revised. Universit>' of Utah Printing Sendee, Salt Lake Gib,'. 663 pp. Baker, W L. 1983. Alpine vegetation of Wheeler Peak, New Me.xico, U.S.A.: gradient analysis, classification, and biogeography. Arctic and Alpine Research 15: 223-240. Billings, W D. 1973. Arctic and alpine vegetations: simi- larities, differences, and susceptibility- to disturbance. Bio Science 23: 697-704. . 1978. Alpine phytogeography across the Great Basin. Great Basin Naturalist Memoirs 2: 105-117. 1988. Alpine vegetation. Pages 391-420 in M. G. Barbour and W. D. Billings, editors. North American terrestrial vegetation. Cambridge University Press, Cambridge. Bliss, L. C. 1985. Alpine. Pages 44-65 in B. F Chabot and H. A. Mooney, editors. Physiological ecology of North American plant communities. Chapman and Hall, New York. Callaghan, E. 1973. Mineral resource potential of Piute County, Utah, and adjoining area. Utah Geological and Mineralogical Survey Bulletin 102. Utah Geo- logical Suwey, Salt Lake Cit\'. 135 pp. Collins, P D. 1980. Comparative life histoiy and floral characteristics of desert and montane plant commu- nities in Utah. Unpublished thesis, Brigham Young University, Provo, UT. 168 pp. Cronquist, a. 1994. Intermountain flora. Volume 5: Asterales. The New York Botanical Garden, Bron.x. 496 pp. Cronquist, A., A. H. Holmgren, N. H. Hol.mgren, and J. L. Reveal. 1972. Intermountain flora. Volume 1. Hafner Publishing Co., New York. 270 pp. Cunningham, C. G., and T. A. Steven. 1979. Mount Belknap and Red Hills calderas and associated rocks, Marysvale volcanic field, west-central Utah. U.S. Geological Suney Bulletin 1468. 34 pp. Cunningham, C. G., T. A. Steven, E D. Rowley, L. B. Gl,\ssgold, and J. J. Anderson. 1983. Geologic map of the Tushar Mountains and adjoining areas, Marys- vale volcanic field. U.S. Geological Survey miscella- neous investigations series map I-1430-A. U.S. Geo- logical Sui-vey, Denver, CO. DORN, R. D. 1988. Vascular plants of Wyoming. VIountain West Publishing, Cheyenne, WY. 340 pp. Goodrich, S. 1994. Written communication. Ashley National Forest, 355 North Venial Ave., Vernal, UT 84078. Goodrich, S., and E. Neese. 1986. Uinta Basin flora. U.S. Department of Agriculture, Forest Service — Inter- mountain Region, Ogden, UT. 320 pp. Hadley, K. S. 1987. Vascular alpine plant distributions with- in the central and southern Rock-y Mountains, U.S.A. Arctic and Alpine Research 19: 242-251. Harper, K. T, D. C. Freeman, W K. Ostler, and L. G. Klikofe 1978. The flora of Great Basin mountain ranges: diversity, sources, and dispersal ecology. Great Basin Naturalist Memoirs 2: 81-103. Hart.man, E. L., and M. L. Rottm.a.n. 1985. The alpine vascular flora of three cirque basins in the San Juan Mountains, Colorado. Madrono 32: 253-272. . 1988. The vegetation and alpine vascular flora of the Sawatch Range, Colorado. Madrono 35: 202-225. Holmgren, N. H. 1972. Plant geograhy of the Inter- mountain region. Pages 77-161 in A. Cronquist, A. H. Holmgren, \. H. Holmgren, and J. L. Reveal, editors, Intermountain flora. Volume 1. Hafner Publishing Co., New York. Hunt, C. B. 1987. Physiography of western Utah. Pages 1-29 in R. S. Kopp and R. E. Cohenour, editors, Cenozoic geology of western Utah — sites for pre- cious metal and hydrocarbon accumulations. Utah Geological Association Publication 16. Utah Geo- logical Association, Salt Lake Cit>'. Johnson, P L., and W D. Billings. 1962. The alpine vegetation of the Beartooth Plateau in relation to 236 Ghkat Basin Natur.\list [Volume 55 cryopedo^fiiic ])r()ct'sses and patterns. Iscolouical Monographs 32: 105-135. KOMARKOVA, V. 1979. Alpine vegetation of the Inchan Peaks area. Front Range, Colorado Rocky Monntains. Strauss & Cramer, Hirschberg II. Cerniany. 650 pp. Leiik, J. H. 1978. A catalogue of the flora of Arizona. Desert Botanical Garden, Phoeni.x, AZ. 203 pp. Lewi.s, M. E. 1970. Alpine rangelands of the Uinta Moun- tains, Ashley and Wasatch National Forests, Region 4. U.S. Department of Agriculture, Forest Service, Ogden, UT 75 pp. . 1971. Flora and major plant communities of the liiib\-East Ilumholdt Mountains with special empha- sis on Lamoille Canyon. Report to Humboldt National Forest, Elko, NV 62 pp. . 1973. Wheeler Peak area species list. Report to Intermountain Region, U.S. Forest Service, Ogden, UT. 17 pp. LooPE, L. L. 1969. Subalpine and alpine vegetation of northeastern Nevada. Unpublished doctoral disser- tation, Duke University, Durham, NC. 287 pp. McMlLLA.N, C. 1948. A taxonomic and ecological study of the flora of the Deep Creek Mountains of central western Utiih. Unpublished thesis, Universit\' of Utah, Salt Lake City. 99 pp. McjORE, T. C. 1965. Origin and disjunction of the alpine tundra flora on San Francisco Mountain, Arizona. Ecology 46: 860-864. MOSELEY, R. K., AND S. Bernatas. 1992. Vascular flora of Kane Lake circjue. Pioneer Mountains, Idaho. Great Basin Naturalist 52: 335-343. MuELLER-DoMBOis, D., AND H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley & Sons, New York. 547 pp. Neese, E. J. 1981. A vascular flora of the Heni-y Mountains, Utah. Unpublished dissertation, Brigham Young Uni- versity, Provo, UT. 370 pp. POLUNIN, N. 1959. Circumpolar Arctic flora. O.xford University Press, London. 514 pp. Re\ EAL, J. L. 1979. Biogeography of the Intermountain region: a speculative appraisal. Mentzelia 4: 1-92. Salisbury, F B. 1964. Soil formation and vegetation on hydrothermally altered rock material in Utah. Ecol- ogy 45: 1-9. S(:h.\.'\ck, C. G. 1983. The alpine vascular flora of Arizona. Madroiio 30: 79-88. Scii,\ack, C. G., and J. D. Morefield. 1985. Noteworthy collections: Arizona. Madroiio 32: 121-122. Smouse, F a., and K. D. Gurgel. 1981. ElevaHon. Page 18 in D. C. Greer, K. D. Gurgel, W. L. Wahlquist, H. A. Christy, and G. B. Peterson, editors. Atlas of Utah. Weber State College, Odgen, UT. [Printed at Brigham Young University Press, Provo, UT] S(ML CoN.SKRX.vnoN Skrvige. 1993. SNOTEL. Snow Sur- vey Section, Salt l^ake City, UT. Spence, J. R., and R. J. Sh.wv. 1981. A checklist of the alpine vascular flora of the Teton Range, Wyoming, with notes on biology and habitat preferences. Great Basin Naturalist 41; 232-242. Spia^, T P 1987. .Alpine annual plant species in the White Mountains of eastern California. Madrono 34: 315-323. Steven, T A., P D. Rowley, and C. G. Cunnlngham. 1984. Calderas of the Marysvale volcanic field, west central Utah. Join^nal of Geophvsical Research 89, No. BIO: 8751-8764. T.\ye, a. C. 1994. Annotated checklist of the vascular plants of the Tushar Mountains, Utah. Unpublished data. Herbarium, M. L. Bean Life Science Museum, Brigham Young University', Provo, UT 84602. Thorne, R. F 1967. A flora of Santa Catalina Island, California. Aliso 6(3): 1-77. U.S. Forest Service. 1993. Tushar-Pa\ant-Canyon soil sur- vey report (preliminaiy unpublished data). Fish Lake National Forest, Richfield, UT. Washburn, A. L. 1979. Geocryology, a survey of peri- glacial processes and environments. Wiley, New York. 406 pp. Webber, P J., J. C. Emerick, D. C. Ebert May, and V. KoviARKON A. 1976. The impact of increased snowfall on alpine vegetation. Pages 201-264 in H. W. Steinhoff and J. D. Ives, editors. Ecological impacts of snow- pack augmentation in the San Juan Mountains of Colorado. Colorado State University, Fort Collins. Weber, W. A. 1965. Plant geography in the southern Rocky Mountains. Pages 453-468 in H. E. Wright and D. G. Frey, editors. The Quaternary in the United States. Princeton University Press, Princeton, NJ. Weber, W A., and R. C. Wittmann. 1992. Catalog of the Colorado flora: a biodiversity baseline. University Press of Colorado, Niwot. 215 pp. Welsh, S. L. 1978. Problems in plant endemism on the Colorado Plateau. Great Basin Naturalist Memoirs 2: 191-195. . 1993. Description of the environment. Pages 6-10 in S. L. Welsh, N. D. Atwood, L. C. Higgins, and S. Goodrich, editors, A Utah flora. Print Services, Brigham Young University, Provo, UT. Welsh, S. L., N. D. Atwood, S. Goodrich, and L. C. Higgins. 1993. A Utah flora. 2d edition revised. Print Senices, Brigham Young Universit>', Provo, UT. 986 pp. Received 1 September 1994 Accepted 7 November 1994 Great Basin Naturalist 55(3), © 1995, pp. 237-248 ECOLOGY OF CELTIS RETICULATA IN IDAHO Ann Marie DeBoltl and Bruce McCune- Abstract. — The small deciduous tree Celtis reticulata (netleaf hackberry) reaches its northern limit in Idaho, where, contraiy to most of its western range, it often occurs as an overstorv' dominant. Two hundred fifty stands of this tree were sampled throughout Idaho. Celtis is slow-growing, averaging 4 m tall at 50 \'r, and long-lived (to 300-400 yr). It occurs in a variety of habitats, from riparian to rocky uplands. Trees grow best where topographically sheltered, such as in draws and narrow canyons, and where soils are loamy. Although plants grow more slowly as surface rock cover increases, stands are often associated with rock, with a mean surface cover of 39% rock. Differences in growth rates were unrelat- ed to parent material and aspect. Most stands are reproducing, in spite of habitat degradation caused by overgrazing, alien plant invasion, and increasing fire frequencies. Stands are typically represented by one dominant cohort; however, young, even-aged stands are rare and are generally found along watenvays on stream terraces or at the high-water line. Although slow-growing, C. reticulata shows promise for land managers interested in site enhancement. This native species is long-lived, produces fruit used by wildlife, and provides structural diversity in a semiarid landscape (with a maximum height of 12 m) in areas that are becoming increasingly dominated by e.xotic plant species. Key words: Celtis reticulata, netleaf hackbernj, ecology, Idaho, growth, longevity, stand structure, recruitment, site characteristics, livestock grazing, rehabilitation. Celtis reticulata Torr. (netleaf hackberry, western hackberry) is a deciduous shrub to small tree in the elm family (Ulmaceae), wide- ly distributed in semiarid regions of the west- ern United States (Fig. 1). It occurs in a diver- sity of habitats, including deciduous riparian woodlands, mountain shrub, wash scrub, and live oak-mixed shrub communities, in rocky canyons, and as scattered individuals in semi- desert grasslands, pinyon-juniper and Joshua tree woodlands (Glinski 1977, Plummer 1977, Brown 1982, Albee et al. 1988). Its elevational range is from 200 to 2000 m (Elias 1980). Populations are often small or highly localized (Daubenmire 1970, Dooley and Collins 1984), particularly at the northerly latitudes in the states of Oregon, Washington, and Idaho (Eliot 1938). Despite its broad distribution, little is known about the plant's ecology, presumably due to its position as a minor component in many of its habitats, and its fragmented occur- rence (Peattie 1953, Lanner 1983). While C. reticulata is sparsely distributed in Idaho, near its northern limit (Fig. 2), it appears to exhibit wide ecological tolerances. However, it tends toward the warmest portions of canyons, especially southerly aspects (Tisdale 1986). It is a member of both riparian and upland communities in Idaho, where it can occur as a locally abundant, overstory dominant (Huschle 1975, Johnson and Simon 1987). Along the Wiley Reach of the middle Snake River, it forms narrow, but extensive, gallery forests of nearly monospecific stands (Bowler 1981). On steep shoreline escarpments of the lower reaches of the Snake River, in the "Douglas" hackberry vegetation type described by Huschle (1975), it forms a dense, nearly closed canopy. On the gentle shoreline slopes, alluvial fans, and colluvial cones of the lower Snake River, it grows in an open savanna (Daubenmire 1970, Huschle 1975). "Open sa- vanna" is perhaps the best way to describe the appearance of a typical Celtis community on an upland site in Idaho, where individuals occur in a random or clumped pattern with exten- sive areas of grassland between. Plants produce a small, fleshy drupe in the fall, favored by a variety of birds and mammals (Hayward 1948, Lanner 1983, C. A. Johnson 1990, personal communication). With as many as 41 species of birds associated with Celtis communities in Idaho, the tree's importance for wildlife cannot be overemphasized (Asherin and Claar 1976). It provides cover for a variety of big game species, including mule deer and bighorn sheep (Asherin and Claar 1976), as well as much-sought-after shade for domestic 'Bureau of Land Management, 3948 Development Avenue, Boise, ID 8.370.5. ^Department of Botany and Plant Pathology, Oregon State University, Conallis, OR 97331-2902. 237 238 Great Basin Naturalist [Volume 55 »" INTERMITTENT DISTRIBUTION — CONTINUOUS DISTRIBUTION Fig. 1. Global distiilnition of Celtis reticulata (revised from Little 1976). livestock along the Snake River (Daubenmire 1970). Due to an apparent tolerance of haish, water- stressed growing conditions, a strong potential to resprout following disturbances such as fire and herbivory, and its high wildlife values, public land managers are interested in using C. reticulata to rehabilitate disturbed habitats. However, we must know more of the growth rate, longevity, stand structure, and ecological tolerances of the species to properly evaluate its potential in site enhancement or rehabilita- tion projects. This study sought to answer the following questions: (1) What are the growth rates and longevities of C. reticulata, and do they differ with aspect, parent material, soil texture, per- cent surface rock cover, topographic position, topographic shelter, and grazing intensity of a stand? (2) How does the size class structure of C reticulata stands differ with the environ- mental parameters listed above? Is the species reproducing in Idaho, and does recruitment diftei- under different environmental conditions? (3) Are environmental conditions related to differences in growth form of the plant (i.e., the formation of single vs. multiple stems)? Fig. 2. Idaho distribution of Celtis reticulata. Methods Field Methods Two hundred thirty stands spread over much of tlie Idaho range of C. reticulata were sampled in 1990 and 1991. Approximately 20 stands on the west side of the Snake River, in adjacent Oregon and Washington, were also sampled (total N = 250). Stands were selected based on within-site homogeneity of apparent history, topography, and parent material, and a mini- mum population size of six individuals (many more individuals were usually present). With these constraints for homogeneity, the sam- pling areas were t\'pically irregularly shaped and small, usually less than 0.25 ha. Stands were chosen to represent a range of sites and disturbance histories. Stands were assigned to topographic posi- tions (Table 1) that included river tenace, high- water line, draw, rocky draw, bench, toe slope, lower slope, broken lower slope, mid-slope, upper slope, and talus. The 11 categories were narrowly defined on the assumption that 1995] Celtis reticulata in Idaho 239 Table 1. Definitions of topographic positions in which Celtis reticulata was sampled. River terrace Relatively flat horizontal sintace cut or built b\' river or stream action High-water line Transition line between flood-tolerant and -intolerant plant species Draw Shallow incision in a slope, with <30% total smface rock cover Rocky draw Shallow incision in a slope, with >30% total surface rock cover Bench Nearly level surface usually well above active floodplains and terraces Toe slope Gently inclined, basal part of a slope continuum that grades to the valley, usually <14° slope Lower slope Lower 1/3 of a hillside (above the toe slope, when present); if steep (> 14°) and toe slope absent, the basal part of the slope that meets the valley floor Broken lower Similar to lower slope but with slope extensive smface cover of large boulders and outcrops Mid-slope Middle 1/3 of a hillside, relative to the surrounding landscape Upper slope Upper 1/3 of a hillside, relative to the surrounding landscape Talus slope Coarse, angular rock fragments derived from and King at the base of cliffs or rock slopes; slopes typically >25° combining them at a later time, if needed, would be possible. Based on field observation and reconnaissance, the number of stands sampled within each topographic position was approxi- mately proportionate to how frequently those topographic positions were occupied by the species. Stand-level data recorded, in addition to topographic position, included elevation; latitude; longitude; aspect; slope; percent sur- face rock cover; surface soil texture; parent material; topographic shelter, grazing intensi- ty, total stand density; density within four structural classes, including seedling, juvenile, mature, and decadent individuals; number of cohort modes; and associated dominant plant species (explained below). Surface soil textures were evaluated by moist- ening in the field according to the Soil Conser- vation Service "Guide for Textiu-al Classificaton" (Brady 1974). When soils were unreachable due to surface rock, the surface rock matrix was classified instead. For example, stands on talus slopes had soils categorized as "talus. " Six categories of parent material were iden- tified initially, including granite, sandstone, basalt, river alluvium, rhyolite, and oolitic lime- stone. However, because of the small sample size of rhyolite (4) and its chemical similarity to granite, the two were combined for analy- sis. A similar situation existed for oolitic lime- stone, an uncommon and geographically restricted coarse-grained rock that typically occurred as a lens within sandstone-dominated strata. Therefore, the eight stands on oolitic limestone were combined with sandstone for analysis. Each stand was categorized by "topograph- ic shelter": open (0), intermediate (1), and shel- tered (2). For example, exposed stands grow- ing within a valley were classified as "interme- diate," while stands growing in a side canyon of the same valley were classified as "sheltered." "Open" stands were those with unobstructed exposure to solar radiation. They were typically not associated with a major, incised drainage; rather, they faced broad, expansive valleys. To evaluate recruitment and growth of C. reticulata under different livestock grazing pressures, we scored grazing intensity within a stand as none to moderate (1) or extreme (2). Stands scored as extreme were recognized by (1) heavy browsing of trees, with a hedged or "pasture-tree" growth form; (2) elimination of vegetation under trees by trampling; (3) tree roots exposed by soil compaction and erosion; and (4) dominance of alien plant species. Thirty-six of the 250 stands were classified as extreme. The overall density of Celtis stands was cat- egorized as (1) widely scattered [mature indi- viduals more than 10 crown widths apart]; (2) scattered [mature individuals separated by gaps of 4-10 individual crown widths]; (3) sub- continuous [breaks in the total canopy exist but mature individuals average no more than 3 crown widths apart]; or (4) continuous [little open space in the canopy; crowns form a con- tinuous matrix with occasional gaps]. Inter- mediate sites were recognized with a mid- point value (e.g., 3.5 for stands approaching a closed canopy). To evaluate the composition of C. reticulata stands, densities in four structural classes were also estimated in a similar fashion. The four structural classes were defined as follows: (1) seedling [individual of the year and < 2 yr old]; (2) juvenile [individual >2 yr old and < 1.5 m tall]; (3) mature [>1.5 m tall]; and (4) decadent [>1.5 m tall and experiencing signif- icant dieback, i.e., at least one major dead branch present]. 240 Great Basin Naturalist [Volume 55 Within each stand at least three indixidu- als, chosen to represent the modal size in the stand, were measured and aj^cd. Modal size was defined as typical size of individuals in the dominant (most abundant) cohort. Measure- ments recorded for each tree included height, age, diameter at core height (typically 20 cm above ground level), number of live and dead stems, and percent rock cover below the can- opy as centered over the main trunk. When two or three modal sizes were present, all modes were sampled for a minimum total of either six or nine individuals. When stands were all-aged with no apparent modal tree size, at least six individuals of the dominant canopy cohort were sampled. The number of modes present, from 1 to 4, with 4 equivalent to an all-aged stand, was recorded as a stand- level variable. Most height measurements were obtained with an 8-m, extendable level rod. For taller trees, height was determined with a clinometer. Increment cores were taken at the same height the diameter was measured (20 cm). Cores were transported in plastic straws, glued onto slotted boards, sanded, and annual growth rings were counted under a dissecting micro- scope. When cores did not reach the tree s center (i.e., because of rot), the number of missing years was extrapolated by first sub- tracting the length of the core from the tree's ladius. This remainder was multiplied by the number of rings counted in the core's inner centimeter, which was then added to the num- ber of rings counted for an estimate of the total age. When cores were off-center, the miss- ing radius was estimated by overlaying a clear transparency with a series of circles of known radii over the core, matching the ring pattern in the core with a circle, and multiplying its radius by the number of rings in the centime- ter nearest the core's center This amount was added to the number of counted years to esti- mate tree age. Small-diameter noncoreable individuals were cut down and a cross section was removed, sanded, and the rings counted as above. Analytical Methods Stands were not included in the analysis if the sample size within a particular topograph- ic position or parent material was too small, or if the majority of cores from a stand were illeg- ible after sanding due to contortions in the radial growth. Nine stands were dropped, for a final sample size of 241. SPSS (1988) was used for all analyses. A heat load index was generated to account for differences in heat load from nortlieast- to southwest-facing slopes (Whittaker 1960, Muir and Lotan 1985). For each stand, index values were calculated with the following e(|uation, where 0 = aspect in radians east of north: heat load = (1 - cos(0 - 7r/4))/2. Index values ranged from 0 (NE slopes) to 1 (SW slopes). To compare C. reticulata growth rate and stand structure differences under various en- vironmental conditions, we developed 50-yr site indices as measures of growth potential (i.e., site quality), as outlined in Husch et al. (1972). Site index is based on average heights of dominant trees at a specified index age (usually 50 or 100 yr) and is the most widely used method of evaluating site quality for tree growth (Husch et al. 1972, Daubenmire 1976). Site index curves are constructed to allow for estimation of site index for stands older or younger than the index age, as index age stands are seldom encountered (Husch et al. 1972). The commonly used relationship of tree height to age formed the basis for one index, and the relationship of tree diameter to age formed the basis for the second (DeBolt 1992). The best linear fit was achieved when log (height, m) and log (diameter, cm) were regressed on the log of tree age (fi- = .25, R^ = .54, respectively; N = 939). The resulting equations were log (height) = 0.428 X log (age) - 0.135 and log (diameter) = 0.764 X log (age) - 0.165. Using these two equations, we obtained the expected (mean) height and diameter at 50 yr, then back-transformed to improve inteipretabilit\', yielding an expected size at 50 yr of 3.9 m tall and 13.6 cm in diam- eter For each tree in the data set, the site index was calculated by first finding its residual from the regression line, then shifting this residual to the 5()->'r point on the line, which fields an estimated height and diameter at 50 >'r Thus, the equations to calculate site index (SI) for each tree were; Log (height SI) = 0.591 + (LOGheight - ((0.428 x LOGage) -0.135)) Log (diameter SI) = 1.134 + (LOGdiam - ((0.764 x LOGage) -0.165)) 1995] Celtis reticulata in Idaho 241 To analyze stioictural class differences under differing environmental conditions, the vari- able TYPE, representing types of stand struc- ture, was created. Based on the density of juvenile, mature, and decadent size classes in a stand, the five TYPEs were defined as fol- lows: (1) young (juvenile); (2) mature, nonre- producing, nondecadent; (3) mature, repro- ducing, nondecadent; (4) mature, reproducing, decadent; and (5) mature, nonreproducing, decadent (Table 2). Based on field obsei-vations, mortality of C reticulata seedlings during year one is extreme- ly high. Because most seedlings were year- lings, seedlings were not used to define TYPE. Stands were classified as reproducing when the juvenile density class was 1 or greater (i.e., > 5 individuals). Celtis reticulata growth rate, expressed by site indices, was analyzed as the dependent vari- able in one-way analyses of variance (ANOVA) against the environmental parameters topo- graphic position, parent material, soil texture, grazing intensity, and topographic shelter. Relationships between site indices and ordered categorical independent variables were ana- lyzed by linear regression. With few excep- tions, height site index was a more sensitive predictor of growth differences than diameter site index. Celtis reticulata growth rates and relationships with topographic position and other environmental parameters were also analyzed with analysis of covariance, to com- bine categorical and continuous factors. Included in the model was the categorical vari- able topographic position, with soil texture, topographic shelter, grazing intensity, and par- ent material as four covariates. Relationships between environmental variables and stand structure (TYPE) and the number of modes were analyzed bv contingency tables and ANOVA. Results Growth Log-log regressions best represented the statistical relationship between height and age (Fig. 3) and diameter and age of C. reticulata individuals. An initial impression that regres- sion lines do not fit the scatter of points at log(age) <1.2 can be reconciled by recognizing that the dense central elliptical clouds of points have controlled the regression results. In both cases the least-squares fit resulted in a good fit to the dense cloud of points representing mid- dle-aged trees, but resulted in almost entirely negative residuals for trees younger than 10-25 yr. Because these younger trees were from a small number of sites, many of which showed battering by floods, distributions of residuals were judged to be acceptable. Celtis reticulata diameter and height were tightly related in a log-log regression {R~ = .75). Mean height and diameter of dominant and codominant C. reticulata, regardless of age, were 5 m and 18 cm, respectively. While diameter is a better predictor of age than height (R- = .53 and .25, respectively), height is more responsive to site characteristics than is diameter, both in the literature and in this study. Thus, height was the preferred basis for the site index. Fifty-year-old C. reticulata trees in Idaho averaged 3.9 m tall and 13.6 cm in diameter. Using height, we constructed site index curves Table 2. Categorization of the Celtis reticulata stand structure variable TYPE. TYPE represents the five types of stand structure that were recognized from the density classification. Within each stand, the three size classes of trees (juve- nile, mature, decadent) were assigned to a density' class based on the following definitions. Mid-point values were used as needed. Juvenile: (1) wideK' scattered — 5 or fewer juveniles present; (2) scattered — >.5 juveniles present in a nonag- gregated distribution averaging > 10 canopies apart; (3) subcontinuous — breaks in the total canopy exist but juveniles average >3 and <10 canopies apart. Mature/Decadent: (1) widely scattered — mature individuals >10 crown widths apart; (2) scattered — mature individuals separated by gaps of >4 and <10 individual crown widths; (3) subcontinuous — breaks in the total canopy exist but mature individuals average < 3 crown widths apart; (4) continuous — mature trees form a continuous matrix with only occasional gaps. TY'PE Description 1 Yoim Density cl ass Juvenile Mature Decadent >1 < .5 < 2 <.5 >.5 < 2 >1 >1 < 2 >1 >1 >2 <1 >.5 >2 HMIH Nonreproducing, nondecadent Reproducing, nondecadent Reproducing, decadent Nonreproducing, decadent 242 Great Basin Natuiulist [Vokinie 55 15 10 Site index ' Height at 50 years 100 200 300 AGE. years 400 Fig. 3. Nontransformed log-log regression of Celtis reticulata height (m) on age and site index cui-ves for the Idaho stands. for Idaho Celtis stands to allow site classifica- tion for a stand at any age (Fig. 3). Site quality of an area can be assessed by determining average height and age of dominant trees and locating the position of these coordinates on the site index graph. The stand's site index is then read from the closest curve. Site quality, as expressed by the height- based site index, differed among the eleven topographic positions identified {P = .0001, F = 4.4) (Table 3). However, variation within topographic positions was large, so that at the .05 significance level, only draws differed from any other specific topographic position. Growth was faster in draws than on talus slopes, upper slopes, mid-slopes, and stream terraces. Although site index means did not differ statistically between most topographic positions, a relatively predictable biological ranking of topographic positions was expressed, with a general trend of faster growth where sheltered and mesic to slower growth on more xeric and exposed sites. For example, site index values were smallest on talus slopes, followed by upper slopes, mid-slopes, and stream tenaces (Table 3). Celtis reticulata occurred infrequently on north- and east-facing slopes (Fig. 4A). Twenty- five percent (60) of stands were found on SW slopes, with a heat load between 0.95 and 1.00, the hottest values of the heat load index; 58% (140) were between 0.74 and 1.00. Only 32 stands (13%) were located on the coolest sites between 0.00 and 0.20, or between 350° and 98° east of north. The mean heat load index was 0.69. No stands were found between 349° and 9° east of north. In spite of C. reticulata s affinity for souther- ly exposures, heat load was not a good predictor of hackbeny growth characteristics. More often than not, stands have an affinity for southerly exposures, but because of topographic shelter- ing, growing conditions are often not as harsh or water stressed as they first appear. Of 241 Celtis stands, 168 (70%) had at least an inter- mediate topographic shelter. In a stepwise regression analysis from a pool of six independent variables (soil texture, rock, grazing intensity, shelter, heat load, and slope), shelter was the most important predictor of site index (R2 = .13, p < .001, F = 35.5). Site index values were largest when shelter was greatest, with well-sheltered stands differ- ing from intermediate and open exposures (Table 4). However, variability in growth rates within a given class of shelter is large, as shown by the low R^. Presence of C. reticulata is correlated with surface rock or rock outcrops. Of the 241 stands sampled, 96 (40%) had a surface rock cover of 50% or more (Fig. 4B). Twenty percent of the stands were extremely rock-y, with rock cover- ing 75-98% of the ground surface. Average rock cover was 39%. A weak, inverse relationship between per- cent surface rock cover and site index was found (fi2 = -.28, P = .0001). As rock cover in- creased, site index tended to decrease slightly. Rock was a statistically significant variable in a stepwise multiple regression as well, following topographic shelter in order of entry. Including rock in the model increased the R^ value from .13 to .20 (F = 28.9, P < .001). On sites classified as draws, where topographic shelter is maximized, surface rock cover is less important. Neither parent material nor grazing inten- sity was a statistically significant predictor of site index (F = .43 and .14, respectively). How- ever, site index values differed with soil tex- ture (F = .023, F = 2.07). As with topographic position (Table 3), means were ranked by Fisher's LSD procedure in an intuitively pre- dictable order. Growth rates were higher on finer-textured soils (clay or loam) than on coarse-textured soils (sand). At alpha = .05, the 1995] Celtis reticulata in Idaho 243 Table 3. Site index values of Celtis reticulata (s — standard desiation) for each topographic position. Mean site index (SI) vakies have been transformed back into the original scale of measurement to aid interpretation. Topographic posi- tions with no overlap of similarit\' grouping letters are different from each other at the .05 significance level (Fisher's LSD). Topographic position Mean SI: transformed Mean SI: back- transformed Similarity grouping Draw High-water line Toe slope Rock\' draw Lower slope Bench Broken lower slope Stream terrace Mid-slope Upper slope Talus slope 0.74 (0.16) 5.5 0.65 (0.15) 4.4 0.61 (0.14) 4.1 0.58 (0.17) 3.8 0.57 (0.19) 3.7 0.56 (0.24) 3.6 0.55 (0.12) 3.5 0.51 (0.31) 3.2 0.50 (0.18) 3.2 0.48 (0.20) 3.0 0.47 (0.13) 2.9 30 A 37 AB 17 AB 15 AB 40 AB 16 AB 20 AB 13 BC 28 BC 12 BC 13 BC only pairs that differed from each other were talus and loam. Interactions between soil te.xture and topo- graphic position were highly significant (Chi- square, P = .001). When the analysis of site index and soil texture was restricted to just upland sites, the effect was slightly more pro- nounced (F = .014, F = 2.49). Growth Form "Shrubbiness" was quantified by counting the number of live and dead main stems or trunks of each individual. Regression analysis of stem number with the variables grazing intensity, topographic shelter, soil texture, heat load, slope, average height, average diameter, and percent surface rock cover produced sev- eral statistically significant, albeit weak, rela- tionships. Live and dead stem density per individual decreased as topographic shelter increased {R^ = .20 and .30, respectively). Average height decreased slightly as the num- ber of live stems increased {R- = .20). In gen- eral, on sheltered sites C. reticulata has a sin- gle stem (treelike) rather than multi-stem (shrublike) growth form. Differences in plant growtli form were found among topographic positions and among par- ent materials. Individuals growing at mid-slope were generally shrubbier, with a greater num- ber of live stems (.T = 2.5), than individuals growing at high-water line {x = 1.4), in draws {x = 1.6), and in rocky draws {x = 1.5) (ANOVA, P = .003, F = 2.71). Dead stems were far less numerous than live stems and were absent from most individuals. The num- ber of dead stems at mid-slope (x = 0.6) was greater than all other topographic positions ex- cept upper slopes (P = .0001, F = 6.5). Stands at high-water line, rock-y draw, stream terrace, draw, and broken lower slope topographic positions averaged only 0.1 dead stems per individual. Growth form did not differ with the number of size modes within a stand. Individuals on sandstone were more com- monly multi-stemmed than those on the three other parent materials, for both living and dead stems (F < .001, F = 8.5; F < .001, F = 14.7, respectively). Longevity The mean age of individuals sampled dur- ing our study was 66 yr, with a range of 1-374 yr (Fig. 5). Old age and large size are not tight- ly related. For example, it is common to find trees 10 m tall but less than 75 yr old. Diameter was often a better predictor of age than was height (F- = .54 and .26, respective- ly, after log-log transformation). The oldest C. reticulata recorded in this study (about 374 yr) grew on an exposed talus slope approximately 300 m above the Salmon River; it was 4.6 m tall and 48 cm in diameter at 20 cm above ground level. Percent surface rock cover of the site was 90%, with the small stand of scattered trees restricted to talus margins where pockets of soil were exposed. Other members of the stand ranged in age from 191 yr (3.35 m tall, 28 cm diam) to 320 yr (5.48 m tall, 46.5 cm diam). 244 Great Basin Naturalist [Volume 55 -inOi LIl -10 -0.2 0.0 0.2 0.4 0.6 0 HEATLOAD J i_ 1.0 1.2 E7i 30 - 20 40 60 80 100 120 PERCENT SURFACE ROCK Fig. 4. Frequenc\' distrihutions of the niiml:)er of Celtis reticulata .stands by (A) heat load and (B) percent surface rock- cover. Stand Structure Of the 241 Celiis stands, 178 (74%) were re- producing and only 4 (1.7%) of diese were deca- dent. Fifty-seven stands (23.5%) were classified as nonreproducing, 6 (2.5%) of them decadent. The remaining 6 stands (2.5%) were recently established (juvenile dominated), with no mature individuals present. Structure of C. reticulata stands, in terms of their relative densities of juvenile and mature size classes, was unrelated to soil texture (Chi- square, P = .31). Structural type was weakly related to topographic position of the stand (Chi-square, P = .08). Of the 11 topographic positions, rocky draws had the highest juve- nile density, or recruitment. In general, juve- nile densities increased as the percent of sur- face rock cover increased. Density of C. retic- ulata juveniles was highest when rock cover was 50% or greater Rocky draws consistent])' had the densest canopies, followed by draws and high-water line. Rocky draws were never assigned an over- all density <2.5, where 3 = subcontinuous. In fact, 75% of rocky draws had closed or nearly closed canopies (overall density = 3.5 or 4). Juveniles were often present on the margins of rocky draws. The few decadent stands were found higher on the slope, on steeper slopes, and in less- sheltered positions than nondecadent stands. Nonreproducing, decadent stands were more steeply sloping than young and nondecadent, nonreproducing stands (ANOVA, P = .003, F = 4.03) (Table 5). Of the 10 decadent stands, 50% were at mid-slope and 20% were on talus. While none of the decadent stands were ex- tremely overgrazed, their distance from water may have confounded this result. Overgrazed stands were typically found on fairly gentle terrain (.v = 14°, S.D. = 8.6) and in close prox- imity to a water source, where livestock tend to concentrate, while decadent stands were on steeper slopes (Table 5) and at higher slope Table 4. Mean site index (SI) vakies for Celtis reticulata for three lexels of topographic sheher, in both transformed and back-transformed scales. Topographic shelters with no o\erlap of similarit) grouping letters are different from each other at the .0.5 significance level (Fisher's LSD). Topographic shelter Sheltered Intermediate E.xposed Mean SI: Mean SI: transformed back- Similarity (s) transformed A' groupmg 0.69 (0.15) 4.9 73 A 0.55 (0.20) 3.6 129 B 0.49 (0,14) .3.1 39 B 1995] Celtis reticulata in Idaho 245 100 r - 80 - n - h- 60 - . O p. q: n 2 3 40 20 n n - 0 UUUU JUuUUUUUL 0D0== o . = Table 5. Average slopes of decadent, nondecadent, and yoinig stands of Celtis reticulata, with the variable TYPE in its original five-categoiy format. TYPE represents the five types of Celtis reticulata stand structure that were recognized. 100 200 TREE AGE (years) Fig. 5. Frequency distribution of the number of Celtis reticulata trees by tree age. positions. Less intensively grazed stands aver- aged 23° (S.D. = 10.1). Grazing level was related to stand structure (TYPE; Chi-square, P = .0002). A larger per- centage of heavily grazed stands (53%) were nonreproducing than were stands with light or moderate grazing intensity (18.5%). Even though sample sizes were veiy different (light or moderate = 205, extreme = 36), the pattern confirms field observations of low recruitment under extreme grazing pressure. However, it is perhaps even more noteworthy that recruit- ment on heavily grazed sites is as high as it is, given how few, if any, other shrub species are present on such sites. Among the four parent materials, 37% of C. reticulata stands growing on sandstone were nonreproducing, as compared to 21%, 22%, and 21% of stands growing on granite, basalt, and river alluvium, respectively (Chi-square, P = .014; Table 6). A greater number of sand- stone-associated stands were nonreproducing than expected (14 and 9, respectively), while fewer were reproducing than expected (22 and 27, respectively). Expected and observed val- ues for the three other parent materials were more similar Newly established C. reticulata stands are apparently rare, as few were observed during the study in spite of efforts to locate them. Only six young (<33 yr) stands were sampled. These were typically on rock)' sites with inter- mediate topographic shelter and gentle slopes TiTE N Mean slope (degrees) S.D. Young (juvenile) 6 12 3.4 Nonreproducing, nondecadent 51 20 9.6 Reproducing, nondecadent 174 22 10.3 Reproducing, decadent 4 27 7.7 Nonreproducing, decadent 6 33 5.0 (x = 12°). All had at least 15% surface rock cover, but most had 75% or greater rock cover (x = 65%). Five of the six stands were on allu- vium, including stream terraces, high-water lines, and benches. All five had sandy soil. The sixth stand was atypical, occurring near a mid- slope, sparsely vegetated band of sandstone with intermediate shelter. All individuals were shrubby, decadent, and old (18-33 yr) relative to the average height of 0.7 m (expected age = 8 yr). Soils were sandy loam in texture. While young stands were only on sites with intermediate topographic shelter, reproducing and nonreproducing stands differed little in the degree of shelter they received (Chi- square, P = .06). Thirty-three percent of repro- ducing stands were sheltered, compared to 25% of nonreproducing stands. The amount of surface rock differed weakly across stand structure (TYPE; ANOVA, P = .038, F = 2.58). Differences were greater when the variable TYPE was restructured to three categories (mature reproducing, mature non- reproducing, young), eliminating decadence as a factor (ANOVA, P = .015, F = 4.26). Under the three-level categorization, young stands were rockier than mature, nonrepro- ducing stands (x = 32%) but did not differ from those that were reproducing. Number of Modes Celtis reticulata stands typically appeared to be unimodal (73%), with one dominant cohort. Stands with two modes were far less common (11%), but a slightly greater number were all- aged (16%). Since only two stands had three modes, they were dropped from analyses; 246 Great Basin Naturalist [Volume 55 Table 6. Cross tabulation ol tlic miiiibi r oi Ccltis reticulata stands b\ stand structure and parent material. The hypothesis of independence of stand structure and parent material is rejected with P — .014. Observed/ expected Mature Row total material Nonreprod. Reproil. Younu Row % Granite obs. 11.0 40.0 0,0 51 21 exp. 12.2 37.6 1.3 Sandstone obs. 14.0 22.0 1.0 37 16 Basalt exp. obs. 8.8 20.0 27.2 71.0 0.9 0.0 91 38 Aihniuni exp. obs. 21.5 12.0 66.3 45.0 2.3 5.0 62 25 exp. 14.5 44.9 1.5 Column total obs. 57.0 178.0 6.0 241 100 Column % 24% 74% 3% thus, the sample size for this portion of the results is based on 239 stands. Growth form or number of stems of the individuals was unre- lated to number of modes. Although of marginal statistical significance, all-aged stands were more common on shel- tered sites (Chi-square, P = .07). For example, 33% of stands in draws, which typically have at least an intermediate topographic shelter, were all-aged. The percent of all-aged stands at other topographic positions ranged from 6% to 16%. Livestock grazing intensity may restrict entry of new cohorts within a C. reticulata stand as shown by the strong tendency for overgrazed stands to be unimodal (92%; Chi-square, P = .0008). In contrast, 70% of light to moderately grazed stands had only one mode, 11% were bimodal, and 18% were all-aged. Size structure of Celtis stands did not differ with topographic position, parent material, soil texture, slope, percent surface rock, or heat load (all P > .2). Discussion In our study, trees were typically tallest and least shrubby when located in draws, on sites with surface or subsurface moisture, and in areas where they received maximum topographic shelter. Similar observations of C. reticulata have been recorded by others (Eliot 1938, Van Dersal 1938, Peattie 1953), and for different species of Celtis as well. For example, Hitch- cock and Cronquist (1964) noted that Celtis reticulata is taller in moist areas in the Pacific Northwest. In Oklahoma, C. laevigata (sugar- berry) is typically a small tree in open areas, but in lowland forests it reaches its maximum development (Schnell et al. 1977). On the eastern Great Plains, C. tcnuijolia (dwarf hack- berry) is a gnarled, shrublike tree when grow- ing on its typical rocky, shallow, calcareous substrate, but in the bottom of ravines it may reach heights of 8-10 m (Stephens 1973). In addition to the influence of an ameliorated environment, sheltered stands may be less prone to repeated disturbances such as fire, to which a vegetative sprouter such as C. reticu- lata will often respond with a shrubbier growth form. In Oklahoma, Celtis occurs almost exclu- sively on loamy bottomland soils (Dooley and Gollins 1984), and in west Texas it is best de- veloped on alluvium (Van Auken et al. 1979). In the canyon grasslands of Idaho, Tisdiile (1986) recognized two types of Cc/f /.s-dominated vege- tation on soils of two principal origins. The C. reticulata- Agropijron spicatiim habitat type occurs on lower valley slopes with rocky (50- 60%), weakly developed loam soils derived from residual and colluvial materials. The sec- ond vegetation type, unclassified because of heavy livestock disturbance and alien plant dominance of the understory, occurs on allu- vial terraces with deep sandy soils. Soil texture appears to have a greater influ- ence on C. reticulata growth on upland sites than on sites associated with a perennial water source. While C. reticulata grows on a range of soil textures in Idaho, we found the tallest trees on loams, possibly because of their greater water-holding capacity and nutrient content. However, 80% of the stands occurred on soils with some sand component, and 30% were on sand or coarse sand. The presence of good drainage may be an imjDortant limiting 1995] Celtis reticulata in Idaho 247 factor for C. reticulata, as finer-textured soils of the uplands were nearly always skeletal. The increased percolation of sandy or skeletal soils provides greater moisture availability for deep-rooted shrub and tree species. In Idaho, C. reticulata is most prevalent on rocky sites with southeast to westerly aspects, although heat load was not an important pre- dictor of growth. The presence of rock, particu- larly bedrock, may in fact be critical for hack- beriy's existence on certain sites. It may also partially explain the fragmented distribution of the species in Idaho. Other rock-associated species have been obseived in semiarid regions as well. In the shrub-steppe region of eastern Montana, Rumble (1987) found that scoria rock outcrops provide a unique habitat for several relatively mesic species. Rhus trilobata (skunkbush sumac), Prunus virginiana (choke- cherry), Ribes spp. (currant), and Jiiniperus spp. (juniper) were found only in association with rock outcrops in that ecosystem. He con- cluded that their occunence is probably related to protection from wind, snowdrift accumula- tion, shading, and mulch effects of rocks. Oppenheimer (1964) and Potter and Green (1964) suggested that the association of mesic species with rocky substrates is due to tempo- rary water reservoirs that rock fissures pro- vide. In Arizona, Johnsen (1962) reported that Juniperus monosperma (one-seed juniper) is largely limited to rock outcrops, where he recorded 2-2.5 times as much available mois- ture. The theoiy of extra moisture availability in rock fissures could also hold true for the deeply rooted C. reticulata, helping explain its frequent presence on southerly aspects. Other plausible explanations for the infre- quency of C. reticulata on northerly aspects and sites with less surface rock cover include its sensitivity to late spring frosts (personal observation) and poor competitive ability with fast-growing species. In Idaho, C. reticulata is the last shrub to break dormancy and expand its leaves in the spring. This strategy, in com- bination with the tendency to grow on warmer slopes, generally prevents frost damage from occurring. The greater effective soil moisture and dense vegetative cover of north slopes probably create an environment too competi- tive for this slow-growing species. In summary, Celtis reticulata can generally be described as slow-growing and small- statured. Fifty-year-old trees averaged 4 m tall and 13.6 cm in diameter in Idaho, with a mean tree height and diameter, regardless of age, of 5 m and 18 cm, respectively. Unlike some shnib and tree species in the Intermountain West, populations are generally maintaining them- selves by vegetative sprouting or seedling recruitment, despite historic and prevailing large-scale habitat alterations resulting from overgrazing, exotic plant invasion, and chang- ing fire frequencies (Tisdale 1986, Whisenant 1990). Hackbeny's general resiliency and abil- ity to resprout following disturbance or injury likely play a role in this, as does its positive asso- ciation with rock. Recruitment, as expressed by the density of juvenile individuals, in- creased as surface rock cover increased. How- ever, under extreme grazing pressure, recruit- ment was significantly lowered and stands were nearly all unimodal. All-aged stands were absent from severely grazed sites. Even though rock favors Celtis recruitment, its growth is favored on less-rocky sites, such as draws. The most likely explanation for relatively slow C. reticulata growth on stream terraces, in spite of the assumed availability of ground- water, is the extreme annual fluctuation of the water level and battering by flood debris. These sites are located below the high-water line. Above the high-water line the mean site index is larger and mechanical stresses are less extreme. While newly established C. reticulata stands were uncommon, they typically occuired on these riparian sites, where establishment conditions occur more frequently than in the uplands. Although individuals are often slow-grow- ing, the variation in site conditions that the species appears to tolerate and its other posi- tive attributes (i.e., wildlife food, cover, land- scape structure, potential large size, tolerance of southerly aspects), are favorable qualities for those seeking rehabilitation species. The species' persistence in heavily degraded ecosystems may speak to its suitability for rehabilitation projects as well. Acknowledgments This study was funded in part by the Boise District Office of the Bureau of Land Manage- ment, witli additional support provided by Idaho Power Company. Nancy Shaw, Ed Tisdale, and Steve Monsen provided insight and 248 Great Basin Naturalist [Volume 55 encouragement during the earliest phase of the research. Roger Rosentreter assisted in the field and pro\'ided helpful suggestions and encouragement throughout the study s dura- tion. We thank Patricia Muir, Boone Kauffinan, and Kermit Cromack for their valuable com- ments on an early version of the manuscript. Thanks arc also due to Stanle>' D. Smith, Sherel Goodrich, and an anonymous reviewer tor their constructive review of the manuscript. Literature Cited ALBEE, B. J., L. M. SCHULTZ, AND S. GOODRICH. 1988. Atlas of the vascular plants of Utah. Utah Museum of Natural History, Occasional Publication No. 7. Salt Lake City, UT 670 pp. Asheri.n, D. a., and J. J. Claar. 1976. Inventoiy of riparian habitats and associated wildlife along the Columbia and Snake rivers. Volume 3A. College of Forestry, Wildlife, and Range Sciences, University of Idaho, Moscow. 556 pp. Bowler, P A. 1981. Natural histoiy studies and an evalua- tion for eligibility of the Wiley Reach of the Snake Ri\er for National Natural Landmark designation. Unpublished report, National Park Service, Seattle, WA. 86 pp. Brady, N. C. 1974. The nature and property of soils. Macmillan Publishing Co., Inc., New York. 639 pp. Brown, D. E. 1982. Great Basin montane scrubland. Pages 8.3-84 in D. E. Brov\n, editor, Biotic commu- nities of the American Southwest — United States and Mexico. Desert plants. Volume 4. Daubenmire, R. 1970. Steppe vegetation of Washington. Agriculture E.xperiment Station Technical Bulletin No. 62. Pullman, WA. 131 pp. . 1976. The use of vegetation in assessing the pro- ductivity of forest lands. Botanical Review 42: 11.5-143.' DeBolt, a. M. 1992. The ecology of Celt is reticulata Torr. (netleaf hackberi-y) in Idaho. Unpublished master's thesis, Oregon State Universit>', Coi-vallis. 161 pp. DooLEY, K. L., AND S. L. COLLINS. 1984. Ordination and classification of western oak forests in Oklahoma. American Journal of Botany 71: 1221-1227. Elias, T. S. 1980. The complete trees of North America. Times Mirror Magazines, Inc., New York. 948 pp. Eliot, W A. 1938. Forest trees of the Pacific Coast. G. R Putnam's Sons, New York. 565 pp. Glinski, R. L. 1977. Regeneration and distribution of syca- more and Cottonwood trees along Sonoita Creek, Santa Cruz County, Arizona. Pages 116-123 in R. R. Johnson and A. Dale, technical coordinators. Impor- tance, preservation, and management of riparian habitat — proceedings. USDA General Technical Report RM-43. Fort Collins, CO. Hayward, C. L. 1948. Biotic communities of the Wasatch chaparral, Utali. Ecological Monographs 18: 473-506. Hitchcock, C. L., and A. Cronquist. 1964. Vascular plants of the Pacific Northwest. Volume 2. Univer- sity of Washington Press, Seattle. .597 pp. HuscH, B., C. I. Miller, and T. W. Beers. 1972. Forest mensuration, fohn Wilev & Sons, Inc., New York. 410 pp. Hlscih.E, Ci. 197.5. Analysis ot the \egetation along the middle and lower Snake River Unpublished master's thesis. University of Idaho, Moscow. 271 pp. JoHXSEN, T N. 1962. One-seed juniper invasion of north- ern Arizona grasslands. Ecological Monographs 32: 187-207. Johnson, C. G., Jr., and S. A. Simon. 1987. Plant associa- tions of the Wallowa-Snake Province, Wallowa- Whitman National Forest. USDA Pacific Northwest Region Publication R6-ECOL-TP-255B-86. 272 pp. Lanner, R. M. 1983. Trees of the Great Basin. University of Nevada Press, Reno. 215 pp. Little, E. L., Jr. 1976. Adas of United States trees. Vol- ume 3-Minor western hardwoods. U.S. Department of Agriculture Miscellaneous Publication No. 1314. Washington, DC. 215 pp. MuiR, P S., AND J. E. LOTAN. 1985. Disturbance histoiy and serotiny of Finns eontoiia in western Montana. Ecology' 66:' 1658-1688. Oppenheimer, H. R. 1964. Adaptation to drought: .xero- phytism. Pages 10.5-135 in Plant-water relationships in arid and semi-arid conditions. The Hebrew Uni- versity, Rehovot, Israel. Peattie, D. C. 19.53. A natural histoi-y of western trees. Houghton Mifflin, New York. 751 pp. Plummer, a. P 1977. Revegetation of disturbed inter- mountain area sites. Pages 302-337 in J. C. Thames, editor. Reclamation and use of disturbed lands of the Southwest. University of Arizona Press, Tucson. Potter, L. D., and D. L. Green. 1964. Ecology of pon- derosa pine in western North Dakota. Ecology' 45: 10-23. Rumble, M. A. 1987. Avian use of scoria rock outcrops. Great Basin Naturalist 47: 62.5-630. Schnell, G. D., P G. Risser, and J. F Helsel. 1977. Factor analvsis of tree distribution patterns in Okla- homa. Ecology 58: 1345-13.55. SPSS. 1988. SPSS/PC -h V2.0 base manual. SPSS Inc., Chicago. Stephens, H. A. 1973. Woody plants of the North Central Plains. University Press of Kansas, LawTence. 530 pp. Tisdale, E. W 1986. Canyon grasslands and associated shrublands of west-central Idaho and adjacent areas. Bulletin No. 40. Forestiy, Wildlife, and Range Exper- iment Station, University of Idaho, Moscow. 42 pp. Van Auken, W W, A. L. Ford, and A. Stein. 1979. A com- parison of some woody upland and ripariiui plant com- munities of the southern Edwards Plateau. South- western Naturalist 24: 165-180. Van Dersal, W. R. 1938. Native woody plants of the United States, their erosion control and wildlife val- ues. U.S. Department of Agriculture Miscellaneous Publication No. 303. Washington, DC. 362 pp. Whisenant, S. G. 1990. Changing fire frequencies on Idaho's Snake River Plains: ecological and manage- ment implications. Pages 4-10 in E. D. McArthur, et al., compilers, Cheatgrass invasion, shrub die-off, and other aspects of shrub biolog\' and manage- ment— proceedings. U.S. Department of Agriculture General Technical Report INT-276. Ogden, UT Whittaker, R. H. 1960. Vegetation of the Siski\ou Mountains, Oregon and California. Ecological Mono- graphs 30; 279-338. Received 22 March 1994 Accepted 29 November 1994 Great Basin Naturalist 55(3), © 1995, pp. 249-257 MIMULUS EVANESCENS (SCROPHULARIACEAE): A NEW ANNUAL SPECIES FROM THE NORTHERN GREAT BASIN Robert J. Meinke^ Abstract. — Recent taxonomic studies in Mhmdiis support the recognition of Mimultts evanescens, a new autoga- mous species moiphologically allied with M. brevijlorus and M. latidens. Initially known only from herbarium speci- mens, the most recent from 1958, M. evanescens was relocated in the field in 1990 in northern Lassen Co., CA. A sec- ond population was found in southern Lake Co., OR, in 1993. Mimidus evanescens is apparendy confined to the Great Basin and its peripheiy where it has been recorded from 10 localities across Idaho, Oregon, and California. Based on collection information and visits to the two extant populations, the new species appears to be restricted to vernally moist sites and fluctuating banks of intermittent streams or pools. Long-term utilization of such sites by livestock may have contributed to the present-day rarity of M. evanescens. The species should be added to federal and state lists of candi- date endangered species pending the results of future field studies and surveys. Key words: Mimulus, Great Basin, faxononiy, Scroplndariaceae, nwnkeyflower, Mimulus breviflorus, Mimulus latidens. Mimuhis breviflorus is a diminutive, self- pollinating, annual monkeyflower occurring primarily east of the Sierra Nevada and Cascade Mountains in the northwestern United States and adjacent British Columbia. Little is known concerning the evolutionaiy or taxonomic rela- tionships of this or most other taxa in the genus, which comprises perhaps 100 predominantly North American species (Thompson 1993). In the only comprehensive monograph of the genus, Grant (1924) placed M. breviflorus in section Paradanthus, an assemblage of small, problematic species groups that are probably paraphyletic and considered difficult to align taxonomically (Argue 1980). Indeed, in a pro- posed phylogenetic chart Grant (1924) affiliated the yellow-flowered M. breviflorus with the M. moschatus alliance, while in the text of her paper she associated the species with mem- bers of the M. inconspicuiis group, particularly the white- to pinkish-flowered M. latidens of California. The proposed relationship between Mimulus breviflorus and M. latidens is largely based on shared features of the corolla and calyx. Both species possess short, inconspicuous corollas and strongly plicate, chartaceous fruiting calyces that inflate with age. Although inflated calyces are also described for some members of the M. moschatus complex (Grant 1924, Munz 1959, Holmgren 1984), the consistently reduced, essentially regular flowers of M. brevi- florus and M. latidens are unlike any species in that group. The calyx moiphology and texture of the two species is also different, being sin- gularly reminiscent of M. inconspicuus and its proposed relatives (Grant 1924). Moreover, the general habit of M. brevijlorus and M. latidens is more comparable to this group than to any other Despite the similarities, Mimulus breviflorus and M. latidens are quite distinct with respect to geography and habitat. Mimulus breviflorus is a basin and range species, principally occur- ring in well-drained, rocky environments near rain pools, rocky meadows, and ephemeral streamsides, often at middle and upper eleva- tions. It has rarely been recorded south of extreme northeastern California, and only then above 2000 m. Mimulus latidens occurs mostly on poorly drained flats or slopes subject to vernal inundation, primarily below 800 m. The species is virtually endemic to California, extending from the Central Valley to northern Baja California. The apparent uncertainty by Grant (1924) over the taxonomic placement of M. breviflorus may have been influenced by geography, in that the range of the species overlaps much of the M. moschatus complex but not M. latidens or the M. inconspicuus ^Restoration Eeolog,' and Plant Conservation Biologv' Cooperative Project. Department of Botany and Plant Pathologx', Oregon State Universit)', Corvallis, OR 97331. (The Restoration Ecolog\' and Plant Consenation Biolog>- Cooperative Project is a collaborati\e reseiucli unit of Oregon State Universit)' and the Oregon Department of Agriculture.) 249 250 Great Basin Naturalist [Volume 55 group, which are restricted to cisniontane Cahfornia. The present study was prompted b\' several unusual herbarium collections identified as Mitiuihis hrcvijloriis, disco\ ered during a taxo- nomic sune> of the Mimiilus washingtonensis complex (Meinke in preparation) in which sev- eral hundred collections (including all rele- vant types) were examined. Despite the evi- dently yellow flowers and the fact that the few collection localities were well within the known range of M. breviflonis, the plants were similar to M. latidens in many respects. The anomalous material originated from several scattered stations across the upper Great Basin and its northern periphery, all within areas belie\'ed historically grazed by livestock. The most recent of these collections is dated 1958, and tliere was concern that the entity may have become extinct in die interim. Unexpected opportunities to obsei've living populations in the field were presented in 1990 and 1993 dur- ing chance visits to two reservoirs in Lassen Co., CA, and Lake Co., OR. The unique and consistent combination of features noted in herbarium collections, including bright yellow corollas, was even more conspicuous in living plants, prompting a taxonomic reevaluation of their relationship with M. breviflonis and M. latidens. After further evaluation, the unusual populations were considered to represent a new species which is here described, illustrated, and contrasted with potentially related taxa. Description of the Species Mimulus evanescens Meinke, sp. nov. (Figs. lA-C).— Type: USA, California, Lassen Co., 20.5 km east of Adin, north side of Ash Valley Rd., ca 0.1 km east of the Lassen National Forest boundary, in broken boulders and heavy gravel abutting Moll Reservoir, T38N RIOE, NWl/4 SWl/4, Sect. 25, ca 1500 m, 27 June 1990, Meinke and Kaije 5900 (holotype, OSC; isotypes, MO, NY, RxM, UC, US, UTC). Herhae annuae, puberulentes, ± viscido-vil- losae; caulis tenuis, erectis, (6-) 10-25 cm altis, internodiis elongatis; /0///5" late ovatis vel lanceolatis, lamina integerrima vel parce den- ticulata, acuta, 1.0-3.8 cm longa, 0.7-2.9 cm lata, 3(-5) nei-vis, base lata, sessili vel subsessili; pedicel foliis brevioribus, tenuibus, ascenden- tibus; calyce in statu florifero 3.5-6.5 mm longo, 1.5-3.5 mm lato, in statu fructifero late urceolato, 7.0-11.0 mm longo, 5.0-8.5 mm lato, valde glabro, dentibus ciliati, late triangu- laribus, ± subaequalibus, acutis; corolla flava, brevi, 4.0-9.5 mm longa, calyce ca 1.5 plo lon- giore, tubo incluso, lobis ± aetiualibus, patulis, erectis; statninibus stylo aequalibus, inclusis, glabris; stylo glabro, 3.0-7.8 mm longo, labiis stigmatis laciniatis, subaequalibus; capsula inclusa, subglobosa, 4.8-9.0 mm longa, sessili vel substipitata; seininibiis late oblongis, ca 0.3-0.6 mm longis. Annual herb, ± succulent, glandular- puberulent throughout (except the calyces), the hairs short and appearing of even length to the naked eye, moist or slimy to the touch, mostly one-celled (excluding the gland); stems slender, (6-) 10-25 cm tall, erect to slightly decumbent in robust individuals, simple or branched from near the base, often sparingly branched above as well, with elongated inter- nodes; leaves acute, broadly ovate to some- what lanceolate, 1.0-3.8 cm long, 0.7-2.9 mm wide, evenly distributed, not much reduced at the upper nodes, not forming a basal rosette, the lower ones abruptly petiolate or subses- sile, petioles 1-3 mm long, blades broadly ses- sile above, with 3(-5) primaiy veins, the mar- gins entire or shallowly denticulate; pedicels slender, 8-18 mm long, ascending in flower and fruit, shorter than the leaves in fruit or longer in depauperate individuals; inflores- cence racemose, flowers axillaiy; ^o«;ers incon- spicuous, autogamous; calyx 3.5-6.5 mm long and 1.5-3.5 mm wide in flower, tubular-cam- panulate at anthesis, green becoming strami- neous and anthocyanic along the angles with age, accrescent and broadly urceolate to oxal in fruit, 7.0-11.0 mm long and 5.0-8.5 mm wide, the tube ehartaceous and glabrous, the orifice narrowing and becoming somewhat oblique, the angles strongly plicate, the teeth broadly triangular, acute, 0.8-1.6 mm in fruit, ciliate on the margins, scarcely unequal, the uppermost lobe occasionally appearing slightly longer in some flowers; corolla short and essentially regular, 4.0-9.5 mm long, clear yel- low or occasionally with a few tiny brownish dots in the throat, the inconspicuous petal lobes rounded or mucronate, the tube includ- ed or barely exserted, the limb exceeding the calyx by 2-3 mm, lobes short and subequal, mostly erect, glabrous externally, bearded in- ternally with a few ± clavate hairs extending in a line from the lower palate into the floral tube; 1995] New Mimulus From Idaho, Oregon and California 251 4 mm 4 mm 4 mm Fig. 1. Mimulus evanescens Meinke: A, habit drawing showing details of leaf moiphology, inflorescence, and relation- ship between pedicel and leaf blade length; B, close-up of fruiting calyx of M. evanescens; C, calyx of M. evanescens opened to show sessile capsule insertion; D, calyx of M. latklens opened to show stipitate capsule insertion. stamens included, about equal with the style, glabrous, pale; style glabrous, included, 3.0-7.8 mm long, stigma lips equal or subequal, shal- lowly laciniate-margined; capsule included, subglobose, 4.8-9.0 mm long, extending to ca 1.0-2.5 mm below the sinuses of the calyx teeth, sessile, or rarely with an abbreviated stipe up to ca 0.5 mm long, the placentae adherent to the apex; seeds ovoid or broadly oblong, brownish, 0.3-0.6 mm long, dormant when first ripe, dispersal often delayed or pro- longed due to the nearly closed, inflated calyx. Para'HPES. — USA, California, Lassen Co., 10 mi south of Ravendale, 9 June 1940, Pennell 25763 (P); 4.8 mi south of Madeline, 17 June 1958, Raven and Solbrig 13298 (JEPS); Modoc Co., along Willow Creek, June 1894, Austin s.n. (UC). Idaho, Owyhee Co., meadow, 3 mi south 252 Great Basin Naturalist [Volume 55 of Riddle, 1 July 1949, Hohngren and Holmgren 7973 (CAS, UC, WS, WTU). Oregon, Crook Co., Grizzly Butte, 18 June 1894, Leiberg 275 (NY, 0RE,'US); Gilliam Co., forks of Cotton- wood Canyon, 6 June 1894, Leiherg 156 (NY, ORE, P, US); Grant Co., Ochoco National Forest, Graylock Butte, 6 July 1912, Ingram s.n. (RM); Harney Co., dry watercourse near Frenchglen, 26 June 1942,' Peck 21389 (CAS, NY, I^ UC, WILLU); Lake Co., moist to muddy margins of receding water, among rocks, Drews Reservoir, 15 June 1993, Meinke and Carlson 6401 (BRY, HSC, NY OSC, RM, RSA, SRR UC, US, UTC, WS, WTU). Distribution and habitat — Mimulus evan- escens is distributed widely along the north- western edge of the Great Basin at elevations of ca 1200-1700 m, ranging from southwest Idaho west through eastern Oregon and south into northeastern California. Mimulus hrevi- florus is more widespread and considerably more common. Although inconspicuous even when in bloom, it has been recorded from numerous collections located throughout much of the northwestern United States east of the Sierra Nevada and Cascade ranges. Outlying populations are known from southern British Columbia and south (rarely) to the mountains near Lake Tahoe. Mimulus breviflorus has a broader elevational range than M. evanescens, occurring from roughly 300 to 2900 m. Mimulus latidens is essentially a California endemic, distributed below 800 m from the northern Central Valley south to San Diego. The most southerly populations are known from northern Baja California, while four his- toric collections from extreme southwestern Oregon, originally identified as M. breviflorus, represent the northern range limits. A recently discovered Great Basin population of M. lati- dens, occurring at ca 1700 m in southern Lake Co., OR (Shelly 1986), is noteworthy as it is the single recorded locality in which the range of this otherwise low-elevation species over- laps either M. evanescens or M. breviflorus. The population was persisting over several acres in a sagebrush-dominated swale as of 1993. This area is along the flyway for various waterfowl species migrating northeast across the Great Basin from central California. The habitat of Mimulus evanescens can be evaluated only from the two extant localities, the first adjacent to Moll Reservoir in Lassen Co., CA, and the second at Drews Resei-voir in Lake Co., OR, l)()th occurring within sage- brush-juniper-dominated vegetation zones. Plants at both sites were scattered among rock fragments and alongside small boulders, in moist, heavy gravel that had been inundated earlier in the spring. The California popula- tion was discovered in 1990 and visited again in 1991, while the Oregon population was first located in 1993. Associated species during these years (for both locations) included Artemisia tridentata, Junipenis occidentalis, Mimulus flori- bundus, M. suksdorfii, Porterella carnosula, Collinsia grandiflora, C. parviflora, Downingia sp., Mimetanthe pilosa, Heterocodon rariflorum, Poa bulbosa, and Bromus spp. The perennials Machaerocarpus californicus and Marsilea vestita were common along the shoreline at the Lassen Co. site. Remaining locations for M. evanescens are known only through scanty herbarium labels, with specimens reportedly taken from rocky stream banks or drying watercourses. Mimulus breviflorus occurs in comparable microsites, frequenting wet, rocky sites that often diy out by late spring or early summer, as well as lush, gravelly meadows. Morphological comparisons. — Monkey- flowers are often phenotypically plastic, and related annual species in particular may be sub- ject to overlapping moi-phological variation de- pending on ecological conditions. In an attempt to objectively evaluate the phenetic relation- ships of the new species and its most similar congeners, a data set was compiled by scoring 18 vegetative and reproductive character states (Table 1) from 114 Mimulus collections representing 38 populations. Measurements were taken from 15 populations each of M. latidens and M. breviflorus, and 8 of the 10 extant and historical populations of M. evan- escens. Three plants were measured per collec- tion to provide population averages for each quantitative trait. Sample populations of M. latidens and M. breviflorus were selected from herbarium collections encompassing the geo- graphic range for each species. Every effort was made to choose individuals of the three species that, based on field experience of the author, represented normally developed plants (i.e., not drought-stressed) from approximately the same life-history stage. Measurements were made on randomly selected individuals where possible, insofar as the limited number of phenologiciilly acceptable collections pemiitted. 1995] New Mimulus From Idaho, Oregon and California 253 Table 1. List of morphological traits measured from Mimtihis plants for use in principal components analysis. Thirty-eight study populations were sampled, including 15 each for M. latidens and M. breviflorus, and 8 for M. evanescens. An average measurement was derived for each trait (from 3 samples per population) for use in the analyses. (1) Presence or absence of a basal rosette (2) Length of initial stem leaf (3) Width of initial stem leaf (4) Length of upper cauline leaf (5) Width of upper cauline leaf (6) Base of upper cauline leaf (sessile versus distincth- petiolate) (7) Peduncle length (in fruit) (8) Caly.x length (in fruit) (9) Calyx width (in fruit) (10) Length of caKx teeth (in fi-uit) (11) Overall corolla length (12) Corolla color (yellow versus rose, whitish, or ochroleucous) (13) Length of lower corolla lip (14) Width of lower corolla lip (15) Length of capsule (16) Width of capsule (17) Capsule insertion (base sessile versus distinctly stipitate) (18) Stem and leaf pubescence (clearly glandular- puberulent versus glabrous or subglabrous) The data set was initially used to analyze moiphological relationships between Mimulus evanescens, M. latidens, and M. breviflorus using a principal components analysis (PC A). Clustering relationships of sample populations were compared along the first two a.\es of vari- ation and graphically displayed. As a second measure of overall dissimilarity, canonical dis- criminant analysis (DA) was performed on the same populations using only the quantitative characters from Table 1 (i.e., eliminating traits 1, 12, 17, and 18). On herbarium specimens the resolution of certain qualitative traits, such as flower color and degree of pubescence, may be open to inteipretation if specimens are poor- ly preserved or mishandled after collection. Since many of the available Mimulus collec- tions were old or otherwise less than optimal for a moiphometric study, the potential existed for errors in judgment of qualitative traits to bias the analysis. As an alternative, DA was utilized to determine whether the elimination of diagnostic qualitative traits would result in a weaker phenetic relationship than that indi- cated by PCA. The multivariate statistical package'in STATGRAPHICS (v. 4.0) was used for the two analyses. PCA clustered the 38 populations into three well-defined groups conforming to a priori determinations of samples as M. evan- escens, M. latidens, or M. breviflorus (Fig. 2). The first two principal components accounted for 88% of total variance (Table 2), indicating that the PCA scatterplot (Fig. 2) is a good gauge of overall morphological differences among the three species. DA resulted in a comparable pattern, although M. evanescens clustered somewhat closer to M. latidens when qualitative characters were excluded (Fig. 2). In both analyses, M. evanescens is clearly and consistently intermediate to M. latidens and M. breviflorus. Upon first inspection Mimulus evanescens appears to be merely a robust version of M. breviflorus. The yellow, nearly regular corollas, essentially nonstipitate capsules, and short- puberulent foliage and stems are traits that are virtually identical in the two species. Since M. evanescens also develops papery, inflated fruit- ing calyces, which is the most prominent fea- ture in most specimens of M. breviflorus, it is understandable that the identity of the new species has been obscured. However, the over- sized habit of M. evanescens is striking, and all floral and vegetative characteristics average larger than in M. breviflorus. The significance of these proportional dif- ferences was first noted when fresh material of M. evanescens from the type locality was com- pared with greenhouse-grown plants of M. bre- viflorus. Although many individuals of the new species suffered from insect predation in the field and others appeared underdeveloped due to drought, undamaged plants from moist microsites commonly grew to 2 dm or more, far exceeding the largest examples of M. brevi- florus. Conversely, M. breviflorus plants culti- vated in the greenhouse (originating from three distinct populations in eastern Oregon) never exceeded 12 cm in height. Rather than grow- ing taller wdth age, they tended to branch out and become unusually floriferous. This obser- vation was confirmed when plants of M. brevi- florus and M. evanescens (32 and 27 individu- als, respectively, from populations in Lake Co., OR) were grown together from seed in a common greenhouse environment. Given iden- tical conditions, all M. evanescens plants grew to over twice the size of M. breviflorus. In addition, all quantitative and qualitative differ- ences for the species originally noted on the 254 Great Basin Natur\list [Volume 55 Principal Components Analysis Discriminant Analysis CO d OJ c o Q. E o O "cd g. o c c o o 4 . 2 Mimulus evanescens A ■ o 0 • • 2 Mimulus breviflorus go Mimulus latidens O 4 1 -5 0 5 First Principal Component (73.9%) u Q ■D c O O CU GO u ■ A 4 A Mimulus evanescens f ? A A A Mimulus breviflorus \ • O 0 V / A° 2 •\ Mimulus latidens 8*8 -10 -5 0 5 First Discriminant Function 10 Fig. 2. Morphometric evaluations (see te.xt for discussion); hvo-dimensional plots depicting principal components (18 qualitative and quantitative characters) and discriminant analyses (14 quantitative characters), contrasting the moipho- logical relationships of A/, evanescens (triangles), M. latidens (open circles), and M. breviflorus (closed circles). herbarium specimens were maintained in cul- ture. Although some herbarium specimens of M. evanescens are not particularly large, it is suspected that this is due to moisture limita- tion rather than genetic potential, based on obsei'vations of living plants. In addition to the overall size disparity, other features readily separate Mimulus evanescens from M. breviflorus. Most evident are the leaves, which are ovate to broadly lanceolate in the new species and rhombic-ovate or nar- rowly lanceolate in M. breviflorus. Moreover, only the very lowest leaves of M. evanescens are petiolate, and these abruptly so (Fig. lA), while leaf blades of M. breviflorus nanow grad- ually into slender, evident petioles at all nodes. In fruit, the pedicels of M. breviflorus general- ly exceed or at least equal the leaf blades. Those of the new species are always shorter in well-developed plants, and in some instances the leaf blade exceeds the calyx as well. Finally, the fruiting calyx of M. evanescens is much more plicate and typically exceeds the length of the mature capsule by 1.5-3.0 mm (Figs. 1B,C). In M. breviflorus ripe capsules are approximately the same length as the calyx. The overall dimensions of Mimulus evanescens, as well as the strongly angled fruiting calyx and broad, sessile leaves, are traits that also imply a relationship with M. latidens. There are sub- stantial differences between these taxa, however, including flower color, pedicel length in rela- tion to leaf length, stem pubescence, capsule insertion (Figs. 1C,D), and the presence of a basal rosette in M. latidens. Dissimilarities among the three species are summarized in Table 3. Table 2. Amount of total variance accounted for by each principal component, in a principal components analysis of moiphological variation among populations of Mimulus evanescens. M. breviflorus, and M. latidens. Component Percent of Ciuuulatixe ninnher variance percentage 1 73.88 73.88 2 14.18 88.06 3 3.75 91.81 4 2.41 94.22 5 2.12 96..34 6 1.22 97..56 7 .,57 98.13 8 .46 98.59 9 .39 98.98 10 .29 99.27 11 .21 99.48 12 .17 99.65 13 .12 99.78 14 .11 99.89 1.5 .06 99.95 16 .05 100.00 1995] New Mimulus From Idaho, Oregon and California Table 3. Diagnostic features oi Mimulus evanescens, M. breviflorus, and M. latidens. 255 Character M. evanescens M. breviflorus M. latidens Plant height (ft-) 10-25 cm 3_10(-14) cm 10-26 cm Pubescence Glandiilar-puberulent Glandular-puberulent Subglalirous Basal rosette No No Yes Leaf base Petiolate at base, sessile above Petiolate throughout Petiolate at base, sessile above Leaf blade shape Ovate to lanceolate Elliptic-lanceolate Broadly ovate Cauline leaf length 1.0-3.8 cm 0.4-1.7 cm 0.8-3.2 cm width 0.7-2.9 cm 0.2-0.5(-0.8) cm 0.5-1.7 cm Pedicels 0.8-1.8 cm long, < the blades 0.5-1.9 cm long, > the blades 1.0-3.3 cm long, > the blades Fruiting calyx length 7-11 mm 4-8 mm 9-12 mm width 5.0-8.5 mm 3.0-4.5 mm 6-8 mm Corolla color Yellow Yellow Whitish, shaded rose or yellow Corolla length 4.0-9.5 mm 3.5-5.5 mm 9.0-11.5 mm Capsule insertion ± sessile ± sessile clearly stipitate Capsule length 4.8-9.0 mm, clearly inserted 4.5-8.0 mm, about equaling calyx 6.0-9.0 mm, clearly inserted Distribution Great Basin and vicinity Great Basin and vicinitA' Cismontane California Elevation -1200-1700 m 300-2900 m <800 m Other small-flowered annuals that might be confused with Mimulus evanescens are primar- ily members of the M. moschatus complex, particularly M. floiibundus, M. patulus, and M. pulsiferae. Of these, only M. floiibundus is ever characterized as having an inflated fruit- ing calyx (Grant 1924), which can be distin- guished from M. evanescens by the multicellu- lar pubescence throughout and narrow, lance- olate sepals. These three species are further differentiated from M. evanescens by distinctly petiolate upper leaves and bilabiate corollas. Depauperate annual forms of M. guttatus also occur in moist sites within the range of M. evanescens. This common yellow-flowered species can be separated by petiolate upper leaves and strongly zygomorphic corollas. Although the calyces of M. guttatus are also markedly inflated, they are distinctly irregular and oriented horizontally in fruit. Mimulus suksdorfii is the only other annual monkey- flower in the Pacific Northwest with features comparable to M. evanescens. Seldom exceeding 6 cm in height, this compact, freely branched species is easily distinguished by obtuse, linear- oblong leaves, a cylindrical fruiting calyx, and flaring, emarginate corolla lobes. Phylogenetic considerations. — Judging from moiphology, Mimulus evanescens appears most closely related to M. breviflorus and M. latidens, and exhibits characteristics of both taxa (Fig. 2). Mimulus latidens, in turn, also seems to have a strong affinity to M. incon- spicuus, M. graiji, and M. acutidens from Cali- fornia, based primarily on flower color, stipi- tate capsules, calyx morphology, leaf shape, 256 Great Basin Natufl\list [Volume 55 and glabrous habit (Grant 1924, Tlionipson 1993). Aside from general \egetati\e and floial similarities, the inflated, plieate fruiting eal>'x is the principal trait linking these six species together. Whether or not this featui-e implies a monoplnietie group is open to debate, however, since inflated calyces have evidently arisen independently in Mimulus on more than one occasion. Nonetheless, the shape and texture of the calyces of these species are distinctive. The recognition of Mimulus evanescens allows for a reevaluation of the relationship between M. hrcviflorus and the rest of the genus. The morphology of M. evanescens, transitional between M. breviflorus and M. latidens, suggests that the new species might have arisen through hybridization. However, this hypothesis conflicts with the current geo- graphical and ecological separation of the putative parents and the fact that M. breviflorus is highly autogamous. An alternative scenario proposes M. evanescens as a descendant of M. latidens. The smaller-flowered and apparentK' more successful M. breviflorus (based on the number of historic collections) may have then arisen from M. evanescens, perhaps as a result of a shift to more xeric conditions in what is now the Great Basin. Mimulus breviflorus is ubiqui- tous and well represented in herbaria while M. evanescens is apparently rare and widely scattered, providing circumstantial support for this concept. The discoveiy of the disjunct M. latidens population in Lake Co., OR (Shelly 1986) is intriguing, because it suggests a mech- anism by which this relationship might have developed. If genotypes of M. latidens capable of survival outside of California's relatively benign Central Valley have been historically transported to the Great Basin by migrating ducks or geese, the means and opportunity for adaptive radiation could have existed. Conservation. — It is not encouraging that only 10 extant or historical populations of Mimulus evanescens are known, with only two sites recorded since 1958. This contrasts with hundreds of collections at dozens of localities for the much less conspicuous M. breviflorus. As with M. breviflorus, the distribution of M. evanescens is apparently limited to damp or wet sites at moderate elevations within open rangeland. Virtually all such sites in the Great Basin are associated with a long history of grazing by domestic livestock. The broad geo- graphic range and relatively unremarkable habitat of Miiiuilus evanescens impK that the compaiativc rarit)' of the species may be the result of habitat loss or disturbance. However, the paucity of herbarium records, especially when contrasted with similar species, suggests that M. evanescens may have never been com- mon, even under pristine, pre-grazing condi- tions. If this is true, the combination of natural scarcity with contemporary grazing or other disturbances ma\ now be jeopardizing the species. As an initial step, M. evanescens should be added to federal and state lists oi candidate endangered species. Although confirmed from Idaho, Oregon, and California, it is expected that northern Nevada is also within the his- toric range of the species. Placing Af. evanescens on candidate lists will bring the species to the attention of land managers in these states and will help justify inventory and research, which may in turn ascertain diat the species is not par- ticularly rare and has merely been overlooked by collectors. However, until this is estab- lished it is piTident to consider the species ex- tremely vulneral:)le, with ample protection given to any sites occurring on public lands. Acknowledgments The author acknowledges field or green- house assistance provided by Thomas Kaye, Matthew Carlson, Steven Gisler, Lisa Lantz, Crista Chadwick, and Melissa Peterson. Line drawings were prepared by John Megahan. The manuscript was reviewed by Kenton Chambers, Robert Frenkel, Mary Barkworth, Teresa Magee, and Edward Guerrant. Financial or logistical support for this study was provided by the Oregon State University herbaria, the USDA (Fremont and Winema National Forests), and the Plant Consei-vation Biology Program of the Oregon Department of Agriculture. Staff of the following herbaria graciously lent specimens or othei'wise provided access to their collections: BRY, CAS, CU, DS, GH, ID, IDF JEPS, M, NY, ORE, OSC, P RM, RSA, UC, US, UTC, WILLU, WS, and VVTU. Literature Cited Argue, C. L. 19' of Cali- fornia Press, Berkeley and London. Shelly, J. S. 1986. Noteworthy collection o{ Mimiihis lati- dens. Madrofio 33: 151. Thompson, D. M. 1993. Mimulus. Pages 1037-1046 in J. C. Hickman, editor. The Jepson manual: higher plants of California. University of California Press, Berkeley and Los Angeles. Received 2 March 1994 Accepted 5 December 1994 Great Basin Nahiralist 55(3), © 1995, pp. 258-266 MORPHOLOGICAL AND HOST-SYMBIONT STUDIES OF TRICHODINA TENUIFORMIS AND APIOSOMA CAMPANULATUM INFESTING MOTTLED SCULPIN {COTTUS BAIRDI) FROM PROVO RIVER, UTAH YingQi' and Kicluird A. Heckmannl'2 Abstiuct. — TrichocliiKi teintijonnis Stein, 1979 and Apiosomci cciinpanitlatiiin Tiinoteev, 1962 were found on j^ills of mottled sculpin {Cotttis bairdi) from two locations in the Provo River, UT. They were studied by light and electron optics. Dimensions and morphology of the adhesive disc and denticles of T. tenuifonnis were differentiated from other Trichodma species. A. campanulatum was characterized by its spindle-shaped cell body. Fine features examined by scan- ning electron microscopy included body shape, pellicle, elements of the adhesive disc, aboral ciliaiy complex, and ado- ral ciliary spiral. Histopathological studies suggested that the organisms are ectocommensals. Ecological aspects of organism infestation between two areas were also investigated. This report establishes a new host and distribution record for these two species in mottled sculpin from the Provo River, UT. Key words: Trichodina tenuiformis, Apiosoma campanulatum, Cottus bairdi, morphology, host-syinbiont relationship, ecological aspects. Provo River High numbers of two ciliated protozoa, Trichodina and Apiosoma, were encountered on the gills of mottled sculpin {Cottus bairdi) during a study of ectoparasites of fishes from the Provo River Trichodina is a mobile ciliate belonging to the subclass Feritrichia, family Trichodinidae (Lorn and Dykova 1992). This protozoan has an adhesive disc characterized by very promi- nent and taxonomically significant denticles (Van As and Basson 1987). More than 140 species of Trichodina have been reported from wild, cultured, and laboratoiy fishes in many parts of the world (Rand 1993). Sessile peritrich ciliates of the genus Apio- soma (syn. GlossateUa) belong to the subclass Feritrichia, family Epistylididae (Lom and Dykova 1992). They are generally attached to fish by a scopula (Lom 1973). They have been largely neglected by fish parasitologists until recently, when more attention has been given to this group. Many species of these two ciliated protozoa have been investigated (Arthur and Margolis 1984, Cone and Odense 1987, Rand 1993); however, a detailed study on mottled sculpin has never been reported. Objectives of this study were to (1) incoiporate different levels of microscopy to study ciliate structure, (2) ob- sei-ve histopathological changes these protozoa may cause to the host, and (3) evaluate the sea- sonal infestation rate to provide ecological infor- mation for the listed ciliates and their host. Materials and Methods Studies were earned out in late summer and fall (August, October 1993), late winter and spring (March, May 1994). Water temperatures in the Frovo River ranged from 14 °C to 4°C and 6°C to 10 °C, respectively. One hundred si.xty sculpin were collected from two sites: one in the city of Frovo (Utah County) munici- pal area, the second in a relatively pristine region near the Jordanelle Reservoir (Wasatch County). Sculpin were collected using elec- trofishing, placed in buckets containing river water, transported to the laboratory, and exam- ined within 24 h after capture. For light microscopy, air-dried smears of gill filament scrapings were prepared from freshly killed fish and treated by Klein's diy sil- ver impregnation technique (Clark and Heck- mann 1984) to examine components of the adhesive disc. Other smears were prepared, fixed, air-dried, and stained with iron hema- toxylin (Carcia and Bruckner 1988) to obsen'e the position and structure of the macro- and micronuclei. Sections of infested gills from the spring sample were fixed, blocked, cut, and 'Ofpartmcnt ol' Zoology', Brigham Young University, Provo, UT 84602. ^Autlior to whom correspondence should be addressed. 258 1995] Ciliated Protozoa in Mottled Sculpin 259 stained with hematoxylin-eosin (Garcia and Bruckner 1988) for histopathological studies. For scanning electron microscopy, gills of freshly killed fish were fixed in 2% buffered glutaraldehyde, followed by repeated washes in a sodium cacodylate buffer and post-fixed in a 1% solution of osmium tetroxide. After that they were washed in the same buffer system. Specimens were dehydrated through a graded alcohol series and critical-point-dried and sputter-coated with gold for examination with a Joel-840 high-resolution scanning electron microscope. For transmission electron microscopy, after fixation and dehydration, gills were embedded in Spurr resin and sectioned with a glass knife. Each section was stained with lead citrate and examined with a Philip EM400 transmission electron microscope. Terminology and methods of measurement follow those given bv Lom (1958), Lorn and Dykova (1992), Wellborn (1967), Arthur and Margolis (1984). Measurements are in micro- meters (/xm) and are based on 30 specimens for each species from each of the four sam- pling periods; range is followed by the mean and ± standard deviation in parentheses. Results Morphology Trichodina temiifonnis Stein, 1979 Host. — Cottus hairdi (Pisces: Cottidae). LOCALITY'. — Provo River, Utah and Wasatch counties, Utah. Site of infestation. — Gill filaments. Light microscopy. — Body 39-53 (44.2 ± 4.0) dia (diameter). Adhesive disc 19-30 (26.3 ± 2.8) dia, surrounded by a border membrane 2—3 (2.5 ± 0.4) wide, with fine transverse stri- ae. Various-sized light forms present in center of adhesive disc when silver-impregnated. Denticular ring 13.5-20 (17.2 ± 1.8) dia, con- sisting of 20-26 (23.7 ± 1.3) denticles with 6-10 (7.8 ± 0.8) radial pins per denticle. Denticle with conical central portions 0.7-1 (0.99 ± 0.06) from which a thorn 2.5-4 (2.9 ± 0.4) extends externally with broadly rounded lobes, tapered slightly to a blunt tip and blade 2-3 (2.3 ± 0.3) attached to central region, some with rounded ends (Figs. 1, 2). Macronucleus horseshoe-shaped 27-48 (39 ± 5.7) dia and approximately 10 ^tm thick. Micronucleus in -Y position (Lom 1958) observed in six specimens, dimension 3x2 (Fig. 3). Scanning electron microscopy. — Body of T. tenuiformis circular in aboral view and aboral surface relatively flat (Fig. 4). Body bell- shaped or domed when viewed from the side (Fig. 5). The aboral ciliaiy complex consists of three distinct ciliary bands: the basal ciliary ring, locomotor ciliaiy wreath, and marginal ciliary ring. The basal ciliary ring, adjacent to the border membrane, has a single row of fine, distally tapering cilia 1-2 /xm long. Separated from the basal ciliary ring by the basal septum is the locomotor ciliaiy wreath, which is com- posed of numerous rows of well-developed, powerful cilia 2-3 fim long whose primary function is locomotion. The precise number of ciliaiy rows composing this wreath could not be ascertained. It is separated anteriorly from the marginal ciliary ring by a poorly developed anterior septum that is evident only when the aboral ciliary complex is uncovered by the velum. The marginal ciliaiy ring is difficult to distinguish from the locomotor ciliary wreath in T. tenniformis. The velum is a thick, well- developed structure covering the bases of the cilia of the aboral ciliaiy complex and separat- ing this complex from the adoral ciliar)' spiral (Figs. 5, 6). The adhesive disc has a smooth pellicular surface beneath which the outline of the den- ticles can be clearly seen. The disc is sur- rounded peripherally by a 2-)U,m-wide border membrane, which functions to seal the margin of the disc during adherence and contains fine vertical striae over its entire surface. These striae on the internal surface of the border membrane are the radial pins that give the membrane rigidity while retaining its ability to conform to the host's surface (Fig. 7). The adoral ciliature forms a counterclock- wise spiral of about 270°. The base of each cil- ium is inserted into a deep furrow and hidden from view when SEM is used (Fig. 8). Deposition of slides. — One slide (HWML 37721) of silver-impregnated specimens and another slide (HWML 37724) of iron-hema- toxylin-stained specimens are deposited in the Harold W. Manter Laboratory, University of Nebraska State Museum. The senior author has additional slides in her collection. 260 Great Basin Natur.\list [Volume 55 Figs. 1-3. Light micrographs ot Triclioclhm tcuuijurmis: 1-2. Silver-impregnated specimens showing body shape and arrangement of components of the adhesive disc. BM, border membrane; D, denticle; RR radial pins. Bar = 10 /im. 3. Iron-hemato.xylin-stained specimen showing horseshoe-shaped macronucleus (MA); arrow points to the micronucleus (MI). Bar = 10/Lim. Apiosoma campanulatum Timofeev, 1962 Host. — Cottus bairdi (Pisces: Cottidae). Locality. — Provo River, Utah and W^isatch counties, UT. Site of infestation. — Gill filaments. Light microscopy. — Body campanulate. Macronucleus round or slightly conical. Size of stained specimens 3L0-66.0 (47.8 ± 7.2) long by 25.0-45.0 (35.6 ± 4.2) wide. Macronucleus lLO-20.0 (15.6 ± 2.4). Micronucleus not ob- served (Fig. 9). Scanning electron microscopy. — The spindle-shaped body is the characteristic fea- ture of this species. Circular striations of pelli- cle conspicuous. Pellicle wrinkled into longi- tudinal furrows. Upper part of body bears the adoral zone, consisting of a tuft of 1-2-^tm-long cilia. Most specimens viewed with SEM have contracted peristomes and contracted peristo- mial lips (Fig. 10). Deposition of slides. — A representative slide of Apiosoma campanulatum (silver stain) is deposited in the Harold W. Manter Labora- tory, University of Nebraska State Museum (HWML 37722). The senior author has addi- tional slides in her collection. 1995] Ciliated Protozoa in Mottled Sculpin 261 Figs. -J-7. Scaumug electron iiueiugraplis of the surface of I temiiformis: 4. Aboral view of entire specnnen of I tenuifonnis. B, bacteria. Bar = 10 fxm. 5. Lateral view of entire specimen of I temiifonnis. ACS, adoral ciliarv spiral; BS, basal septum; BCR, basil ciliaiy ring; GE, gill epithelium; LCW, locomotor ciliary weath; VEL, velum. Bar = 10 fim. 6. Higher magnification of Figure 4 showing the structure of aboral ciliaiy complex. ACS, adoral ciliary spiral; BCR, basal ciliary ring; K, kinetosomes; LCW, locomotor ciliary wreath; MCR, marginal cilian,- ring. Bar = 1 ^tm. 7. Adhesive disc of T tenuifonnis. D, denticle; PR peripheral pins; RR radial pins. Bar = 1 ^tm. 262 Great Basin Natur.\list [Volume 55 Fig. 8. Adoral view of I tenuifonnis showing how the adoral cihatiire (ACS) forms a counterclockwise spiral of about 270°. Bar = 10 /xm. Fig. 9. Light micrograph of Apiosoma campanulatum. Note conicle-shaped body. MA, macronucle- us. Bar = 1 /xm. Fig. 10. Scanning electron micrograph of A. campanulatum attached to the gill epithelium (GE). Note transverse striations of pellicle and its longitudinal furrows (aiTOw). PL, peristomal lip; PS, peristome. Bar = 5^im. Fig. 11. Light micrographs showing Thchodina (T) and Apiosomu (A) infested gill epithelium. Bar = 20 /xm. Host-Symbiont Relationships Light microscopy. — Sections of mottled sculpin gills had no apparent pathological damage. The conical body of some A. campan- ulatum appeared to be attached to host gill surfaces by the scopula, while others were freely distributed over the epithelial surflice. Most T. tenuifonnis glide over the surface; only a few ciliates adhere to the host epithelial cells (Fig. 11). Transmission electron microscopy. — Sections of the interface between the host epithelial cell and T. tenuifonnis were pre- pared. No permanent or temporary structure could be detected between the adhesive disc, adoral zone of cilia, and gill epithelial cells (Fig. 12). However, injury to the epithelium due to 7^ tenuifonnis can be detected by the number of mitochondria, which decrease and disappear in the immediate host cell. Host necrotic tissue, mucous layers from gill epithe- lium, and particles dispersed in the water were on the surface of T! tenuifonnis (Fig. 13). No ultrastructural damage was observed for A. campanulatum. Presence of this ciliate in- flicts no serious damage to the host cell. There was some change in number of mitochondria, with cristae showing major changes (Fig. 14). Ecological Aspects of Infestation In the Provo River near the Provo residential area, T. tenuifonnis reached the highest infes- tation rate in April and May. It was uncommon during summer and autumn and appeared to be absent in the winter With the increase of water temperature in spring, ciliates reinfest- ed the fish. Apiosoma campanulatum at this site maintained an average of 35% infestation rate (no. of infested fish vs. no. of total exam- ined fish) for all seasons. In the upper Provo River the tendency of in- festation of T. tenuifonnis corresponded closely 1995] Ciliated Protozoa in Mottled Sculpin 263 ^^. Y 4^ ^m •••• fife' ^^ Figs. 12-13. Transmission electron micrographs of gill epithelium infested by T. teniiiformis. 12. Host necrotic tissue (arrows) sloughs off for parasite's food. ACC, aboral cilian' complex. Bar - 1 /j,m. 13. Interface between T. teniiiformis and mucous layer (ML) of epithelial cells. Note damage to mitochondria (M). C, cilia. Bar — 1 fim. 264 Great Basin Naturalist [Volume 55 14 y^^;:^^!^^ \- Nu ''«^' iy ^/fon^t. =^!-, \ r Fig. 14. Transmission electron micrograph of gill epithelium infested hy A. campanulatum. A. campanulatwn (A) caus- es number of mitochondria (M) to decrease and cristae to disappear. M, mitonchondria; Nu, nucleus of epithelial cell. Bar = Ifim. to that of the lower area. The highest infesta- tion rate occurred in May and then decreased until the next spring. Percentage of fish infest- ed by T. tenuiformis in the lower river area was 20.5% vs. 12.5% in the upper Provo River. Similar to that of the lower river, A. campanu- latum at the upper site had an average of 37% infestation in all four seasons. In general, Apio- soma did not show measurable fluctuations with seasons. Discussion Taxonomy and host-symbiont studies of Trichodina and Apiosoma infesting fishes in the United States have received surprisingly little attention considering the frequency with which these organisms have been associated with fish diseases (Khan et al. 1974, Cone and Odense 1987, Khan 1991). Wellborn (1967) described 13 species of Trichodina in south- eastern United States, but few reports have been published for this ciliate west of the Mississippi River (Hechmann et al. 1987). Little information is available on Apiosoma studies in this country, which is not the case in the former Soviet Union (Bauer 1984). Cottus hairdi represents a new host record for Trichodina tenuiformis and Apiosoma campanulatum. Comparative Morphology At the LM level comparison of the adhesive disc of r tenuiformis with that of other species oi Trichodina reveals a few similarities. Tricho- dina reticulata Hirschmann and Partsch, 1955 described from Carassius auratus has denti- cles similar to T tenuiformis (Bauer 1984). The adhesive disc of the former has a central light zone separated into reticulated structures. But T. reticulata differs in having larger overall dimensions (average adhesive disc diameter is 60 /xm vs. 25^tm for our material). T tenuiformis has a close affinity to T elegans described by Stein (1979) from fish in Russia. The latter is characterized by an unbroken light zone in the adhesive disc. Our specimens have various- sized light forms in the center of the adhesive disc. To a lesser extent T tenuiformis is similar to T puijtoraci Lom, 1962 and T domerguei Dogel, 1940; however, denticle shape and structure of the adhesive disc clearly distin- guish T tenuiformis from these species. Surface features of the adhesive disc and arrangement of the aboral ciliaiy complex of T 1995] Ciliated Protozoa in Mottled Sculpin 265 tenuifonnis seen by SEM were generally simi- lar to those described for T! truttae, an ecto- parasite on pacific salmon {Oncorliyncluis spp.) and steelhead trout {OncoHiynclius inykiss; Arthur and Margolis 1984), and T. labrisomi, an ectoparasite on hairy blenny {Labrisomas nuchipinnis; Rand 1993). However, in T. tenui- fonnis, aboral cilia length is generally shorter than in those previously described. Further- more, comparison of the aboral ciliature of T. tenuifonnis with these species of Trichodina showed some differences in the extent of development of the anterior and basal septa, in velum structure, and in the degree of evi- dence of the marginal ciliary ring. The anteri- or septum is relatively large and the basal one is small in T. tnittae, whereas in T. tenuifonnis the basal septvmi is prominent. The velum is well developed in both T. labrisomi and T. tenuifonnis, but T. tenuifonnis lacks any protu- berances (Rand 1993). Similar to T. labrisomi, the marginal ciliary ring of T. tenuijormis is poorly developed and cannot be distinguished from the locomotor ciliaiy ring, whereas in T. truttae the marginal ciliary ring is well devel- oped (Arthur and Margolis 1984). Rand (1993) has suggested these marginal ciliature are sen- sory structures associated with feeding and orientation. Unlike T! labrisomi and T. truttae, T. tenuifonnis has no pellicular pores between denticles and the pellicular ridges on the oral surface, which might be a species-specific characteristic for these two species respective- ly (Arthur and Margolis 1984, Rand 1993). Over 50 species of Apiosoma have been recorded from fishes, the majority of which have been described by Russian authors (Bauer 1984). Although some are common fish para- sites in some parts of the world, only one ref- erence concerning Apiosoma piscicola on Salve- linus fontinalis was reported in North America (Cone and Odense 1987). There is a paucity of data pertaining to Apiosoma over the last two decades, likely reflecting taxonomic difficul- ties due to variability in ciliaiy structure and lack of strict host-specificity. Apiosoma conica has a body shape similar to A. campanulatum. But our specimens com- pared more closely to the original description of A. campanulatum. The species identifications were based on original descriptions from Europe; there is a possibility that the two species described in this content are not absolutely identical on both continents. Host-Symbiont Relationships Trichodina tenuifonnis is an ectocommensal with a tendency to be parasitic in mottled sculpin. There were no visible pathological symptoms with light microscopy; however, electron microscopy disclosed changes in the organelles of host epithelial cells infested by T. tenuifonnis. Mitochondria decreased in num- ber and disappeared, which might indicate respiratory blockage due to lack of oxygen. This change in mitochondria was obsei-ved in Trichophrya infesting other fish (Heckmann and Carroll 1985). Necrotic host epithelial tis- sue sloughs off following organelle loss, sup- plying sustenance for the parasite. No serious damage to mottled sculpin could be obsei-ved for A. campanulatum. Lom (1973) suggested that this simple ectocommensal relationship could change to parasitism in case of heavy invasions, although this tendency is much less pronounced than in trichodinids. Ecological Aspects of Infestation This study shows that the infestation of Tri- chodina has both seasonal and regional fluctu- ations. The higher infestation rate on fish came from the Provo residential area during the spring sampling period. Heavy impact from the local human population may contribute to this infestation. After summer, the number of T. tenuiformis gradually reduces with the decrease of water temperature and reaches the highest number the following spring. This may be related to the ciliate life cycle. Unlike T. tenuifonnis in this study, A. cam- panulatum maintained a fairly constant infes- tation on mottled sculpin from the two sites on the Provo River in all four seasons. Acknowledgments The authors thank the Utah Fish and Game Department for their cooperation on this study. Dr. Dennis K. Shiozowa provided help with fish collections. Technical assistance from Dr John Gardner and staff members of the Elec- tron Optics Laboratoiy, Brigham Young Univer- sity, was greatly appreciated. Literature Cited Arthur, J. R., and L. Margolis. 1984. Trichodina truttae Mueller, 1937 (Ciliophora: Peritriehida), a common pathogenic ectoparasite of cultured ju\ enile salmonid fishes in British Columbia: redescription and exami- 266 Great Basin Naturalist [Volume 55 nation I)y scanning electron microscop\'. C-'aiiadiau Journal ofZooloj,^ 62: 1842-1848. Bauer, N., EDITOH. 1984. Parasitic protozoa of freshwater fislies in the USSR. Volume 1. Academy of Sciences and Zoological Institute, St. Petersburg. 4.30 pp. ClAHK G. W., AND R. A. Hf.c;kni.\nn. 1984. An atlas of ani- mal parasites. Brigliam Young University' Press, Proxo, UT 218 pp. Co.\E, D. K., AND R M. Odense. 1987. Occurrence of Het- eropolaria hvoffi (F"aure-Fremiet, 1943) and Apio- sorna piscicola Blanchard, 188.5 (ciliata) on Salvelinus fontinaUs (Mitchill) in Nova Scotia, Canada. Cana- dian Journal of Zoology 65: 2426-2429. Garcia, L. S., and D. A. Bruckner. 1988. Diagnostic medical parasitology. Elsevier Science Publishing Co., Inc., New York. 500 pp. Heckm.wn, R. A., and T. Carroll. 1985. Host-parasite studies of Trichophnja infesting cutthroat trout {Salmo clarki) and longnose suckers {Catostotniis catostomtis) from Yellowstone Lake, Wyoming. Great Basin Naturalist 45: 255-265. Heck.\l\nn, R. a., a. K. Kimball, and J. A. Short. 1987. Parasites of mottled sculpin, Cottus bairdi Girard, from five locations in Utah and Wasatch comities, Utah. Great Basin Naturalist 47: 1.3-21. Khan, R. A. 1991. Mortality in Atlantic salmon associated with trichodinid ciliates. Journal of Wildlife Diseases 27: 153-155. Khan, R. A., V C. Barber, and S. McCann. 1974. A scan- ning electron microscopical study of the surface topograpin of a tricliodinid ciliate. Transactions of the American Microscopical Society 93: 131-134. LoM, J. 1958. A contriliution to the systematics and mor- phology of endoparasitic trichodinids from amphib- ians, with a proposal of uniform species characteris- tics. Journal of Protozoology 5: 251-263. . 1973. The mode of attachment and relation to the host in Apiosoma piscicola Blanchard and Epistylis hcojfi Faure-Fremiet, ectocommensals of freshwater fish. Folia Parasitologia (Prague) 20: 105-112. LoM, J., AND I. Dykova. 1992. Protozoan parasites of fishes. Elsevier Science Publishers B. V., New York. 316 pp. R^ND, T. G. 1993. Light and scanning electron microscopic studies on Tnclwdina labrisomi n. sp. from Labrisomas niichipinnis (Osteichthyes: Labrisomidae) from Mangrove Lake, Bermuda. Canadian Jomnal of Zool- ogy 71: 1855-1860. Stein, G. A. 1979. Variability of the ciliates of the famiK Urceolariidae (Peritricha, Mobilina) in the Baikal Lake. Journal of Protozoology 26 (No. 3, Part 1, Programs and Abstracts): 36A-37A (abstract). Van As, J. G., and I. Basson. 1987. Host specificity of tri- chodinid ectoparasites of fi^eshwater fish. Parasitology Today 3: 88-90. Wellborn, T. L. 1967. Thchodina (Ciliata: Urceolariidae) of freshwater fishes of the southeastern United States. Journal of Protozoolog>' 14: 399-412. Received 13 Janiian/ 1995 Accep1ed'l2 April 1995 Great Basin Naturalist 55(3), © 1995. pp. 267-270 EFFECTS OF HORSE GRAZING IN SPRING ON SURVIVAL, RECRUITMENT, AND WINTER INJURY DAMAGE OF SHRUBS Dennis D. Austin^ and Philip J. Urness^ Abstract. — The use of domestic grazers to shift the growth advantage toward shrubs is a commonly applied tool on winter ranges managed primarily for big game. Results from horses grazing in spring indicated grazing also benefits shrub survival, seedling reciaiitment, and reduced winter injuiy damage on some species of shrubs. Key words: winter range, range management, mule deer horses, shrubs, browse, Utah, revegetation. n^ountain big sagebrush, Douglas rabhitbrush, true mountain mahogany. On winter ranges managed primarily for big game, the management alternative often selected to maintain the desired mixture of shrubs and understory herbage is grazing by livestock in spring. Numerous studies have re- ported the benefits of spring livestock grazing to maintain and improve stands of shrubs on winter ranges (Christensen and Johnson 1964, Smith and Doell 1968, Jensen et al. 1972, Hull and Hull 1974, Reiner and Umess 1982, Austin et al. 1994, and others). However, information is limited concerning shrub responses to the effects of livestock grazing with respect to (1) survival of individual mature plants, (2) seed- ling recruitment, and (3) winter injuiy damage. In this study the responses of shrubs to domestic horse grazing treatments in spring are reported for Artemisia thdentota var. vaseyana [Rydb.] Beetle (mountain big sagebrush), Chrysothcun- nits viscidiflonis [Hook.] Nutt. (Douglas rabbit- brush), and Cercocarpus montamis Raf (true mountain mahogany). Methods The studv site, located on the foothills east of Logan, UX 4r46' N latitude, lir47' W longitude, at 1600 m elevation, contained three 50 X 50-m adjoining paddocks. Within each paddock the three browse species were hand- planted from transplants in spring 1983 in 5 X 5 clusters of 25 plants, with 1 m between plant centers. Seven clusters were planted in each paddock, with each cluster separated by a min- imum of 20 m. Before planting, all vegetation was removed by root plowing; for two growing seasons following planting, all seedlings were removed by hand and rototiller weeding. Between 1983 and 1987 the three paddocks received equal use by mule deer {Odocoileus honionus) in winter and no livestock grazing. A detailed description of the site is found in Olsen-Rutz and Urness (1987). This study was conducted during the six growing seasons between spring 1987 and fall 1992. In spring 1987 all shrub seedhngs that had become established from seeds were re- moved from each paddock by hand pulling to minimize soil disturbance. The number of sur- viving, previously transplanted shrubs within each cluster was counted. Paddocks were randomly assigned a grazing treatment by horses as heavy, moderate, or protected. Three to seven horses were used, de- pending upon herbage production, to obtain utilization levels of 35-50% and 65-80% for moderate and heavy treatments, respectively. Horses were selected as grazers because of tlieir high foraging selectivity for grasses and avoid- ance of shrubs, and the managerial opportunity to manipulate the herbaceous understory to improve shrub growing conditions (Reiner and Umess 1982). Paddocks were grazed yearly be- tween 1 May and 30 June 1987-1991. In 1992 all paddocks were rested from grazing. In the moderately and heavily grazed paddocks, herbage production, comprised almost entire- ly of annual grasses, and percent utilization were determined from four paired 1-m^ bas- keted and unprotected plots, randomly placed in spaces between clusters. Baskets were con- structed fiom 1.2-m-high netting wire supported iRangeland Resources, Utah State Universit>', Logan, UT 84322-5230. 267 268 Great Basin Naturalist [Volume 55 by steel fence posts. Plots were reestablished and relocated yearly before grazing. In fiill 1992 all surviving shrubs were counted by cluster, all seedlings within 10 m of each cluster were counted, and percent winter injui") damage was visually estimated. Winter injury was defined as the amount of dead stems and twigs as a percentage of total dead plus live stems and twigs. Damage was estimated at five- unit increments from 0 to 95%. Because we were not able to replicate the three paddocks established for the previous study, we considered clusters as experimental units. We agree with Hurlbert (1984), who described this experimental design as "simple pseudoreplication," but because of constraints of time, space, and costs, this design was the only option. Consequently, we recognize that differences between treatments could be caused by inherent differences between paddocks, but argue that potential spatial error is low due to adjoining paddocks, identical use during the three years preceding our experiment, simple grazing treatments applied, and lack of differ- ences in the number of surviving seedlings among paddocks for each species (P > .10) at the beginning of the experiment. T tests of the means were used to deter- mine differences between grazed and protect- ed plots within and among paddocks. For plant survival a split-plot design using repeated mea- sures (1987 and 1992 data) analysis of variance determined treatment and year effects. One- way ANOVAs assessed differences among treatments for species within years. For seed- ling reciTjitment, because all seedlings were re- moved in 1987, one-way ANOVAs were used for species within years. For winter injury damage, differences between treatments were analyzed using chi-square tests. A significance level of P < .10 was used for all tests. Results and Discussion Horse Grazing Horse use reduced herbage at the end of the grazing period during all years in both the moderately and heavily grazed treatments (P < .10), except in 1987 when neither treatment was different from protected plots. Mean herb- age utilization during all years was 46% in the moderately grazed treatment and 71% in the heavily grazed treatment. Following grazing, remaining herbage was different between the moderately and heavily grazed treatments din- ing all years except 1987. Herbage production in protected plots was not different from the moderately and hea\'ily grazed treatments dur- ing the first three years. However, the heavily grazed treatment had lower production during the last two years, suggesting that heavy graz- ing by horses reduced production of herbage. Shrub Sunival Horse grazing increased suwival oil Artemisia (P = .01) and Cercocarpus (P = .10) but had no effect on Chnjsothamnus (Table 1). All three species declined in numbers between 1987 and 1992 (P = .001). In 1987 the number of surviving plants among treatments for Aiionisia, ChrysotJiam- mis, and Cercocarpus was not different (Table 1). However, in 1992 the number of sui-viving Artemisia plants among treatments was differ- ent (P = .005). The protected treatment had lower survival than both the moderately and heavily grazed treatments (P = .001), but the moderately and heavily grazed treatments were not different. Similarly, for Cercocarpus the number of sui^viving plants among treatments was different (P = .03). The protected treat- ment had lower sui-vixal than both the heavily (P = .005) and moderately (P = .10) grazed treatments, but the moderately and heavily grazed treatments were not different. For Chrysothammis, no differences were found. Seedling Recruitment For Artemisia, seedling recruitment was sig- nificantly different among treatments (P = .08). The heavily grazed treatment had more seed- lings than the protected and moderately grazed treatments (P = .05). No differences among treatments were found for Chnjsothamnus, and no seedlings were counted for Cercocar- pus (Table 1). Although the low numbers of seedlings counted in this study require inteipretive cau- tion, results are consistent with other studies in which livestock grazing was reported to in- crease shiiib densitv (Stewart 1941, Christensen and Johnson 1964, Hull and Hull 1974). Furthermore, the results from this study, that horse grazing in spring resulted in higher sur- vival of mature plants and increased seedling establishment for several species of shrubs, are consistent with reports of increased pro- duction of shrubs following livestock grazing 1995] Shrub Responses to Grazing 269 Table 1. Plant sundval (total niiinber/paddock), seedling recruitment (total niimber/paddock), and winter injin-\' dam- age (mean % per shmb) o{ Artemisia tridentata (ARTR), Chnjsothainniis viscidiflorus (CHVI), and Cercocarpus montanus (CEMO), as affected by heavy (H), moderate (M), and protected (P) horse grazing treatments'. Species Treatment 1987 1992 Shrub suni\al ARTR CHVI CEMO Seedling recnjitment^ ARTR CHVI Winter injur}' damage'^ CEMO Number/paddock 119 91« 128 93^' 120 42'' 125 110 86 78 106 101 164 158" 161 140ab 168 119'^ Number/paddock 18" — 5'' — 5h 2 — 2 — 3 Mean % per shiTib 11" — 241' — 41c ^Data with different superscripted, lowercase letters witliin year and species were different at P < .05. ^No seedlings of CEMO were found. ''No winter injury damage on ARTR or CHVI was found. with horses (Reiner and Urness 1982, Austin et al. 1994), sheep (Jensen et al. 1972), cattle (Smith and Doell 1968), or goats (Riggs and Urness 1989). Winter Injury Winter injuiy was not found on either Artem- isia or Chrysothamntis (Table 1). For Cercocar- pus, winter injury among treatments was dif- ferent (F = .001), with highest damage occur- ring on the protected treatment, medium on the moderately grazed treatment, and lowest dam- age on the heavily grazed treatment. All treat- ments were different fi-om each other (P = .001). Winter injury has been reported for many shrub species, including Cercocarpus (Nelson and Tiernan 1983). However, only one known report compared winter injury to grazing. Contrary to our results, Jensen and Urness (1979) compared heavy (70%) and moderate (35%) levels of grazing of grasses and forbs by sheep and reported that injuiy to Purshia tri- dentata (antelope bitterbrush) was indepen- dent of grazing intensity or time of use. Summary Our results support the use of grazing by horses of herbaceous understoiy in spring to maintain and improve stands of browse for winter use by big game. Herbage production was reduced by heavy grazing, survival of mature plants of Artemisia and Cercocarpus was increased, recruitment oi Artemisia was increased, and winter injury to Cercocarpus was decreased. No negative effects on shrubs from grazing by horses were found. Acknowledgments This report is a contribution of the Utah State Division of Wildlife Resources, Pittman- Robertson, Federal Aid Project W-105-R. 270 Great Basin Naturalist [Volume 55 Literature Cited Austin, D. D., F. J. Urness, and S. L. Di hiiam. 1994. Impacts of mule deer and horse grazing on trans- planted shrubs for revcgetation. Journal of Range Management 47: 8-11. CHR1STEN.SEN, E. M., AND H. B. JoHNSON. 1964. Presettle- ment vegetation and vegetational change in three valleys in central Utah. Brigham Young University Science Bulletin, Biological Series IV(4): 1-15. Hull, A. C, Jr., and M. K. Hull. 1974. Presettlement vegetation of Cache Valley, Utah and Idaho. Journal of Range Management 27: 27-29. HURLHERT, S. H. 1984. Pseudoreplication and the design of ecological field e.xperiments. Ecological Mono- graphs 54: 187-211. Jensen, C. H., A. D. Smith, .\nd G. W. Scotter. 1972. Guidelines for grazing sheep on rangelands used by big game in winter. Journal of Range Management 25: 346-352. Jensen, C. H., and P J. Urness. 1979. Winter cold damage to bitterbnish related to spring sheep grazing. Joimial of Range Management 32: 214-216. Nelson, D. L., and C. E Tiernan. 1983. Winter injurv- of sagebrush and other wildland shiiibs in the western United States. USDA Intermountain Eorest and Range E.xperiment Station Research Paper INT-314. Olsen-Rutz, K. M., and R J. Urness. 1987. Comparability of foraging behavior and diet selection of tractable and wild mule deer. Utah Division of Wildlife Resources Publication No. 88-3. Reiner, R. J., and P J. Urness. 1982. Effects of grazing horses managed as manipulators of big game winter range. JoiuTial of Range Management 35: 567-571. RiGGS, R. A., and P J. Urness. 1989. Effects of goat brows- ing on Gambel oak communities in northern Utah. Journal of Range Management 42: 354—360. Smith, A. D., and D. Doell. 1968. Guides to allocating forage between cattle and big game on winter range. Utah Division of Fish and Game Publication No. 68- 11. Stewart, G. 1941. Historic records bearing on agriculture and grazing ecology in Utah. Ecolog>' 39: 362-375. Received 7 Fehnuinj 1994 Accepted 24 January 1995 Great Basin Naturalist 55(3), © 1995, pp. 271-281 NORTH AMERICAN TYPES OF OXYTROPIS DC. (LEGUMINOSAE) AT THE NATURAL HISTORY MUSEUM AND ROYAL BOTANIC GARDEN, ENGLAND, WITH NOMENCLATURAL COMMENTS AND A NEW VARIETY S. L. Welshi Abstract. — Specimens of Oxytropis in the herbaria of The Natural Histoiy Museum and Royal Botanic Garden were examined to interpret their role in nomenclature. This is the first attempt at a systematic ovei'view of specimens so important in understanding the genus as it occurs in North America. The review of specimens at BM and K during the present research has resulted in realignment of names of some of the ta.\a. Oxytropis cainpestris var. gracilis (A. Nelson) Barneby is recognized herein as being predated by O. campestris var. spicata Hook., O. sericea var. spicata (Hook.) Barneby is replaced by O. sericea var. speciosa (Torn & A. Gray) Welsh comb, nov., and O. campestris var. terrae-novae (Fern.) Barneby is superseded by O. campestris var. minor (Hook.) Welsh comb. nov. One new taxon is proposed; Oxytropis deflexa (Pall.) DC. var. pidcherrima Welsh & A. Huber, var. nov. Lectotypes are designated for the following taxa: Astragalus retroflcxus Pall., Oxytropis arctica R. Br., O. arctica var a subumbellata Hook., O. arctica var. [J uniflora Hook., O. campestris var. P speciosa Torr. & A. Gray, O. campestris var. ^ melanocephala Hook., O. campestris var. 5 spi- cata Hook., O. multiceps var. minor A. Gray, O. splendens Douglas ex Hook., O. splcndens (3 richardsonii Hook., O. uralensis (3 subsucculenta Hook., and O. uralensis y minor Hook. Key words: Legiiminosae, Ox\'tropis, nomenclature. Natural History Museum, Royal Botanic Garden. Concepts of species within a genus undergo an evolutionary progression through time as additional information is obtained. A review of the history of botanical treatments of the genus Oxytropis parallels that of other genera in North America, wherein the early explora- tions, researches, and publications were under- taken by explorers and scientists from the Old World, especially from England. Specimens arriving from the New World were compared to those of the same genera from Eurasia. Specimens of Oxijtropis from Russia and other regions with arctic, subarctic, or boreal floras arrived piecemeal at herbaria in Europe, where important collections accumulated, par- ticularly at the Royal Botanical Garden at Kew (K) and the Natural History Museum (BM; fomierly die British Museum [Natural Histoiy]) in London. These materials formed the basis for comparison with North American speci- mens. Some American plants were similar and were given the same names as some Old World species. The earliest revision of Oxy- tropis based on NoiiJi American specimens was that of William Jackson Hooker (1785-1865), whose concepts of species, set forth in the Flora horeali-americana (Hooker 1831), were to have a profound effect on all later inteipre- tations of the genus. Annotations of the speci- mens at K and BM present a history of the use of epithets by various botanists interested in this fascinating genus. However, of the revi- sionaiy workers on North American members of the genus Oxytropis, only Asa Gray appears to have systematically studied the historical collections at the Royal Botanical Garden, Kew, and no one has examined all materials of the genus in North America at the Natural History Museum. Various workers on regional floras, M. L. Fernald, A. E. Porsild, and N. Folunin, have annotated part of the specimens, and R. C. Barneby, whose revision (1952) is a classic presentation of the genus in North America, has examined selected material. The pui-poses of this paper are to clarify the status of historical specimens, to record their place of deposit, and to trace their nomencla- tural histoiy as it affects interpretation of the genus in North America. Plants at BM and at K are the center of focus for this treatment. However, the location of duplicate types in various herbaria in the United States is also included where that information is known. Names and synonyms of the North American 'Life Science Museum and Department of Botany and Life Science, Brigham Young University', Provo, UT 84602. 271 272 Great Basin Naturalist [Volume 55 taxa were reviewed previously by Welsh (1991). Abbreviations of the repositories follow Holmgren et al. (1990). The synopsis of Oxytropis in Flora horcali- amevicana (hereinafter Flora) by Hooker is relevant to an understanding of many of the names in the following list. This is true even though Hooker had not seen the materials in the field, and even though his concepts were based on limited and often inadequate materi- als for a definitive understanding of the taxa. The Flora has no introduction outlining the scope and rationale for the treatment. It does, however, contain a ven* detailed title page: Flora Boreali-AnieriL'ana; or, the Botany of the Northern Parts of British America: compiled prin- cipally from the plants collected by Dr Richardson & Mr Drummond on the late northern expedi- tions, under command of Captain Sir John Franklin, R.N. to which are added (by permission of the Horticultural Society of London,) those of Mr Douglas, from North-West America, and of other Naturalists. That Hooker does not mention the collections of Captain William Edward Parry and his associates is not to be considered an oversight; their materials had been published previously by Hooker (1825), and their specimens are by no means neglected in the Flora. It is unfortu- nate that the specimens on which the names were based were not routinely so noted by the authors. The dedication in the Flora by Hooker honors both Franklin and Richardson, under whose auspices, as Commander and Naturalist of two separate expeditions to The Polar Seas, a great portion of the more rare and interest- ing plants that ornament this volume were collect- ed, under circumstances of singular difficulty, hardship, and danger. In this important pioneer work. Hooker recog- nized only 10 species of Oxytropis but regard- ed several of them as consisting of variants designated by Greek letters, some of which were followed by epithets. The names in order of their appearance are O. horealis; O. Uralensis a, ^ siihsucciilenta, jfninor; O. arcti- ca a, |3 minor, 5 inflata; O. foliolosa; O. argen- tata; O. lambertii; O. nigrescens; O. carnpestris a, y siilphiirca, 6 spicata, £ glahrata, and t, melanocepliala; O. splendens a vestita and (3 richardsonii; and O. deflexa. These names have occurred in subsequent literature and have been accounted for by various workers of the genus, often without citation of type mate- rial or place of deposition. An attempt is made here to associate all names noted by Hooker with their 20th-century equivalents. Hooker's personal herbarium, containing many of the Oxytropis types, is at Royal Botanic Garden (K Hooker); those of Richardson, Douglas, Drummond, and Parry and his associates are represented in some part at both K and BM. Richardson (1823) cited species oi' Oxytropis based on his owai findings, but he did not name any as new. The species treated by him include O. oxyphylla, O. deflexa, O. carnpestris, O. argentata, and O. uralensis. Of these, only O. deflexa stands almost in the same sense today. The treatment of Oxytropis by Torrey and Gray (1838) followed Hooker's account in nearly all details, but it added six new species collected and described by Thomas Nuttall (1786-1859) on his journey across the conti- nent with Wyeth in 1834. These were the first American species described by a botanist who had seen them in the field. The diversity of the species in the American West must have seemed overwhelming even to Nuttall, who proposed several additional species of Oxy- tropis not published in Torrey and Gray's monumental work. Specimens in Nuttall's personal herbarium, which is deposited at BM, are evidence of his belief in a greater number of species than would be published subsequently. Some of the Nuttall names were later cited as synonyms, but some were not mentioned at all. Nuttall was unfamiliar with most meml)ers of the genus in the Old World, and some of his proposals reflect that lack of understanding. The difficulty in comprehend- ing a genus as large and complex as Oxytropis is understandable and is not confined to the 19th centuiy Asa Gray (1810-1888) became the most important 19th-century North American devo- tee to the genus, revising it twice, once in 1863 and again in 1884. Concepts in the revi- sion of Oxytropis by Gray (1884) were influ- enced by his examination in 1880 of specimens at Kew, which had formed the basis of the treatment by Hooker in die Flora, and fiom the large number of specimens in eastern American herbaria collected during the inten'ening se\'- eral decades. Gray's 1884 publication includ- ed plants from a broader geographic area than those examined by Hooker and contained descriptions of 16 species. Specimens collected 1995] North American Types of Oxytropis 273 by Nuttall, especially, and several other west- ern American explorers formed the basis of additional species not included in his and Hooker's earlier works. Gray accounted for some proposals, bringing the concepts of the genus to date as new materials had accrued. He accounted for some but not all taxa treated by Hooker, e.g., O. borealis DC. under O. leii- cantha (Pall.) Pers. The name O. Iciicantha (Pall.) Pers. was long considered a potential replacement for viscid members of the genus in North America. The type was examined by Welsh (1977) and the name excluded from inteipretation of North American taxa. It is a portion of the O. campestris complex in Siberia. Most of the other names treated by Hooker were ignored by Gray, cited in synonymy, or provisionally included in other entities. Apparently Gray did not see all pertinent his- torical material in England, particularly not that at BM. It is evident, likewise, that not all ambiguities are resolved by the research lead- ing to this paper. Nevertheless, as noted by Barneby (1952), "the resulting synopsis of the genus [by Gray] in 1884 stands as a small but enduring monument to his genius." Barneby (1952), in a classic account, recog- nized 22 species and 21 additional infraspecif- ic entities as occurring in North America. He accounted for all names used previously in North America, with problematical names being discussed in a list of excluded and imperfectly known species. A summary treatment of the genus by Welsh (1994) for the Flora North America pro- ject likewise treats 22 species, somewhat realigned from those of Barneby, but recog- nizes 35 infraspecific taxa. Many names pro- posed at infraspecific rank are from Arctic regions of the continent, areas whose collec- tions were not well represented in herbaria prior to 1950. Students of plant taxonomy must examine authentic materials of all previously named taxa, whether currently recognized or not. In the 19th centuiy David Douglas (1798-1834), Thomas Drummond (ca 1780-1835), Thomas Nuttall, and Sir John Richardson (1787-1865) were the most important contributors of speci- mens on which North American names in Oxytropis are based. Repositories for their specimens, later designated as types, are Philadelphia Academy of Sciences (PH), New York Botanical Garden (NY), and Gray Herbarium (GH) in the United States, and Royal Botanic Garden at Kew (K) and Natural History Museum (BM) in England. Later in the 19th centuiy several other workers gath- ered specimens that were considered new to science; these were deposited in numerous other herbaria in addition to those cited above. David Douglas was an intrepid Scottish botanical explorer of North America whose contributions to Oxytropis came from his jour- ney across the continent mainly in 1826-27. He collected the specimens on which the con- cept of O. splendens was based. According to Stafleu and Cowan (1976), the first set of his North American plants is at K; his own her- barium is partly at BM and partly at CGE. Hooker based his treatment of Oxytropis in the Flora in part on collections by Douglas (Stafleu and Cowan 1979), as evidenced by specimens at K. Sir John Richardson, British (Scottish) explorer and naturalist with the Royal Navy, was a medical doctor who accompanied Sir John Franklin on two expeditions, 1819-22 and 1825-27, and later (1848-49) commanded an expedition in search of Franklin who was lost on an ill-fated sea voyage of 1845-47 in search of the Northwest Passage. Richardson's herbarium of vascular plants is mainly at BM, with further material at K and elsewhere (Stafleu and Cowan 1983). The first expedi- tion, in 1819-22, was from Great Slave Lake to the Coppermine River, down which they traveled to Coronation Gulf, and then cross- country to the Coppermine in winter of 1821-22. Only 9 of 21 on the expedition sur- vived the ordeals of hunger, cold, and exposure; that anyone sumved is a tribute to persistence of the men and aid of local aborigines who res- cued them from certain death (Houston 1984). The expedition is remembered as one of the most deadly in the histoiy of biological inves- tigations in North America. Both Richardson and Franklin barely escaped with their lives. The second expedition was down the Mackenzie River to the Polar Sea, with Franklin exploring westward and Richardson eastward along the coast to Coronation Gulf and return. The many plant and animal names proposed by Richardson, and those named after him, serve as a tribute to the genius and perseverance of this remarkable man. 274 Great Basin Naturalist [Volume 55 Edward Sabine (1788-1883), John Edwards, James Clark Ross (1800-1862), William Edward Parry (1790-1855), Alexander Fisher, and Charles James Beverley collected plants on the first Parry journey in search of the North- west Passage (Parry 1821). It is evident from the introduction to Brown's (1824) treatment that each of those named, mainly medical doc- tors with various ships, made their own collec- tions, which were placed initially in their pri- vate herbaria. Sabine was astronomer to the Arctic expeditions led by Pany and collected plants in Melville Island and Greenland. Edwards was surgeon to the Hecla on Pany s voyages of 1819, 1820, 1821-23, and Fisher was assistant surgeon on the Hecla in the 1819-20 voyage, while Beverley was assistant surgeon on the Gripper. Ross was in the Royal Navy with the Parry expeditions in 1819-20, 1821-23, 1824-25, and 1827. Parry was a British explorer who commanded expeditions to the Arctic in search of the Northwest Passage. Specimens from the Parry expedi- tions are deposited at both K and BM, with some of the collectors' private herbaria better represented at BM and some at K. Peter Simon Pallas (1741-1811), important for his exploration of Russia and his work on the Russian flora, named Astragalus deflexus and Astragalus retroflexus, later included in Oxijtropis, names whose interpretations bear on North American species of the genus. According to Stafleu and Cowan (1983) the main personal herbarium of Pallas was sold at a London auction in 1808 to A. B. Lambert, who subsequently sold one part to Robert Brown and another to William Robertson. Both parts presently are at BM and are perti- nent to this paper; additional parts reside else- where, but they are not the basis of this treat- ment except for an important set at LE, which evidently contains the type specimen of Astragalus deflexus, which has not been exam- ined for this treatment. The author wishes to thank the curators at BM and K for their cooperation in providing specimens on loan, and for their hospitality during a visit to London. Also acknowledged is Dr. Rupert Barneby, who read critical por- tions of the manuscript and who provided sug- gestions and encouragement. 1. Aragallus abhreviatits Greene, Proc. Biol. Soc. Wash. 18: 12. 1905. = ()xytr(>i)is liiinhciiii Fiirsh \ar. articulata (Greene) Harneh}' Type; "Te.xas, near Dallas, Limestone prairie, Dallas County, J. Reverchon 603, May 1876"; holotype NDG!; isotypes NY!, BM!; "Dry calcareous soil near Dallas, Texas. Curtis 603, April, May"; paratype GH!, NDG!, NY! 2. Ar«^rt//».v aven-nel.sonii Lunell, Bull. Leeds Herb. 2: 6. 1908. = Oxijtropis lainbeiiii Pursh var. Uimbertii Type: North Dakota, "Aragallus Aven-Nelsoni Lunell, n. sp. Butte, Benson Count\', N. Dak., June 14, 21, July 2,1908," legit J. Lunell; holotype .MIN?; isotypes BM!, NDA!, MIN!, NY!, US!, WTC! ' 3. Arafiallus invenustus Greene, Proc. Biol. Soc. Wash., 18: 12. 1905. — Oxijtropis sericea Nutt. var. sericea Type: "South Dakota, about Fort Meade, Meade County, W H. Fonvood 96a, 96b, 3 June 1887, 96b, 7 June 1887";'syntypes US!, photo BRY!, K Hooker! 4. Aragallus majusculus Greene, Proc. Biol. Soc. Wash., 18: 12. 1905. = Oxijtropis sericea Nutt. var sericea Type: Utah, Heniy Mts., Gai-field County, Utah, M. E. Jones 5674, July 1894; holotype US!; isotypes NY!, MO!, BM!, photo BRY! 5. Aragallus mctcalfei Greene, Proc. Biol. Soc. Wash., IS: 12. 1905. = Oxijtropis lambertii Pursh var. bigelovii A. Gray Type: New Mexico, Sawyer's Peak, Grant County, open glade, ca 10,000 ft, O. B. Metcalf 1079, 7 Julv 1904; holotype US!; isot>'pes NY!. CAS!, GH!, POM, WO^C, BM! 6. Astragalus deflexus Pall., Acta Acad. Sci. Imp. Petrop. 2: 268. 1779." = Oxijtropis deflexa (Pall.) DC. var. deflexa Type: "ad nivalia Dauriae . . . circa Balyra rivum aliosque Ononem influentibus" [Siberia], P S. Pallas s.n.; holotype LE. 7. Astragalus retroflexus Pall, Sp. Astragal., 33, tab. 27. 1801. = Oxijtropis deflexa var. deflexa Original location: Provenit haec species tantuni in alpinis transbaicalensibus et circa lacum Baical, praeser- tim in scaturiginosis frigidus, circa fontes rivulorum Baltschikan, Carol, Bargusin et Chilik; verosimillime quoque per omnem alpestrem tractum, Sibiriam a Sinarum Imperio diffinientem. Type: Pallas, s.n.; lectotype (here selected) BM (#45444)! There are two Pallas collections labeled Astragalus deflexus at BM, 45443 and 45444; the former (paratype) with three stems bears juvenile to mature fruit and a label in Russian script, the latter (lectotype) with a complete plant in flower and two racemes (one in flower and one in immature fruit). Appearing on the sheet with the BM number 45444 on the label are "No. 10" and illegible script. Both collections simulate what has passed in North America as O. deflexa var. sericea Torr. & A. Gray. Sheet 45444, here designated as lectot>pe for Astragalus retro- flexus Pall., is a close match for the illustration in Table 27 of Pallas' Species Astragalorum (1800-1803), except that the drawing is a mirror image of the actual specimen. The 1995] North American Types of Oxytropis 275 reversal of the image comes from the use of the copper phite on which the original drawing was produced. Flower buds, leaves, and leaflets are the same. Only the flowers and their proportions are slightly different; those of the drawing are much too distinct and perhaps too large. The fruiting branch represented in Table 27 is from sheet 45443, again in reverse image, and taken from the branch at the upper left. Pallas (1800-1803) notes that the plant was growii in a garden at St. Petersburg and flow- ered and produced fniit the second year It seems that one specimen or both cited above are from plants grown in the garden at St. Petersburg. 8. Oxytropis argentata sensu Richardson, Frankl. 1st Jour., Bot. Append. 745. 1823. = O. sericea var. speciosa (Torn & A. Gray) Welsh, pro parte Authentic specimen; "British North America. Dr Richardson 1819-22. Astr. argentatus Pallas Astr. Carlton" (BM #45476!). 9. Oxytropis arctica R. Br., Chloris Melvill. 20. 1823. T>'pe locality: Canada, "Melville Island, Pariy's First Vo\age, Sabine, Edwards, Ross, and others, 1819-1820" (R. Brown I.e.). Type: "Melville Island, coll. b\' Mr. Beverley"; lecto- type (here selected) K Hooker!; probable isolectoty-pes S!, GH! The reverse of the Beverley sheet contains the nota- tions, "Winter Harbour, 23 July 1820 [flowering materi- al?]" and "Winter Harbour, 4th July 1820 [fruiting speci- men?, the portion here selected as lectotype]." The Beverley material is the most complete for the species of any of those collected by the Parry expedition members and includes both flowering and fruiting material; the Sabine paratype cited below also has a flowering and a fruiting branch. Brown (1823) did not designate any specimens to support his new species. However, in the introduction to his list of plants collected in Melville Island, he lists the herbaria of the officers of the expedition on which the list was based. Included are "Captain Sabine, Mr. Edwards, Mr. James Ross, Captain Parry, Mr. Fisher, and Mr. Beverley, whose names are here given in order of the extent of their collections." The following are paratypes of Oxytropis arctica at BM and K: "36. Oxytropis arctica. [illegible]. Melville Id. Mr. James Ross" (BM!); "Parry's First Voyage 1819-20. Melville Island. 11th August 1820, (BM!)"; "Melville Id. Capt. Pany. 36. O.xytropis arctica" (45446 BM!); "Melville Isld., Sabine" (K Hooker!); and "O.xytropis arctica. Mr. Edwards" (BM!). The literature citation for O. arctica has traditionally been given as "Parry's First Voy., Append. 9: 278. 1824." The list of plants by Robert Brown was published twice, however, once as "Chloris melvilliana," in 1823 and subse- quently in the appendix to Parry's first voyage. Evidently, Chloris melvilliana was published a year prior to the appearance of identical material under different pagina- tion in the Parry appendix. On an introductory page in Chloris melvilliana is written tlie following: "The follow- ing List is printed as Nl. XI of the Appendix to Captain Parry's Journal of the First Voyage, commencing at page cclix." It is evident that at least page proofs of the Appendix were a\'ailable at the time the Chloris was read\' for printing, and that they formed the basis of the Chloris. 9a. Oxytropis arctica van siihwnbeUata Hook, in W. E. Pairy, Sec. Voy. 4: 396. 1825. = O. nigrescens var. nigrescens Type locality: "Arctic shores and Islands of North America. Capt. Sir E. Parry; Dr. Richardson; Capt. Sir John Franklin; Capt. Back, & c." (Hooker 1831: 146). Type: "Dr Richardson. 1/146. Oxytropis arctica Br. var sul^umbellata. Coast"; lectotype (here selected) BM! The lectotype at BM consists of four specimens of Oxytropis nigrescens var. nigrescens, and they are mount- ed on one sheet with four specimens of O. nigrescens var. uniflora bearing the label, "Oxytropis arctica (i, Frankl. Exp. Dr Richardson. " 9b. Oxytropis arctica var. uniflora Hook, in W E. Pany, Append. Parry Sec. Voy. 4: 396. 1825. = O. nigrescens (Pall.) Fischer var uniflora (Hook.) Barneby Type: "Barrow River, E coast Melville Peninsula, lat. 67°2rN, on Parrv's Second Voyage, Edwards s.n. 1821-3"; lectotvpe (here selected) BM!; isolectotypes K Hooker!, GH!, NY! The Barrow River lectotype at BM, a mere fragment with three flowers, is mounted with a second much more complete collection, "Igloolik. Mr. Edwards, Pany's 2nd Voyage" (a paratype). The collection from Igloolik consists of several flowering specimens and one with a solitary fruit. Additionally, there are four almost mature fruits, two of which have been opened displa\ing the septum. There are several sheets of this variety at BM taken on various Parry voyages (e.g., BM 45452, "O.xytropis arctica, Barrow River, 1822"). The isolectotype at K Hooker is similarly a mere fragment. 9c. Oxytropis arctica 6 inflata Hook., Fl. Bor-Amer 1: 146. 1831. = O. podocarpa A. Gray Type: "Highest summits of the Rocky Mts. Drummond' ; holotype K Hooker! The Drummond material at K consists of six plants, one of which is in young fruit; they are mounted with two specimens by Beechey from Kotzebue Sound, both of which appear to be O. nigrescens. Gray cited the Drummond material with several other specimens when he described O. podocarpa (q.v.). 9d. Oxytropis arctica, "varietas notabilis," R. Br, Chlor. Melvill. 51. 1823. = O. nigrescens (Pall.) Fischer var uniflora (Hook.) Barneby The name as noted above was cited by Barneby (1952) as not validly published; it is here included for con- sistency of use in North American literature on the genus Oxytropis. Hooker (1825), in the botanical appendix to Parr\'"s Journal of a Second Voyage, notes in his discussion of O. arctica: This variet\' (|3.) is noticed by Mr. Brown at the end of his valuable Remarks on the Flora of Melville Island, as discovered by the gentlemen of the present expedition, and says of it, "Varietas notabilis, vix enim distincta videtur species, statura minor, scapo unifloro passemque umbella biflora, dentibus calycis respecti tubi paulo longioribus, foliolis saepe 7, quan- doque 7, \illis persistentibus utrinque argento-seri- ceis." ... To these remarks I may add, that the plants are not above half the size of a [subiimbellata], the stems less woolly, the leaflets fewer, denser, and cov- 276 Great Basin Naturalist [Volume 55 ered with short, ven white silk\ hairs. Thi- pwlvincle scarcely rises above the leaves, and each subtended by small bractea. The corolla is of the most beautiful deep purplish blue; the calyx and legume black from the (juautitN of black hairs; but these are mixed with .several longer white ones. The contrast between the deep blue of the corolla and the dense, white and sil- very lea\'es render this a most loveK little plant. The .specimens arc still silvery white and beautiful after more than 17 decades. 10a. Oxytropis arctohia Biinge, Mem. Acad. Imp. Sci. Saint-Petersbourg 22: 114. 1874. = O. nicens (Pall.) Fisch. \ar. uniflora (Hook.) Barneby Type: "Habitat in arcticis Americae borealis; v. s. sp. in herb, olim Fischerano nunc h. bot. Petrop. " (Bunge I.e.). This was based b\' Bimge (1874) exactly on a speci- men of O. arctica (3 uniflora Hook., in the Fischer herbari- um at LE. The name was utilized by some (Polunin 1940) for North American specimens. 10b. \ar. hyperarcticd Pohmin, Bot. Canad. E. Arctic, 293. 1940. = O. nigrescens (Pall.) Fisch. var uniflora (Hook.) Barneby Type: "Franklin district, Baffin Island, Arctic Bay, N. Polunin 2583, 8-11 Sept. 1936"; holotype CAN; isotypes GH!, BM!, OXF! 11. Oxytropis borealis DC. Prodr. 2: 275. 1925. It appears that the name was not used by Hooker in the modern sense as interpreted by Welsh (1990). The name appears on a sheet of O. maydelliana Trauts'. in the Hooker herbarium at K. It is, however the earliest name available at species rank for the viscid-glandular material that has passed under O. viscicla Nutt., and other taxa (see various uses elsewhere in this paper). 12a. Oxytropis campestris (L.) DC. var. davisii Welsh, Leafl. W Bot. 10: 25. 1963. Tvpe: "British Columbia, mi 403.4, Alaska Hwv', R. J. Davis 6076, 19 JuK 1962"; holotype BRY!; isotv'pe IDS! Distribution; SW Alberta, NE British Columbia. This ta.xon has been represented in herbaria since early in the 19th century. Specimens by Douglas and Drummond, almost assuredh' assignable to this taxon, are present at BM and K. Their collections formed portions of the concepts O. lamhertii (3. Hook, (c^.v.) and O. uralensis y minor Hook. (q.v). This plant is readily distinguished by its colorful flowers, fasciculate leaflets or tendency to fasciculate leaflets, and elongate inflorescences. Specimens have been considered as intermediates between members of the Oxytropis campestris complex and O. borealis var vis- cida, or they have been misidentified as O. splendens because of the fasciculate leaflets. Plants of van davisii are localK' abimdant on stream gravels and adjacent slopes in the focjthills mainly of the Alberta Rockies and in north- eastern British Columbia. Intermediates between var. davisii and O. sericea var. speciosa occur in northern British Colinnbia. Another variety with pink-puiple flowers, var roaldii (Lindstr.) Welsh, occurs in northern Yukon Territory and adjacent Alaska. The northern plant is of low growth and has fewer flowers than van davisii. A similar l)air of low \ersus tall varieties, partialK s\mpatric, exists in eastern Canada, i.e., van johannensis Fernald and van minor (Hook.) Welsh [var. terrae-novae (Fernald) Barneby]. 12b. Oxytropis cam))estris e glabrata Hook., Fl. Bor- Amen 1: 147. 1831. = O. maydelliana Trautv. Type locality: "Bear Lake to the Arctic Shores and Islands, Dn Richardson; Capt. Sir J. Franklin and Capt. Back; Capt Sir E. Parry, &c." (Hooker I.e.). Type: Bear Lake to the Arctic Shores and Islands; neotype GH! Hooker characterized the ta.xon "foliolis glabriuscu- lus subsucculentis." No authentic material of van glabrata was discovered at either K or BM. Hooker's statement of the locality' infomiation applied to both vars. glabrata and melanocephala. The specimen at GH, annotated by Gray indicating that it was based on Hooker's van glabrata, is here designated as neotype. Thus, the name is fixed in the sense used by Barneby (1952). 12c. Oxytropis campestris van johannensis Fernald, Rhodora 1: 88. 1899. Type: "Maine, gravelly shores, valley of St. John River, Fort Kent, Aroostook County, M. L. Femald 2289, 15 June 1898"; holotype GH!; isotypes CAN!, US!, NY!, BRY!, BM! Distribution: Newfoundland, New Brunswick, Nova Scotia, Quebec, Ontario, and Maine. Plants of this variety from the Farm Riven south of James Bay, Ontario, have fascicidate leaflets and short pods. In the latter feature they simulate van chartacea (Fassett) Banneby, which might best be regarded as only a disjunct phase of this varietv'. 12d. Oxytropis campestris ^ melanocephala Hook., Fl. Bon-Amen 1: 147. 1831. = O. maydelliana Trautw Synonym: O. maydelliana ssp. melanocephala (Hook.)' Porsiid Type locality: "Bear Lake to the Arctic Shores and Islands, Dn Richardson; Capt. Sir J. Franklin and Capt. Back, Capt. Sir E. Pany, &c." (Hooker I.e.). Type: "24 July 2-Augt [?] 1826. O. camp. ^. Dn Richardson"; lectotype (here selected) BM!; isolectotvpes GH!, K Hooker! ("^. Arctic Sea. Richardson. Oxytropis campestris, "). Paratypes: "Capt. Parn'. 2nd Voy. [and] Parn 's 2nd Voy.," both at K Hooker!; and "O.xytropis campestris. Duke of York's Bay. Arctic Regions. Parry's 2nd Voyage," BM 45449! Several additional specimens from Pane's sec- ond voyage are also present at BM. The Richardson isolectotype at K consists of two specimens, both with stipules pale and merel\- mottled with piuple instead of puiplish overall as in most speci- mens of O. maydelliana. Because they are, however, somewhat unusual among specimens of the species, they are not chosen as lectotvpe. Specimens at BM more close- ly represent the concept of the ta.xon. The isolectotype at K is mounted on a sheet with two collections by Capt. Parry from the second voyage, and a third collection by Simpson from the "Polar Sea." The Party collections are apparent paratv'pes of t, melanocephala Hook. An extrane- ous stem of O. borealis van hudsonica, mounted on the same sheet, appears to belong to one of the Pany collec- 1995] North American Types of Oxytropis 277 tions. On the lectot\pe specimen at BM the name "Dr. Richardson" was obviously added later as it is on most specimens attributed to him, and the date 1826 is coiTect for the second Franklin expedition. There is a second sheet at BM!, "1/147. O.xytropis campestris DC. River Rae, ' with the name Dr. Richardson written below the label, and bearing the notation, "0.\\tropis campestris. DC. P sordida Lin." Possibly it is a paratope of van rnelanocephala. Many of the specimens from Parn s second voyage desig- nated as O. campestris are O. inaydclliana. but some are O. horealis van hudsonica. 12e. Oxytropis campestris var. minor (Hook.) Welsh, comb, nov., based on "Oxytropis uralensis y minor Hook., Fl. Bor.-Amer 1: 146. 183i. Synonym: O. terrae-novae Fernald; O. campestris \ar. terrae-novae (Fernald) Barneby (see O. uralensis for discussion of the reasons for this combination) I2f Oxytropis campestris var speciosa Torr. & A. Gray, Fl. X..\mer 1; 341. 1838. Based on; O. campestris y sulphurea sensu Hook., Fl. Bor Amer. 1: 147. 1831 (see below). — O. sericea Nutt. var. speciosa (Torr. & A. Grav) Welsh Type locality; "British America, west to the Pacific" (Torrey and Gray I.e.). Type; "Dr. Hooker"; lectotype (here selected) NY! Parat\-pes; "O.x. campestris y sulphurea. Dr Hooker's Fl. B. Am. Rocky Mountains. Drummond," at K Hooker! and "O.xvtropis campestris y. Frankl. Exp. Dr. Richardson," atBM! ' The sheet at NY, communicated by Dr. Hooker, bears the notation, "Carlton House on the Saskatchewan to the Rock^ Mountains, Dnimmond," and die initials E. E S[heldon'r']. The quote is identical to that for varieties a and 5 in the Flora, but since Hooker did not cite a localitx' for y sulphurea. the material sent to Gray by Hooker could ha\e been taken by either Douglas or Richardson and not necessarily by Drummond. There are authentic specimens bearing the name "sulphurea" at BM, one by Douglas and the other by Richardson, and possibly a third by Flichardson with the simple designation, "Oxytropis campestris y." Probably the lectotype at NY was taken by one or the other of the two collectors, and not by Drummond. However, the specimen at K cited as paratype was collect- ed by Drummond. Hooker's brief description of y sulphurea, "foliolis latioribus, spicis capitatis, floribus majoribus speciosis," characterizes this common plant of the western plains and foothills of the Rockies from far north in Canada south to Montana, Wyoming, and Idaho. 12g. Oxytropis campestris 5 spicata Hook., Fl. Bor.-Amer 1; 147. 1831. = O. campestris (L.) DC. var. spicata Hook. Type locality; "Between Carlton House on the Saskatchewan to the Rocky Mountains [Alberta], T. Dnnnmond' (Hooker I.e.). Type; "O.xytn campestris 8. Carlton House. Fl. Bon Am. "; lectotype (here selected) K Hooker! Hooker (1831) characterized 8 spicata as "spicis elongatis, floribus remotioribus. That description, although short, matches the lectotvpe. There is only one sheet of Oxytropis at Hooker's herbarium at K that bears the designation 8 and the locality information "Carlton House. ' The fact that Drummond is not indicated as col- lector on the sheet is apparently of little consequence as Hooker frequently failed to record collectors on his small herbariiuii labels or in his annotation of the specimens. The plants on that sheet clearly belong to what has passed in recent times imder the name of O. campestris van gra- cilis (A. Nelson) Barneby. That some plants sent by Hooker to Gray represented taxa (jther than van spicata is unfortunate. It appears that Hooker clearly had two enti- ties in mind when he described van spicata and discussed van sulphurea. Indeed, Barneby (1952; 279) masterfully summarized the problems of identification of specimens of van spicata (as O. campestris van gracilis) and van spe- ciosa (as O. sericea van spicata). Much of the material exchanged by Hooker is, indeed, the larger-flowered phase (here termed van speciosa) with fewer leaflets. Whether Hooker was responsible for sending wrongly labeled material that was segregated prior to its distribu- tion is not known. If Drummond mixed his collection so as to include both varieties, he was not alone in mixing the two pale-flowered species of the plains of western Canada. The van spicata, with its numerous leaflets and small flowers, grows in the general area occupied by the larger-flowered phase with fewer leaflets. Many botanists have made similar misinteipretations. The need to replace van spicata with van speciosa is an unfortunate but neces- sary' change mandated as the result of study of the types. 12h. Oxytropis campestris van sulphurea sensu Hook., Fl. Bon Amen 1; 147. 1831. non DC. = O. sericea van speciosa (Torn & A. Gray) Welsh Authentic specimens of O. campestris P speciosa Torn & A. Gnay; "2. O. campestris. P sulphurea on the ned deer and eagle hills of the [illegible]. 1827," North America. D. Douglas, and "Oxytropis campestris y, Frankl. Exp. Dn Richardson" (both at BM!); and "Ox. campestris y sulphurea. Dn Hooker s Fl. B. Am. Rocky Mountains. Drummond" (at K Hooker!). 13. Oxytropis deflexa var. pulcherrima Welsh & A. Huber, var. nov. O. deflexae var foliolosae (Hook.) Barneby aspectu similis sed in floribus majoribus, racemis latioribus et dense pilosis, et a van sericeae Torr. & A. Gray in racemis compactis et floribus majoribus et purpureis differt. Type; USA, Utah, Duchesne County; "T2N, R7W, S31 NW/NW UB&M, head of Log Hollow, 1.8 mi due SW of Upper Stillwater Reservoin Uinta Mts., gravelly subalpine meadow, common, limestone substrate, at ca 3294 m, 12 July 1994, A. Huber 1673"; holotype BRY!, duplicates to be distributed. Additional materials (paratvpes); Utah, Duchesne County; Uinta Mountains, T2N, 'R7W, S31, 11 mi N of Tabiona, 12 July 1972; do, T2N, R6W, S18, divide between Rock Creek and Brown Duck Basin, at 3447 m, 28 August 1981, S. Goodrich & D. Atwood 16163; do, T2N, R7W, S30, 3 July 1978, S. Goodrich & L. Hart 11705; do, TIN, R8W, SI, head of Wedge Hollow, 6 July 1990, D. Atwood 13934. Utah, Summit Count>'; Lost Creek Park, road from Hoop Lake to Spirit Lake, at 3050 m, 23 August 1977, K. Ostler 932. Utah, Daggett CountA'; T2N, R17E, Sll, 22.2 km SW of Manila, at 2815 m, 11 August 1983, S. Goodrich 19661. Colorado, Chaffee County: 0.7 mi NE of Cumberland Pass, T12 N, R4E, S12, at 3730 m, 17 August 1982, J. Peterson et al. 82-58. Colorado, Gunnison County; Virgina Basin, at 3691 m, 8 July 1946, C. L. Hax'ward 148; do, Cumberland Pass, at 3874 m, 23 July 1970, L. C. Higgins 3755; do, 9 July 1969, L. C. Higgins 2103; do. North ridge 278 Great Basin Naturalist [Volume 55 of Mount Belview, at 3750 m, 5 July 198S. L. C. Manin 3105. All specimens at BRY! Plants of var. piilcherritna may he distinguished from all other North American materials of the deflexa complex by their compact racemes (remaining so in fruit), larger flowers (hence broader racemes), and short, plump pods that tend to have an abrupt shoulder beyond the stipe. Their bipartite distribution from alpine sites in the Uinta Mountains and similar, but higher, areas in the southwestern Colorado Rockies is unique among Oxtjfro- pis species. This is material that has been regarded in con- temporary treatments of the genus in North America as var. deflexa. Specimens of var. deflexa from Siberia, at least those examined by me, have elongate racemes, smaller flowers, and more slender pods. 14. Oxytropis foUolosa Hook., Fl. Bor.-Amer. 1: 146. 1831. = O. deflexa \'m. foliolosa (Hook.) Barneby Type locality: "From Carlton-House to the Rocky Mountains, in lat. 54°" (Hooker I.e.). Type: "Ox. foliolosa Hook. Fl. Bor. Am. Rocky Mountains, Drumniond" Alberta, Canada; holotype K Hooker!; isotype NY! The holotype consists of two flowering specimens representing the common acaulescent or subcaulescent phase of var. foliolosa. The upper specimen displays immature fruit. There are collections of O. deflexa var sericea Torr & A. Gray at K by Nuttall ("Oxytropis defle.xa. R. Mts.") and by Drummond ("Saskatchewan. Drummond. Ox. deflexa. DC"). 15. Oxytropis hookcriana Nutt. in Torr. & A. Gray, Fl. N. Amer. 1:340. 1838. = O. lamhei'tii Pursh var. lainhertii T\pe: "Oxytropis * Hookeri. O. Lambeiti P Hook. 147. Platte plains," Nuttall s.n. 1834; holotype BM Nuttall! The specimen at BM has "Platte plains" written on the back of the sheet. It seems apparent that xNuttall hoped to honor Hooker by providing a name for what he took to be the concept of O. lambertii p, a plant that is a phase of O. carnpesiris scarcely related to O. lainhertii. 16. Oxytropis ixodes Butters ik Abbe, Rhodora 45: 2, tab. 745, fig. 1-6. 1943. = O. borealis DC. var viscida (Nutt.) Welsh Type: Minnesota, "slate cliffs on north side of a high hill 1/2 mi. west of the outlet of South Fowl Lake, Cook County, F K. Butters, E. C. Abbe, & G. W. Bums 611, 27 June 1940"; holotype MIN; isotype GH!, NY!, PH!, US!, UC!, DAO!, BRY!, BM! 17. Oxytropis lagopiis Nutt., J. Acad. Nat. Sci. Philadelphia 7: 17. 1834. Type: Sources of the Missouri, N. B. Wveth s.n. 1833; holotype BM Nuttall!; isotopes NY!. PH!, K! The specimen at BM has a label of a typical Nuttall collection, "O.xytropis * lagopus — Sources of the Missouri." It is mounted on a sheet with a collection from "Rocky Mts, near the Platte," by Dr. Parry. The Royal Botanic Garden material consists of a single caudex branch and two inflorescences. 18. Oxytropis lambertii Pursh, Fl. Amer. Sept., 740. 1813. Type: "On the ^Missouri, on the liluffs from the Maha \illage to the Poncars, Louisiana [NE Nebraska or adjacent South Dakota or Iowa], Bradbury s.n. 1811; lectotype PH! (Barneby, Proe. Calif Acad. Sci. IV. 27: 285. 1952, as type); isotype BM! The sheet at BM is labeled "Louisiana. |. Bradbury 1811-12." 18a. Oxytropis lainheiiii (3. Hook., Fl. Bor -Amen 1: 147. 1831. — O. campe.sfris (L.) DC. var davisii Welsh Type locality: "Dry banks on Red River and Saskatch- awan, (Douglas,) to the Prairies in the vallies of the Rock'v Mountains. Drummond" (Hooker I.e.). Type: "O.xytropis lambertii (3. Dr Hooker"; holotype K! (see discussion below); isotype NY! Authentic specimen: "O.xytropis Lambertii (3. Frankl. E.xp. Dr. Richardson" (BM!). .Neither Douglas s nor Drummond s materials with the unequivocal designation O. lainhertii (3 have been seen at either K or BM. There is at K Hooker, however, a collection by Drummond (cited below as a possible paratype of O. uralensis y minor Hooker, q.v.) that bears the penciled notation "Ox. Lamberti??" It appears to be O. campestris var davisii Welsh. The Richardson collec- tion at BM (authentic specimen cited above) almost cer- tainly is that ta.xon also. A second sheet at K Hooker bears three collections, two small plants by Drummond, two taller plants presumed to be by Douglas, and a third extraneous collection by Percival. The first and second are van davisii, and tliere is a penciled notation, "Ox. Lambertii Pursh." That Hooker misunderstood that at least some part of his var minor and his var (3 were conspecific indi- cates the problem he had in dealing with plants from such diverse areas as those found in North America, plants he had not seen in the field. Designation of a lectotype for the material is probably moot, since the material was not given more than alphabetical designation. Perhaps these sheets, as interpreted herein, will lay to rest the name O. lambertii (3. Had there been an epithet applied, it would have precluded the use of the name davisii. 18b. Oxytropis lambertii P leucoplnjlla Nutt. in Torr & A. Gray, Fl. N. Amer. 1: 339. 1838, pro syn. = O. lambertii e. (see below) 18c. Oxytropis lambertii £. Torn & A. Gray, Fl. N. Amen 1:339.1838. = O. lagopus van atropiirpiirea (Rydb.) Barneby pro parte et O. nana Nutt. pro parte Synonym: O. lambertii P leucoplnjlla Nutt. Authentic specimen: "Oxytropis * leucophylla. R. Mts.," Nuttall s.n. 1834, BM Nuttall! The authentic specimen has the information "Rocky Mts. Nuttall's Herb." written on the reverse side, and the following notation on the front: "Oxytropis leucophylla Nuttall! [Oxytropis] Lamberti Pursh van glabrata Torn & Gn Fl. N. America 1: p. 339." The name was published as a synonym by Ton-ey & A. Gray (I.e.), based on a manu- script provided b\' Nuttall. The>- characterize the plant by the following description, but evidenth' did not see the material: "e. very dwarf, canescently woolly; the leaflets shorter and about 5 pairs; scape scarceh' longer than the leaves; flowers capitate or nearly so; calyx densely woolly; bracts small and short; wings emarginate. — O. Lamberti P leucophylla, Nutt. mss," from "Plains of the Platte." Barneby (1952: 304) noted that there "seem to be no cor- 1995] North American Types of Oxytropis 279 responding specimens either at Philadelphia or in the herbaria of Gray and Torrey." He therefore surmises that the plant in question belongs to O. lagopus van atroptir- piirea. There is a sheet at BM bearing Nuttall s character- istic label with the name "Oxytropis * lettcophylla" in Nuttalls handwriting. The two plants on the sheet appar- ently belong to two different taxa, the smaller one to O. lagopus var atropurpurea as surmised by Barneby, and the second larger one to O. nana of Nuttall (or perhaps, but unlikely, a dwarf specimen of O. sericea, my first impres- sion). 19. Oxytropis mollis Nutt. ex A. Gray, Proc. Amer Acad. Arts 6: 235. 1864. pro syn. — O. borealis var. viscida (Nutt.) Welsh Authentic specimen: "Oxytropis mollis. O. Ochro- leuca Led. Altai proxima . . . R. Mts. Oregon," Nuttall s.n. 1834 (BM Nuttall!). 20. Oxytropis multiceps Nutt. in Ton; & A. Gray, Fl. N. Amen 1:341. 1838. Type locality: "Summit of lofty hills in the Rocky Mountain range, towards Lewis's River [S. Wyoming], Rock-y Mts. Nuttall" (Toney and Gray I.e.). Type: "O.xytropis (Physocalyx) multiceps. R. Mts." T. Nuttall s.n. 1834; holotype BM Nuttall!; isotypes NY!, GH!, K Hooker! The two specimens at BM are mounted on a sheet with collections by J. M. Coulter and M. E. Jones. The Nuttall material bears three labels: "*Physocalyx * multi- ceps R. Mts. and two odiers. The second label makes com- parisons with Old World species that the proposed new genus and species could not be, and the third label con- tains a brief description, "Gal. inflatus, apice 5-fidus, legu- mine includens." Nuttall was at least entertaining the idea that the plant represented a new genus. Distribution: Colorado, NE Utah, SW Wyoming, and W Nebraska. The accrescent calyces, broad bracts, and few flow- ers are characteristic for the species, which stands alone in the genus in North America in its morphology. 20a. Oxytropis multiceps vai: minor A. Gray, Proc. Amer Acad. Arts 20: 2. 1884. = O. multiceps Ton: & A. Gray Type: Clear Greek County, Colorado, C. C. Parry 991, 1861; lectotA'pe GH! (designated by Barneby, Proc. Calif Acad. Sci. IV 27: 220. 1952); isolectotv-pe NY! Paratvpe: "Rockv Mountain Alpine Flora, Lat. 39°-4r. No. 144. E. Hall & J. P Harbour, Colls. 1862"; BM! 21. Oxytropis nana Nutt. in Torr. &: A. Gray, Fl. N. Amer 1:340. 1838. Type: Plains of the Platte in the Rocky Mountain Range [Wyoming], T. Nuttall s.n. 1834; holotype BM Nuttall!; isotypes NY!, PH! Distribution: Drainage of the North Platte and Cheyenne rivers, westward to the Wind River Mountains, Albany, Carbon, Converse, Fremont, Natrona, Platte, and Sweetwater counties, Wyoming; endemic. This is a beautiflil species of cla\s, shales, and gravelh' bluffs and ridge tops endemic to Wyoming. Barneby (1952) postulated that it might have arisen through hybridization oi Oxytropis sericea and O. multiceps, a likely supposition. Flower colors are \'ariable in a gi\'en population from pale pinks through lavender and purple, and white-flowered populations are known. A contribution from O. lamhertii is also suggested by the presence of incipiently malpighi- an hairs in some specimens. The relationship to segre- gates of O. hesseyi postulated by Isely (1983) seems tenu- ous at best. The relegation of O. nana to that species might require a realignment of other taxa as well, includ- ing combination oi lamhertii, sericea, campestris, and even multiceps. Such a proposal is, of course, absurd. Taxonomy must be both practical and reflect biological reality. 22. Oxytropis plattensis Nutt. in Torr. & A. Gray, Fl. N. Amer 1:340. 1838. = O. lamhertii \"m: lamhertii Type: "Oxytropis * Plattensis, R. Mts. Platte," Nuttall s.n. 1834; holotype BM Nuttall!; isotxpe NY! The holotype at BM consists of a single plant cut fiom another sheet. It is topical of the Great Plains phase of O. lamhertii. 23. Oxytropis podocarpa A. Gray, Proc. Amer. Acad. Arts 6: 234. 1864. Type locality: "Labrador, Arctic regions, and Rocky Mountains, lat. 49°" (Gray I.e.). Types: Labrador (Schweinitz), Arctic America (Richardson?), O. arctica 5 injlata Hook. (Drummond), and Alberta (Bourgeau). The Schweinitz and Bourgeau specimens at GH! are cotypical, both having been used by Gray in characterization of the species. However, the species was lectotypified by Fernald (Rhodora 30; 154. 1928) on the Schweinitz collection from Laborador The remaining specimens are considered to be paratopes. Paratypes: "Highest summits of the Rocky Mts., Drummond" s.n, K!, type of O. arctica 6 injlata Hook., q.v.; "Oxytropis arctica 8 R. Br. Arctic America. Frankl. Exp., K Hooker! Distribution: Rock)' alpine ridges and coastal shores in Colorado, Wyoming, Montana, Alberta, Northwest Terri- tories, Ungava Peninsula, Labrador, and Baffin Island. The bladdery-inflated stipitate pods of Oxytropis podocarpa are characteristic of this and few other oxy- tropes. The folded, falcate leaflets are useful in distin- guishing this from other closely related mat- or mound- forming species, such as O. nigrescens, in vegetative con- dition. 24. Oxytropis sericea Nutt. in Torr & A. Gray, Fl. N. Amer 1: 339. 1838. Type: Rocky Mountains toward the sources of the Oregon [S Wvoming], T Nuttall s.n., 1834; lectotype NY! (Barneby, Proc. Calif Acad. Sci. IV, 27: 272. 1952). 24a. Oxytropis sericea var. speciosa (Torr. & A. Gray) Welsh, comb, nov., based on "Oxytropis campestris P spe- ciosa Torr. & A. Gray, Fl. N. Amer 1: 341. 1838, this in turn based on O. campestris y .mlphurea sensu Hook., Fl. Bor Amen 1; 147. 1831. Missapplied name: O. sericea van spicata sensu Barneby, Leafl. W Bot. 5: 111. 1951. Distribution: Yukon, British Columbia, Alberta, Sas- katchewan, Manitoba, Montana, Idaho, and Wyoming. Members of this variety are characterized by ochroleucous flowers with immaculate keel. In general aspect they simulate the partially sympatric Oxytropis campestris van spicata, from which they may be distin- guished by fewer leaflets and generally larger flowers. Alpine phases of O. campestris van cusickii approach van 280 Great Basin Naturalist [Volume 55 speciosii l)()tli ill llower size and color. MaiiiK var. speciosa does not occur in hijihlancls inhabited In var. cusickil l)ut the similarities of the two varieties should not l)e discounted. Apparent hybrids are known between this and O. cam- pestris var. davLsii in northeastern British (Columbia. 25. Oxijtropis spiendens Douglas e.\ Hook., Fl. Bor.- Amer. 1: 147. 1831. Type locality: "On limestone rocks ol the Red River, and south toward Pembina [S Manitoba]. Douglas" (I.e.). Type: "On Limestone rocks oi the Red River and on the south towards Pembina, 1827, a Ox. spiendens. Douglas"; lectotype (here selected) K Hooker!; isolecto- types OXF! (photo BRY!), BM! ("O. spiendens. Dry soils on the plains of Red River. 1827. Douglas s.n."; 2 sheets). The two sheets at BM are both by Douglas and represent a vestita, the tvpical phase of the species. The lectotype at K bears a label with almost the exact information as the published type locality. It is mounted on a sheet with a specimen designated "(3," and with the label information "Rocky Mts. Richardson." A better choice for lectotype of var. |3 is on a second sheet (see below). 25a. Oxytropis spiendens a vestita Hook., Fl. Bor.-Amer. 1: 148. 1831 = O. spiendens Douglas Type locality and type: As for the species. 25b. Oxytropis spiendens (3 richardsonii Hook., Fl. Bor.- Amer. 1: 148. 1831. = O. spiendens Douglas Type locality: "From Cumberland-House on the Saskatchewan, north to Fort Franklin and the Bear Lake, and West to the dry prairies of the Rocky Mountains. Dr. Richardson; Drummond" (Hooker I.e.). Type: "278. O.xytropis o.wphylla. Dr. Richardson" s.n. in 1821; lectotype (here designated) K Hooker!; isolecto- types NY!, O. oxyphylla of Richardson, GH! Paratype: "Fort Franklin to the Rock-y Mts. Drum- mond. p. Ox. Spiendens. Dougl. Hook. Fl. B. Am.," K Hooker! The lectotype has three specimens, each designated "P." The label "278. Oxytropis o.xyphylla" is affixed across the base of the middle specimen; adjacent to the left one is "Bear Lake," and below the specimen at the right is the collector's name, "Dr. Richardson." The name "O. oxy- phylla," in the sense utilized by Richardson in the botani- cal appendix to Franklin's first journey (1823), is clearK the basis for Hooker's p. richardsonii. The Drummond syntype consists of a beautiful pkuit with several flowering stems and numerous leaves, and a fniiting raceme and peduncle. Bameby (1952) notes: Hooker recognized from the first a typical a vestita, "valde hirsuto-sericea, bracteis hirsutissiniis calyce multo longiorihus. ' described from Douglas s Red River plants, and a p richardsonii. "minus hirsuta, bracteis vi.\ longitudine caixcis, collected between the Saskatchewan River and the Rocky Moimtains. Plants of the two types pass into each other by degree and have not been recognized at taxonomic rank in recent times. 26. Oxytropis uralensis sensu American authors, non (L.)DC." North American specimens at BM and K bearing this name are a mixture of (a) O. arctica R. Brown var. arc- tica (eglandular, with large, pink-purple flowers); (b) O. horealis var. hudsonica (Greene) Welsh (glandular, the calyx teeth short and purjilish flowers (see "7/1834.7 0.xy- tropis Uralensis. British North America. Dr. Richardson 1819-22, BM!," and "Repulse Bay. Parry's 2nd Voyage, BM!"); (c) "Arctic Regions. Oxytropis Uralensis. Repulse Bay. Parry's 2nd Voyage"); (d) O. maydelliana Trautv. (ochroleucous flowers, with stipules castaneous); and (e) O. canipestris var. minor (Hook.) Welsh (including var. ter- rae-novae, flowers pink puiple, eglandular). 26a. Oxytropis uralensis a in Hook., Fl. Bor.-Amer. 1: 146. 1831. = O. arctica R. Brown var. arctica Locality: "Arctic regions and islands. Dr. Richardson; Capt. Pariy &c." (Hooker I.e.). Authentic specimen: "O. uralensis a Frankl. Exp. Dr. Richardson, "BM! Hooker s (1831) use of Oxytropis uralensis a in his discussion of habitat merely indicated acknowledgment of the taxon in the sense of Old World materials. However, an authentic specimen at BM with that label information is O. arctica sens. str. (see above). 26b. Oxytropis uralensis var. subsucculenta Hook., Fl. Bor.-Amer. 1: 146. 1831. = O. horealis DC. var. horealis Type locality: "Arctic seashore, to the east of the Mackenzie River" (Hooker I.e.). Tyi^e: "O. uralensis p. 126. H. Sea Coast. Dr. Richard- son"; lectotype (here designated) BM! The lectotype at BM is mounted with O. uralensis a, i.e., O. arctica R. Br. var. arctica. A possible syntvpe of P suhsucculenta Hook, is also present at BM, with the label "British North America. Dr. Richardson 1819-22. " Above the label is a pencil notation, "cut from sheet of Oxytropis canipestris." 26c. Oxijtropis uralensis var. arctica (R. Br.) Ledebour, Fl. Ross. 1: 594. 1842. Basionym: Oxytropis arctica R. Br. = O. arctica R. Br. var. arctica 26d. Oxytropis uralensis y minor Hook., Fl. Bor.-Amer. 1: 146. 1831, = O. campestris var. minor (Hook.) Welsh (see 12e above) Type locality: "Dr\ hills and prairies of the Rock-> Mountains. Mr. Drummond. Labrador. Mr. Morrison' (Hooker I.e.). Type: "Labrador. O. uralensis y. Momson"; lectotype (here designated) K Hooker! The International Code of Botanical Nomenclature allows recognition of a taxon based on discordant material where the name can be applied to at least one of its parts. Hence, var. minor is not to be rejected simply because the specimens on which it was based represent more than one taxon. Evidence to support the assignment of the name to the Labrador material is unequivocal, while its application to materials from western Canada is problematical. The lectotype at K consists of two specimens, one flowering and the other in fi-uit. They were mounted pre- viously with plants of another species, which have been removed by cutting the sheet. The specimens both bear 1995] North American Types of Oxytropis 281 the notation "y, and the sheet contains the annotation "O. uralensis y Fl. Bor. Am. sed certi friictu ab Uralensi diver- sum," the author unknown. Barneby (1952) was unable to resolve the application of the name but noted: "The prob- lem is nomenclaturally important in that var. minor could prove to be the earliest name in its category for either O. viscida var. hudsonica or O. campestris van terrae-novae." The Labrador specimens cited with the original descrip- tion are O. campestris var terrae-novae in a modern sense and are here selected as the lectotype for the ta.xon. Two small specimens at K Hooker! bearing the label "Saskatchewan. Dnmimond" appear to be var. davisii, but one cannot be certain of their provenance or that they represent the material designated by Hooker as var. minor They are mounted on a sheet with two additional specimens, apparently var davisii also, but probably col- lected by Douglas. Possibly all four specimens fonned the basis for still another of Hooker's plants, O. lambertii P (q.v.). A more convincing collection possibly included by Hooker within var. minor, at K, is labeled "Astragalus uralensis. Dr\' mountain prairies & low hills. Drummond. " It is a possible syntype of var minor and appears to be var davisii Welsh. There is no certainty, however, that the specimen is part of what Hooker indicated as var minor References Bar.neby, R. C. 1952. A revision of the North American species of Oxytropis DC. Proceedings of the Cali- fornia Academy of Sciences IV, 27: 177-312. Brow.n, R. 1823. Chloris melvilliana. A list of plants col- lected in Melville Island (latitude 74°-75° N. longi- tude 110° -112° W.) in the year 1820; by the officers of the voygage of discovery under the orders of Captain Parry. With characters and descriptions of the new genera and species by Robert Brown. W Clowes, London. . 1824. A supplement to the appendi.x of Captain Parry's first voyage for the discover}- of a North-West passage, in the years 1819-20. Containing an account of the subjects of natural histoiy. IX. Botan\'. A list of plants collected in Melville Island, by the officers of the expedition; with characters and descriptions of the new species. Pages 261-309 in Journal of a voy- age for the discovery of a North-West Passage fi"om the Atlantic to the Pacific; performed in the year 1819-20, in His Majesty's ships Hecla and Gripper under the orders of Captain William Edward Parry, R.N., ER.S., and commander of the expedition. John Murray, London. BUNGE, A. 1874. Species generis Oxytropis, DC. Memoirs of the Academy of Imperial Sciences Saint-Peters- bourg, VII, 22: 1-166. Gray, A. 1863. A revision and arrangement (mainly by the fiaiit) of the Noilh American species of Astragalus and Oxytropis. Proceedings of the American Academ\- of Arts and Sciences 6: 188-236. . 1884. A revision of the North American species of the genus Oxytropis, DC. Proceedings of the American Academy of Arts and Sciences 20: 1-7. Holmgren, P K., N. H. Holmgren, and L. C. Barnett. 1990. Index herbariorum. Part I: The herbaria of the world. Regnum Vegetabile 120: 1-693. Hooker, W J. 1825. Botanical appendix. Pages 381-430 in Appendix to Captain Parry's journal of a second voyage for the discovery of a North-West Passage from the Atlantic to the Pacific performed in his Majesty's ships Fury and Hecla, in the years 1821-22-23. John Murray, London. . 1831. Flora boreali-americana; or, the botany of the northern parts of British America: compiled principally from the plants collected by Dr. Richardson & VI r Dnmimond on the late northern expeditions, under command of captain Sir John Franklin, R.N., to which are added (by permission of the Horticultural Society of London) those of Mr. Douglas, from North-West America, and other natu- ralists. Treuttel & Wurz, London. [Published as fas- cicles, 1829-1840.] Houston, C. S. 1984. Arctic ordeal. The journal of John Richardson surgeon-naturalist with Franklin 1820-1822. McGill-Queen's University Press, Montreal, Quebec. Canada. IsELY, D. 1983. New combinations and two new varieties in Astragalus, Orophaca, and Oxytropis (Legumi- nosae). Systematic Botany 8: 420-426. Pallas, P S. 1800-1803. Species Astragalomm descriptae et iconibus coloratis illustratae. Godofredi Martini, Lipsiae. Parry, W E. 1821. Journal of a voyage for the discoveiy of a North-West Passage fi-om the Atlantic to the Pacific: perfomied in the years 1819-20, in his Majesty's ships Hecla and Gripper, under orders of William Edward Parr>-, R.N., ERS., and commander of the expedition. John Murray, London. POLUNIN, N. 1940. Botany of the Canadian Eastern Arctic. Part I. Pteridophyta and Spermatophyta. National Museum of Canada Bulletin 92: 1^08. RiGHARDSON, J. 1823. Botanical appendix. Pages 729-768 in J. Franklin, Narrative of a journey to the shores of the Polar Sea. W. Clowes, London. Staflei', E a., and R. S. Cowan. 1976. Taxonomic litera- ture. Volume I: A-G. Regnum Vegetabile 94: 1-1136. . 1979. Ta.xonomic literature. Volume II: H-Le. Regnum Vegetabile 98: 1-991. . 1983. Taxonomic literature. Volume IV: P-Sak. Regnum Vegetabile 110. 1-1214. TORREY, J., and a. GR/\y. 1838. A flora of North America. Volume 1. Wiley & Putnam, New York. Welsh, S. L. 1977. On the typification of Oxytropis leu- cantha (Pallas) Pers. Taxon 21: 155-157. . 1990. On the txpification of Oxytropis horcalis DC. Great Basin Naturalist 50: 355-360. . 1991. Oxytropis DC. — names, basion\ms, types, and synonyms — Flora North America Project. Great Basin Naturalist 51: 377-396. . 1994. Oxytropis de Candolle. Flora North .\jnerica. In press. Received 20 September 1994 Accepted 2 December 1994 Great Basin Naturalist 55(3), © 1995, pp. 282-283 SALTATION IN SNAKES WITH A NOTE ON ESCAPE SALTATION IN A CROTALUS SCUTULATUS Breck Bartholomew^ and Robert D. Nohavec^ Key words: Crotalus scutulatus, escape saltation, behavior. Escape saltation and aggressive saltation have been reported in relatively few snakes (Gans and Mendelssohn 1971, Klauber 1972, Gans 1974, Armstrong and Muiph\' 1979, Gasc 1994). These reports range fi-om the incredulous to the well documented. Gasc (1994) relates an unbelievable case of jumping in Atropoides {=Porthiclimn) nwnmifer in which individuals "tend to jump, either when they hit a prey or to clear a height of up to 3 ft (1 m), starting from a low point [emphasis added]." Certainly A. nummifer may jump; however, in their decades of experience with hundreds of these snakes in both the wild and captivity, W. Lamar, L. Porras, and A. Solorzano have never seen nor heard of this behavior (personal communi- cation). It is possible that A. nwnmifer may appear to jump as they strike from an arboreal perch (i.e., a log) and fall to the ground (L. Porras personal communication). The best reports of ophidian saltation are those of Gans and Mendelssohn (1971) and Gans (1974). These authors analyzed Bitis cau- dalis jumping behavior in terms of stimulus and biomechanics. They determined that B. caudalis weighing less than 23.5 g, with a body temperature between 31 °C and 37 °C, were able to jump using sidewinding locomotion. This type of saltation is energetically expen- sive, and jumping snakes tire quickly. Believable reports of rattlesnake saltation are relatively few. Klauber's (1972) reports con- sist primarily of exaggerated accounts of rattle- snakes jumping while striking at either prey or man. However, one of his reports cannot be ignored: Dr. R. B. Cowles told me that he was always skep- tical of stories of rattlesnakes leaving the ground in the course of a strike, until he saw this done two or three times by an angry southwestern speckled rattler {Crotalus mitchellii pyrrhus). The snake was on pavement and struck for more than its full length. The only other report of rattlesnakes jumping is of C sciitiilatm salvini which "struck so violent- ly that their entire body appeared to be momen- tarily air borne" (Armstrong and Muiphy 1979). Neither of these reports discusses the biome- chanics of how these snakes jumped. Since both accounts are of aggressive saltation, and neither of the species typically utilizes side- winding locomotion (Cowles 1956, Klauber 1972), the biomechanics involved in rattle- snake saltation is likely different fiom that in B. caudalis. Here we report an instance of escape saltation in a wild C. s. scutuhitus. Although our observations are anecdotal, we believe they offer important insight into the biomechanics of rattlesnake jumping. On 4 September 1993 we observed an un- usual flight behavior by a wild C. s. scutulatus in the Hualapai Mountains, Mojave County, AZ. When approached, the snake lunged forward using its tail as the origin of force. This lunge was powerful enough to cause the snake's en- tire body to lift off the ground (Fig. 1). Actual forward movement from this "jump" was mini- mal, and the snake recoiled into a series of tight S-cun'es and jumped again. This type of saltation was observed a total of four times. None of the four jumps were directed toward a person, and the snake s mouth appeared to remain closed. Of the four types of snake locomotion, this jumping behavior could only be accomplished using concertina, in which the tail is the main point of force during foward movement. Klau- ber (1972) noted that rattlesnakes use conceiUna movement for slow progression in open areas and where restraints are involved (i.e., smooth '195 Wf.st 200 Ncirtli, Logan, UTS4321-.390.5. ^Venoni Researcli Laboratoir, Veterans Administration Mc-dical Center. Salt Lake Cit\. UT S414S-151H. 282 1995] Notes 283 :jt~- •'J^r^^m-^ Fig. 1. Crotalus s. sciitulatus exhibiting escape saltation. Photograph taken just before the tail left the ground. surface or narrow channel). Neither of these cir- cumstances was apphcable to this particular situation. The use of concertina locomotion rather than sidewinding as a basis for jumping in this snake is understandable as an anti- predator response. Crotalus s. sciitidotiis typi- cally utilize quick seipentine locomotion during flight; Klauber (1972) noted they are rather clumsy sidewinders. By reducing the number of pressure points to one (e.g., the tail), die snake changed from serpentine to concertina loco- motion. Whether this change is an effective use of energy remains to be tested. However, given the short distance the snake traveled, it would appear the relative energy cost would be high. Acknowledgments We thank Louis Porras, William Lamar, and Alejandro Solorzano for the information they provided about Atropoides niimmifer. James Glenn offered the financial assistance that made these observations possible. Literature Cited Armstrong, B. L., and J. B. Murphy. 1979. The natural histoiy of Mexican rattlesnakes. University' of Kansas Museum of Natural Histor\' Special Publications 5: 1-88. CowLES, R. B. 1956. Sidewinding locomotion in snakes. Copeia 1956: 211-214. Cans, C. 1974. Biomechanics; an approach to vertebrate biology. University of Michigan Press, Ann Arbor 261 pp. Cans, C, and H. Mendelssohn. 1971. Sidewinding and jumping progression of vipers. Pages 17-38 in A. de Vries and E. Kochva, editors. Toxins of animal and plant origin. Gordon and Breach, New York. Gasc, J.-P 1994. Locomotion. Pages 60-75 in R. Bauchot, editor, Snakes: a natural histoiy Sterling Publishing, New York. Klauber, L. M. 1972. Rattlesnakes: their habits, life histo- ries, and influence on mankind. University' of Cali- fornia Press, Berkeley. 1536 pp. Received 14 September 1994 Accepted 29 November 1994 Great Basin Naturalist 55(3), © 1995, pp. 284-285 A TRAP FOR BLUE GROUSE Eric C. Pelreii' and Jolm A. Crawford' Key tconls: Blue Grouse, Dendrauai^iis ()l)sciiriis, interception trap, Ore' 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; Wilford M. Hess, Botany and Range Science; Richard R. Tolman, Zoology. All are at Brigham Young University. Ex Officio Editorial Board members include Steven L. Taylor, College of Biolog>' and Agriculture; H. Duane Smith, 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 1995 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 Copyriglit © 1995 by Brigham Young University ISSN 0017-3614 Official publication date: 31 October 1995 10-95 750 15818 The Great Basin Naturalist Published at Provo, Utah, by Brigham Young UwivERsiri' ISSN 0017-3614 Volume 55 31 October 1995 No. 4 Great Basin Naturalist 55(4), © 1995. pp. 287-303 CLASSIFICATION OF THE RIPARIAN VEGETATION ALONG A 6-KM REACH OF THE ANIMAS RIVER, SOUTHWESTERN COLORADO Gillian M. WalforcU and William L. Baker^-^ Abstract. — Riparian ecosystems are important components of landscapes, particularly because of their role in biodi- versitx'. A first step in using a "coarse-filter" approach to riparian biodiversity conservation is to determine the kinds of riparian ecosystems. These ecosystems vaiy substantialh' in plant species composition along a single river reach, as well as between rivers, and yet the river-reach scale has received little attention. We sampled the vascular plant composition of 67 contiguous patches of riparian vegetation along a reach of the Animas River, in southwestern Colorado's San Juan .Mountains, that is relatively undisturbed by human land uses. Using cluster analysis and detrended correspondence analysis, we identified eight riparian community' types along the reach. Using a new technique, we combined overstory size-class data and understoiy cover data to identify community types. The eight community types, which are in part the products of past floods, are spatially arranged along the reach in relation to variation in valley morphology, tributaiy location, and geomorphic landforms. These eight community types do not necessarih' represent successional stages of a single potential vegetation t\'pe. This study at the river-reach scale suggests that sampling and analysis, as well as con- sei-vation, may need to be tuned to the scale of patchiness produced b> flood disturbances in the riverine landscape, since xegetation varies significantly at this scale. Key words: riparian vegetation, Rocky Mountains, Colorado, multivariate analysis. Riparian vegetation provides several impor- will be preserved (Hunter 1991, O'Connell tant functions in landscapes, and riparian com- and Noss 1992). munities have thus been a focus for conserva- Classification of vegetation communities is tion. Riparian vegetation contributes to water an essential first step in implementing this quality, stream bank stability, and healthy fish coarse-filter conser\'ation approach, but classi- habitat (Johnson et al. 1985, Malanson 1993). fication of riparian vegetation in the Rocky Riparian vegetation also provides cover and Mountains is incomplete. The montane and forage for wildlife that is particularly important subalpine riparian zones of Colorado's west- in the arid portions of North America (Knopf ern slope have been classified (Baker 1989). 1985). The idea of a "coarse-filter ' approach to Riparian community type classifications for biodiversity conservation is that by preserving U.S. Forest Service lands in Utah and parts of viable communities, associated species also Idaho and Wyoming are available (Mutz and ^Wyoming Natural Diversity Database, The Nature Conservancy, 1604 Grand A\e., Laramie, \VT 82070, USA. "Department of Geography, University of Wyoming, Laramie, \VT 82071, USA. "^ Author to whom correspondence should be addressed. , 287 288 Great Basin Naturalist [Volume 55 Gnihani 1982, Youngliloocl ct al. 1985a, 19851), Padgett et al. 1989 ). A riparian-wetland classi- fication and key were produced for Montana (Hansen et al. 1991), and other classifications are available for small parts of the Rocky Mountains (Mutz and Graham 1982, Cooper and Cottrell 1990). Methods of riparian vegetation classification vary with the researcher and project goals. The U.S. Forest Service classifies forest eco- systems into "habitat t\'pes" based on potential climax vegetation (Daubenmire 1952, Pfistcr and Arno 1980). Climax vegetation represents the stable, self-perpetuating community pre- dicted on the basis of climate, topography, and soils in the absence of distmbance. Often, how- ever, riparian vegetation never reaches climax due to frequent floods (Campbell and Green 1968). Therefore, some researchers classify vegetation into "community types" according to existing structure and composition without reference to successional stage following dis- turbance. This approach, however, is not con- cerned with whether vegetation consists pri- marily of native or e.xotic species or has been disturbed by human land uses (e.g.. Young- blood et al. 1985b, Padgett et al. 1989). Another classification approach (1) recognizes that cli- max vegetation is seldom reached due to nat- ural disturbances, but focuses on the more mature successional stages; and (2) empha- sizes that classification of "natural vegetation," vegetation that is as free as possible of exotic species and the effects of human land uses, provides essential information for effective biodiversity conservation (Baker 1989). Vegetation types classified using either a habi- tat type approach or Baker s approach are referred to as "associations." Even if there is only one association along a river reach, there may be several community types. Riparian community types along a river reach comprise a complex which Winward and Padgett (1989) name on the basis of the most prominent community type plus geographical features describing where it occurs. This spa- tially complex mosaic of community types cre- ates difficulties for classification, but the diversity of communities is an important com- ponent of liiodiversity (Hunter 1991). An additional problem is that classification may use only overstory species, or it may be based on the entire flora. Classification tradi- tionally uses one technique or a combination of tcclmi(|ues including subjective grouping, evaluating and scjrting of stand tables, cluster analysis, or ordination (Whittaker 1962, Gauch 1982). However, the dominance of the over- story in forests may skew mathematical analy- ses that use the entire flora (Padgett et al. 1989). In northern regions, where the under- stoiy flora often is more sensitive to environ- mentiil variation than is tlie overstoiy (Whittaker 1962), quantitative techniques that give equal weight to the undcrstory and overstory may not be ideal. Along the Animas River in southwestern Colorado s San Juan Mountains, we investigat- ed variation in plant species composition of contiguous patches of riparian vegetation. We classified riparian vegetation along a 6-km reach using both understory and overstory vegetation. Goals of this paper are to identify community types found along the river reach, to describe the community types in relation to their environment, and to explain a new approach to balance the use of both overstory and understoiy vegetation data in quantitative classification. This approach, we suggest, may lead to community classifications more useful for consei"vation and management. Study Area The Animas River starts in the San Juan Mountains of southwesteiTi Colorado and flows south to the San Juan River in New Mexico. The study area is approximately 40 km north- east of Durango in LaPlata County (Fig. 1), along a continuous 6-km reach of the Animas River between 2430 and 2550 m in elevation. This is one of the least disturbed montane river reaches in western Colorado (Baker 1990). A narrow-gauge railroad track and a wilder- ness access trail occur along the reach, but the reach has probably never been grazed b\' cat- tle or sheep. There are some silver and gold mines upstream. The Animas is an unregulated gravelbed river with a mean annual peak flow of 145 m^s"!. Within the stud\' reach the river has a mean gradient of 0.0193, a mean channel widtli of 34.3 m, and a mean channel depth of 3.6 m. The river is entrenched in a deep canyon sur- rounded by the Needle Mountains. The valley, varying in width from less than 100 m to about 400 m, is lined with alluvial deposits fomied fi-om Precambrian granites in the north and south 1995] Animas Riner Riparian Vegetation 289 Community Types Populus angustifolia I Alnus incana Populus angustifolia I Agrostis scabra Populus angustifolia - Pseudotsuga menziesii I Pyrola asarifoUa Picea pungens - Populus angustifolia I Antennaria parvifolia Picea pungens / Alnus incana I Equisehtm arvense Picea pungens- Populus tremuloides I Mahonia repens Pseudotsuga menziesii I Acer glabrum I Pyrola asarifoUa Populus tremuloides - Pseudotsuga menziesii I Bromus ciliatus \:^:y\ Not sampled Animas River fi.S 1.0 kilometer Fig. 1. Animas River study area and its location in Colorado. Patches are shaded according to their comniunit\ t\pe and numbered for identification. sections of the study area and from Precambrian tion to floods and climatic fluctuations (Baker gneiss and schist in the central section of the 1988, 1990). Baker (1988) identified 57 vegeta- study area (Osterwald 1989). tion patches having distinct boundaries recog- nizable on aerial photographs and on the Methods ground. This patchiness is largely the result of tree regeneration after past floods (Baker 1990). Previous work on the study reach focused on In the field we refined the boundaries of some the structure of riparian tree populations in rela- of the 57 patches and identified new patches 290 Great Basin Naturalist [Volume 55 for a total of 67 patches available for sampling. Patches were mapped on aerial photographs in the field, then digitized and rectified using the GRASS geographic information system (USA-CERL 1991). A final map of patch loca- tions and vegetation (Fig. 1) was produced with ATLAS Draw (Strategic Mapping 1991). Vegetation Sampling and Environmental Data Within each of the 67 patches, one 20 X 50-m plot (0.1 ha) was subjectively placed par- allel to the river to obtain a representative sample of herbaceous and woody vegetation. Methods of herbaceous vegetation sampling followed Peet (1981). Percent cover of each vascular plant species (except trees) present was estimated, during mid-growing season, in 25 contiguous 0.5 X 2-m quadrats along the 50-m center line of each plot. Percent cover of 0-10% was estimated to the nearest 1%; per- cent cover of 10-100% was estimated to the nearest 5%. Species located during a sui-vey of the plot, but not found in the quadrats, were assigned 0.1% cover. Nomenclature follows Kartesz and Kartesz (1980). Baker (1988) collected data on diameter-at- breast-height (dbh) size classes of tree species in his original 57 stands. We added to this data set by tallying trees (>2.5 cm dbh in 10-cm classes), saplings (<2.5 cm dbh and >1 m tall), and seedlings (<2.5 cm dbh and <1 m tall) of each species in plots of the 10 addition- al patches as Baker had done. Increment cores were extracted from the bases of 5-15 of the largest trees in each patch for estimating patch age. Ages of the largest trees tend to be simi- lar, reflecting a common origin following floods (Baker 1990). Each patch was assigned to a 10-year age class according to the maxi- mum age of the 5-15 cored trees. Age zero is A.D. 1990. A set of environmental variables was mea- sured in the field in each patch. Patch slope was measured using an Abney level and sur- vey rod. Aspect of the patch was measured in degrees with a compass. We surveyed the dis- tance to the channel and the height above the channel using the rod, level, and a distance meter. Patches were identified as either on terraces or depositional bars. The depositional bar is the lowest prominent feature higher than, but within, the channel bed, while terraces are older, higher fluvial landforms (Osterkamp and Hupp 1984). At eveiy 0.5 m along the 50- m center line of each plot, we measured the intermediate axis of the surface particle at that point and assigned it to a size class, in a varia- tion of the Wolman (1954) technique. Later, using Rodriguez's (1986) MOMENTS pro- gram, we calculated mean size, %<1 mm, %<2 mm, and sorting value for each patch. Soil samples of the upper 15 cm of the profile were taken in onK' 20 of the 67 patches, due to the cost of chemical analyses. These 20 sam- ples spanned the spectrum of patch ages and floristic and environmental variation. Samples were later analyzed for standard fertility (organic matter, pH, N, F, and electrical con- ductivity) by the University of Wyoming Soil Testing Lab. Quantitative Analyses We used the SPSS/PC + cluster analysis program (SPSS 1990) to determine groups of patches similar in overstory and understory vegetation composition (Romesburg 1984). After experimenting with several clustering methods, we identified the BAVERAGE method (aver- age linkage between groups) and the cosine distance measure (angular separation of vec- tors of variables) as the best clustering combi- nation. This combination emphasizes relative abundances within a plot and de-emphasizes absolute abundance differences between plots (SPSS 1990). Species composition data were also ordi- nated by detrended correspondence analysis (DCA) using CANOCO (Canonical Community Ordination), a multivariate statistical program for applications in community ecology (Ter Braak 1988). Correspondence analysis pro- vides a geometrical representation of the rela- tionships among samples and species in a data set and identifies the dominant trend of \ aria- tion in community composition. Initial ordination and cluster anal> sis of the combined overstory and understory data set resulted in groupings primarily reflecting just the high cover values of overstory tree species rather than the joint pattern of both overstory and understor)' species. To counteract this, we analyzed the overstoiy tree species size-class data and the understory shrub and herbaceous species cover data separately, and then merged the two results. The overstor\ size-class data of each plot were first clustered; then the per- cent cover data of understorv shrubbv and 1995] Animas River Riparian Vegetation 291 herbaceous species in each pk)t were clus- tered. These understory cover data were also ordinated using DCA. Final classification groups were the result of intersections of over- story cluster groups with understory cluster groups overlain on the understory DCA ordi- nation diagram. We calculated the mean value for several environmental variables in each community type. Environmental variables we used are those found to be important to vege- tational variation along the reach based on a separate, but related, gradient analysis (Baker and Walford 1995). Classification groups referred to here are "community types" because they represent existing rather than potential natural vegeta- tion. Each community type is based on the entire flora but is named based on tlie dominant species in the overstoiy and the dominant or most diagnostic indicator species in the understory (Mueller-Dombois and Ellenberg 1974). When there are co-dominants in a layei; both species are included in the name and are separated by a hyphen. Results Classification The cluster analysis and DCA ordination of the plot understoiy cover data (grasses, forbs, and shrubs) suggested four major groups and one outlier (Fig. 2). The similarity cut level was kept coarse so that overstory cluster groups could be incoiporated later. This specific level was chosen after considering alternative cut levels at slightly greater or lesser similarity (Fig. 2). Groups A and B, for example, would become one group if the cut level were at a slightly lower similarity, yet these two groups are quite different (Fig. 2). Ordination of the same data set is represented by the DCA Axis 1 vs. Axis 2 ordination diagram (Fig. 3a). The distinctiveness of understoiy groups produced by cluster analysis is supported by the com- paratively distinct location of the groups on this ordination diagram. Understory groups identified by cluster analysis and ordination are compositionally distinct and occur in different environmental settings. Group A was dominated by Ainus incana and Eqiiisetum arvense. These patches were predominantly located on bars. Group B was located entirely on bars and had the fewest species of any group. Agrostis scahra was always Fig. 2. Understory cluster analysis dendrogram based on percent cover of herbaceous and shrubby species. Plot numbers correspond with patch numbers on the study area map (see Fig. 1). The dashed line indicates the simi- larity level at which miderston- groups were separated. present in Group B patches. The third group (C) was dominated by Rosa woodsii and Pijrola asarifolia. Patches of Group D are almost always on terraces and generally have the highest species richness. Mahonia repens is always present, and Rosa woodsii, Bromus ciliatus, and Onjzopsis asperifolia are usually well rep- resented. Overstory size-class data were clustered by the same method. Three overstory groups were identified at approximately 25% similarity (Fig. 4). Each of these major groups has mem- bers from at least three different understory groups. Group I is recognized by a dominance of Populus angiistifolia seedlings, saplings, and small trees (Table 1). All understoiy Group B members are found within diis overstoiy group. But other members of this overstory group have the understoiy of Groups A, C, D, or E. Overstoiy Group II is characterized by Picea pimgens of all sizes and larger P. angiistifolia 292 Great Basin Natur.\list [Volume 55 (Table 1). Most of its menilHMs ha\e an uiider- story of Groups A or D. The third overstory group tends to have a mixed canopy dominat- ed by all sizes of Pseudotsuga menziesii and small Abies voncolor (Table 1). Fopuhis tremu- loicles and Picea pungens are often present. Half of the members of understory Groups C and D have this mixed overstor)' composition. These overstoiy groups are indicated on the same DCA ordination diagram (Fig. 3b). Since this diagram represents the ordination of shrubby and herbaceous species in plots, and the understory composition varies within the overstoiy groups, it is not sui-prising that these overstory cluster groups are scattered within the ordination diagram. This suggests that the understory is to some extent independent of the overstoiy. Final classification groups resulted from the intersection of the understoiy groups and overstoiy groups overlain on the ordination diagram (Fig. 3c). This results in eight final classification groups plus two single-member groups and one outlier plot. The symbol for each classification group is a combination of its overstoiy cluster group (I, II, or III; Fig. 4) and its understory cluster group (A, B, C, or D; Fig. 2). Groups are presented in an age sequence within their overstoiy group, from youngest (IB) to oldest (HID). Community Types The following paragraphs summarize tree composition and structure, understoiy species composition, and environment of each of the eight community types (Tables 1, 2). In prior analyses (Baker and Walford 1995) the gradi- ent controlling spatial variation of the shrubby and herbaceous vegetation mosaic was found to be age and disturbance related. Variables most affected by disturbance events are illus- trated for each community type (Fig. 5). A map of the patches and their community type is in Figure 1. [IB] Populus (ingustifolia / Agrostis scahra. — The 14 patches constituting this community type are found on bars close to the channel in both height and distance (Fig. 5). Of the eight major types, this type was most recently estab- lished (mean age = 31 years) and has the largest mean surficial sediment size. Its soils have low organic matter. Patches of this type are most common in the middle parts of the study reach (Fig. 1). This type is characterized by an abun- (b) I , oi ■* III** , Of o Overstory Cluster Groups o Fig. 3. Ordination diagrams obtained by detrended cor- respondence analysis of data on percent cover of shruliby and herbaceous species in the plots: (a) plots are coded according to their understory cluster analysis group (see Fig. 2); (b) plots are coded according to their overstory cluster analysis group (see Fig. 4); (c) plots are coded according to their final community type, based on the intersection of understory cluster groups with overstoiy cluster groups. dance of P. angiistifolia seedlings and saplings (Table 1) and sometimes small to medium-size trees. Picea pungens seedlings are almost always present and are sometimes abundant along with saplings and small trees. Small Pseudotsuga menziesii and Abies concolor may be present. Herbaceous vegetation is veiy sparse (Table 2). Graminoids dominate the understor)' with both Agrostis scabra and either Trisetem niontanuni or T. spicatum always present. Epilobiwn lati- foliiun often occurs in significant amounts. [lA] Popuhis (ingustifolia I Alnus incana. — Patches of this type are on average 10 years older than those of Type IB (Fig. 5). This com- munity type is found on bars slightly higher above the channel than those of IB. Surface sediment sizes are diverse, but soils have little organic matter. Like Type IB, these patches 1995] Animas River Riparian Vegetation 293 Fig. 4. Overstoiy cluster analysis dendrogram based on size-class data of tree species. Plot numbers correspond with patch numbers on the study area map (see Fig 1). The dashed line indicates the similarity level at which overstoiy groups were separated. are most common in the middle part of the study reach (Fig. 1). Populus angiistifolia char- acterizes stands of this txpe; many small and some medium- size trees are present along with abundant seedlings and saplings (Table 1). Picea pungens and Pseudotsiiga menziesii seed- lings and saplings are usually present. The understoiy composition distinguishes this type from IB (Table 2). Shrubs are more common. Alnus incana is always present and Salix drum- mondiana is usually present. Agrostis scahra is occasionalh' present in minor amounts. [IC] Populus angustifoIia-Pseudotsuga men- ziesii I Pyrola asarifoUa. — This small commu- nity type comprises two patches on bars and one on a terrace, with an average surface par- ticle size <10 mm and soils with low organic matter (Fig. 5). Patches in this community type are scattered along the study reach (Fig. 1). The largest trees and most abundant seedlings of this type are P. angustifolia (Table 1). Pseudotsuga menziesii are always present as seedlings dirough medium-size trees. Medium-size Picea pimgens or Abies concolor may also be pres- ent. Pinus strobiformis seedlings or saplings are always present in this type. The under- stoiy of this type is not dense, and all but one of the understory species have cover values <0.8% (Table 2). Three shrub species occur in small amounts. Pyrola asarifoUa is always pres- ent in the highest amount of any understory species. [ID] Piceo pungens-PopuIus angustifolia I Antennaria parvjifolia. — Three terrace patches and two patches on bars make up this commu- nity tyi^e, which occurs on surfaces < 1 m above the channel that have soils with low organic matter content (Fig. 5). Patches in this type are scattered along the study reach (Fig. 1). They have strong similarities in overstory composition and weak ones in understoiy composition. Picea pungens and P. angustifolia are the largest trees of this type and are always present as seedlings, saplings, and small trees (Table 1). Abies concolor and Pseudotsuga menziesii seed- lings can always be found. The lack of a con- stant understory is reflected in the low simi- larity level at which patches 65 and 71 are joined in the dendrogram (Fig. 2). Six species are present at 80% constancy, Antennaria parvifolia having the greatest mean cover in the type. No single herbaceous or shrubby species is present in all five patches of this type, but in general there is much more herbaceous and shrubby vegetation present than in types lA, IB, and IC (Table 2). [IIA] Picea pungens I Alnus incana / Equi- setum arvense. — Patches of this community type occur in more persistently moist areas. They span several age classes, can be found on bars or terraces, and have developed finer sur- ficial sediments than might be expected for their age class (Fig. 5). Their soils typically contain only a little more organic matter than soils in patches of Type I. Patches in this type are scattered along the study reach (Fig. 1). Picea pungens seedlings, saplings, and small to medium-size trees as well as P. angustifolia of various sizes characterize the type (Table 1). Few other tree species occur, although Pseudo- tsuga menziesii may be present in small amounts. Alnus incana and Salix drummondi- ana are the dominant shrubs of this type, both occurring in greater amounts here than in any other types (Table 2). Equisetum arvense is always present in substantial amounts. Sedges 294 Great BasiiN Naturalist [Volume 55 i * § c^ 1- K io ID I- \n> K u r- ~ o 3 ^ C CO o 1 =*t oj CO -r s -r -r yi i * i * i^ * i =*t 5 =**: 5 =tfe 10 lO -- O ID ac JO ;o ^H -f cc oi ;c -H lo t~^ oi o o o M Ol Ol Ol oi — ^ d o o d d ;r) o cr; o p o o t~^ CO lo -r d d d I- o o o o o d d d d d d t- d d d d d d -r t^ o o o o -r '^ di d (6 d> di 2 m j- p c c j: a— c ^cococo '"-' ^T^plScotMCOrrcO £ O ^'^ CO CO CO 7' •~ X i^cO^tNcoA cc -H CO io -r — ; lo ^ -t c-i —H d -^ d 00 CO CO CO CO CO O -i d '>\ d d d d- O CO I^ OJ CO -C d CO t- d CO ^ CO oi --r cc -r p -f ^ -t t--^ d c^i d d CT) -* CO 00 oi p cc -r o oi oi -r ^ in -t CO oi O CO CO CO t^ o o i-^ —i -H — * -H d d t— DO t- O p p p in t-^ in ^ d d d 05 -H -H p p p p t^ 00 00 d d d d CO §i bt c P S £ p SiS C ^COCOCO ^ ^'^ .^ CO CM c^ "^ CO S 5j ^ 1 CO CO c^ ' j; 3^ -y! CO -^ ci CO A 35 O O 05 00 05 05 CO CO in CO p CC Ol o o d d d d o o o o o o o d d d d d d d o o o -r fM o o d d d d d d d t^ O O CO CO o o ^ d ci d d d o r^ 1-- i^ t^ o o o oq t~- c^ CO o o d d d d d d d C^ O fM o o o o — < d d d d d d £;= '-. '^cococo '-' s.-a i= CO cj o o c> o o o o o o o o -r o o -* o o o CO —5 -r -r d d d O CO o o o o o c4 cN d d d d d o o o o o o o o o o o o o o di d> d> d> di d> d> — I c^ o o o o o -r -H o o o o o — < d d d d d d o o CO in o CO in p p CO 00 00 O CO 'H oi oi iri t--^ ^ ^ (N ^ j ej CO p p d -* in CO d d d [£;in^a50ooo .—.oicg^dddd s~— corMco-fco f^ bjo ^ c c c ^ -C::C^?cjcjoc CO— - '^cococo "^ s"0— coe-ico-fco S 1; ^ I CO CO CO T ;s,5; Jico^cgcoA t2 -X t/j ^ C C C r- C "^ .1^ m oi '^ "T c^ i; n: tir' I ro r^ r^ ^ ^ -^ Cr CO ^ r>i CO A 296 Great Basin Naturalist [Volume 55 Table 2. Mean |)c'icfnt cover \aliies lor main species in coiniiiinnt\ t\pes (includes species with at least 19( cover in an\ plot). Species with double underlined co\er \alues had 100% constancx in that coniniunit) t\pe. Species with single underlined cover values had >S0% and <100% constancy in that conunnnity type. C^onnnunitx t\ pe IB lA IC ID IIA III) IIIC: HID — .5.1 — 0.6 1.0 1.2 — 2.4 — 3^ 0.4 |J 0.4 02 — M M M 1.6 14 0.5 M M M — 0.8 — 0.2 0.4 0.5 Shrubs Acer ^hihniiii Ton: — — — 1.7 — 0.2 3^ Alntis incana (L.) Moench ssp. temiifolia (Nun.) Breitnng OJ. 0^ — 0.1 5A) Amelanchier alnifolia (Nutt.) Nutt. — — O.I 0.6 — Corniis sericea L. — — — — — Jitiiipenis communis L. — — 0.1 2. .3 — Lonicera involiicrata (Richars.) Banks ex Sprang. — — 0.1 0.1 0.1 Mahonia repens (Lindl.) G. Don — — — L3 — Prumifi vir^inia)i=S) II A (n=S) II D III c 4«i » H.T 10 15 .■■■T.l.lf. ■■»..■ .I..f..>i 10 15 .. .lilMII UuU 4...^ 10 15 III D ^^ (r>=n) -Wt-J 10 15 (a) Age Class (lOyr) io_ 5 J f Ipp ■ ¥ 5 J t#- ^ 2 :^ S « S !!! § ^ a 1 oj 5 J 0 . u* ^ 2 !5 S S S !i? 9 ^ 8 10. 5- 0. T I I I I I I I I I "?2£S!(i8;!?§^S 10- 5 J 0. lOJ 5 J 0. "?2£S!i38^9«8 °'"2£S!^SSS^ (b) Surface Sediment Size (mm) 10_ 5 J 0. w ■ yip lO rN I/) c*) lO ^ O - ^n CM lO O 4^0^ ^#- -^^#- 4* , ' - ' N ' n ' Q in I lO . I/) I tO - - M tN n n (c) Height Above Channel (m) 0.4 % I n = 3 I (n = 0) 0.5 % ( n = 2) 1.0 % (n = 1 ) 6.7 % (n = 2) 4 4.0 % (n = 2 ) 2 0.3 % (n = 61 Mean (d) Organic Matter Fig. .5. En\'iionmental attributes within each community type (see Fig. vertical axis. 1). Histograms (a-c) show the frequenc\' on the age-class span as patches in Type IID, and have the largest mean organic matter content of community types along the reach (Fig. 5). The overstory is a mixed forest similar to Type HID, but the understory is not as rich. Patches in this type are restricted to the lower one- third of the study reach (Fig. 1). Pseudotsuga menziesii, Popiilm angustifolia, and usually Abies concolor are the large trees of these patches (Table 1). Popiilus angustifolia seedlings or sap- lings are rare. Regeneration appears strongest in Abies concolor and P. menziesii (Table 1). Acer glabrum is the dominant shrub usually associated with lesser amounts of A/nu,s 300 Ci{i' AT Basin Naturalist [Volume 55 incaruL Priiiius lir^iiiidiid. and liosa uoodsii (Table 2). Pyrola asarifolia is the most preva- lent forb with cover values averaging 18%. Geranium richardsonii, Smilacina stellata, Haplop(i])])us parnji, and Arlctiiisia franseri- oides al\va>'s occur. Onjzop.sis a.spcrifolia is the dominant graminoid. with Broiniis cilidfus in lesser amounts. [IIIDJ Fopidus trcmidoides-Pseudoisuga nienziesii I Bromus ciliatiis. — Examples of this community' type are found on terraces an aver- age of 1.75 m above the channel (Fig. 5). Surface particles are predominantly <1 mm, while soils have about 20% organic matter content (Fig. 5). The 11 patches comprising this type have the oldest average age and are most com- mon in the lower one-third and upper one- third of the study reach (Fig. 1). This group of patches is a mixed forest type, with the dens- est underston' of all the types. Popuhis treinu- loides and P. menziesii tend to be the largest and the most abundant trees of these patches. Some patches have very large Populus angusti- folia as well. Abies concolor seedlings and sap- lings are iilways present, sometimes in veiy large numbers. A few Picea pungens of various sizes usually can be found. Mahonia repens\ Rosa woodsii, and Jiiniperus conununis are co-domi- nant shrubs. Five Erigeron species were found in the 1 1 patches of this type, with one to three species present in each patch. Antennaria parci- folia and Geranium richardsonii had high cover values in most patches. Dominant grasses were Bromus ciliatus and Orijzopsis asperifolia, their quantities being distinctive from any other type. Discussion The New Classification Technique Ecologists working in northern climates have long recognized that individual stratal layers (e.g., tree and shrub) or "synusia" within forest communities may be distributed some- what independently and may not all have the same value in distinguishing vegetation units (Whittaker 1962). Classification approaches of the northern European schools, such as the "sociations" of the Uppsala school and "site- types" of Cajander (Whittaker 1962), empha- size that lower strata often are more useful in classification, although the tree layer is of some importance as well. Yet, conunon multivariate techniques for analyzing \'egetation data ignore the synusial affiliations of the species in a community (e.g., Gauch 1982). Since cover values of overstory tree species are often large relative to cover values of understoiy species, overstoiy species may prevent understory species from having much influence on the outcome of multivariate analyses (e.g., Padgett et al. 1989). Moreover, these techni(|ues do not take advantage of the different kinds of data that are useful in describing the structure of different synusia. For example, it is much easier and more useful to obtain size-class structure data than cover data for overstory trees in forests, as size-class data can indicate tree composition and abun- dance as well as population structure (e.g., regeneration status). The value of both kinds of data in forests has long been recognized (e.g., Pfister et al. 1977). Yet, size-class data and cover data are incompatible and cannot both be used readily in a single multi\ariate analysis. The technique used here allows an equal consideration of both the understoiy and overstoiy data and data of different types from different synusia. Community types that are produced have homogeneous tree populations combined with homogeneous understories. Community types identified using this tech- nique can be useful in consei-vation and man- agement. The eight communitx' t\pes reflect major variations in vegetation and environment along the reach. A preserve could be designed, shorter than the full length of the reach, that contained all eight connnunity types; it is like- ly that much of the floristic variation along the reach would then be protected. Moreover, the community types should be useful in manage- ment because they are (1) functionalK' homo- geneous, in the sense that tree populations within a type might respond similarK to dis- turbances, and (2) environmentalh' sensitive, in the sense that the finer en\'ironinental dis- crimination of understory synusia common in northern regions has been incoiporated. The Community Types in a Regional Context Community types identified in this study have not been widely reported. This is proba- bly due to the river-reach scale of the study and the tendency to exclude \'eiy >'oung vege- tation in developing regional classifications. This is also one of a very few river systems in die 1995] Animas River Riparian Vegetation 301 southern Rocky Mountains with a complete mosaic of riparian vegetation relatively free from human land uses; thus, there have been few other opportunities for this kind of studv'. Our Type II IC is very similar to Baker s (1989) Abies concolor-Picea pungens-Popidus angiistifolia / Acer glabrum association, previ- ously found along the Animas River and the San Juan River, as well as in northern New Mexico (DeVelice et al. 1986). Baker collected data from within the study reach, which ex- plains the similarity of his association with our Type IIIC, but his goal was to classify vegeta- tion regionally based upon the similarity of the more mature vegetation patches along sepa- rate rivers. Baker did not sample the mature stands containing Popiihis tremidokles that are found in our Type IID and HID, thinking they were earlier successional stages of our Type IIIC. However, the age-class data (Fig. 5) sug- gest that stands within Types IID and HID are just as old as or older than those in Type IIIC. Thus, although it may be a necessary shortcut for regional classification efforts, sam- pling and classifying only the mature vegeta- tion may result in errors if the successional se- quence along a reach is not clearly understood. Sampling and Classification of Riparian Vegetation Complexes In riparian areas, and perhaps anywhere vegetation classification is being approached, it is important to sample and classify not only mature vegetation stands but younger stands as well. The diverse patch structin-e along rivers may only reach a homogeneous mature com- position similar to that in the older patches if the fluvial disturbances that have produced the mosaic are controlled. Moreover, younger stands may not all be leading to the same mature community; there may instead be more than one serai sequence. This spatial and temporal complexity at the river-reach scale compounds the difficulty of developing regional classifications. One solu- tion to this problem is to adjust the scale of sampling to the scale of patchiness produced l^y the primaiy ecological processes (e.g., fires, floods) in a particular landscape. An area such as this free-flowing river requires fine-scale sampling as there is a fine-scale mosaic pro- duced by disturbances and geomorphic varia- tion. A river with less geomoiphic complexity or a coarser, more infrequent flood-produced patchiness may require only a coarse sampling focused on the more mature vegetation. Spatial Variation in Vegetation Along the Reach The sampling and vegetation analysis sug- gest that substantial landscape diversity is pro- duced by floods and geomorphic variation along this reach of the Animas River (Fig. 1). The spatial aiTangement of this diversity is con- trolled in part by location of tributaries and width of the valley floor, both of which influ- ence how and where floods create new patches. Needle Creek flows into the Animas River in approximately the middle of the study reach (Fig. 1). Smaller tributaries enter above and below this point, but none carries as great a volume of water. The valley floor also widens approximately 0.5 km below the entry of Needle Creek. The first four community types (lA, IB, IC, and ID) with Populus angiistifolia in their over- stoiy are found primarily in this section at the outer river curves or mid-channel where scour- ing is greatest. The wet environment IIA patches also are found in this wider section, often away from the main channel on side channels that dissect major patches. None of the largest trees is found in this middle section of the reach. In contrast, community types with a more mature overstory are more common in the lower one-third and upper one-third of the study reach. These parts are narrower and have fewer substantial tributaries. Many of the mature vegetation patches are located on ter- races quite high above the channel in these parts of the reach. Conclusions This study of riparian vegetation on the river-reach scale revealed considerable spatial and temporal complexity. Flood disturbances, modulated by variation in valley morphology and tributary location, have created distinct patchiness in the vegetation. A new technique, based on both overstoiy and understory species, offers an improved quantitative method for identifying community types. If classification is to be used effectively to aid in consenation, greater attention to younger, less mature stands of vegetation may be needed. These young stands are a major component of the biodiver- sity on the river-reach scale and can represent 302 Great Basin Naturalist [Volume 55 serai stages of new vegetation associations unlike the association represented by present mature stands. Spatial complexity' along a single river may make the development of regional classifications, based on many ri\ers, more dif- ficult. However, regional classifications can still be completed, and will be more valuable, if sampling efforts are tuned to the scale of patchiness and complexit> along river reaches. Acknowledgments This research was completed with funds from the Ecological Research Division, Office of Health and Environmental Research, U.S. Department of Energy (Grant No. DE-FG02- 90ER6()977). This support does not constitute an endorsement by DOE of the views expressed in this article. Comments of Sherman Swanson and an anonymous reviewer improved the manuscript. Literature Cited Baker, W. L. 1988. Size-class structure of contiguous riparian woodlands along a Rocky Mountain river. Physical Geography 9: 1-14. . 1989. Classification of the riparian vegetation ot the montane and subalpine zones in western Colorado. Great Basin Naturalist 49; 214-228. . 1990. Climatic and hydrologic effects on the regen- eration of Populus angustifolia James along the Animas River, Colorado. Journal of Biogeography 17: 59-7.3. Baker, W. L., and G. M. Walford. 199.5. Multiple stable states and models of riparian vegetation succession on the Animas River, Colorado. Annals of the Associ- ation of American Geographers 85; .320-.3.38. Campbell, C. J., and W. Green. 1968. Perpetual succes- sion of stream-channel vegetation in a semiarid region, journal of the Arizona Academy of Science 5; 86-98. Cooper, D. J., and T. R. Cottrell. 1990. Classification of riparian vegetation in the northern Colorado Front Range. Unpublished report. The Nature Consen'ancy, Colorado Field Office. Boulder, CO. 85 pp. Daubenmire, R. 1952. Forest vegetation of northern Idaho and adjacent Washington, and its hearing on concepts of vegetation classification. Ecological Monographs 22; .301-330. DeVelice, R. L., J. A. Ludwig, W. H. Moir, and F Ronco, Jr. 1986. A classification of forest habitat types of northern New Mexico and southern Colorado. US DA Forest Service General Technical Report RM-131. Rocky Mountain Forest and Range E.xperi- ment Station, Fort Collins, CO. 59 pp. Gauch, H. G., Jr. 1982. Multivariate analysis in commu- nity ecology. Cambridge University Press, Cambridge, UK. 298 pp. Hansen, R, K. Boggs, R. Pfister, and J. joy. 1991. Classification and management of riparian and wet- land sites in Montana. Draft version 1. Montana Riparian .Association, Montana Forest and Conserva- tion E.xperiment Station, School of Forestn; Univer- sity of Montana, Missoula. 478 pp. HlNTER, M. L., Jr. 1991. Coping with ignorance; the coarse-filter strategy for maintaining biodiversity. Pages 266-281 in K. A. Kohn, editor. Balancing on the brink of extinction; the Endangered Species Act and lessons for the future. Island Press, Washington, DC. JOFINSON, R. R., C. D. ZlEBELL, D. R. P.ATTON, R F Ffollkjtt, and R. H. Hamre. 1985. Riparian ecosystems and their management; reconciling con- flicting uses. US DA Forest Service General Tech- nical Report RM-120. Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. 523 pp. K.'\RTESZ, J. T., AND R. K.\RTESZ. 1980. A synonymized checklist of the vascular flora of the United States, Canada, and Greenland. University' of North Carolina Press, Chapel Hill. 498 pp. Knope F L. 1985. Significance of riparian vegetation to breeding birds across an altitudinal cline. Pages 105-111 in R. R. Johnson, C. D. Ziebell, D. R. Patton, P. F Ffolliott, and R. H. Hamre, editors. Riparian ecosystems and their management: reconciling con- flicting uses. USDA Forest Service General Tech- nical Report RM-120, Rock-y Mountain Forest and Range Experiment Station, Fort Collins, CO. Malanson, G. P 1993. Riparian landscapes. Cambridge University Press, Cambridge, UK. 296 pp. Mueller-Dombois, D., and H. Ellenberg. 1974. Aims and methods of vegetation ecology. John Wiley and Sons, New York. 547 pp. MUTZ, K. M., and R. Graham. 1982. Riparian community- type classification — Big Piney Ranger District, Wyoming. Unpublished report, USDA Forest Semce Region 4, Ogden, UT. 92 pp. O'CONNELL, M. A., and R. E Noss. 1992. Private land management for biodiversity consenation. En\'iron- mental Management 16: 43.5—450. Osterkamr W. R., and C. R. Hupp. 1984. Geomorphic and vegetative characteristics along three northern Virginia streams. Geological Society of America Bulletin 95: 109.3-1101. Osterw'ald, D. B. 1989. Cinders and smoke. A mile by mile guide for the Durango to Silverton narroyv gauge trip. Westei-n Guideways, Lakewood, CO. 142 pp. Padgett, W. G., A. P Youngblood, and A. H. Winward. 1989. Riparian community tyise classification of Utali and southeastern Idaho. Unpublished report R4- Ecol-89-01, USDA Forest Service Intermountain Region, Ogden, UT. 191 pp. Peet, R. K. 1981. Forest \egetation of the Colorado Front Range: composition and d>namics. Vegetatio 45: 3-75. Pfister, R. D., and S. F Arno. 1980. Classifying forest habitat types based on potential clima.x vegetation. Forest Science 26; 52-70. Pfister, R. D., B. L. Kovalchik, S. E Arno, and R. C. Presby. 1977. Forest habitat types of Montana. USDA Forest Service General Technical Report INT-34, Intennountain Forest and Range Experiment Station, Ogden, UT 174 pp. Rodriguez, J. 1986. User's guide to MOMENTS. Page 21 in H. R. Burger, editor. Personal computer softyvare for geological education. National Association of Geolog>' Teachers, USA. Romesburc;, H. C. 1984. Cluster analysis for researchers. Lifetime Learning Publishers, Belmont, CA. .334 pp. 1995] Animas River Riparian Vegetation 303 SPSS. 1990. SPSS/PC + . Statistical package for the social sciences. SPSS, Inc., Chicago. Strategic Mapping, Inc. 1991. Atlas Draw user's guide. Strategic Mapping, Inc., San Jose, CA. Ter Braak^C. J. K 1988. CANOCO— a FORTRAN pro- giam for ciinonical comniunit)' ordination by [partial] [detrended] [canonical] correspondence analysis, principal components analysis and redundancy analysis. Technical Report LVVA-88-02, Agricultural Mathematics Group, Wageningen, The Netherlands. Microcomputer Power, Ithaca, NY. 95 pp. USA-CERL. 1991. GRASS 4.0 user's reference manual. United States Army Construction Engineering Research Laboraton-, Champaign, IL. 280 pp. VVhittaker, R. H. 1962. Classification of natural commu- nities. Botanical Review 28: 1-239. WiNVVARD, A. H., AND Padgett, W. G. 1989. Special con- siderations when classifying riparian areas. Pages 176-179 in Proceedings — land classification based on vegetation: applications for resource management. US DA Forest Service General Technical Report INT-257, Intermountain Forest and Range E.xperi- ment Station, Ogden, UT. WOLMAN, M. G. 1954. A method of sampling coarse river- bed material. Transactions of the American Geo- physical Union 35: 951-956. YouNGBLOOD, A. P, VV. G. Padgett, and A. H. Winward. 1985a. Riparian community type classification of northern Utah and adjacent Idaho. Unpublished report, USDA Forest Sei-vice, Intermountain Region, Ogden, UT 103 pp. . 1985b. Riparian community type classification of eastern Idaho-western Wyoming. Unpublished report R4-Ecol-85-01, USDA Forest Service, Intermountain Region, Ogden, UT. Received 24 June 1994 Accepted 24 January 1995 Great Basin Naturalist 55(3), © 1995, pp. 304-3 14 ADl^n IONS TO KNOWLEDGE OF PALEOCENE MAMMALS FROM THE NOm II HORN FORMATION, CENTRAL UTAH Kichard L. C>itclli', Nicliolas J. Czaplewski', and Kenneth D. Rose^ Abstkact. — The distinctive but inadeejuately known Paleocene faunas of central Utah are significant in that they sample a time interval not well represented l)y secjucnces in other areas. New materials from the Wagon Road (late Puercan) and Dragon (earl\' Torrejonian) local faunas. North Horn Formation, provide additional information on the composition of the assemblages and systematics of included mannnal ta.xa. The proteutherian tPropalaeosinopa is recorded, for the first time, from the Wagon Road fauna, indicating a significant extension for the enigmatic family Pantolestidae, othei-wise first known from the Torrejonian. Associated premolars oi Aphnmorus sirnpsoni, a pentacodon- tid proteutherian from the Dragon fauna, indicate that the species is more distinct from its Torrejonian congener, A. fmudalor, than previousK' suspected. New materials of Desmatuclaenus hermaeus uphold the synonymy of this species with D. paracreodus and permit more adequate definition of the genus with respect to the arctocyonid Loxolophua and the phenacodontid Tetraclaenodon; because Desinatoclaeniis appears to share derived morphology with Loxolophu.s, we refer it to the basal condylarth family Arctocyonidae. The periptychid condylarth Haploconus, represented in the Wagon Road fauna by the geologically oldest described species of the genus, H. elachistus, is shown to be distinctive in the con- figuration of lower molars and premolars; H. elachistus appears to be more primitive than species known from the Torrejonian of New Mexico. Some features oi Haploconus are suggestive of the Conacodontinae, a distinctive clade of diminutive periptychids. Kcij words: Paleocene, North Horn Fonnation, Puercan, Torrejonian, Dragon local fauna. Wagon Road local fauna, Mainnuitiii. Paleocene mammals were first reported fi-om the North Honi Formation, Emery and Sanpete connties, UT, by Gazin (1938). Further field- work resulted in the recoveiy of additional taxa, interpreted as representing two faunas, from two main localities (Gazin 1939, 1941). In sub- sequent years, additional sites in the region have yielded finther specimens, including more taxa and a third faunal assemblage (Spieker 1960, Van Valen 1978, Tomida and Butler 1980, Tomida 1982, Robison 1986, Archibald, Rigby, and Robison 1983). Three assemblages are cur- rently recognized, the Gas Tank, Wagon Road, and Dragon local faunas (Robison 1986). On the basis of the latter two, a "Dragonian" land- mammal age was initially established (Wood et al. 1941). Later work, including magnetic stratigraphy and biostratigraphic comparisons, suggests that the Gas Tank and Wagon Road faunas are Puercan and the Dragon fauna Torre- jonian in age (Tomida and Butler 1980, Tomida 1981, Robison 1986). Archibald et al. (1987) tentatively assigned the Gas Tank to Pu2 {Ectoconus I Taeniolahis taocnsis interval zone). Wagon Road to Pu3 {Taeniolahis taocnsis I Periptyclws interval-zone), and Dragon to Tol {Periptijchus I Tetraclaenodon interval-zone). Both Pu2 and Pu3 are interpreted to occur within magnetic polarit>' chron 29N (Butler and Lindsay 1985); the Dragon faima is considered to lie within anomaly 27N (Tomida and Butler 1980). The Paleocene mammals of central Utah are of special interest in both temporal and geo- graphic contexts: they fall within a time inter- val not well represented elsewhere, and they lie geographically between the classic sequence of the San Juan Basin, NiM, and faunas from more northerly parts of the Western Interior (cf. Archibald et al. 1987: fig. 3.1). Mammals from the Paleocene of the North Horn Formation are not, in general, well known. We describe herein newly collected materials that provide further details on the moiphology and SNstem- atics of some of the included taxa. The approximate locations of the major mammal sites in the Paleocene part of the North Horn Formation, taken from data pre- sented by Gazin (1941) and Robison (1986), are given in Figure 1. The materials described OklaliDiiK. Mil )r\'atiinil llistiin and Drpartnicnt ofZiKiIdRv, University nfOkhihoma. Norman, OK 73019. ^Department (il Cell HiclciKy ami Anatomy, Johns llopknis I niversil\ Seliool of Medicine. 725 North \\c 304 1995] Paleocene Mammals, Utah 305 / — 1 Fig. 1. Approximate locations of mammal-bearing sites in Paleocene part of North Horn Formation, Emeiy and Sanpete counties, UT; data from Gazin (1941) and Robison (1986). Localities, Dragon local fauna: Dragon Canyon (1). Wagon Road local fauna: Wagon Road (2), \Vagon Road Ridge (3). Gas Tank local fauna: Gas Tank Hill (4), Dairy Creek (5), Jason Spring (6), Ferron Mountain (7; probably equivalent to OMNH V829), Blue Lake (8), and Sage Flat (9). herein were collected in 1993-94, through sur- face prospecting methods. With one excep- tion, all specimens are from the classic Dragon Canyon (Dragon local fauna; ?Tol) and Wagon Road (Wagon Road local fauna; ?Pu3) sites described by Gazin (1941). The exception is a specimen assigned to Ectoconus ditrigoniis (OMNH 28111), collected by Jon Judd of Castle Dale, UT, at a site south of Ferron Mountain. The site, OMNH V829, is probably the same as Robison's (1986) Ferron Mountain locality (Gas Tank local fauna; ?Pu2). The following abbreviations are used for in- stitutions cited in the text: BYU, Brigham Young University, Provo, UT; OMNH, Oklahoma Museum of Natural Histoiy, Norman; USNM, National Museum of Natural Histoiy, Washing- ton, DC. Measurements, in mm, are as follows: L, anteroposterior length; W, transverse width; WTal, transverse width of talonid; WTri, trans- verse width of trigonid. Descriptive Accounts Order Proteutheria Family Pantolestidae Cope, 1884 ?Propalaeosinopa sp. Figs. 2A-B Material.— OMNH 27681, fragment of right dentaiy bearing the talonid of P4 (WTil = 1.5) and complete M^ (L = 2.8, WTri = 1.8, WTd = 1.8). Locality and horizon. — OMNH V800, "Wagon Road" locality (Gazin 1941, Robison 1986); Wagon Road local fauna, late Puercan (early Paleocene). Joes Valley Member, North Horn Formation, Emeiy County, UT. Description and discussion. — The den- taiy fragment includes the anterior root of P4 and the anterior root of M2. The anterior root of P4 is bowed forward as in pentacodontids and most pantolestids, and its placement indi- cates that P4 was relatively long, longer than Mj. The posterior mental foramen is large and is positioned between the posterior root of P4 and the anterior root of M j. The talonid of P4 includes a large hypoconid and a small entoco- nid; these two cusps are united by a small, thin postcristid, forming a small talonid basin. The apex of the hypoconid is on the midline of the tooth, at the posterior termination of a cristid obliqua that angles lingually toward the front; the postcristid is oriented almost peipendicu- lar to the cristid obliqua. Posterior to the post- cristid and separated from it by a tiny trans- verse basin, a small cuspule (hypoconulid?) is present; this cuspule is connected to the hypo- conid b>' a thin ridge. A tiny entoconulid, not connected to the other cusps, is present at the lingual base of the talonid basin. The trigonid and talonid of M ^ are of equal width; the trigonid is distinctly higher than the talonid, though the tooth is moderately worn. The protoconid and metaconid are both triangular in occlusal outline and of equal occlusal area; the protoconid is the taller of the two cusps. The paraconid is small, low, and transversely oriented. Anterior and posterior carnassial notches are present in the para- cristid and protocristid, respectively. Because of the transverse orientation of the paraconid, the paracristid forms an obtuse angle, with its apex at the anterior carnassial notch. A short anterior cingulum, which disappears at the anterolingual corner of the tooth, is present. 306 Great Basin Naturalist [Volume 55 Fig. 2. Proteutheria from tlie North Horn Formation. A, B, P4-M1 of Propalaeosinopa sp. (OMNH 27681) in occlusal (A) and labial (B) views. C-G, Aplmmonis simpsoni (OMNH 27667); C, D, right P4 in occlusal and labial views, respec- tively; E, F right P3 in occlusal and labial views, respectively; G, left P"* in occlusal view. Scale bar represents 2 mm; tooth roots and jaw fragments have been eliminated to improve clarity. 1995] Paleocene Mammals, Utah 307 The posterior wall of the trigonid is planar; the cristid obliqua meets the base of the posterior wall of the trigonid below the posterior carnas- sial notch. Although it has been mostly obliter- ated by wear, an entoconulid (or at least an ento- cristid) appears to have been present anterior to the entoconid. Of described species, OMNH 27681 most resembles the Torrejonian Propalaeosinopa di- luculi (which we tentatively regard as distinct from P. albertensis following Rose 1981; see discussion in Van Valen 1967). However, the Utah taxon differs in several respects. The pos- terior mental foramen is more anteriorly located than in figured specimens of P. diluculi (Simp- son 1936: fig. 3; Krause and Gingerich 1983: figs. 8, 9). Ml of OMNH 27681 is long and narrow relative to the corresponding tooth of P. diluculi: it slightly exceeds published size ranges (Simpson 1937a, 1937b, Krause and Gingerich 1983) in length but not width. In the Utah taxon the cusps of M^ are somewhat more robust and the postvallid wall more obliquely oriented with respect to the long axis of the tooth; the paracristid is higher, and the metaconid lower, than in P. diluculi. The talonid of P4 is broader and more basined than in P. diluculi (or other species of the genus). We regard the specimen from the North Horn Formation as representing a distinct species, but materials in hand are inadequate to proper- ly diagnose and circumscribe it. Gazin (1941) briefly described two morphs, represented by upper molars, as generically undetermined pantolestids; both were from die Dragon local fauna. Of these, he found pantolestid "a" to compare favorably with Bessoecetor {=Propo- laeosinopa), differing from "J5. thomsom' {=P. diluculi) in being slightly larger and in a few morphological details. It is possible that pan- tolestid "a" and OMNH 27681 represent the same species, although we point out that they derive from different horizons in the North Horn Formation. Differential representation precludes direct comparison with OMNH 27681. If referral of the newly recovered specimen to Propalaeosinopa is correct, it represents the oldest record of the genus and of the family Fantolestidae, a somewhat aberrant group of enigmatic affinities. The new occurrence is estimated to be late Puercan (Pu3) in age; the genus and family are otherwise first known from the late Torrejonian (To3; Archibald et al. 1987). In this context, we note that several morphological details show the North Horn taxon to be distinct, at the species level at least, from described species; when better known, it may prove to be generically separable. Family Pentacodontidae (Simpson, 1937) Van Valen, 1967 ApJironorus simpsoni Gazin, 1938 Figs. 2C-G Newly referred material. — OMNH 27667, right dentaiy fragment with P3^ (P3L = 2.4, W = 1.4; P4L = 4.0, W = 2.5) and asso- ciated left P^ (L = 3.3, W = 4.1). Locality and horizon. — OMNH V799, "Dragon" locality (locality 2 of Gazin 1941: p. 7, fig. 1), Dragon local fauna, early Torrejonian (early or middle Paleocene). Joes Valley Mem- ber, North Horn Formation, Emery Gounty, UT Description and discussion. — OMNH 27667 differs from the type of A. simpsoni (USNM 15539) in minor ways but is clearly referable to the species. P4 is slightly larger than in the type and differs in having a weaker anterior cingulum, which is barely indicated on the anterolingual part of the tooth and is completely absent labial to the keel extending down the anterior face of the protoconid. The minute ridge that extends down the posterior wall of the metaconid (to meet with the cristid obliqua) is lacking; however, the development of this ridge in the type may be due partly to the advanced wear in that specimen. P4 of OMNH 27667 bears a small but distinct ento- conid; this region of the tooth is broken in USNM 15539. The anterior end of P4 in OMNH 27667 is slightly more developed downward than in USNM 15539, vaguely recalling the more advanced condition seen in Pentacodon (Simpson 1937a: 124). Unlike either species of Pentacodon, however, the P4 lacks a basal para- conid, the protoconid is not as inclined poste- riorly from base to apex, and the talonid is bet- ter developed. P3 has not been previously figured or de- scribed for Aphronorus simpsoni, though this tooth is known for A. fraudator (illustrated in outline by Simpson 1937a, Gazin 1941). P3 of OMNH 27667 is more anteroposteriorly elon- gate than in A. fraudator. The tooth is distinct- ly two-rooted and is much smaller than P4; maximum width occurs just posterior to the 308 Great Basin Naturalist [Volume 55 protocoaid. A small talonid basin is developed, with a minute hypoconid and a "cristid oblicjua connected to a ridge running down the poste- rior flank of the protoconid. A small, short ridge and swelling on the posterolingual (lank of the protoconid are suggestive of a metaconid. A faint cingulum is present anterolingually. No associated upper teeth have been previ- ously described for Aphronorus siinpsoiii, al- though a few isolated specimens may belong to the species (Gazin 1941, Robison 1986). P^* of OMNH 27667 is broken near the paraconule and at the lingual edge of the tooth, between the cingulum at the base of the protocone and the lingual root; the labial side of the meta- conid is also damaged. Three roots are pres- ent. The tooth, although similar to P"* of A. fraudator, differs in several respects. The para- style is absent; a small paraconule is present; a metaconule as such is lacking, although there is a vague swelling of enamel in this position. The basal protoconal cingula show no tendency to develop cuspules, as they do in A. frauda- tor, and the metacone is much smaller in size, relative to the paracone, than in that species. The labial cingulum of p4 in OMNH 27667 is also less developed than in A. fraudator Aphronorus situpsoni was diagnosed as dis- tinct from the comparatively well-known A. fraudator mainly on the basis of differences in proportions of F_^ and the lower molars (Gazin 1941). OMNH 27667, which includes teeth pre- viously unreported for A. simpsoni, shows that it is further distinct in having a somewhat more elongate P3; P4 has a narrower, smaller-basined talonid. P'^ differs from that of A. fraudator in several respects, including the lack of a meta- conule and parastyle, and the much lesser de- velopment of the metacone. Considering the specializations of the posterior premolars in pentacodontids (Simpson 1937a) and the pos- sibility that they represent a relatively archaic group (Van Valen 1967), it is difficult to judge which conditions are apomorphous, although some of the states possessed by A. simpsoni (e.g., smaller P4 talonid; P^ with small metacone and no metaconule) would appear — by com- paiison to more primitive Eutheria — to be prim- itive. The Tiffanian species A. orieli, known by remarkably complete specimens (Gingerich et al. 1983), appears to be considerably more advanced, with greatly expanded crushing sur- faces (particularly the protocone) on P"*. Order Cond\ larthra Family PArctocyonidae (Giebel, 1855) Murray, 1866 Desniatoclaenus hennaeus Gazin, 1941 Fig. 3A Newly referred material. — OMNH 27682, associated skull and jaw fragments with broken right and left P^ (right P^L = 6.5), right Ml-3 (MlL = 7.3, W = 8.6; M^L = 7.3, W = 11.0; M^L = 6.2, W = 8.7), left M2-3 (M^ bro- ken, L = 7.4; M^L = 6.0, W = 8.8), left M^ (L = 8.8, WTri = 7.2, WTal = 7.4), talonid of right Mo (W = 6.4), trigonid of left M2 (W = 6.0), and talonid of right M3 (W = 5.2)." Locality and horizon. — OMNH V800, "Wagon Road ' locality (Gazin 1941, Robison 1986); Wagon Road local fauna, late Puercan (early Paleocene). Joes Valley Member, North Horn Formation, Emery County, UT. Description and discussion. — P-* has dis- tinct conules, with the paraconule being taller than the metaconule. These cusps have not previously been noted for P"^ of the species, perhaps because of wear on the type specimen (USNM 16202; see Gazin 1941:' fig. 19; West 1976: fig. 2). The upper molars have a labial cingulum that is continuous. Interruption of the ectocingulum at the base of the paracone was cited as a generic character o{ Desmatodaenus . However, the cingulum is complete in other specimens, such as BYU 3800 (Robison 1986: pi. 2, fig. 10), and we regard this as a feature that is intraspecifically variable. M'^ bears a small but distinct cingular hypocone, another character that is apparently variable in the species (Gazin 1941: figs. 19, 20; Robison 1986). The only variation worthy of note in the lower dentition of OMNH 27682 is the hypoconulid of M3, which apparently projected posteriorly as a distinct lobe, unlike the condition seen in USNM 16202 (Gazin 1941: fig. 19). Gazin (1941) originally described tvvo species of Desniatoclaenus, D. hennaeus and D. para- creodus, both fiom the Wagon Road fiiuna. West (1976) synouNinized the two, a view apparent- ly shared by Tomida and Butler (1980), but Robison (1986) recognized them as distinct and reported additional materials of both species from other localities. In the original diagnosis (Gazin 1941), D. paracreodus was said to be larger than D. hermaeus, with the lingual portion of upper molars more inflated and widi a relatively larger M'^, bearing a better- Paleocene Mammals, Utah 309 5 mm t 1 ^ ■ ■ I Fig. 3. ?Arctoc\oniclae and Periptychinae fioni the North Horn Formation. A, right P"*-M^ of Desmatoclaenm her- maeiis (OMNH 27682) from the North Horn Formation; base of M- restored from contralateral tooth of same specimen, and maxilla eliminated to improve clarity-; B, left dP^-^ and Ml of Ectoconus ditrigonus (OMNH 28111) in occlusal view; maxilla eliminated to improve clarit)'. developed hypocone. As shown by West (1976), these differences in size and morphology are both minor and inconsistent. In this context, we note that M^-^ of OMNH 27682 are rela- tively small (a supposed character of D. her- maeus), yet M-^ is proportionately large, with a well-developed hypocone (characters cited for D. paracreodiis). We follow West (1976) in regarding the species as synonymous. In the original diagnosis and discussion of Desmatoclaenus, Gazin (1941) compared the genus with Tetraclaenodon and Frotogonodon, as the latter taxon was tlien conceived (Matdiew 1937, Simpson 1937a). Van Valen (1978) placed "?F. " protogonioides (cf. Matthew 1937) — originally referred (Cope 1882a), in part, to the genus Mioclaenus — in Desmatoclaenus , adding to the genus two additional species, D. diaiiae and D. mearae; Frotogonodon was syn- onymized with Loxolophus. We are in agree- ment with these assignments; D. protogo- nioides is relatively well represented and adds 310 Great Basin Natur.\list [Volume 55 significantly to knowledge of the genus. Tlius recognized, Desniatoclaenus is distinct from Loxolophus in having stronger protocones on P*^^; better-developed, more lingually placed hxpocone on \l'~-, with hypocone occasionally distinct on M'^; and paraconid of lower molars placed more posterolingually and closely appressed to the metaconid. Desmatoclaenus differs from Tetraclaenodon in having less molarized premolars (a metacone is lacking on P'^""^; the trigonid is poorly developed and a talonid basin is lacking on P4), upper molars lacking mesostyle and with lesser develop- ment of the hypocone; and lower molars with more distinct, anteriorly placed paraconid. Gazin (1941) considered Desmatoclaenus to be stmcturally inteimediate between the archa- ic ungulate "Protof^onodou' (then considered a creodont) and Tetraclaenodon, a primitive phenacodontid; the differential comparisons presented above uphold this view. Subsequent workers have referred Desmatoclaenus to the Ai^ctocyonidae on the one hand (Van Valen 1978, Cifelli 1983) or the Phenacodontidae on the other (Simpson 1945, West 1976, Robison 1986). The positioning of the upper molar h)TDO- cone somewhat more lingually in Desmatoclae- nus than in Loxolophus is vaguely reminiscent of the presumably derived condition in the Phenacodontidae; similarly, the low, bunodont cusps bearing mainly flat, apical wear are sim- ilar to conditions generally obtained in mem- bers of that family. Desmatoclaenus may well be a transitional taxon, but in the absence of compelling evidence in the form of synapo- morphies, we here tentatively retain it in the Arctocyonidae. In this context, we note that the referred species D. protogonioides apparently has a reduced anterior dentition, a condition shared with loxolophine arctocyonids (Cifelli 1983). Family Periptychidae Cope, 1882 Anisonchus ?oligistus Gazin, 1941 Fig. 4A Newly referred material. — OMNH 27679, right M'l Locality and horizon. — OMNH V800, "Wagon Road" locality (Gazin 1941, Robison 1986); Wagon Road local fauna, late Puercan (early Paleocene). Joes Valley Member, North Horn Formation, Emeiy County, UT. Description. — OMNH 27679 is missing tlie lingual base of the crown and enamel from the posterior margin of the tooth; its estimated L is 3.1. This specimen is appropriate in size for only two of the four species of Anisonchus re- ported from the North Horn Formation (Gazin 1941, Robison 1986); OMNH 27679 differs from M'^ referred to A. athelae (including A. eowijnae; Robison 1986) and is tentatively referred to A. oligistus, for which M"^ was not previously known. Although the tooth is incom- plete and worn, it can be seen that the anter- ocingulum was relatively weak and lacked a pericone. Similarly, the hypocone was weak in comparison to the condition in A. athelae, being more similar to the larger A. dracus in this respect. The pattern of wear suggests that both paraconule and metaconule were present, placed near the base of paracone and meta- cone, respectively. Haploconus elachistus Gazin, 1941 Figs. 4B-F Newly referred material. — OMNH 27670, fragments of mandible with left M^_2 (MiL = 3.8, WTri = 2.7, WTal = 2.8; M2L = 3.9, WTri = 3.2, WTal = 2.9) and right M2 (L = 4.0, WTri = 3.1, WTal = 3.0); 27713, frag- ments of left mandible with P3 (L = 4.5, W = 2.8) and a heavily encrusted molar; OMNH 27680, right P4 (L = 4.5, W = 3.3). Locality and horizon. — OMNH V800, "Wagon Road" locality (Gazin 1941, Robison 1986); Wagon Road local fauna, late Puercan (early Paleocene). Joes Valley Member, North Horn Formation, Emery County, UT. Description and discussion. — Available lower premolars (OMNH 27680, 27713) are too worn to detennine whedier a paraconid was present; Gazin (1941) reported the presence of this cusp on P3 but not P4 of Haploconus ela- chistus. The protoconid is a large, inflated cusp, particularly on P4. A talonid crescent extends posteriorly from the lingual base of the proto- conid, curving labially at the posterior margin of both P3 and P4. The metaconid of lower molars is nearly as tall as the protoconid and is transversely aligned with that cusp; a weak paracristid descends anterolingually from the protoconid, teniiinating in a small paraconid, which lies in a median position. As described b\' Gazin (1941), the pre-entocristid is taller than the cristid oblicjua. The entoconid forms a distinct pillar and projects somewhat on the 1995] Paleocene Mammals, Utah 311 Fig. 4. Anisonchinae fiom the North Horn Formation. A, Anisonclut.s 'fuli^istus (UMNH 27679, right M^ in occkisal view). B-F Haploconus elachistus: B, left Mi_2 (OMNH 27670) in occlusal view; C, E, left P3 (OMNH 27713) in occlusal and labial views, respectively; D, F, right P4 (OMNH 27680) in occlusal and labial views, respectively. Scale bar repre- sents 2 mm; tooth roots and jaw fragments have been eliminated to improve clarit)'. 312 Gril\t Basin Naturalist [Volume 55 lingual side of the tooth; the in poeonulid forms a fingerlike projeetiou at the baek of the tooth and is somewhat lingual in position, an appearance emphasized in later wear stages. Two species of Haploconus, H. angustus and the larger H. coniictilaiiis, are recognized from the Torrejonian (To2; Archibald et al. 1987) of the San Juan Basin, NM (Matthew 1937). The apparent last record of Haploconm is represented by a single molar, of uncertain specific affinities, from Swain Quarry (To3?; Archibald et al. 1987), WY (Rigby 1980). The genus is othenvise known only from the North Horn Formation. Gazin (1939) described H. inopinatiLS fiom the Dragon fauna, later adding a second species, ?//. elachistm, fi-om die Wagon Road (Gazin 1941). More recently, Robison (1986) has reported specimens of Haploconus sp. from the Gas Tank local fauna; these mate- rials are of interest in documenting the first appearance of the genus, but unfortunately they are not specificalh' diagnostic. H. inopina- tiis, of Tol age, is similar in size to the later H. angiistus but differs from that species in pro- portions of the upper molars (Gazin 1939). H. elachistus, the geologically oldest described species, is smaller than the Tonejonian species and, as noted by Gazin (1941), differs from them in a number of respects. In the lower dentition, P3_4 are less inflated than in H. angmtus. Similarh, the trigonids of lower molars in H. elachistus lack the inflated appearance seen in Torrejonian species; a small paraconid is still present, whereas in remaining species the paracristid forms a bladelike surface extending anteriorly from the protoconid and bears no cusp. Lower molars of H. elachistus also lack the crenulated or striated enamel and promi- nent labial cingulum seen in other species. As might be expected, the geologically older H. elachistus appears to be more primitive than the Torrejonian species for the characters cited. In this context the apparent presence of a more derived species in the Gas Tank local fauna (Robison 1986) is somewhat surprising. Haploconus is distinctive in the extreme modification of lower molar trigonids (with reduction to loss of the paraconid) and in the unusual configuration of the talonid in posteri- or lower premolars (with a lingual rather than labial crescent), characters that are both ex- pressed in H. elachistus. The affinities of the genus are puzzling; Gazin (1941), noting the primitiveness of some features of H. elachistus. considered the species to be transitional be- tween Conacodon and more derived species of Haploconus. In retaining unreduced lower molar trigonids and relatively unspecialized lower premolars, species of Conacodon are primitive with respect to Haploconus. In terms of characters that are probably derived within tlie context of Condylarthra, Conacodon, Haplo- conus, and Oxyacodon have a lingually placed hypoconulid and hypertrophied postmeta- cristid on lower molars, lingually placed hypo- cone on upper molars, loss of protocone on P^, and, possibly, a columnar, lingually placed entoconid on lower molars (not clearly seen in all species of Oxyacodon). However, the exclu- siveness of these characters and their potential status as synapomoiphies remain to be estab- lished. Archibald, Schoch, and Rigby (1983) have shown that Conacodon and Oxyacodon represent a distinctive subfamily, Conacodon- tinae, whose relationship to other periptychids is unclear; further investigation of the position of Haploconus with respect to this clade is clearly warranted. Ectoconus ditrigonus (Gope, 1882) Fig. 3B Newly referred material. — OMNH 28111, fragment of left maxilla with dP'^^ and Ml (dp3L = 7.5, W = 7.0; dP-^L = 7.5, W = 8.4; MlL = 9.6, W = 13.5). Locality and horizon. — OMNH V829, probably the same as Robison's (1986) Ferron Mountain localit)'; Gas Tank local fauna, middle Puercan (early Paleocene). Joes Valley Member, North Horn Formation, Emeiy Gounty, UT. Description and discussion. — The decid- uous teeth, dP'^~^, are markedly smaller than M^; both have conspicuous parast\'lar and meta- stylar lobes. The third deciduous premolar has a roughly triangular occlusal profile and is longer than it is wide. The paracone and meta- cone are sube(]ual in height; a large parastyle is present almost directly anterior to the para- cone. A prominent ridge extends lingually from the parastyle to the protocone, which is nearly as tall as the paracone and metacone; another ridge descends the labial slope of the parastyle, continuing posteriorly as a weak ecto- cingulum. Labial to the metacone, the stylar shelf broadens; a small cusp, serially analo- gous (if not homologous) to a similar cusp on upper molars of Ectoconus ditrigonus (Osborn 1995] Paleocene Mammals, Utah 313 and Earle 1895), is present labial to the meta- cone. A salient postmetacrista descends pos- terolabially from the apex of the metacone, extending to the posterolabial corner of the tooth. Weak paraconule and metaconule are present on the pre- and postprotocrista, respectively. Faint pre- and postcingulae are present on the lingual slopes of the protocone. The fourth deciduous premolar is more molar- iform than dP'^, differing from M^ in having smaller conules and associated cristae, and in the lesser development of the protocone region. The parastyle of dP^ is more labially placed than on dP'^, and the ectocingulum and cingular cusp better developed than on that tooth; a small mesostyle is also present. The lingual cingulae are strong; pericone and hypo- cone are present. M^ is typical o{ Ectoconus and complete description is unnecessaiy The ectocingulum is strong and bears both a meso- st)'le and posterior stvlar cusp. The latter is sub- conical and is connected to the base of the metacone by a low ridge. Paracone, metacone, and protocone are subequal in height; conules are strongly developed and are only slightly lower than the principal cusps. Ectoconus ditrigonus, the type species, was first described on the basis of material from the San Juan Basin, NM (Cope 1882b). Matthew (1937) reported a second species from the San Juan Basin, E. majusculus, considered by Van Valen (1978) to be synonymous with E. ditri- gonus. The genus is known from several locali- ties, including both Pu2 and Pu3 horizons, in that area (Archibald et al. 1987). Gazin (1941) described the species E. sijmbolus from the Wagon Road (?Pu3) fauna. North Horn Forma- tion. Robison (1986) described additional mate- rials of E. sijmbolus from localities of the Gas Tank fauna, thereby extending the range of the species to ?Pu2, and reported E. ditrigonus from two Gas Tank localities. OMNH 28111 can be referred to the latter species on the basis of size (larger than E. symbohis) and the pres- ence of a relatively small posterior cusp, con- nected to the base of the metacone b\' a low ridge, on the ectocingulum of M^ (Robison 1986). Deciduous teeth of archaic ungulates have not been widely described or illustrated, a notable exception being the deciduous premo- lars of Phenacodontidae (West 1971). To our knowledge, deciduous teeth of Periptychidae have not been previously described, so that there is no basis for comparison with dP'^~l of Ectoconus ditrigonus . Acknowledgments We are especially grateful to Dale Harber for the cooperation of the U. S. Forest Service. We thank Jon Judd, Monte Swasey, and Scott Madsen for help in the field; Dr Scott Russell, Noble Electron Microscopy Laboratory, for access to equipment and facilities; and Estelle Miller for preparing the SEM photographs. Drs. David W. Krause, J. David Archibald, and Jeffrey G. Eaton provided invaluable com- ments that improved the manuscript. Field- work was supported by grant number 5021-93 from the National Geographic Society. Literature Cited Archibald, J. D.. E D. Gingerich, E. H. Lindsay, W. A. Clemens, Jr., D. W. Krause, and K. D. Rose. 1987. First North American land mammal ages of the Cenozoic Era. Pages 24-76 in M. O. Woodburne, editor, Cenozoic mammals of North America: geo- chronolog\' and biostratigraphy. University of Cali- fornia Press, Berkeley. Archibald, J. D., J. K. Rigbv, Jr., and S. E Robison. 1983. Systematic revision of Oxijacodon (Condylarthra, Peript\'chidae) and a description of O. ferronensis n. sp. Journal of Paleontology 57: 5.3-72. Archibald, J. D., R. M. Schoch, and J. K. Rigby, Jr. 1983. A new subfamily, Conacodontinae, and a new species, Conacodon kohlbergeri, of the Peript)chidae (Condylarthra, Mammalia). Postilla 191: 1-24. Butler, R. E, and E. H. Lindsay. 1985. Mineralogy of magnetic minerals and revised magnetic polarity stratigraphy of continental sediments, San Juan Basin, New Mexico. Journal of Geology 94: 53.5-.554. ClFELLi, R. L. 1983. The origin and affinities of the South American Condylarthra and earh' Tertiaiy Litoptema (Mammalia). American Museum Novitates 2772: 1-49. Cope, E. D. 1882a. Some new forms from the Puerco Eocene. American Naturalist 16: 83.3-8.34. . 1882b. Synopsis of the Vertebrata of the Puerco Eocene epoch. Proceedings of the American Philo- sophical Society' 20: 461-471. Gazin, C. L. 1938. A Paleocene mammalian fauna from central Utah. Journal of the Washington Academv' of Science 28: 271-277. . 1939. A further contribution to the Dragon Paleo- cene fauna of central Utah. Journal of the \^ashington Academy of Science 29: 273-286. . 1941. The mammalian faunas of the Paleocene of central L'tah, with notes on the geolog)'. Proceedings of the Lhiited States National Museum 91: 1-53. Gingerich, R D., P Houde, and D. W. Kr.4USE. 1983. A new earliest Tiffanian (late Paleocene) mammalian fauna fiom Bangtail Plateau, western Craz\- Mountain Basin, Montana. Journal of Paleontology' 57: 957-970. 314 Great Basin Naturalist [Volume 55 Kr\use, D. VV., ano E D. C;i\(;khi(;ii. 1983. Mammalian fauna from Douglass Quany, earliest Tiffanian (late Paleocene) of the eastern Crazy Mountain Basin, Montana. Contributions from the Museum of Paleon- tology, University of Miehigan 26; 157-196. M.vnilEW, W. D. 1937. Paleoeene faunas of the San Juan Basin, New Mexico. Transactions of the American Philosophical Societ>', new series 30: 1-510. OsBORN, H. K, AND C. Earle. 1895. Fossil mammals of the Puerco beds. Collection of 1892. Bulletin of the American Museum of Natural Ilistoiy 7: 1-70. RiGBV, J. K., JH. 1980. Swain Quarr\' of the Fort Union Formation, middle Paleocene (Torrejonian), Carbon Count)', Wyoming; geologic setting and mammalian fauna. Evolutionar\' Monograph 3. 178 pp. RoBiSON, S. F 1986. Paleocene (Puercan-Torrejonian) mammalian faunas of the North Horn Formation, central Utah. Brigham Young University Geology Studies 33; 87-133. Rose, K. D. 1981. The Clarkforkian land-mammal age and mammalian faimal composition across the Paleocene- Eocene boundarv'. University of Michigan Papers on Paleontology 26; 1-196. Simpson, G. G. 1936. A new fauna from the Fort Union of Montana. American Museum Novitates 873; 1-27. . 1937a. The Fort Union of the Crazy ^Mountain Field, Montana, and its mammalian faunas. Bulletin of the United States National Museum 169; 1-2S7. . 1937b. Additions to the upper Paleocene fauna of the Crazy Mountain Field. American Museum Novitates 940; 1-15. . 1945. The principles of classification and a classi- fication of mammals. Bulletin of the American Museum of Natural Histoi-v 85; 1-350. Spieker, E. M. 1960. The Cretaceous-Tertiaiy boundary in Utah. 21st International Geological Congress, Copenhagen 5; 14-24. ToMiDA, Y. 1981. "Dragonian" fossils from the San Juan Basin and status of the "Dragonian" land mammal "age." Pages 222-241 in S. G. Lucas, J. K. Rigby, Jr., and B. S. Kues, editors. Advances in San Juan Basin paleontology. University of New Mexico Press, Albuquerque. . 1982. A new genus of picrodontid primate from the Paleocene of Utah. Folia Primatologica 37; 37—43. TOMIDA, Y. AND R. F Butler. 1980. Dragonian mammals and Paleocene magnetic polarit\' stratigraphy of the North Horn Formation, central Utah. American Journal of Science 280; 787-811. Van Valen, L. 1967. New Paleocene insectivores and insectivore classification. Bulletin of the American Museum of Natural Histoid 135; 217-284. . 1978. The beginning of the Age of Mammals. Evolutionary Theory 4; 45-80. West, R. M. 1971. Deciduous dentition of the early Tertian' Phenacodontidae (Condylarthra, Mammalia). American Museum Novitates 2461; 1-37. . 1976. The North American Phenacodontidae (Mammalia, Condylarthra). Contributions to Bio- logical Geology, Milwaukee Public Museum 6; 1-78. Wood, H. E., II, R. W. Chaney. J. Cl.\rk, E. H. Colbert, G. L. Jepsen, J. B. Reeside, Jr., and C. Stock. 1941. Nomenclature and correlation of the North American continental Tertiaiy Bulletin of the Geolog- ical Society' of America 52: 1-48. Received 6 May 1994 Accepted 12 December 1994 Great Basin Naturalist 55(4), © 1995, pp. 315-321 ■ SPRINGTIME MOVEMENTS, ROOST USE, AND FORAGING ACTIVITY OF TOWNSEND'S BIG-EARED BAT {PLECOTUS TOWNSENDII) IN GENTRAL OREGON David S. Dolikinl, Ronald D. Gettinger^, and Michael G. Gerdes^ Abstract. — Seasonal movements, roost-site fidelity, and foraging activity patterns are largely unknown for western populations of Townsend's big-eared hat [Plecotus toicnsendii). We used miniature radiotelemetry units to track spring- time movements of si.x bats inhabiting forested lava flows in central Oregon, and found that bats moved up to 24 km from hibemacula to foraging areas. Individual bats returned to the same foraging area on successive nights but shifted to different areas in presumed response to changes in insect availabilit>-. Both se.xes apparently use a series of interim roost sites between emergence from hibernation and the time females enter into maternitv' colonies, with little individual fidelity to these sites. In regions characterized by extensive lava-flow topography, suitable daytime roosts are numerous and dispersed over a large area, allowing bats to move relatively great distances to locate foraging ranges. Hence, the actual area of concern for effective management of individual populations can be considerably larger than indicated solely by locations of hibemacula and maternity caves of this declining species. Key icords: Toicnsend s big-eared bat. Plecotus townsendii,/orag»ig movements, roost sites, roost fidelity, hibemacula, caves, central Oregon, radiotelemetry, lavaflous, candidate species. Townsend's big-eared bat {Plecotus town- sendii) is distributed over much of western North America (Hall 1981), although popula- tions may be widely scattered within its range. The species appears to be a habitat generalist, reportedly inhabiting coniferous forests in nortli- ern New Mexico (Jones 1965), mixed meso- phytic forests in Kentucky (Adam et al. 1994), deserts in Arizona (Hoffmeister 1970), native prairie in Kansas and Oklalioma (Humphrey and Kunz 1976), riparian communities in north- eastern Montana (Swenson and Shanks 1979), Kansas, and Oklahoma (Humphrey and Kunz 1976), and agricultural areas and coastal regions in California and Washington (Dahlquest 1947, 1948, Pearson et al. 1952). In Oregon the dis- tribution of Townsend s big-eared bat is dis- continuous and highly local across forest and shrubsteppe habitats throughout the state (Perkins and Levesque 1987). Two disjunct subspecies occur in eastern North America, both of which are listed as en- dangered under the U.S. Endangered Species Act. Kunz and Martin (1982) suggested that western populations also are vulnerable, espe- cially to disturbance in winter hibemacula and summer maternity caves. Both subspecies found in the western United States are declin- ing markedly (Perkins and Levesque 1987, Pierson et al. 1991), and the species is listed as endangered, sensitive, or of special concern by several western states and federal land management agencies. Big-eared bats feed almost exclusively on Lepidoptera (Ross 1967, Whitaker et al. 1977, 1981, Dalton et al. 1986, Sample and Whitmore 1993) and are viewed as moth specialists (Dalton et al. 1986, Sample and Whitmore 1993). Probably most limiting to their distribution, however, is availability of suitable sites for roosting, hibernation, and reproduction, which consist primarily of caves and abandoned mines. These three activities require different microclimatic conditions (Dahlquest 1947, Pear- son et al. 1952, Twente 1955, Barbour and Davis 1969, Martin and Hawks 1972, Humphrey and Kunz 1976, Marcot 1984, Center 1986, Perkins and Levesque 1987, Pierson 1989, Pierson et al. 1991, Lacki et al. 1993, Clark et al. 1995). Any single site generally is unsuitable for more than one function, although microclimates in different regions of the same cave sometimes differ sufficiently to accommodate more than one activity (e.g., Clark et al. 1995). 'High Desert Ecological Research Institute, 15 S.W. Colorado Avenue, Suite 300, Bend, OR 97702. ^Biologs' Department, Randolph-Macon Woman's College, Lynchburg, VA 24.503. ■'Deschutes National Forest, United States Forest Service, 1645 Highway 20 East, Bend, OR 97701. 315 316 Great Basin Naturalist [Vokime 55 Big-eaix'd bats are colonial tor most of the year, but colony dynamics and seasonal move- ments have not been studied in the Inter- mountain West. In central Oregon, P. town- sendii undergoes arousal from hibernation and movement from hibernacula in April, although the precise timing of these events appears to vary with weather conditions and topography (U.S. Forest Service, Deschutes National Forest, inipublished data). Females form maternity colonies in late spring or early sunnner (USFS unpublished data), but the timing of their amval at mateniit)' roosts is poorly documented, and it is unclear whether they move immedi- ately to maternity roosts upon departure from hibernacula. A much better understanding of seasonal movements among roost sites is necessaiy for effective management of populations. Although it is clear that traditional site use (sensu Dobkin et al. 1986) occurs for specific hibernacula and matemit)' roosts, the extent of roost site fidelity' by individual bats is unknown. Recent teleme- try studies have been conducted for both endangered subspecies found in eastern North America (Clark et al. 1993, Adam et al. 1994, Lacki et A. 1994), lout no telemetiy studies have examined the movements of western sub- species. Our primaiy objective was to acquire information concerning the extent of move- ments by individual Townsend's big-eared bats during the period following arousal from hibernation in an area containing a significant proportion of Oregon's known population. Study Area and Methods Fieldwork was conducted in Deschutes County on the Fort Rock Ranger District of the Deschutes National Forest and adjacent lands administered by the Bureau of Land Management. The primaiy study area (Fig. 1) consists of a NW-SE-oriented basin contain- ing extensive forested lava flows, and the sur- rounding buttes from 44° 25' to 43° 37' N, and 121° 15' to 120° 48' W. Elevations range from 1400 m on the basin floor to nearly 2000 m atop Pine Mountain. Forests are open stands of ponderosa pine {Piniis ponderosa) with bitter- brush {Purshia tridentata), manzanita {Arcto- staphijlos spp.), and bunchgrass understories. Scattered, relatively closed, stands of lodge- pole pine {P. contorta) also occur throughout. Areas adjacent to lava flows consist of shrub- steppe habitat dominated by sagebrush {Artemisia tridenlata). The study area lies within the broad zone of intergradation between the western interior form {P. t. pallescens) and the coastal Pacific form {P. t. townsendii) of Townsend's big-eared bat (Handley 1959). We concur with Handley's (1959: 199) assessment that "allocation . . . from much of this area to one race or the other is largely a matter of personal opinion. " Based on USFS cave surveys conducted from 1985 to 1991, two hibernacula (SI and S2, Fig. 1) but no maternity caves were known from the southern end of the basin. The north- em end of the basin contained a series of hiber- nacula and one maternity cave (N3, which was gated), as well as one other cave (N2) that reportedly was used as a maternity roost in the past. The maternity cave and the northern- most hibernaculum in the southern portion of the basin are separated by 30 km, which prompted the assumption that big-eared bats in the basin consisted of two separate popula- tions (J. M. Perkins, unpublished report to USFS). Subsequent to completion of our field- work, a previously unknown maternity cave was discovered beyond the southern end of the basin, 17 km southeast of SI. Fieldwork in 1992 commenced on 7 April and continued through 9 June. Six big-eared bats (5 females, 1 male) were captured by hand between 1100 and 1730 h from four different caves in April and May (Table 1). Each bat was fitted with a battery-powered (14-21 da>' bat- tery longevity), miniature radiotransmitter (0.6-0.7 g; Model BD-2B, Holohil Systems, Ltd.) affixed to the dorsal, interscapular fur (Dobkin et al. in press) with eyelash cement. Transmitter units averaged 6% of bat body mass (x = 10.6 g. Table 1), which should have had minimal effect on maneuverability and energy costs for this species (Davis and Cock- rum 1964, Aldridge and Brigham 1988). Bats carrying transmitters were tracked with portable receivers (Telonics) equipped with directional antennae (Wilkinson and Brad- bury 1988). Bats were monitored for nearly 850 observer hours over the 64-day period through a combination of daytime ground searches and nighttime triangulations from fi.xed locations. Two or three observers with receivers were located on the tops of buttes widely separated aroimd the basin (Fig. 1) to provide the directional data necessary for 1995] Radiotelemetry of Townsend's Big-eared Bats 317 :hina HAT S2 • • • EAST SI • BUTTE • QUARTZ MTN. • FOX BUTTE Fig. 1. Map of the study area in central Oregon showing locations of the four caves in whicli Townsend's big-eared bats were captured and fitted with radiotransmitters (SI, S2, Nl, and N2), and location of the onl\- known maternit\' cave (N3) in the basin. Telemetered bats were monitored from atop Pine and Quartz mountains, Coyote, East, Fox, and China Hat buttes. determining bat locations. The monitoring protocol for fixed-point triangulation consisted of scanning all active frequencies for the initial five minutes of each quarter hour. If contact was made, tlie other observers were notified by radio and the bat was tracked continuously. In addition, seven flights were made at night by fixed-wing aircraft carrying a receiver and wing-mounted antennae and equipped with a LORAN system. LORAN fixes were integrated with simultaneous directional information ob- tained from ground-based receivers. We conducted ground searches on foot and from moving vehicles. Efforts were concen- trated in the vicinity of caves known to be used by bats, including caves in which teleme- tered bats originally were captured. These searches continued for 7-14 days following attachment of transmitters. Due to rugged top- ography and the distances between northern and southern ends of the basin, only southern caves were checked systematically following tagging of the first three bats, all of which were from the southern basin. Likewise, only northern caves were checked systematically following tagging of the last three bats, all of which came from the northern basin. All caves were checked as opportunity permitted, re- sulting in essentially complete coverage of all known cave sites in the basin at least weekly. Results Movements and Roost Site Fidelity All marked females left their caves within two nights of capture and neither returned to these caves nor entered the known maternity cave (N3) during the remainder of transmitter battery life. Upon emergence from their hibernacula, all three females from the south- ern end of the basin moved 11-12.5 km north- east to the western slopes of Pine Mountain and did not return to the vicinity of their hibernacula in the southern end of the basin. Only female #579 was located subsequently, again on the western slope of Pine Mountain. Faint signals were received briefly from one of these females on 3 May on a precise bearing toward the then-unknown maternity cave southeast of the study area. The most extensive telemetry data were collected for female #707, which left Nl on the second night following capture. She was located again five nights later and was tracked for the following five nights (including a series of LORAN fixes made from the air), and then to a day roost located just east of the crest of Pine Mountain, ca 20 km from Nl but only 2-4 km from where she had been foraging on the preceding five nights. Although we do not know whether she had used this roost previ- ously, she was not found there subsequently. This bat went undetected over the next three nights and was then located for the last time on the following night. All foraging locations beyond the immediate vicinity of Nl were on the western slope of Pine Mountain, 17-24 km from Nl. Foraging locations for the fifth female (#728) were within 2-5 km of N2: southwest of N2 on one night and northeast of N2 two nights later This bat dropped her transmitter, which we recovered 15 days after attachment, at a location 5 km west of N2 and within 1 km 318 Grkat Basin Naturalist [Volume 55 Table 1. Suninuin of radiotelemetrv' contacts witli Townsentl's big-eared hats earning transmitter nnits on the Deschutes National Forest in central Oregon, 1992. Max . distance'' Bat# L. )cation-' (km) Sex Mass (g) Dates of contact 558 SI 11 F 11.0 17-19 April 568 S2 11 F 11.2 17-19 April 579 SI 12 F 12.0 19 April-2 May 707 Nl 24 F 10..3 28 April-10 May 728 N2 5 F 10.6 12-25 Mayt 768 \2 S M 9.0 20-26 Ma>' ^Indicates location of cave where bat v\as captured and fitted witli radiotransniitter. To maintain site security, caves are designated by alphanumeric codes; S and N indicate cave location in southern and northern portions of the stud> area, respectively. SI, S2, and Nl were winter hibernacula; N2 was an interim roost site. ''MiLKinium distance moved from cave of initial capture, as cletiTUiinetl by radio contact with foraging bat. ^Transmitter dropped from bat on 26 or 27 May and recovered on 27 May. of unnamed caves known to have harbored big-eared bats occasionally in the past (L. Becker, unpublished USFS sui-vey data). The single telemetered male (#768) for- aged extensively in the immediate vicinitx' of N2 upon evening emergence, then moved 6-8 km east to forage over Horse Ridge. This bat was not located again until five nights later, when he returned to N2, and was recorded over the next two nights foraging in and around the sinkhole immediately in front of N2. Although male #768 returned to roost for two consecu- tive days in the cave where originally captured, he then left and did not return again prior to the end of fieldwork 12 days later. The habitat used for foraging consisted of sagebrush shrubsteppe (western slopes of Pine Mountain and Horse Ridge) and very open ponderosa pine woodland with extensive bit- terbrush and interspersed areas (<5 ha) of sagebrush. Relatively little time appeared to be spent foraging in more densely forested areas. Times of Activity Big-eared bats emerged from their cave roosts to forage shortly after sunset, with time of emergence becoming later as day length increased in the spring (Fig. 2). Although our data are ver\' limited, an interpretable pattern of activity can be seen in the May data. Big- eared bats foraged in the immediate vicinity of their cave roosts during the first few hours of darkness, moved to areas farther from their roost to forage (perhaps intermittently) from around midnight to within an hour or two of sunrise, and then once again returned to for- age in the vicinity of their day roost. Discussion Continuous monitoring of movements and activity in small, cave-dwelling bats like Town- send s big-eared bat is constrained by (1) the need to minimize load mass carried by an ani- mal, which strongK' limits both strengdi of trans- mitter signal output and batteiy longevity, and (2) the difficulty of signal detection in land- scapes of rugged, rock>' topograph}' and from witliin caves. Despite diese limitations, a number of salient points can be deduced fi'om our study. Our data clearly indicated that female big- eared bats in central Oregon did not move directly from their winter hibernacula to mater- nity caves, but instead utilized a series of interim roost sites over a period of perhaps as much as two months. The four females marked in April were captured in winter hibernacula in the company of other roosting conspecifics. In con- trast, the two bats captured in May were the only big-eared bats roosting in the cave on the dates of capture; we assumed that neither of these bats hibernated in N2 during the pre- ceding winter, although we cannot exclude this possibility. None of the four caves in which bats were captured was used as a maternity cave. We suggest that little fidelity to interim roost sites occurs because neither of die two females found in day roosts returned to these roosts on subsequent days. In addition, male #768 left his roost cave, returned five days later, re- mained for two days, and then left again for at 1995] Radiotelemetry of Townsend's Big-eared Bats 319 Bat I.D. Date 558 4/19/92 568 4/19/92 579 4/19/92 5/2/92 707 4/28/92 5/3/92 5/4/92 5/5/92 5/6/92 5/7/92 5/10/92 728 5/14/92 5/16/92 5/25/92 768 5/20/92 5/25/92 5/26/92 LEGEND < 0.2 km from cave > 1 .0 km from cave ^^■H V/////////M Y////////////A 2300 2400 Time of Night (H) 0400 Fig. 2. Temporal distribution (Pacific Daylight Savings Time) of foraging activit)' by Townsend's big-eared bats in rela- tion to distance from daytime roost sites in forested lava flows of central Oregon. least the next 12 clays. With such small sample sizes, we cannot say whether males and females differ in their use of roosts during this period or whether both sexes exhibit the same pattern of periodic use. We believe that the most reasonable interpretation of the data is that both sexes opportunistically use interim roost sites during this period, and that the choice of roost area is most likely determined by spatial and temporal variation in prey avail- ability. Even species that exhibit strong indi- vidual fidelity to day roosts and repeated use of the same foraging areas on successive nights (e.g., Euderma maculatum) shift both roost site and foraging area seasonally (Wai-Ping and Fenton 1989). Big-eared bats in our study moved up to 24 km from hibernacula to foraging areas, al- though our data suggest that distances moved from interim day roosts to foraging areas are typically 2-8 km during the period prior to entry into maternity colonies. These shorter moves between roosts and foraging areas are consistent with research on eastern subspecies of big-eared bats in which females foraged at distances of 2-7 km from their roosts (Clark et al. 1993, Adam et al. 1994). Repeated use of the same foraging area on successive nights or alternation among several sites appears to characterize both eastern subspecies of big- eared bats (Anonymous 1991, Clark et al. 1993, Adam et al. 1994), as well as big-eared bats in central Oregon (e.g., bat #707). Although big-eared bat diets are composed primarily of forest Lepidoptera, bats in eastern Oklahoma foraged preferentially at the inter- face between forested and open pasture habi- tats (Clark et al. 1993). Nevertheless, bats ex- tensively used open, forest, and edge habitats, and significant shifts in relative habitat use were recorded by Clark et al. (1993). Similarly, in central Oregon we found that Townsend's big-eared bats foraged primarily (but not ex- clusively) in the more open habitats provided by shrubsteppe and forest-shiTib ecotones. 320 Great Basin Natufl\li.st [Volume 55 In our study, activih' patterns of hiu-eared bats in spring most closely resembled patterns documented for females of eastern subspecies during late lactation and prior to parturition (Clark et al. 1993, Lacki et al. 1994), i.e., por- tions of the annual cycle when females are less constrained in the amount of time they can spend away from tlie maternity cave. Flight initi- ation inside caves and subsequent emergence documented by Clark et al. (1993) and by Lacki et al. (1994) were identical to the pat- terns exhibited in our study. Primaiy determinants of habitat suitability for Ozark big-eared bats are the availal)ilit\ of an adequate food supply and appropriate roost sites (Clark et al. 1995). Unlike areas where big-eared bats are limited by a small number of suitable roost sites, the extensive forested lava flows found in the Pacific Northwest offer numerous potential temporaiy roost sites that enable individual bats to forage over a consid- erable area by using a succession of roost sites during the period following emergence from their hibernaculum. Bats still are limited sea- sonally, however, to a very small number of sites that provide suitable microclimatic con- ditions for hibernacula and maternity caves. Such an inteqDretation of potential movement patterns is consistent with our tracking data and the loss of contact with telemetered bats for successive days followed by subsequent con- tact. Even our seven attempts to locate bats by aircraft, which should have avoided problems arising from topographic interference with transmitter signals, succeeded only once, indi- cating that bats may well have left the basin entirely, as was apparently the case for at least the one bat we detected southeast of the study area in the vicinity of the previously unknown maternity cave. Populations of Townsend's big-eared bats inhabiting regions with extensive lava flows likely use many roost sites dispersed over large areas. The extent of movements that we docu- mented and the use of the same foraging areas by bats from both ends of the basin make it unlikely that bats from southern and northern hibernacula represent separate populations. A better understanding of movements among seasonal and interim roost sites is urgently needed for successful conservation of dwin- dling populations. Our data demonstrate that the actual area of concern for management of individual populations is considerably greater than indicated solely by locations of hibernac- ula and maternity caves. Acknowledgments We tliank Bijaya Kattel and Jamie Haskins for their invaluable field assistance; this stud\' could not have been completed without their con- siderable help. Lew Becker of the Deschutes National Forest and Chris Carey of the Oregon Department of Fish and Wildlife contributed in many ways to the success of this project. Helpful discussions with Brad Sample and Bruce Wunder and review of earlier versions of the manuscript by Brenda Clark and William Clark improved the final manuscript. This project was carried out in part with funding provided by the United States Forest Sei-vice under Contract No. 43-04GG-2-69020. Literature Cited Adam, M. D., M. J. Lacki, and T. G. Barnes. 1994. Forag- ing areas and habitat use of the Virginia big-eared bat in Kentucky Journal of WikUife Management 58: 462-469. Aldridge. H. D. J. N., AND R. M. Brigham. 1988. Load carrying and maneuverabihty in an insectivorous bat: a test of the 5% "rule" of radio-telenietiy Journal of Manmialogy 69: 379-382. Anonymous. 1991. Endangered Species Technical Bulletin 16: 14. Barbour. R. W., and W. H. Davis. 1969. Bats of America. Universit)' Press, Lexington, O'. CL.4RK, B. K., B. S. Clark, D. M. Leslie, Jr., and M. S. Gregory. 1995. Characteristics of caves used by the endangered Ozark big-eared bat. Wildlife Society Bulletin; in press. Clark, B. S., D. M. Leslie, Jr., and T. S. Carter. 1993. Foraging activity of adult female Ozark big-eared bats {Plecutus townsendii ingem) in summer. Journal of Mammalogy 74: 422—427. Dahlquest, W. W. 1947. Notes on the natural histoiy of the bat Corynorhimis rafinesqiiii in California. Joimial of Mammalogy 28: 17-30. . 1948. Mamniiils of Washington. Universib.' of Kimsas Publications of the Museum of Natural History 2: 1-444. Dalton, V. M., V. Brack, Jr., and E M. McTeer. 1986. Food habits of the big-eared bat, Plecotm townsendii virginianii.s. in Virginia. N'irginia Journal of Science 37:248-254. D.WIS, R., and E. L. Cockrum. 1964. E.xpcrimentally determined weight lifting capacity' in indi\ iduals of five species of western bats. Journal of Mammalogy 45: 64.3-644. DoBKiN, D. S., J. A. Holmes, and B. A. Wilcox. 1986. Traditional nest-site use by White-throated Swifts. Condor 88: 252-253. DoBKiN, D. S., B. K.\TTEL, AND R. D. GETriNGER. Com- parative retention of radiotransmitters by fur-clipped 1995] Radiotelemetry of Townsend's Big-eared Bats 321 and undipped Townsend's big-eared hats and pallid bats. Bat Research News: in press. Center, D. L. 1986. Wintering bats of the Upper Snake River Plain, occurrence in lava-tube caves. Great Basin Naturalist 46; 241-244. Hall, E. R. 1981. The mammals of North America. Volume 1. John Wiley & Sons, New York. Handley, C. O., Jr. 1959. A revision of American bats of the genera Eitderma and Plecotiis. Proceedings of the United States National Museum 110: 95-246. HOFFMEISTER, D. E 1970. The seasonal distribution of bats in Arizona: a case for improving mammal range maps. Southwestern Naturalist 15: 11-22. Humphrey, S. R., and T. H. Kunz. 1976. Ecology of a Pleistocene relict, the western big-eared bat {Plecotiis townsendii), in the southern Great Plains. Journal of Mammalogy 57: 470-494. Jones, C. 1965. Ecological distribution and activit}' periods of bats of the Mogollon Mountains area of New Mexico and adjacent Arizona. Tulane Studies in Zoology 12: 93-100. Kunz, T. H., and R. A. Martln. 1982. Plecotus townsendii. Mammalian Species 175: 1-6. Lacki, M. J., M. D. Adam, and L. G. Shoemaker. 1993. Characteristics of feeding roosts of Virginia big- eared bats in Daniel Boone National Forest. Journal of Wildlife Management 57: 539-543. . 1994. Observations on seasonal cycle, population patterns and roost selection in summer colonies of Plecotus townsendii virginianiis in Kentucky American Midland Naturalist 131: 34-42. Marcot, B. C. 1984. Winter use of some northwestern California caves by western big-eared bats and long- eared Mijotis. MuiTclet 65: 46. Martln, R. A., and B. C. Hawks. 1972. Hibernating bats of the Black Hills of South Dakota. I. Distribution and habitat selection. Proceedings of the New Jersey Academy of Science 17: 24—30. Pearson, O. R, M. R. Koford, and A. K. Pearson. 1952. Reproduction of the lump-nosed bat [Conjnorhiniis rafinesqiiii) in California. Journal of Mammalogy 33: 273-320. Perkins, J. M., and C. Levesque. 1987. Distribution, sta- tus, and habitat affinities of Townsend's big-eared bat {Plecutits townsendii) in Oregon. Oregon Depart- ment of Fish ik Wildlife Technical Report 86-5-01. Pierson, E. D. 1989. Help for Townsend's big-eared bats in California. Bats 7: 5-8. Pierson, E. D., W. E. Rainey, and D. M. Koontz. 1991. Bats and mines: experimental mitigation for Town- send's big-eared bat at the McLaughlin Mine in California. Pages 31-42 in Proceedings V: Issues and technology in the management of impacted wildlife. Thome Ecological Institute, Boulder, CO. Ross, A. 1967. Ecological aspects of the food habits of in- sectivorous bats. Proceedings of the Western Founda- tion of Vertebrate Zoology 1: 205-264. Sample, B. E., and R. C. Whitmore. 1993. Food habits of the endangered Virginia big-eared bat in West Virginia. Journal of Mammalogy 74: 428-435. SwENSON, J. E., AND G. F Shanks, Jr. 1979. Noteworthy records of bats from northeastern Montana. Journal of Mammalogy 60: 650-652. TwENTE, J. W, Jr. 1955. Some aspects of habitat selection and other behavior of cavern-dwelling bats. Ecology 36: 706-732. Wai-Ping, V, AND M. B. Fenton. 1989. Ecology of spot- ted bat (Eiidenna maculatwn) roosting and foraging behavior. Journal of Mammalogy 70: 617-622. Whitaker, J. 6.. Jr., C. Maser, and S. R Cross. 1981. Food habits of eastern Oregon bats, based on stomach and scat analyses. Northwest Science 55: 281-292. Whitaker, J. O., Jr., C. Maser, and L. E. Keller. 1977. Food habits of bats of western Oregon. Northwest Science 51: 46- 55. Wilkinson, C. S., and J. W. Bradbury. 1988. Radio- telemetiy: techniques and analysis. Pages 105-124 in T. H. Kunz, editor. Ecological and behavioral methods for the study of bats. Smithsonian Institution Press, Washington, DC. Received 5 January 1995 Accepted 28 April 1995 Great Basin Naturalist 55(4), © 1995, pp. 322-334 NAMES AND TYPES IN PERENNIAL ATK/PLEX LINNAEUS (CHENOPODIACEAE) IN NORTH AMERICA SELECTIVELY EXCLUSIVE OF MEXICO Stanley L. VVelshl and Clifford Cromptoii- Absthact. — Cited are names and eoniliinatioiis within the woody species of Atriplcx as the\' occur in North America. Tv'pes and tlieir repositories are inchided ibr all ta.\a except those for which that information could not be located. New nomcnclatural proposals include Atriplcx gardneri var aptera (A. Nelson) Welsh & Crompton, comb, nov.; A. garrettii van navajoensis (C. A. Hanson) Welsh & C>rompton, comb, nov.; Atriplcx acanthocarpa vai. coahiiilcnsis (Henrickson) Welsh & Crompton, comb. nov. A lectotvpe is designated for A. breweri S. Watson. Key words: Chcnopodiaccac, Atriplcx types. North America. This list of names and synonyms of peren- nial and woody Atriplex taxa is preliminaiy to the preparation of a taxonomic treatment for the woody species of Atriplex as they occur in North America, both indigenous and intro- duced species. All names, whether treated as taxa recognized by me or as mere synonyms, are included. The taxonomic treatment that will appear subsequently in the publication of the Flora North America Project will distin- guish between the names of taxa per se and their included synonyms. The relatively large number of names and synonyms for this small group of plants is indicative of the changes in generic concepts, the ever-changing inteip re- lation of the status of a taxon, and the general phenotypic plasticity of this amazing group of shrubs, subshrubs, and perennial herbs, which hybridize freely among themselves and some- times with other taxa not apparently closely allied. They grow on a surprising array of sub- strates in the American West, from the cold temperate of northern Alberta to the much warmer climates of Mexico. Often they are among the most important shrub species on saline, fine-textured substrates, and some- times they are the only shrubby inhabitants. Their ability to survive and even thrive in saline sites has placed them in a position of importance for browsing animals where other browse is scarce or lacking. They cover huge areas where geomoiphological processes have exposed raw, saline strata in vast expanses. Niobrara Shale, Mancos Shale, Morrison Formation, and numerous other geological for- mations support these plants. Saline pans and other poorly drained lowlands are occupied by these species. Despite the affinity for saline areas, where they have little competition (except from other halophytes), some of the species thrive where total soluble salts are low. The four-wing saltbush, Atriplex canescens (Pursh) Nuttall, is such a plant. It grows from the edge of saline areas up gradient into far less saline substrates, often in grasslands or in shrublands dominated by sagebrush and other shrubby species. Hybridization is an important factor con- tributing to the diversity of woody Atriplex species. There are at least two main taxa around which many of the remainder are placed, and with which most form at least occasional hybrids, i.e., A. canescens (Pursh) Nuttall and A. confertifolia (Torrey & Fremont) S. Watson. Some of the hybrids have received names and formal taxonomic recognition. Most of them are of occasional occurrence, and some of the taxa treated in contemporaneous taxonomic works are apparently of hybrid derivation — now more or less stabilized as pop- ulations, mainly on veiy peculiar substrates. The following list is thought to be exhaus- tive for woody and perennial Atriplex names in North America, especially for those north of Mexico. A few taxa represented in Mexico are included where thev roimd out the names for HMc ScicDcc Miiscuin and Department (i( Botaiu' and Kan);e Science, Brighani Young Universit>', Piovo, UT 84602. ^Biosysteniatics Research Centre, W'ni. Saunders Bnilding. C;. E. F. Ottawa, Ontario KIA PC6, Canada. 322 1995] North American Perennial Atr/plex Types 323 species complexes largely confined north of that country. Pertinent types have been received on loan by the gracious kindness of curators of the herbaria cited with the specimens. Abbre- viations for the herbaria are those standard ones cited in Index Herbariorum, except that the origin of the collection is indicated by use of such designations as NY Torrey for historic specimens. This is thought to be important because it authenticates the antiquity of the specimen and might prove important in cer- tain cases in judging whether a particular author had access to a given specimen. The type information is presented below in dual format for some taxa, with the type locali- ty or collector information (herein arbitrarily designated "Type locality") as recorded with the protologue cited first and with the label data of the type specimen (herein designated "Type") cited second where there is a substan- tial difference in the two accounts. Atriplex acanthocarpa (Torrey) S. Watson, Proc. Amer. Acad. Arts 9: 117. 1S74. Basionym: Obione acanthocarpa Toney This is a shrub or subshriib, generally less than 1 m tall, characterized by spong)' fruiting bracteoles S-15 mm long, borne on slender to stout pedicels 4-20 mm long. Leaves are variable, but often sinuate-dentate to undu- late-crisped and with hastately lobed base. The species occurs from west Texas and southern New Me.xico south to Mexico. It is represented in the United States by two varieties, i.e., van acanthocarpa from western Texas west through southern New Mexico to southeastern Arizona, and var coahiiih'nsis in southern Texas. Atriplex acanthocarpa ssp. coahiiilensis Henrickson, Southw. Naturalist 33: 4.58. 1988. = A. acanthocarpa var coahuilensis (Henrickson) Welsh & Crompton (cited below). Type: Mexico, Coahuila, ca 2 km W of Nadadores in saline pastured flats near El Porvinir along Hwy. 30, with Suaeda, Sporobolus, Distichlis, nar 27° 03' N lat, 10r37'W long, .540 m, 6 Dec 1975, J. Henrickson 14784; holotype TEX; isotypes MEXU, NY!, RSA. This ta.xon is distinguished by its fruiting bracteoles bearing radiating processes, and stems with at least the medial leaf blades hastate-lanceolate, and with mature fi-uiting inflorescences ver>' long. Its range is from south- ern Texas to southeast Coahuila and coastal Tamaulipas. Atriplex acanthocarpa var. coahuilensis (Henrickson) Welsh & Crompton, comb, now Basionym: A. acanthocarpa ssp. coahuilensis Henrick- son, Southwest. Nat. .33: 458. 1988. Atriplex acanthocarpa var. cuneata (A. Nelson) M. E. Jones, Contr West. Bot. 11: 20. 1903. = A. gardneri var cuneata (A. Nelson) Welsh Atriplex acanthocarpa var. pringlei (Standley) Henrick- son, Southwest. Nat. 33: 461. 1988 Basionym: A. pringlei Standley The taxon is endemic to Mexico, from northern Zacatecas and southern Nuevo Leon south to San Luis Potosi. Atriplex acanthocarpa ssp. stewartii (I. M. Johnston) Henrickson, Southwest. Nat. 33: 457. 1988. Basionym: A. stewartii 1. M. Johnston The taxon is endemic to Coahuila, Mexico, and is dis- tinguished by its 4-winged fruiting bracteoles, although specimens are transitional to A. acanthocarpa var acan- thocarpa. Atriplex amnicola P. G. Wilson, Flora of Australia 4: 322. 1984. TyiDe: "Yalgoo, W. A." [western Australia], 10 Oct. 1945, C. A.' Gardner 7751a; holotype PERTH! Atriplex angustior Cockerell, Proc. Davenport Acad. Nat. Sci. 9: 7. 1902. = A. canescens (Pursh) Nuttall Type: New Mexico, Dona Ana Co., Mesilla Park, Cockerell in 1900; holotyi^e US! The US specimen bears the following label data: "Atriplex angustior, n. sp. Distinguished fi-om A. canescens by the \'ery nanow (3 mm broad) leaves. Apparently = A. canescens angustifolia but that name is preoccupied. Sand Hills Mesilla Park, N. M. 1900. TD.A. Cockerell." Hall and Clements (1923) cited this as a new name for A. canescens var angustifolia, but it seems obvious that while Cockerell recognized the equivalency of the taxa, he was proposing a new taxon, not merely a new name. Atriplex aptera A. Nelson, Bot. Gaz. 34: 356. 1902. = A. gardneri var. aptera (A. Nelson) Welsh & Crompton Type locality: Wyoming, Laramie, Sept. 1901, E. Nelson 738; A. Nelson (1902). Type: "Atriplex aptera A. Nels. n. sp. Moist saline soil. Laramie, Albany Co., Wyoming. Sept. 1901. Elias Nelson No. 738"; holotype RM!;'isotype GH!, UC (frag.)! Hanson (1962) suggested that his ta.xon was of hybrid derivation involving A. canescens and A. huxifolia as parental taxa. Distribution of specimens assignable to the concept is sporadic, possibly indicating multiple origins, and it cannot be considered a taxon in the usual sense. The type specimen of A. aptera has definite wings aligned in four rows similar to some A. canescens but agrees in aspect, size, and general features with A. gardneri. Atriplex berlandieri Moquin-Tandon, Chenop. Enum. 65. 1840. = A. canescens (Pursh) Nuttall Type: "In regno Mexicano. Berlandier 1828"; holotype ? Moquin-Tandon enlarges on the type information in his treatment in Prodromus (13[2]: 114. 1849), "In regno Mexicano inter Laverdo et Bejar (Berland.! n. 1450)." Atriplex bonnevillensis C. A. Hanson, Stud. Syst. Bot. Brigham Young Univ. 1: 2. 1962. = A. gardneri var bonnevillensis (C. A. Hanson) Welsh Type: Utah, Millard Co., "diy lake bed 1.5 miles north- east of headquarters. Desert Range Experiment Station (dominant plant)," 12 July 1961, C. A. Hanson 354; holo- type BRY!; isot>'pes GH!,'mO!, NY!, UTC! The sheets at GH and NY have the date printed as 13 July 1961, probably representing t\'pographical errors. 324 Great Basin Natuiulist [Volume 55 Atriplex brcweri S. Watson, Proc. Anicr. Acad. Arts 9: 1 19. 1874. Type locality: "Fremont; 459 Torrcy; 75 Brewer" (I.e.). Paratypes: "Fremont's 2nd Expedn. Atriplex Breweri S. Wats.'"; NY Torre\'!; "No. 459. Santa Barbara County, California. J. Torrey 1865"; NY! and NY LeRoy!, GH! Type: "Geological Survey of California, 1863. Coll. H. Brewer. No. 75. Atriplex Breweri n. sp. Sea Shore — Sta Monica. 6"|ft] high or more"; lectotype GH!, here desig- nated; isolectotypes NY!, UC, US! Since the plant was described by Watson on the basis of at least three collections, and as there are duplicates of the Brewer collection, it is proper to designate the materi- al at GH as lectotype. The sheet at US bears a sketch and notes by John Torrey: "75. Obione — near the Sea, at Sta Monica: Probably O. lentiformis (large fruited) in an abnormal state. The bracts appear to have been changed by galls. Recognition of A. breweri at taxonomic level as either a variety or subspecies of A. lentiformis is not without merit. Indeed, the plants have typically larger leaves and fruiting bracteoles that average larger However, there is a series of intermediates that connect the robust coastal material with the less robust plants in the interior. Plants designated as belonging to A. breweri are considered by me as ta.xonomically negligible. Atriplex buxifolia Rydberg, Bull. Tone\ Bot. Club 39: 311. 1912. = A. gardneri van aptera (A. Nelson) Welsh Type locality: Wyoming, Sheridan Co., Dayton, 1220 m altitude, September 1899, Tweedy 2456; holotype NY! Type; "F Tweedy 2656 (2456 in publication), Dayton, 4000 ft, Sheridan Co., "Wyoming, September 1899"; lecto- type NY! (Basset et al. Genus Atriplex in Canada 58. 1983). Atriplex canescens (Pursh) Nuttall, Genera N. Amer PI. 1: 197. 1818. Basionym: Calligonum canescens Pursh Putative or actual hybrids are known between A. canescens and A. confertifolia or A. gardneri (various vari- eties). Such hybrids are only occasiontd; tliey do not swamp the characteristics of the taxa nor persist as populations. The two e.xceptions to the sporadic nature of the hybrids involving A. canescens as one of the parental types are A. gardneri var. bonnevillensis and A. gardneri var aptera. Neither of these ovei-whelms the parental taxa, but being long-lived, they persist for long periods of time and occu- py rather large areas in specific habitats. Bracts with four wings appear to have arisen independently at several places within the woody atriplexes. Such a condition is not necessarily an indication of close genetic affinities. Indeed, the garrettii and acanthocarpa complexes seem to be more distantly removed from A. canescens than from other taxa. Atriplex canescens var. angtistifolia (Torrey) S. Watson, Proc. Amer Acad. Arts 9: 121. 1874. — A. canescens (Pursh) Nuttall Basionym: Obione occidentale var angiistifolia Torrey Narrow-leaved shrubs from west Texas are transitional with broader-leaved materials both there and elsewhere. They do not seem to constitute a taxon. Atriplex canescens ssp. aptera (A. Nelson) Hall & Cle- ments, Publ. Carnegie Inst. Wash. 326: 343, pi. 58. 1923. Basion\in: A. aptera A. Nelson = A. gardneri \ ar. a])tera (A. Nelson) Welsh Atriplex canescens var. aptera (A. Nelson) C. L. Hilchc., Vase. Pis. Pacif NW. 2: 186. 1964. Basionym: A. aptera A. Nelson = A. gardneri var aptera (A. Nelson) Welsh Atriplex canescens ssp. garrettii (Rydberg) Hall & Clements. Publ. Carnegie Inst. Wash. 326: 344. 1923. Basionym: A. garrettii Rydberg Atriplex canescens var. garrettii (Rydberg) Benson, Amer J. Bot. 30: 236. 1943. Basionym: A. garrettii Rydberg Atriplex canescens var. gigantea Welsh & Stutz, Great Basin Nat. 44: 189. 1984. Type: Utah, Juab Co., Lynndyl sand dunes, T35S, R4W; 8 Sept. 1965, S. L. Welsh & G. Moore 5126; holo- type BRY!; isotype NY! The variety is based on its very broad bracts, stems that produce roots b\' layering, thus accommodating bur- ial in dimes, and diploid chromosome number Atriplex canescens var. laciniata Parish, in Jepson, Fl. Calif. 442. 1914. = A. canescens X A. polycarpa? as to possible origin. Type locality: California, Imperial Co., "Caleb, Colorado Desert, Parish 8256" (Jepson I.e.). Type: California, Imperial Co., "Plants of Southern California, Salton Basin, Caleb. About 200 feet below sea level. No. 8256. Coll. S. B. Parish. Oct 11. 1911"; holotype UC J E PS!; isotype GH! This variety has been suggested as based on speci- mens intermediate between A. canescens and A. linearis (C. A. Hanson I.e.), although Stutz (personal communica- tion 1994) poses quite another possibility, i.e., that a chro- mosomal race of A. polycarpa forming hybrids with A. canescens has resulted in at least partially stabilized popu- lations of var laciniata within the Salton Basin. The type is characterized by deeply laciniate, 4-lobed bracteoles within the size range of A. canescens. It has slender branch lets and narrow leaves approaching those of both A. linearis and A. canescens var niacilenta. which had a similar origin from a separate chromosomal race of A. polycarpa forming hybrids with A. canescens. Atriplex canescens ssp. linearis (S. Watson) Hall & Cle- ments, Publ. Carnegie Inst. Wash. 326: 344, pi. 58. 1923. Basionym: A. linearis S. Watson = A. linearis S. Watson Atriplex canescens var. linearis (S. Watson) Munz, Manual S. Calif Bot. 141. 1935. Basionym: A. linearis S. Watson = A. linearis S. Watson Atriplex canescens ssp. macropoda (Rose & Standley) Hall & Clements, Ph> log. Meth. 11«on 344. 1923. Basionym: A. macropoda Rose & Standley This ta.\on is known from Baja California. Atriplex canescens var. macilenta Jepson, Fl. Calif 1: 442. 1914. Tvpe locality: California, Imperial Co., "Holhille, Colo- rado Desert, Parish 8258" (I.e.). 1995] North American Perennial Atr/plex Types 325 Type: California, Imperial Co., "Plants of Soutliern California. Salton Basin. Bluffs of Alamo River, Halhartle. About 15 feet below Sea Level, S. B. Parish 8258, Oet. 18, 1912"; holotype UC JEPS!; isotypes DS (Xale.vico"), GH!, POM! The type has leaves to 4 mm wide, narrowly oblanceo- late and obtuse apically. Bracts are small, as in A. linearis, and toothed along the margin of the wings. The toothed margin of the wings hints at the laciniate nature of bracts on plants from the nearby Salton Basin and named van laciniata Parish. Plants called var. macilento approach but do not e.xactly match the more characteristic specimens of A. linearis from southern Arizona and northern Mexico. According to Stutz (personal commimication 1994), the var. macilenfa is a high polyploid, while A. linearis is a diploid. The relatively broader, thicker leaves of var. maci- lenta are apparently diagnostic. The specimen at DS, Parish 8258, Oct. 1912, is labeled as having been taken on "Bluffs of the Alamo, Calexico." It is one of three localities cited under Parish's number 8253, and the specimens other than the one taken at Holtville are probably best considered as paratypes. Parish made a series of collections from the Salton Basin in October 1912. His numbers 8255 and 8256 were collected on October 11; 8255 is a small-bracteoled, nar- row-lea\'ed plant assignable to van uiacileuta, the type of which (8258) was taken on 18 October Parish's number 8256, the type of var. laciniata, is evidently closely placed geographically within the Salton Basin, which also sup- ports A. pohjcarpa, which is potentially involved in the origin of both vars. macilenta and laciniata through hybridization with different chromosome races of A. pohj- carpa through hybridization with A. canescens. Number 8255 approaches A. linearis in size of bracts and width of leaves, and possibly that species is also involved in the derivation of both vars. laciniata and macilenta. Atriplex canescens var. occidentale (Torrey & Fremont) Welsh & Stutz, Great Basin Nat. 44: 188. 1984. Basion\'m: Pterochiton occidentale Torrey & Fremont = A. canescens (Pursh) Nuttall var. canescens This name was resurrected on false supposition that tlie type of A. canescens sensu stricto differed from the tall phases of the plant so widely distributed in the American West. It is an unfoiiunate later synonym. Atriplex collina Wooton & Standley, Contr. U.S. Natl. Herb. 16: 119. 1913. = A. confeiiifolia (Torrey & Fremont) S. Watson Tyjje: Aiizona, Apache Co., "diy hills near the north end of the Carrizo Mountains," P C. Standley 7481, 31 July 1911; holotype US! Atriplex confertifolia (Torrey & Fremont) S. Watson, Proc. Amer. Acad. Arts 9: 119. 1874. Basionym: Obione confertifolia Torrey & Fremont, in Fremont Atriplex corrugata S. Watson, Bot. Gaz. 16; 341. 1891. T\pe locality: "Nearly allied to A. nutiallii. Discovered by Miss Alice Eastwood at Grand Junction, Colorado, in well formed fruit on 20th May, 1891. Miss Eastwood notes it as the earliest in fruit of several perennial species of the genus growing in the same locality (I.e.). lype: "Atriplex conaigata Watson, n. sp. Grand Junction, Colorado. Miss Alice Eastwood — May 20/1891 "; holotype GH!; isotypes UC (fiagments taken from holotspe bv H. M. Hall)!, K,' MO, US! The type consists of two fertile branches, one pistillate and the other staminate. Both have the small, narrow leaves characteristic of the taxon throughout its rather small range. The species is almost exclusively restricted to saline substrates of such fine-textured strata as the members of the Cretaceous Mancos Shale and Jurassic Morrison Formation, inter alia, where it often occurs as a monotype. It forms occasional hybrids with A. confertifolia and A. gardneri var. cuneata, with whom its ecology is sporadic. The taxon is probabh' most closely allied to the latter, with which it shares large land areas, but from which its aute- cology is restricted. It is regarded herein at species rank because of the maintenance of morphological integrity despite occasional contact with the other taxa over much of its area. Additionally, there are hints in its morphology of close ties in still another direction, i.e., with A. ohovata. Atriplex cuneata A. Nelson, Bot. Gaz. 34: 357. 1902. — A. gardneri var. cuneata (A. Nelson) Welsh Type locality: "M. E. Jones 5443, Emery, Utah, 1894," Nelson (1902). Type: M. E. Jones 5443, Emeiy, 7000 ft., Emeiy Co., Utah, 16 June 1894; holotype RM!; isotypes MO!, NY! (3 sheets), US! Atriplex cuneata ssp. introgressa C. A. Hanson, Stud. Syst. Bot. Brigham Young Univ. 1: 4. 1962. = A. gardneri var. cuneata X var. tridentata Type: Utah, Carbon Co., "Wellington, ca 0.1 mi S of Price River, in clay hills along road leading to city dump," 9 July 1961, Hanson 346; holotype BRY!; isotypes GH!, POM! The specimens on which this taxon are based demon- strate intermediacy between the cuneata and tridentata phases of A. gardneri. Their recognition at any taxonomic level is problematical. Atriplex curvidens T. S. Brandegee, Proc. Calif Acad. Sci. II, 2: 201. 1889. = A. pohjcarpa (Torrey) Watson Type: Baja California, Comondu, four feet high, rounded April 24, 1889, Brandegee sn; holotype UC! Atriplex decumbens S. Watson, Proc. Amer. Acad. Arts 12: 275. 1877. = A. watsonii A. Nelson. Type locality: California, "Near San Diego; Dr. E. Palmer, 1875 (n. 334)" (Watson 1877). Type: "Southern part of San Diego Co., California. Coll. Edward Palmer, M.D., 1875. No. 334. Atriplex decumbens, Watson n. sp. San Diego"; holotype GH!; iso- type NY! (2 sheets). The type consists of a small and a large branch, both staminate. The large branch is evidently from a sprawling herbaceous perennial. Leaves are luostly opposite, becom- ing subopposite above, elliptic to ovate-lanceolate, obtuse to roinided apicalK'; the glomerules are 3-5 mm thick and are arranged in terminal spikes 1—4 cm long. Atriplex eremicola Osterhout, Bull. Torrey Bot. Club 25: 284. 1898a. nom. no\-. Basionym: A. fruticulosa Osterhout. = A. gardneri (Moquin-Tandon) Dietrich var. gardneri 326 Great Basin Naturalist [Volume 55 Atriplex falcata (M. E. Jones) Standley, N. Anier. Fl. 21; 68. 1916. Ba.sionvni: A. mittallii \dr. fiilcata M. K. Jones, Coiitr. W. Bot. ll! 19. 1903. = A. gardneri vnr. falcata (M. E. Jones) Welsh Atriplex fruticosa Nuttall ex Moquiii-Tandon, in de Candolle, Prodr. 13(2): 112. 1849. pro syn. = A. gardneri var. gardneri Type: "Atriplex * fruticosa. A. Haliinuni afiinis. R. Mts." Nuttall; holotypcBM! The type oi A. fruticosa is mounted with collections with the notation "British North America. Dr. Richardson 1819-28, " and designated as A. caiu'scens. In Inde.x Kewen- sis the name fruticosa is noted as a synonym of A. canes- cens, a supposition possibly based on the identity of the Richardson material, but more probaljly on the publica- tion of the name as a synonym of A. canescens by Moquin- Tandon. The epithets /n/licosa and heterophyUa, both herbarium names of Nuttall, were published as synonyms and are not to be regarded in considerations oi priority. Atriplex fruticulosa Jepson, Pittonia 2: 306. 1892. Type: California, "Little Oak, Solano Co., Aug. 16, 1892. Willis L. Jepson"; holotyiDe UC!; isotype MO! This plant functions mostK' as an annual but is appar- ently capable of a longer life span, extending to become a short-lived perennial. The name has priority' over the later homonym, A. fruticulosa Osterhout (1898). Atriplex fruticulosa Osterhout, Bull. Torrey Bot. Club 2.5; 207. 1898. non A. fruticidosa Jepson 1892. Basionym for: A. eremicola Osterhout = A. gardneri (Moquin-Tandon) Dietrich var. gardneri Type locality: Wyoming, Albany Co., Steamboat Lake, "The type was collected near a small alkaline lake in South- ern Wyoming," G. Osterhout s.n. 2 July 1896; holotype (no. 1324) RM!; isotype NY!, RM! (this second sheet, pre- sumably an isot\'pe, lacks the collector's number). A col- lector's number was not cited with the protologue, but the holobt'pe sheet at RM bears the number 1324. Mateiial on which this entity was based differs in no particular way from A. gardneri var. gardneri. Atriplex gardneri (Moquin-Tandon) Dietrich, Syn. Pi. 5; 537. 1852. Basionym: Ohione gardneri Mocjuin-Tandon There is a sheet, possibly identifiable as belonging to this species and not bearing on the nomenclature of the species, in the Lewis and Clark herbarium at PH; "A half shrub from the high plains of Missouri. July 20th 1806." It is cited here to demonstrate that the species was known from the earliest collections into the western plains. Atriplex gardneri var. aptera (A. Nelson) Welsh & Crompton, comb. nov. Basionym: Atriplex aptera A. Nelson, Bot. Gaz. 34: 356. 1902. ' This entity was treated by Hanson (1962) as a proba- ble derivative of hybridization between A. canescens and A. buxifolia (A. gardneri sens. kit.). It is a low subshrub most similar to the latter, but with bracteoles winged as in A. canescens or with tubercles aligned in foiu' rows, and with yellow staminate flowers. It is likely that the condi- tion of 4-winged fruits has arisen independently on many occasions and that the resulting populations are not asso- ciated genetically as in a typical taxon. Regardless of origin. however, the resultant plants are readiK recognizable and are widely distributed from southern Canada south along the plains to Nebraska and Wyoming. A. canescens also forms hybrids with other phases of the gardneri complex (see below). Atriplex gardneri var. honnevillensis (C. A. Hanson) Welsh, Great Basin Nat. 44: 190. 1984. Basionym: A. honnevillensis C. A. Hanson This \ariety is more or less intermediate between A. gardneri \ar falcata and A. canescens, but it most nearly resembles the former in habit. The bracteoles are 5-8 mm long and 3-9 mm wide, ovoid, with four lateral wings or rows of flattened tubercules to 3 mm wide, or the wings rarely absent. The plants are confined to playas and saline pans in the valleys of western Utah and across Nevada. Atriplex gardneri var. cuneata (A. Nelson) Welsh, Great Basin Nat. 44; 191. 1984. Basionym: A. cuneata A. Nelson Atriplex gardneri \ar. falcata (M. E. Jones) Welsh, Great Basin Nat. 44: 191. 1984. Basionym: A. inittaUii vdi: falcata M. E. Jones Atriplex gardneri var. tridentata (Kuntze) Macbride, Contr Gray Herb. 3: 11. 1918. = A. gardneri var utahensis (M. E. Jones) Dorn Basionym: A. tridentata Kuntze Atriplex gardneri var. welshii (C. A. Hanson) Welsh, Great Basin Nat. 44: 191. 1984. Basionym: A. welshii C. A. Hanson Atriplex gardneri var. utahensis (M. E. Jones) Dorn, Vase. PI. Wyo. 130. 1988. Basionym: A. nuttallii \m: utahensis M. E. Jones Atriplex garrettii Rydberg, Bull. Tone)' Bot. Club 39; 312. 1912. T\pe; Utah, Grand Co., "Vicinit>' of Moab, " JuK* 1-2, 1911, P A. Rvdberg & A. O. Garrett 8465; holot>pe NY!; isotypesGH!;US!, UT! Despite earlier treatments in which this taxon was regarded at infraspecific status within A. canescens, the nearest allies appear to be in the gardneri complex. Apparent hybrids are known between A. garrettii and A. confertifolia (C. A. Hanson 1962), but not witli A. canescens. Atriplex garrettii var. navajoensis (C. A. Hanson) Welsh & Crompton, comb. nov. Basion\m: A. navajoensis C. A. Hanson, Stud. Syst. Bot. Brigham Young Univ. 1; 3. 1962. This variet\' differs from the type material in plant size, length of staminate inflorescences, color of staminate flowers, and other intangibles. Generalh' the plants are very similar. The few known localities, from the vicinit>- of Lee's Ferry to Navajo Bridge in Coconino County, AZ, are only disjunct by about 100 km from the nearest popula- tions of \'ar. garrettii. Atriplex gordoni Hooker, J. Bot. 5; 261. 1853. nom. nov. pro A. gardneri McMiuin-Tandon. = A. gardneri (Mocjuin-Tandon) Dietrich var gardneri Atriplex greggii S. Watson, Proc. Amer. Acad. Arts 9; 118. 1874. 1995] North American Perennial Atr/plex Types 327 = A. ohovata Moquin-Taiidon Type locality: "New Mexico to Sonora. Collectors: — 1346 Berlandier; 462 Gregg; Emorv; Thurber; Bigelovv; 572, 1137, 1138 Wright" (Watson I.e.). Paratypes: "No. 462. Atriple.x obovata Moc]. Perros Bravos, Coahiiila, Mexico, Dr. J. Gregg, leg. 1S48-49" (GH Lowell!); "Berlandier, No. 1346. Bae de del Salad. San Luis Potosi, 1827" (Gil!). Type: "462. Atriplex. Perros Bravos, north of Saltillo. 1 ft. tall. Abundant. State of Coahiiila, Mexico. Dr. J. Gregg, leg. Sept. 20, 1848"; lectotype GH! (I. M. Johnston, J. Arnold Arb. 25(2): 147. 1944); isolectot\'pe GH Lowell! Atriplex griffithsii Standley, N. Amer Fl. 21: 63. 1916. = A. lentifonnis \ar. grijfithsii (Standley) L. Benson Type: Arizona, Cochise Co.: "Wilcox," Griffiths sn. 1895,' Oct. 12, 1900; holotype NY!; isotype US! This is a distinctive tiixon with silveiy, thick leaves. It is disjunct fiom die remainder of the species. Atriplex heterophylla Nuttall ex Moquin-Tandon, in de Candolle, Prodr. 13(2): 112. 1849. pro syn. = A. gardneri (Moquin-Tandon) Dietrich var. gardneri Type: "Atriplex * heterophylla. R. Mts." Nuttall; in- tended type BM! This is yet another herbarium name by Nuttall cited as a synomym of A. canescens by Moquin-Tandon in de Candolle's Prodromus. It again demonstrates that the species was well represented in collections prior to the collection of the type material of A. gardneri. Atriplex hymeneltjtra (Torrey) S. Watson, Proc. Amer. Acad. Arts 9: 119. 1874. Basionym: Obione hijmenelijtra Torrey Atriplex johnstonii C. B. Wolf, Occas. Pap. Rancho Santa Ana Bot. Card. 1:3. 1935. = A. numimilaria Lindl. Type: California, Los Angeles County, Coastal cliffs, Playa del Rey. C. B. Wolf 1821, 23 Dec! 1930; isotvpes CAS!, GH!, NY! The isotype at GH consists of four woody, leiily branches, two ot them with fruiting bracts. Leaves are short-petio- late, with blades 1.2-3.5 cm long and 1-3 cm wide. Atriplex jonesii Standley, N. Amer. Fl. 21: 65. 1916. nom. nov. pro A. sahitlosa M. E. Jones. = A. ohovata Moquin-Tandon Atriplex lentifonnis (Torrey) S. Watson, Proc. Amer Acad. Arts 9: 118. 1874. Basionym: Obione lentifonnis Torrey, in Sitgreaves This is a wami-desert species, important in saline pans along drainages at low elevations in the valleys of the Colorado and Gila rivers and Salton Sink. The species is distributed from western and southern Arizona, through southern Nevada and California, and also in Mexico. Hanson (1962) notes that A. lentifonnis sens. lat. forms liybrids with A. leiicophijlla (Moquin-Tandon) Dietrich, a perennial not especially woody species, and possibly even with an annual species. Such hybridizations might indi- cate that A. lentifonnis and its near relative A. torreiji have alliances elsewhere than with the other wood)' species treated herein. Atriplex lentiformis ssp. hreueri (S. Watson) Hall & Clements, Publ. Carnegie Inst. Wash. 326: 335, pi. 54. 1923. Basionym: A. breiveri S. Watson = A. lentifonnis sens lat? Atriplex lentiformis var. breweri (S. Watson) McMinn, Man. Calif Shrubs 113. 1939. Basionjm: A. breweri S. Watson = A. lentifonnis sens lat? Atriplex lentiformis ssp. griffithsii (Standley) Hall & Clements, Publ. Carnegie Inst. Wash. 326: 336, pi. 55. 1923. Basionym: A. grijfithsii Standley = A. lentifonnis sens, lat? Atriplex lentiformis var. griffithsii (Standley) Benson, Amer. J. Bot. 30: 236. 1943. Basionym: A. griffithsii Standley — A. lentifonnis sens, lat? Atriplex lentiformis ssp. torreiji (S. Watson) Hall & Clements, Publ. Carnegie Inst. Wish. 326: 335. 1923. Basionym: Obione torreiji S. Watson Atriplex lentiformis var. torreiji (S. Watson) McMinn, Man. Calif Shrubs 113. 1939. Basionym: Obione torreiji S. Watson Atriplex linearis S. Watson, Proc. Amer Acad. Arts 24: 72. 1889. T^pe locality'; Mexico, Sonora, alkaline soil about Guav- mas. Palmer 120, 121, 235; s>'nt>pes GH. Parat>pes: "Flora of Gua>anas, Mex. Dr. Edward Palmer, 1887. No. 120. Atriplex linearis Watson, n. sp. Garden fences in alkaline soil. July"; GH! and "Flora of Guaymas, Mex. Dr. Edward Palmer, 1887. No. 121. Atriplex Linearis, Watson, n. sp. Garden fences, alkaline soil. July"; GH! Type: "Flora of Guaymas, Mex. Dr. Edward Palmer, 1887. No. 235. Atriplex linearis Watson, n. sp. Plains in alkaline soil. Sept."; lectot\pe GH! (G. D. Brown, Amer. Midi. Nat. 55: 210. 1956). Paratypes 120 and 121 are immature, the former pistil- late, the latter staminate. The lectotype sheet #235 has at least four branches with more or less mature fruiting bracteoles. The bracteoles are 4-winged, rather deeply laciniately lobed to merely toothed along the wings, and are 3-6 mm wide. Hanson (1962) regarded A. linearis as the most sub- stantial variant within the canescens complex but recog- nized that it forms hybrids with A. canescens. The plants are certainly moiphologically distinct from most phases of that entity. The slender, short to elongate leaves (seldom more than 4 mm wide and to 3.8 cm long), fi-uiting brac- toles seldom over 6 or 7 mm wide, and very slender branchlets are apparently diagnostic in most instances. Atriplex macropoda Rose & Standley, N. Amer. Fl. 21: 72. 1916. = A. linearis S. Watson (sens lat?, but the fruiting bracteoles are long pedicellate, unlike A. canescens) T\pe localit}': "T\pe collected on Pinchillinque Island, Lower California, March 27, 1911, J. N. Rose 16518 (U.S. Nat. Herb. no. 638567)." Type: Lower California, Pinchilinque Island, Gulf of California, J. N. Rose 16518, March 27, 1911; holotype US! Atriplex matamorensis A. Nelson, Proc. Biol. Soc. Wash. 17: 99. 1904. 328 Great Basin Naturalist [Volume 55 Nom. ii()\. pro. A. opixi.sitifolia S. Watson Atriplex navajoemus C. A. Hanson, Stud. S> st. Hot. Brig- ham \bung Univ. 1: 3. 1962. = A. t^airettii var. navajoensis (C. A. Hanson) Welsh 6c Crompton Type; "Arizona: Coconino Co., east side of the Navajo Bridge, July 21, 1961," C. A. Hanson 388; Iiolotype BRY!; isoh'pe CH! Atriplex X neomexicana Standley, N. Amer Fl. 21: 67. 1916. = A. gardneri van cimeata X A. confertifolia Type locaht)'; "Type collected on dry hills near Rmning- ton, New Mexico, altitude 1550-1650 m, July 19, 1911, Paul C. Standley 7066 (U.S. Nat. Herb. no. 686089)." Tvpe: New Mexico, "Diy hills near Farmington," San Juan Co., New Mexico, July 19, 1911, E C. Standley 7066; holot>'pe US! The name is evidenth' based on plants intermediate between A. gardncri var. cuneata and A. confertifolia. Atriplex nummularia Lindley, Mitch. J. Exped. Trop. Australia 64. 1848. T\'pe: Australia, "Cultivated in Italy, seed from South Australia"; holot\'pe not seen. Atriplex mtttallii S. Watson, Proc. Amer Acad. Arts 9: 116. 1874. nom. nov. = A. canescens (Pursh) Nuttall sens. str. It is unfortunate that one must at this late date attempt to analyze Watson's use of the name mittallii for a portion of the woody atriplexes in the American West. From its publication in 1874 the name has been the source of much confusion, sei-ving to clutter Atriplex nomenclature for all subsequent time. It seems certain fiom a study of Watsons proposal, justification for which can only be inferred, that he was merely presenting a new name for material that he thought to be misinteriDreted by contemporary' botanists. The evolution of botanical thought with regard to the perennial atriplex species parallels that for other newly discovered ta.xa in the American West and was initiated when the first of the woody specimens arrived from west- ern botanical explorers. Few names were available, speci- mens were few and often fragmentan, literature was diffi- cult to obtain, and it was easy to misapply concepts and mix names, a symptomology not of that era alone. Supposed sensu names cited by Watson (1874) within the synomymy of A. nuttall ii include Atriplex canescens as used by Nuttall and an assortment of other historical authors, Obione canescens of Moquin-Tandon and other authors, and still another synonym, i.e., "A. gordonii Hook.," with the citation "Pi. Geyer in Lond. Jour. Bot. 5: 261?," and l)\ implication the type of A. gordonii (i.e., A. gardneri). Watson first cited the name A. canescens as published by Nuttall (1818), the implication being that Calligonwn canescens Pin'sh, basionym of A. canescens, could not appK'. Nuttall is indeed author of the combination Atriplex canescens, and the place of citation is his 1818 publication, wherein he cites C. canescens as the basionym of his com- bination; furthermore, Nuttall's description is clearly C. canescens Pursh, sensu stricto. It is Watson's understand- ing of Nuttall's use of the epithet that is in error. Thus, A. canescens of Nuttall is certainly not a mere sensu name, however one might wish to interpret the application of the epithet. Both the name and the concept as supplied by Nuttall are A. canescens, including its basionym. A. mtttallii of Watson thus includes the type oi Calligonwn canescens, and the epitiiet nntlallii is illegitimate under stipulations of the International CJode. Hence, from a nomenclatural viewjDoint there is no problem. Nuttall based his Atriplex canescens squarely on CaUigonum canescens Pursh, and Watson quoted A. canescens Nuttall as the name-bringing synonym of A. mittallii, which was stillborn. The lectotype of Caligomiin canescens Pursh is at PH and is therefore the lectotype of both Obione canescens and A. mittallii, which cannot be transferred to a different species or brought to life by sophisticated arguments. Hence, the proposal for lectotypification by McNeill et al. (1983) is illegitimate. Atriplex mittallii var. anomala M. E. Jones, Contr W. Bot. 11: 19. 1903. = A. gardneri var. falcata (M. E. Jones) Welsh Type localitv: "The type is my specimens from Dolly Varden Smelten E. Nevada, July 1894 [1891]." Type: Nevada, Elko Co., "Marcus E. Jones Herbarium. Atriplex nuttallii var anomata [sic] Jones n. var Dolly Varden at the Smelter, VII-24-91. N.W of Ibapah, Utah." M. E. Jones sn; holotype POM!; isotype UC (frag.)! Jones was clearly in error in citing the date of the col- lection as 1894. His itinerai-y cited in Leaflets of Western Botany (10: 189-236) places him at the Dolly Varden Smelter on 24 July 1891, not 1894. Atriplex mtttallii ssp. buxifolia (Rydberg) Hall & Clements, Phylog. Meth. Taxon. 325. 1923. Basionym; A. buxifolia Rydberg = A. gardneri (Moquin-Tandon) Dietrich var gardneri Atriplex mittallii corrugata (S. Watson) A. Nelson, in Coulter & Nelson, New. Man. Bot. Rocky Mts. 168. 1909. = A. corrugata S. Watson Atriplex mittallii ssp. cuneata (A. Nelson) Hall & Cle- ments, Publ. Carnegie Inst. Wash. 326; 324, f 45. 1923. Basionym; A. cuneata A. Nelson = A. gardneri var cuneata (A. Nelson) Welsh Atriplex nuttallii ssp. falcata (M. E. Jones) Hall & Cle- ments, Publ. Carnegie Inst. Wash. 326; 324. f 45. 1923. Basionym: A. nuttallii \m: falcata M. E. Jones = A. gardneri vm: falcata (M. E. Jones) Welsh Atriplex nuttallii vm: falcata M. E. Jones, Contr W Bot. 11; 19. 1903. = A. gardneri vm: falcata (M. E. Jones) Welsh Type locality: "Weiser, Idaho, July 1899, Jones" (I.e.). Type; Idaho, Washington Co., "Flora of Idaho. Type material. Atriplex nuttallii var falcata Jones n. \ar. Weiser, Wash. Co. July 7 1899. Alt. 2200 Ft." M. E. Jones sn; holo- type POM!; i.sotype UC! Atriplex nuttallii ssp. gardneri (Moquin-Tandon) Hall & Clements, Publ. Carnegie Inst. Wash. .326; 324. 1923. = A, gardneri (Mociuin-Tandon) Dietricli var gardneri Basionym: Obione gardneri Moquin-Txndon Atriplex nuttallii ssp. tridentata (Kuntze) Hall & Cle- ments, Publ. Carnegie Inst. Wash. 326: 324. 1923. = A. gardneri var iitahensis (M. E. Jones) Dorn Basionym: A. tridentata Kuntze Atriplex nuttallii van gardneri (Moquin-Tandon) R. J. Davis, Fl. Idaho. 261. 1952. 1995] North American Perennial Atr/plea Types 329 = A. gardneri (Moquin-Tandon) Dietrich van gardneri Btisionym: Obione gardneri Mncjuin-Tandon Atriplex inittaUii van tridentata (Kuntze) R. J. Davis, Fl. Idalio 261. 1952. = A. gardneri var. utaJiensis (M. E. Jones) Dom Basionym: A. tridentata Kuntze Atriplex nuttaUii var. titahensis M. E. Jones, Contr. VV. Hot. 11: 19. 1903. = A. gardneri var. utahensis (M. E. Jones) Dom T\pe locality-; "This is No. 1760 Jones from Salt Lake Cit}', and is the more common form in Utah." Type: Utah, Salt Lake Cit\-, Salt Lake Co., M. E. Jones 1760, 16 June 1894: holot>pe POM?; isot>'pe UC (frag.)! Atriplex oblanceolata Rydberg, Bull. Toney Bot. Club 31: 403. 1904. = A. gardneri var. ciineata (A. Nelson) Welsh Type locality: Colorado, Delta Co., Delta, Cowen 4071 (Rydberg 1904). Type: "Plants of Colorado. No. 4071. Atriplex oblance- olata Rydb. Delta, J. H. Cowen. Sept 3, 1897"; holotype NY!; isotypes GH!, RM! (2 sheets), US! Atriplex obovata Moquin-Tandon, Chenop. Enum. 61. 1840. Type locality: "In Peruvia. (v. s. in herb. Mus. Paris) ' (I.e.). Type: "No. 1346. Bae del Salad, Saint Louis Potosi. Dbre. 1827," and "Herbarium Berlandierianum Te.xano-mexi- canum. No. 1346. Atriplex obovata, Moq.! O. canescens, var? Torr. San Luis Potosi; Mexico, State of San Luis Potosi, Berlandier 1346"; lectotype P? (I. M. Johnston, J. Arnold Arbor 25[2]: 148. 1944); isolectotype GH! The isolectotype sheet at GH consists of three leafy branches, now lacking fruiting bracteoles or staminate flowers. The material is certainly a match for what has tra- ditionally passed under the name obovata; hence, there is no problem with its interpretation. Atriplex obovata var. tuberata Macbride, Contr Gray Herb. 3: 11. 1918. — A. obovata Moquin-Tandon Type locality: Texas, El Paso Co., Fornillo Creek, Harberd 103. Type: "No. 103 (see specimen of male). 1-2° [feet] high — Foliage & specially fruit different from that of A. acan- thocarpa. Tornillo Creek. W. Texas. Aug. [1S]S3. V. Havard, U.S.A."; holotvi^e GH!; isotype US! The sheet at GH has two branches, one staminate and one with fruiting bracteoles. The bracteoles are rather stronglv' tuberculate, a feature not unusual witliin the species as a whole. Atriplex occidentalis (Torrey & Fremont) Dietrich, Svn. PI. 5: 537. 1852. Basionym: Pterochiton occidentalc Torre\' 6f Fremont = A. canescens (Pursh) Nuttall Atriplex odontoptera Rydberg, Bull. Torrey Bot. Club 31: 404. 1904. — A. canescens X A. gardneri var gardneri Type: Wyoming, Johnson Co., "3302. Atriplex canescens (Pursh) James. A. odontoptera Rydb. (Type) Buffalo. Ele- vation 4000-5000 feet. Frank Tweed>'. September 1900"; holotype NY!; isofype RM! This is a coarse specimen, very woody and obviously intermediate between A. canescens and A. gardneri var gardneri Atriplex oppositifolia S. Watson, Proc. Amer Acad. Arts 9: 118. 1874, non DC. = A. matamorensis A. Nelson; Obione oppositifolia (S. Watson) Ulbrich, in Engler & Prantl Type locality': "In the Rio Grande Valley on the Mexican side, collected only by Berlandier (No. 3201, 'Matamoras to San Fernando ) (Watson I.e.). Type: "de Matamaras a San Fernando circa Guijano, Oct. 1830," and "Herbarium Berlandierianum Texano- Me.xicanum. No. 3201. A. oppositifolia n. sp. S.W! [initials are Sereno Watson's on sheet at GH]," Berlandier; holo- type GH!; isotype NY! The specimen at GH is doubly mounted with Palmer 1160, 1879. It is a portion of a herbaceous perennial with minute leaves ca 2-3 mm long and 1 mm wide. Bracteoles are conspicuously veined on the faces and prominently toothed lateral to the apical tooth. Atriplex orbicularis S. Watson, Proc. Amer Acad. Arts 17: 377. 1882. = A. lentifonnis (Torrev) S. Watson (the A. breweri S. Watson phase) Type locality: "At Santa Monica, California, on the sea- shore at the base of the bluffs; S. B. & W. F. Parish, October, 1881" (Watson 1882). Type: "Flora of Southern California, S. B. & F W. Parish, No. 1126, perennial, somewhat woody at base, 3-4 ft high, base of bluffs, sea shore, Sta Monica, Oct. 1881"; holotype GH!; isotypes DS!, NY!, US! The fruiting bracts are ca 3 mm high and 4 mm wide. Leaves are elliptical and obtuse, tapering basally to a short petiole. Atriplex pabularis A. Nelson, Bull. Torrey Bot. Club 25: 203. 1898. = A. gardneri var utahensis (M. E. Jones) Dom Type locality: Wyoming, Sweetwater Co., Point of Rocks, A. Nelson 4429, Aug. 30, 1897. Tv'pe: "A. Nelson 4429, Bitter Cr, Point of Rocks, 6500 ft, Sweetwater Co., Wyoming, 30 August 1897"; lectot\pe at RM! (Hall & Clements, Publ. Carnegie Inst. Wash. 326: 324. 1923); isolectotypes GH! (two sheets, male and female), NY!, US! Atriplex pabularis var. eremicola (Osterhout) A. Nelson, Coulter & Nelson, New Man. Bot. Rocky Mts. 168. 1909. Basionym: A. eremicola Osterhout = A. gardneri (Moquin-Tandon) Dietrich var gardneri Atriplex parrtji S. Watson, Proc. Amer Acad. Arts 17: 378. 1882. Type locality: "Near Colton [actually at Lancaster according to Parish in Zoe 5: 113, 1901], California; Dn C. C. Parry 1881"; holot\pe (Pam- 221) GH!; isotypes NY!, UC (frag.)! The type consists of a branched stem, with lateral spinescent stems to 4 cm long. The leaves are ovate-orbic- ular The plant is obviously allied to A. conferiifolia, but distinct. Atriplex polycarpa (Torrey) S. Watson, Proc. Amer Acad. Arts 9: 117. 1874. Basionym: Obione polycarpa Tomey 330 Great Basin Naturalist [Volume 55 Atriplex pringlei Standley, N. Amcr. Flora 21: 68. 1916. = A. acanthocarpa s.sp. pringlei (Standley) Henrickson Type locality: "Type collected on alkaline plains. Hacienda de Ango.stiiia, San Lui.s Potosi, Mexico, July 15, 1891, C. G. Prinj^le 3775 (U.S. Nat. Herb. no. 48298)."' Type: "Mexico, San Luis Potosi, alkaline plain, Hacienda de Angostura, 15 Jul 1891," C. C Pringle 3775; holotype US!;isotypeC;H! Atriplex sahidosa M. E.Jones, Contn \\'. Bot. 11: 21. 1903. non A. sdbulosd \\m\\\ 1890. Basioin in of: A. jonesii Standley = A. ohovata Moquin-Tandon Type locality: Arizona, Navajo Co., "No. 4109 Jones, Winslow, Ariz., Sept., 1884, distributed as A. Greggii" (Jones 1903). Type: "Flora of Arizona. 4109. Atriplex Greggii, Watson. Winslow, M. E. Jones, September 1, 1884"; holotype US!; i.sotypesGH!, NY!, POM! The isotype at GH consists of three branches, two sta- niinate and one pistillate. Atriplex spinifera Macbride, Contr. Gray Herb. 53: 11. 1918. Type locality: California, Kern Co., Maricopa Hills, May 15, 1913, East\vood 3269 (Macbride 1918). Type: "3269. Flora of California. Atriplex. Mai^copa hills, Kern Co., Alice Eastwood May 15. 1913"; holotype GH!; isotype CAS!, US! The holobt'pe at GH consists of a branched stem bear- ing lateral spinescent branches to 4.2 cm long; that at US consists of spinose branchlets and two packets of fruiting bracteoles. Atriplex spinosa (Moquin-Tandon) D. Dietrich, S\n. Pi. 5: 536. 1852. Basionym: Obione spinosa Moquin-Tandon, in de Candolle = A. canescens (Pursh) Nuttall Atriplex stewartii I. M. Johnston, J. Arnold Arbor. 22: 110. 1941. — A. acanthocarpa ssp. steuartii (I. M. Johnston) Henrickson Type locality: Mexico. Type: "Mexico: western Coahuila. Atriplex stewartii n. sp. Jour Am. Arb. 22: 110. 1941. Eastern border of the Llano de Guaje, along road from Tancjue del Aparejo 20 miles southeast of Tanque Armendais. Abundant on flats margining playa at base of Lomas del Aparego (3 miles south of Tangue Asparejo). Plant erect, 10-15 inches tall. I. M. Johnston, C. H. Muller No. 777. Aug. 28, 1940"; holotype GH! The plant is obviously allied to A. anthocarpa. the \ari- ably 4-winged fruiting bracteoles having been derived independently or possibly through introgression from A. canescens. Henrickson (1988) does not suggest the latter possibility but does note that the 4-winged condition is not consistent, that there is a transition from that condi- tion to those wheie the wings are replaced by radiating processes. Atriplex subconferta Rydberg, Fl. Rock-v Mts. 248. 1917 [1918]. = A. confertifolia (Torrey & Fremont) S. Watson Type locality: Idaho, between Twin and Sliosiione Falls, Nelson & Macbride 1379; holotvpe NY; isotvpes POM, UC. Type: "No. 1379. Atriplex confertifolia (Torr.) Wats. Dry bench lands, alt. 3700. Twin Falls and Shoshone Falls, 3700 ft., Idaho, July 27, 1911, Aven Nelson & J. F Macbride"; holotype NY!; isotypes GH!, MO!, PO.M, RM!, UC, US! This appears to be a small-leaved phase of A. confcrti- jolia of little or no taxonomic significance. Atriplex tetraptera (Bentham) Rydberg, Bull. Torre\ Bot. Club 39: 311. 1912. Basionym: Obione tetraptera Bentham = A. canescens (Pursh) Nuttall Atriplex torreiji (S. Watson) S. Watson, Proc. Amer. Acad. Arts 9; 119. 1874. Basionym; Obione torrcyi S. Watson Atriplex torreiji var. griffithsii (Standley) G. D. Brown, Amer. Midi. Nat. 55: 205. 1956. Basionym: A. grijfithsii Standley = A. lentifonnis (Torrey) S. Watson Atriplex tridentata Kuntze, Rev. Gen. Pi. 2: 546. 1891. = A. gardneri var. utahensis (M. E. Jones) Doni Type locality: Utah, Box Elder Co., Corrine, Kuntze ,3084^ 1874. Type: O. Kuntze .3084, "Bei Corinne am Salzsee, 7000 [much too higli] ft, [Box Elder Co.], Utah, September 1874"; holotype NY!; isotype? K! The specimen at K, labeled "Atriplex tridentata OKze n. sp. U.S. N. Am. zw. Cheyenne & Corinne. 7000'. Sept. 74. 3084. Herbarium Otto Kuntze, is perhaps best regarded as a paratype. Atriplex watsonii A. Nelson, Proc. Biol. Soc. Wash. 17: 99. 1904. nom. nov. pro A. decwnbens. Basionym: A. decwnbens S. Watson Atriplex welshii C. A. Hanson, Stud. S) st. Bot. Brigham Young Univ. 1:1. 1962. = A. gardneri var. welshii (C. A. Hanson) Welsh Type: "Utah: Grand Co., 4 mi south of Cisco along state highway 128, July 5, 1961"; C. A. Hanson .322; holotype BRY!; isotypes GH!, ISC! Calligontim canescens Pursh, Fl. Amer. Sept. 2: 370. 1814. = A. canescens (Pursh) Nuttall Type locality: Lyman or Buffalo counties. South Dakota, M. Lewis in 1804. Type; "Big Bend of the Missouri, Sept. 21, 1804," Lewis and Clark Herbarium; lectoype PH!, G. D. Brown, Airier. Midi. Naturalist 55: 209. 1956. The original description of Calligonum canescens Pursh is "C. dioicum, pidverulento-fruticulosum; folis lanceolatis, floribus iixillaribus glomeratis in apice ramulo- rum subspicatis, fructibus alatis, alis venosis cristato-den- tatis. In the plains of the Missouri, near the Big bend. H. July, Aug. v.s. in Herb. Lewis. Flowers exceeding small. Goats delight to feed upon this shrub." The sheet at PH contains three branches, the left one with immature friiit, the middle one sterile, and the one at right with mature fruiting bracteoles. This latter specimen was designated specifically as the lectotype by McNeill et al. (1983); it clearly fits the concept of the species as inter- preted by contemporary authors, except for Stutz and Sanderson (1979), who claim that the type belongs to what was subsequently named A. aptera A. Nelson, based on the 1995] North American Perennial Atr/plex Types 331 assumption that the Lyman County, South Dakota, type locahty is not within the current range of A. canescens as presently accepted but is within the range of A. aptera. Examination of a great many specimens from throughout the western plains has failed to yield a plant of A. aptera with fruiting bracteoles identical to the lectot\'pe, which is matched many times among the specimens traditionally passing as A. canescens. The type sheet bears the designation "Sept. 2L 1804," and the site of the Lewis and Clark camp on that date is adjacent to present Lower Brule, Lyman or Buffalo comi- ties, a short distance above the confluence of the White River. That portion of the Missouri River has been inun- dated by waters behind the Fort Randall Dam, far dovvTi- stream. Nuttall had traversed the river corridor in 1811, going upri\'er as far as Fort Mandan. The description and discussion by Nuttall (1818) of the species is pertinent to the inteipretation of die Lewis Wpe material. He describes the plant as about 3 or 4 feet high, with the "cali.x (i.e., fruiting bracteoles) 2-partecl, becoming indurated, acute, with 4 unequal cristated or dentated angles "; the habitat was designated: "On the denudated saline hills of the Missouri [possibly a reference to the lower-growing, vari- able, gc/rc/;j('n'-like A. aptera]; commencing about 15 miles below the confluence of the White River, and continuing to the mountains [i.e., to the Mandan, as near as he went toward the mountains]. Much of the habitat where plants typical of A. canescens, as traditionally inteipreted, could ha\e grown is beneath the waters of Fort Randall Dam, and a valid assumption that bi'pical A. canescens did not occur there cannot be made. Some plants from areas of South Dakota adjacent to Lower Brule clearly approach hpical A. canescens. There is no justification for inteipre- tation of the name differently from that used in the his- toric past. Obione acanthocarpa Torrey, U.S. & Mex. Bound. Bot. 2: 183. 1859. = A. acantliocarpa (Torrey) S. Watson Type locality: "Plains between the Burro mountains; September, Bigeloiv. (in fruit.) On the Rio Grande, below Presidio del Norte; Parry. Near the Piloncilla, Sonora, September"; Thurber (No. 1739; Wright. His No. 1737 seems to be a slender form of the same.) Type: "Rio Grande below Presidio del Norte (El Paso), Aug." Pan-y s.n.; lectotvpe NY! (Henrickson Southwest. Nat. 33: 454. 1988); isolectotype NY! Obione berlandieri (Moquin-Tandon) Moquin-Tandon, in de Candolle, Prodr. 13(2): 114. 1849. Basionym: A. berlandieri Moquin-Tandon = A. canescens (Pursli) Nuttall Obione canescens (Pursh) Moquin-Tandon, Chenop. Enuni. 74. 1840. Basionym: Calligonuin canescens Pursh — A. canescens (Pursh) Nuttall Obione confertifolia Torrey & Fremont, in Fremont, Rep. Explor. Exped. Oregon & California 318. 1845. = Atriplex confertifolia (Torre)' & Fremont) S. Watson Type localits': "On the borders of the Great Salt Lake" (I.e.).' Type: "Obione confertifolia. Torn if Frem. in Freni. 2nd Reprt. (1845). Borders of the Great Salt Lake [near mouth of Weber River, Weber Co.], Utah. 761. 1843"; Fremont s.n. probably 10 September 1843; holotjpe NY! This species is noted by Fremont (1845) in his journal entry for 10 September 1843, on his return trip from Disappointment [Fremont] Island. The plant was probably collected on the trip from the water's edge to the camp on the lower Weber River, in Weber Co., Utah. The holot\pe consists of a single branch in young fruit. The sheet bears the notation in Torreys handwriting, "Obione rigida var. confertifolia n.sp. (crossed out) T. & F" Below the notation is a drawing of a fioiiting bract, with one side folded back, and an ovary. This is clearly the specimen on which the species was based. There is a second sheet at NY!: "Fremont's 2nd Expedn." with the notation "Grayia or near it." The specimen has male inflorescence fragments and clearly is not a portion of the t>'pe collection. Obione coriacea (Forssk.) Moquin-Tandon, Chenop. Enum. 71. 1840. This Egyptian species was compared by Torrey and Fremont (Fremont 1845) with Obione confertifolia (see abo\'e). It does not occur in North America. Obione gardneri Moquin-Tandon, in de Candolle, Prodr. 13(2): 114. 1849. = A. gardneri (Moquin-Tandon) Dietrich var. gardneri Type locality: SE Wyoming or W Nebraska, "Ad La Platte, Gardn. n. 250 " (Moquin-Tandon in de Candolle I.e.). Tvpe: "Gordon 250. La Platte. Obione Gardneri Moq. A low female plant, lax spike in fruit," possibly 1843; holo- t>'pe K?; isot\pe GH! The fragments at GH consist of a leaf and two imma- ture fiiiiting bracteoles, probably taken from the t\pe at K (Hooker herbarium), which we have not seen. Writing on the fragment enxelope is in ink, but partly illegible. The name of the collector is subject to inteipretation, but is presumed to be "Gordon. " Moc]uin-Tandon inteipreted it as "Gardner, and named the species after the person assumed by him to be the collector. The epithet was spelled gardneri on purpose and is not an orthographic variant. It is legitimate under stipulations of the Inter- national Code. Obione hymenelytra Torrey, in Whipple, Pacif R. R. Rep. 4: 129. 1857. = Atriplex hymenelytra (Torrey) S. Watson Type locality: "Hills and gravelly places on the William's River [Bigelow]. This species was found by Dr. Parr)' and by Colonel Fremont on the Gila" (I.e.). Type: "Fremont's Expedition to California, 1849. Obione hymenelytra, n. sp. " (lectotype NY'!, Brown, Anier. Midi. Nat. 55: 203. 1956.). "Fremont's Expedition to California, Gila" [1849] (presumed isolectotypes NY Crooke!, GH!). A third sheet, "Fremont's 2nd Expedition," is at NY! Except for the sheet designated as lectot>pe, the Fremont materials fi-oiii 1849 are scant}', consisting mainly of fh^iiting bracts (presumed isolectotypes NY!, GH!) and a branchlet of equivocal source (GH!). The lectotype at NY bears all of the accoutrements of a Torrey type specimen, except for lack of illustrations, but includes a descriptive label in Torre\ 's handwaiting and the name Obione hymenelytra, 11. sp., on the label. Obione lentiformis Torrey, in Sitgreaves Rep. 169. 1854. = Atriplex lentiformis (Torrey) S. Watson T\pe localit)': Ciilifoniia, along the Colorado River, S. W. Woodhouse s.n., 6 November 1851 (Sitgreaves E.xpedition, November 1851) (I.e.). 332 Great Basin Naturalist [Volume 55 Type: "Sitgreaves Report hSol. ()l)i()iic Iciitifonnis Toit. in Sitgreaves ex Torrey. Rio CJoIoracIo, (^alil. — Nev. ex Torrey "; "Rio Colorado. Nov. 6th 1851. Dr. Woodhoiise "; lectotype NY! (.selected by E V. Covillc, Contr. U.S. Natl. Herb. 4: 181. 1894); isolectotype GH! Obione lentifonnis (i rhomhifolia Ibrrey, Pacific R. R. Rep. 4: 129. 1857. Type: Arizona, NY? I have been unable to locate material at NY with this designation nor make a detennination as to its disposition tiLxonomically. Obione leucophylla Moquin-Tandon, in de CandoUe, Prodr 13(2): 109. 1849. = Atriplex leucophylla (Mo(iuin-Tandon) D. Dietrich Type locality: "In California (Chamisso!), San-Fran- cisco (Barclay!)" (I.e.). Type: "San Francisco. Barclay ; holotype K! Obione obovata (Moquin-Tandon) Ulbrich, Natm. Pfl. ed. 2. 16(c): 508. 1934. = A. obovata Mocjuin-Tandon Obione occidentalis (Torrey & Fremont) Moquin-Tandon, in de Candolle, Prodr. 13(2): 112. 1849. Basionym: Pterochiton occidentale Torrey & Fremont == A. canescens (Pursh) Nuttall Obione occidentale van angtistifolia Torrey, in Emoiy, Bot. Mex. Bound. 2(1); 189. 1848. = A. canescens (Pursh) Nuttall Type localit\': Texas, Valley of the Rio Grande, Wright in 1852. Type: "Field No. 394. Obione, Sandy ridge on Rio Grande, 3-5 ft tall, much branching, June 17, 1852. Rio Grande below El Paso Te.xas. [Wright] 1742 = 394"; holo- type NY!; isotypes GHl (3 sheets). All three sheets at GH bear the number 1742 on the laliel. One of them also has the number 394, which was evidently the field collection number The number 1742 was subsequently applied. The specimens all have veiy narrow leaves to ca 4 mm wide and immature fruiting bracteoles. The sheet at GH with the number 394 is dou- bly mounted with a second Wright collection (1741 = No. 24), which has 4-winged fruiting bracteoles to 7 mm wide. Notes appear above both labels on the sheet. That above 24 reads: "24, Chenop. hills near Erontera, 3-4 ft tall, branching widely, July 19, 1851. El Paso Co., Texas"; above 324 is, "324. Obione, sandy ridges on Rio Grande, 3-5 feet tall, much branching, June 17, 1852, Rio Grande below El Paso, Texas.' The latter is an isotype. All of the specimens appear to be A. canescens, sens. lat. Specimens with narrow leaves occur here and there throughout the range of the species. Those from western Texas that fit within the concept of van angustifolia seem not to represent a taxon worthy of consideration. Obione oppositifolia (S. Watson) Ulbrich, in Engler & Prand, Die Natun Pflanzenf Ed. 2. 16c: 508. 1934. Basionym: Atriplex oppositifolia S. Watson Obione polycarfm Torrey, in Whipple, Pacific R. R. Rep. 4: 130. 1857. = A. polycarpa (Torrey) S. Watson Type: Arizona, Graham Co., "With the preceding," i.e., "Hills and gravelly places, on William's River valley of the Gila River [near base of .\lt. Graham, ca 13 mi SW of Staiford]," October 28, 1846, Enioiy s.n.; holotype NY! Obione rigida Torrey & Fremont, in Fremont, Rep. Explon Exped. Oregon &: C'alifornia 318. 1845 (nom. nud.). = Atrij)lex confeiiifolia (Torrey & Fremont) S. Watson Authentic specimen: "Obione rigida T. & F On an island [Fremont Island] in Great Salt Lake, [Weber Co., Utah], Fremont 767, 1843"; Fremont s.n., 9 September 1843 (NY!, ToiTey!). The name was published without a description and is a nomen nudum. The specimen was taken on 9 September 1843 when Fremont and his boating party were on Disappointment [Fremont] Island in the Great Salt Lake. It seems clear from the notation that Torrey intended, at least initial!)', to name the species O. rigida, with the spec- imen taken later on "borders of the Great Salt Lake" as van confertifolia of that species. Reasons for change of mind are not apparent, but Torrey abandoned the epithet rigida in favor oi confertifolia. The application of the same number in this case 767, to two sheets of the same ta.xon is in keeping with the practice of Fremont, at least occa- sionalK', of using the number to indicate a species and not a collection. The sheet bears drawings of bracts, fruit, seed, and embiyo, roughly sketched by Dn Torrey. Obione spinosa Moquin-Tandon, in de Candolle, Prodr 13(2): 108. 1849. = A. canescens (Pursh) Nuttall Type locality; "In Columbia (Nutt.!). Phyllocaipa spin- osa Nutt.! in herb. Hook." (I.e.). T\pe; "Lophocan'a * Pterocarya (crossed out) * spinosa. R. Mts of the Colimibia. Pt. canescens. Atriplex canescens?," Nuttall s.n.; holotype B\l! This name has consistently been treated as a synonym of A. confertifolia. but the Nuttall specimen at BM is A. canescens. Obione tetraptera Bentham, Bot. Voyage Sulph. 48. 1844. = A. canescens (Pursh) Nuttall Type locality: California, San Diego. Type; "Ex Herbariae Musei Brittannici Voyage of H.M.S. Sulphur Capt. E W Beechey 1836-37, Capt. E. Belchen 1837-41. (Type collection of Obione tetraptera Benth.) California, San Diego. Sept-Oct. 1839. Straggling shrub 7-9 ft. Hills San Diego. George W Barcla\' 3060"; holot>'pe BM!; isot>pes GH!, K!, MO! The isotypes at GH and K each consist of a large branch witli few leaves and fruiting bracteoles still attached. The leaves are up to 4 mm wide and the bracts somewhat laciniate. Specimens approach the "laciniata" phase of A. canescens and possibly represent intergradation of A. lin- earis with A. canescens. The specimen at K bears the label information, "Oliione tetiptera. California. Barkle\'. Hooker 1844." Obione torreyi S. Watson, Rep. Geol. Explon 40th Parallel 5; 290. 1871. = A. torreyi (S. Watson) S. Watson Type locality: Nevada, Humboldt Co., diy valleys bor- dering the Trukee and Carson rivers, ToiTey 463 (Watson 1871). Type: "Herbarium of Columbia College, New York, No. 463. Obione torreyi S. Wats. Sterile saline plains, Hum- boldt Co., Nevada. Collected by J. Toney 1865"; lectotype GH! (G. D. Brown, Amen Midi. Nat. 55; 205. 1956); isolectotype NY! 1995] North American Perennial Atr/plex Types 333 The holot\'pe at GH is doubly mounted with Parry 280, 1881. It is staniinate, with glonierules ca 2 mm thick aiTanged on short lateral spikes (to ca 1.5 cm long) on lat- eral branches of a much larger paniculate cluster to 28 cm long. Branches are longitudinally striate and ridged with low, acute ridges. PhyUocarpa spinosa Nuttall ex Moquin-Tandon, in de Candolle, Prodr, 13(2); 108. 1849. pro syn. = A. cdiu'scens (Pursh) Nuttall Pterochiton canescens (Pursh) Nuttall, J. Acad. Nat. Sci. Philadelphia 1: 184. 1847. Basionym; Calligonum canescens Pursh = A. canescens (Pursh) Nuttall Pterochiton occidentale Torrey & Fieniont, in Fremont, Rep. E.xplor. E.xped. Oregon & California 318. 1845. A. occidcntalis (Ton-e>' & Fremont) Dietrich; A. canes- cens var. occidentalis (Torrey &: Fremont) Welsh & Stutz = Atriplex canescens (Pursh) Nuttall var. canescens T>pe locality; "The precise locality- of this plant we cannot indicate, as the label was illegible; l)ut it was prob- ably ft'om the borders of the Great Salt lake" (I.e.). Tvpe; "Pterochiton occidentale Toix & Frem." Fremont, probably 10 September 1843 [locality data are missing from the type specimen] (holotype NY!; microfiche BRY'!). The herbarium sheet bears a folded sheet of paper with the usual careful and detailed drawings of bracts, embiyo, and seed, and the designation "Pterochiton. " In the lower right corner of the sheet is written "Fremont, N. Gen. Pterocaly.x," and at the bottom center the words "Pterochiton occidentale, Torr. & Frem." The sheet con- tains three branches, with the bracts mainly fallen away. This sheet was designated as lectoype by G. D. Brown. Amer. Midi. Nat. 55; 209. 1956, but no other specimens are cited with the protologue and the designation should be holotype. References Brown, G. D. 1956. Taxonomy of American Atriplex. American Midland Naturalist 55; 199-210. Davis, R. J. 1952. Flora of Idaho. W. M. C. Brown Com- pany, Dubuque, lA. Dietrich, N. F. D. 1852. Synopsis Plantarum. Sec. V. \'imariae; Frieder. Voigtii. Hall, H. M., and F E. Clements. 1923. The phylogenetic method in taxonomy. Publications of the Carnegie Institute, Washington 326; 1-355. Hanson, C. A. 1962. New species of perennial Atriplex fiom the western United States. Studies in Systematic Botany, Brigham Young Universit\' 1; 1-4. Henrickson, J. 1988. Revision oi Atriplex acanthocarpa complex. Southwestern Naturalist 33; 451-463. Jepson, W. L. 1914. Chenopodiaceae. Flora of California, 1(4); 428-448. Associated Students Store, University of California, Berkele\-, J(JHNST()N, I. M. 1941. New phanerogams from Mexico. IV. Journal of the Arnold Arboretum 22; 110-124. . 1944. Plants of Coahuila, eastern Chihuahua, and adjoining Zacatecas and Durango. \. Journal of the Arnold Arboretum 25; 135-182. Jones, M. E. 1903. Chenopodiaceae. Contributions to Western Botany 11: 18-22. Kuntze, D. E. O. 1891. Revisit) Genera Plantarum 1; 1-374. Mac:bride, J. F 1918. New or othei-wise interesting plants, mostly North American Liliaceae and Chenopodia- ceae. Contril)utions from the Gray Herbarium 3: 1-22. McMinn, H. E. 19.39. An illustrated manual of California shrubs. J. W Stacey, Inc., San Franciso. McNeill, J., I. J. Bassett, C. W. Crompton, and E M. Taschereau. 1983. Ta.xonomic and nomenclatural notes on Atriplex L. (Chenopodiaceae). Tiixon 34; 549-556. Moquin-Tandon, A. 1840. Chenopodeanim Monographica Enumerato. R J. Loss, Paris. . 1849. Atriplex L. In: A. P de Candolle, Prodro- mus Svstematis Naturalis Regni Vegetabilis 13(2): 90-115. Nelson, A. 1898. New plants from Wyoming, I. Bulletin of the Torrey Botany Club 25; 202-206. . 1902. Contributions fi^om die Rocky Mountain Her- barium, IV. Botanical Gazette 34: 355-371. . 1909. Atriplex L. In: J. Coulter and A. Nelson, New manual of botany of the central Rocky Mountains. American Book Compan\', New York. Nuttall, T. 1818. The genera of North American plants 1; 1-312. D. Heartt, Philadelphia, PA. . 1847. Plantes Gambler (?). Journal of the Academy of Natural Sciences, Philadelphia 1; 184. Osterhout, G. 1898a. A new Atriplex. Bulletin of the Toney Botany Club 25; 207. . 1898b. A correction. Bulletin of the Toney Botany Club 25: 284. PuRSll, F T. 1814. Flora Americae Septentrionalis 2: 359-751. White, Cocchrane and Company, London. Rydberg, R a. 1904. Studies on the Rocky Mountain flora XI. Bulletin of die Torrey Botanical Club 31; 399-410. . 1912. Studies on the Rocky Mountain flora XXVI. Bulletin of the Torre>' Botanical Club 39: 99-113. . 1917 [1918]. Flora of the Rocky Mountains and adjacent plains. Published by the author, New York. . 1932. Flora of the prairies and plains of central North America. The New York Botanical Garden, New York. Standley, P C. 1916. Chenopodiales. North American Flora 21; 1-93. . 1917. The Chenopodiaceae of the North American flora. Bulletin of the Torre\' Botanical Club 44; 411-429. . 1922. Trees and shrubs of Mexico. Contributions from the U.S. National Herbarium 23: 1-690. Stutz, H. C. 1978. Explosive evolution of perennial A//7'/j/fx in western North America. Pages 161-168 in K. T. Harper and J. L. Reveal, editors, Intermountain biography: a symposium. Great Basin Naturalist Memoirs 2. Stutz, H. C, and S. C. Sanderson. 1979. The role of polyploidy in the evolution of Atriplex canescens. In: J. R. Godin and D. K. Northington, editors. Arid plant resources. International Center for Arid and Semi-arid Land Studies, Lubbock, TX. Torrey, J. C. 1848. Botany In: W Emorv', Notes of a mili- tary reconnaissance from Fort Leavenworth, in Missouri, to San Diego, in California. U.S. Government Printing Office, Washington, DC. . 1852. Appendix D, Botany. 7/j; H. Stansbuni', Ex- ploration and survey of the valley of the Great 334 Great Basin Naturalist [Volume 55 Salt Lake of Utali. Lippencott, (Jiainlio (S: (^o., Philack'lpliia. . 1854. Botany. In: L. Sitgreavcs, lii'port ol an ex- pedition down the Zuni and Colorado ri\er,s. U.S. Governinent Printing Office, Washington, DC. . 1857. Description of the general botanical collec- tions. In: A. VV. Whipple, E.xplorations and surveys to ascertain the practical)le and economic route for a railroad from the Mississippi River to the Pacific Ocean (Pacific Railroad Report). A. O. R Nicholson printer, Washington, DC. 1859. Botany of the honndaiy Pages 30-259 in W Emory, Report of the U.S. and Mexican boundary survey. Volume II. U.S. Department of the Interior, Washington, DC. TORREY, J. C, AND J. C. FREMONT. 1845. Descriptions of some new genera and species of plants, collected in Captain J. C. Fremont's exploring expedition to Oregon and North California, in the years lS43-'44. Pages 311-319 in J. C. Fremont, A report of the ex- ploring expedition to Oregon and North California, in the years 1843-'44. Senate Document 174, 28th Congress, 2nd Session. Blair & Rivers Printers, Washington, DC. Ulbrich, E. 1934. Chenopodiaceae. In: A. Engler and K. Prantl, Die Naturlichen Pflanzenfamilien. Ed 2. 16c; 500-519. W.vrsON, S. 1871. Botany. In: C. King, Report of geological exploration of the fortietli parallel 5: 1-525. U.S. Government Printing Office, Washington, DC. . 1874. A revision of the North American Cheno- podiaceae. Proceedings of the American Academy of Arts 9: 82-126. . 1877. Descriptions of new species of plants, with revisions of certain genera. Proceedings of the Aineri- can Academy of Arts 12: 246-278. . 1882. Contributions to American botany. XVII. List of plants from southwestern Texas and northern Mexico, collected chiefly by Dr. E. Palmer in 1879-80. — PoK'petalae. Proceedings of the American Academy ofArtsl7:316-.378. . 1891. Atiii)I('x in articles. Botanical Gazette 16: 345-346. WooTEN, E. O., AND R C. Standley. 1913. Descriptions of new plants preliminar\' to a report on the flora of New Mexico. Contributions of the U.S. National Herbarium 16: 109-196. Received 15 Fchnmnj 1995 Accepted 25 April 1995 Great Basin Naturalist 55(4), © 1995, pp. 335-341 ■ NEW RECORDS OF SCOLYTIDAE FROM WASHINGTON STATE Malcolm M. Furniss^ and James B. Johnson^ Abstract. — Eighteen species of Scolytidae are reported from Washington state for the first time or raised from obscurity: Scieriis annectens LeConte, Hijlesiniis californicus (Swaine), Phloeotribus lecontei Schedl, Carphoborus vandykei Bruck, Polygraphiis rufipennis (Kirby), Cnjpturgus borealis Swaine, Pityogenes knechteli Swaine, Ips rnexicanus (Hopkins), Ips pertiirbatus (Eichhoff), Ips plastographiis plasfographus (LeConte), Ips woodi Thatcher, Trypodendron betidae Swaine, Trypophloeus striattdus (Mannerheim), Procryphalus mucronatus (LeConte), Procryphalus iitahensis Hopkins, Pseudopityophdwrus piibipcnnis (LeConte), Pityophdionis alpinensts G. Hopping, and Pityophdwnis grandis Blackman. Host tree and collection data are given for these species. A total of 105 scolytid species known from Washington are listed. Key words: Scolytidae, fanned list, Washington state. Washington is a large state with seven physi- ographic provinces (Franklin and D\niess 1973), ranging from sea level (Fuget Trough) to over 4450 m on Mount Rainier (southern Washington Cascades). Under the influence of moisture, temperature, and substrate, natural vegetation types range from coniferous forests through woodland to shrubsteppe. Along Washington's western edge, the Coast Range and Olympic Mountains intercept the moisture-laden pre- vailing winds from the Pacific Ocean, helping to make the temperate forests of western Washington (and northern Oregon) the most dense in the world. They are composed almost exclusively of conifers and in that respect are also unique among temperate forests. Eastward lies the Cascade Range that contains Mount Rainier and other volcanic peaks. Mixed conifers prevail in these mountain ranges. Farther east is the Columbia Rasin, largest and most arid of the provinces, occupying virtually the southeast quarter of the state, except for a bulge of the Rlue Mountains extending north- ward from Oregon. Trees of this province are restricted mainly to water courses and urban areas. North of the Columbia Basin is the Okanogan Highlands province, bordering on British Columbia and Idaho, which provides a vegetational bridge to the more diverse north- ern Rocky Mountain flora. The provinces of Washington vary greatly in their climate, resulting from complex inter- play between maritime and continental air masses and the mountain ranges, particularly the Cascade Range that divides the state into east- em and western parts. For example, Quinalt on the Pacific side of the Coast Range receives 337 cm of precipitation annually, whereas Yakima, in the rain shadow to the east of the Cascade Range, has only 20 cm. Average January and July temperatures for Seattle (Puget Trough) are 4.5°C and 18.7°C, whereas those for Yakima (Columbia Basin) are -2.5 °C and 21.7°C. The Scolytidae of Washington are host spe- cific to vaiying degrees, and the extent of their diversity is related to the diversity of their woody host plants. Conifers are hosts of 87 species listed herein. A majority of these (81 species) are restricted to one or a few species of Pinaceae in the genera Abies, Larix, Picea, Piniis, Pseudotsuga, and Tsiiga, while six species infest Cupressaceae {Thuja, Chamaecyparis, and Juniperus). The remaining 19 species infest angiosperms {Popidus, Salix, Alnus, etc.). By their habits, Washington Scolytidae are charac- terized as true bark beetles, living in phloem (90 species); ambrosia beetles, living in xylem where they may feed entirely or partly on symbiotic fungi that they transmit (13 species), living in pine cones {Conophthorus ponder- osae Hopkins), or living in the roots of red clover {Hijlastiniis obscurus [Marsham]). Patterson and Hatch (1945) listed 73 species of Washington Scolytidae, adjusted to present-day synonymy. Wood (1971, 1982) lists lDi\ision ot Entomolog), Universih- of Idaho. Moscow, ID 83S44-2339. 335 336 Great Basin Naturalist [Volume 55 Washington in the distribution ol 82 speeies of Scolytidae; six adchtional speeies are hsted h\ Wood and Bright (1992). We herein update those pubhcations with 15 new state records collected by us or found in museum collec- tions, and three species collected by M. A. Deyrup (personal communication). Similar lists have been published for Idaho (Furniss and Johnson 1987), Montana (Cast et al. 1989), and Oregon (Furniss et al. 1992). Additional species of Scolytidae are likely to be collected in Washington in the future. The\' may include species known to occur in adjacent states or British Columbia, hosts of which occur in contiguous areas of Washington. Also, commerce from foreign countries enter- ing Puget Sound and the Columbia River may bring exotic species accidentally. Species that infest xylem (ambrosia beetles) are especially well adapted to such transport. The establish- ment of ambrosia beetles, which typically are not very host-specific, is enhanced by the moderate climate and great diversity of native and e.xotic flora in the Seattle area. Indeed, it is probable that such introduced scolytids may have already gained a foothold there and have not yet been detected. The following are abbreviations for reposi- tories listed for specimens new to Washington: ABS = Aichbold Biological Station, Lake Placid, FL; FS-Rl = Forest Sei-vice, USDA, Region 1, Missoula, MT; PNW = Pacific Northwest Forest and Range Experiment Station, Forest Service, USDA, Coi-vallis, OR; SLW = S. L. Wood, Brigham Young University, Provo, UT; WFBM = W F Ban- Entomological Museum, University of Idaho, Moscow, ID. Species New to Washington Subfamily Hylesininae Scierus annectens LeConte Biology. — Monogynous. Infests lower bole and roots of felled Picea spp., rarely Pinus con- torta, often by entering a galleiy of Demlroc- tonus rufipennis (Kirby). The parent galleiy is 3-4 cm long, inclined diagonally across grain. One generation per year (Stewart 1965). Distribution and notes. — Canada: Alta., B.C., N.B., Newf., Ont., Que., NWT; USA: Alas., Ariz., Calif., Colo., Ida., Me., Mont., N.H., N.M., Ore., Ut.; Washington: Tieton Ranger Station, Yakima Co., 17-VIII-1955, Picea engel- mannii, K. H. Wright (4 PNW, 1 WFBM). Hylesuuis califoniiciis (Swaine) Biology. — Monogynous. Infests the bole and limbs of Fraxinus spp. Egg galleries are transverse and deepK' engrave the wood. Over- wintering beetles evidently form feeding timnels in green bark o( Fraxinus spp. (Wood 1982). Distribution and notes. — Mexico: Chih.; USA: Ariz., Cahf., Colo., N.D., N.M., Okla., Ore. Tex., Ut.; WASHINGTON: Pack Forest, La Grande, Pierce Co., lO-V-1941, Fraxinus latifo- lia (=()reg.ona), R. L. Furniss. Two trap trees, 4" and 7" diameter, felled 4-II1-1941. Pairs of beetles and eggs present in 2.5-cm galleries lO-V-1941. Ten km N Adna, Lewis Co., 14- VII-1991, Fraxinus latifolia, M. M. Furniss and J. B. Johnson (approx. 100 WFBM, 2 SLW). Infesting underside of a 12-cm-diameter bro- ken-off branch on ground. Galleries each with a female and male parent, eggs present. Adult progeny reared, some larvae tunneled into xylem for a depth of four annual growth rings before transforming to adults. Phloeotribus lecontei Schedl Biology. — Monogamous. Male constnicts an entrance tunnel and the bases of two egg gal- leries that are then completed by the female. Egg galleries run obliquely across the grain of shaded-out branches in merchantable-size liv- ing trees. Adults and larvae may be present throughout the year; ovei"wintering adults may occur in brood galleries, special hibernation or maturation tunnels, or newly formed parental galleries (Wood 1982). Distribution and notes. — Canada: Alta., B.C.; USA: Aiiz., Calif, Colo., Ida., Mont, N.M., Ore., Ut., Wyo.; WASHINGTON: 7 km S Harts Pass, Okanogan Co., 5-VII-1988, Picea engel- mannii, M. M. Furniss (1 WFBM). Collected from a branch of a 60-cm-diameter wind- felled tree. Horseshoe Lake, Skamania Co., 17-VII-1991, Picea engelmannii, M. M. Furniss and J. B. Johnson (9 WFBM). New attacks in 1-cm-diameter shaded-out branch, without needles, attached to live tree. Swank Pass, Blewett, Chelan Co., ll-V-1975, Abies grandis, M. A. Deyrup (ABS). In a small branch. Same locality and date, Pseiidotsuga nienziesii, M. A. Deyrup (ABS). In a shaded-out branch. Carphoboriis vandykei Bruck Biology. — Polyg>'nous, unstudied. Members of the genus infest small, shaded-out branches of living trees or boles of small, suppressed. 1995] New Records of Washington Scolytidae 337 unthrifty trees. Most species live in host tissue that is drier than is typical for bark beetles (Wood 1982). Distribution and notes. — Canada: B.C.; USA: Calif., Ore.; Washington: Heritage Campground, Olympia, Thurston Co., 14-VII- 1991, Pseudotsuga tnenziesii, M. M. Fumiss and J. B. Johnson (approx. 200 WFBM). Infesting a 2.3-m-long, 6-cm-diameter, broken-off branch with red foliage. Also present was Pseudohyle- sinus nebulosus LeConte. Two to four egg gal- leries radiated from the central nuptial cham- ber, deeply etching the sapwood. Egg galleries each extended 2-5 cm, their length inversely dependent upon attack density. Eggs present, laid alternately on opposite sides (not opposite each other) in deep niches at a rate of 6 per cm and sealed with a reddish brown coating of frass. Hatched larvae fed in the phloem, not etching the wood. Some lai-val mines equaled or exceeded the length of egg galleries but most were shorter and very broad, apparently influenced by brood density. Kept at room temperature, adult brood pulverized the bark and deeply scored the xylem before emerging from veiy diy branch-wood one and one-half years later. The scored xylem had a powdeiy white appearance, perhaps due to presence of associated yeast. Carson, Skamania Co., 18-Vll- 1991, Pseudotsuga menziesii, M. M. Fumiss and J. B. Johnson (approx. 100 WFBM). Infesting 1-2 V2-cm-diameter branches of a 25-cm- diameter standing tree that had discolored foliage (dying). Galleries with parent beetles and larvae. Phloem very dry. Little Rock, Thurston Co., 30-IV-1975, Pseudotsuga men- ziesii, M. A. Deyrup (ABS). In a dead branch. Tahuya, Mason Co., 21-VI-1975, Pseudotsuga menziesii, M. A. and N. Deyrup (ABS). In a small, suppressed tree. Polygraphus rufipennis (Kirby) Biology. — Folygynous. Recorded common- ly from Picea spp., especially P. glauca and P. engelmannii, rarely from other genera of Pina- ceae. Occasionally kills small-diameter, sup- pressed trees, commonly occurs as a secondaiy species in trunks of felled or dying trees. Two to five egg galleries radiate from each nuptial chamber, most commonly two, each made by a different female. One generation per vear (Hilton 1968). Distribution and notes. — Canada: all provinces; USA: Alas., Ariz., Colo., D.C., Ida., Me., Mass., Mich., Minn., Mont., N.H., N.M., N.Y., N.C., N.D., Ore., Penn., S.D., Tenn., Ut., Ven, W.V., Wise, Wyo.; WASHINGTON: Evans Creek, King Co.; Nacotta, Pacific Co. (Hilton 1968). Kooskooskie, Walla Walla Co., 28-IX- 1955, Picea engelmamiii, W J. Buckhorn. Lake Wenatchee, Chelan Co., 22-IX-1955, Picea engelmannii, P W Orr. Metaline Falls, Pend Oreille Co., 1929-1931, Picea engelmannii and Pseudotsuga menziesii, H. J. Rust and W D. Bedard. Park-way, Pierce Co., 17-V-1934, Pinus contoi-ta, J. A. Beal. Plain, Chelan Co., 19-IX- 1955, Picea engelmannii, P W Orr. Mt. Rainier N.P, 29-X-1930, Picea engelmannii, F P Keen and W J. Buckhorn. Winthrop, Okanogan Co., 22-X-1935, Picea engehnannii, R. L. Fumiss (all PNW). Horseshoe Lake, Skamania Co., 17-VII- 1991, Picea engelmannii, M. M. Fumiss and J. B. Johnson. Infesting shaded-out branches of a 60-cm-diameter, wind-felled tree (3 WFBM). Comment. — This common beetle is certain to occur throughout the range of P. engelman- nii in the Cascade Range and Okanogan High- lands. The Pacific Co. record is likely to be in P. sitchensis; if so, it is a new host record. Subfamily Scolytinae Crypturgus borealis Swaine Biology. — Monogamous. This smallest Washington scolytid enters galleries of other bark beetles in stems of conifers {Abies, Picea, Pinus). They then tunnel irregularly into the phloem. Apparently one generation per year, ovei-wintering as adults in the brood galleries (Wood 1982). Distribution and notes. — Canada: Alta., B.C., Man., N.B., NWT, N.S., Ont., Que., Sask.; USA: Ai-iz., Colo., Ida., Me., Mich., Mo., Mont., N.M., N.Y., Ore., Penn., S.D., Ut.; Washington: Harts Pass, Okanogan Co., 5-VII-1988, Abies lasiocarpa, M. M. Furniss (6 WFBM). Infest- ing lower trunk of a 30-cm-diameter standing tree having orangish red foliage and new attacks by Pityokteines sp. Seventeen km W Mazama,' Okanogan Co., 12-VII-1991, Abies lasiocarpa, M. M. Furniss and J. B. Johnson (3 WFBM). Infesting lower trunk of a 25-cm- diameter standing tree having red foliage and abandoned galleries of another scolytid, either Pityophthonis sp. or Pityokteines sp. Pityogenes knechteli Swaine Biology. — Polygynous. The egg gallery is stellate with 4 to 6 branches radiating from the 338 Great Basin Naturalist [Volume 55 nuptial chamber. Ovenvintering stages include larvae, pupae, and adults (Alberta, Canada). One and a partial second generation occur per year at that latitude (Reid 1955). Distribution and notes. — Canada: Alta., B.C., Sask.; USA: Aiiz., Calif., Ida., Mont., Ore., Ut., Wyo.; Washington: Twisp, Okanogan Co., 12-Vin-193(), Pimis contorta, E R Keen (2 PNW). Ips mexicanus (Hopkins) Biology. — Polygynous. Not studied. Infests Piniis spp.; egg galleries cune outward from a central chamber (Wood 1982). Distribution and notes. — Canada: Alta., B.C.; Mexico: Baja Calif, Distrito Federal, Chiapas, Dgo., Hildago, Mex., Mich., Pue., Vera.; GUATEMALA; USA: Alas., Ariz., Calif, Colo., Ida., Mont., Ore., Ut., Wyo.; Washing- ton: Tieton Ranger Station, Yakima Co., 18- VI- 1956, Pinus alhicaulis (new host record), P W On- (15 PNW, 2 WFBM). Horseshoe Lake, Skamania Co., 17-VII-1991, Pinus contorfa, M. M. Furniss and J. B. Johnson (3 WFBM). Sparse galleries in 30-cm-diameter standing tree with dead top and mottled (dying) foliage. Hyhtrgops porosiis (LeConte) also sparse in base. Umatilla National Forest, 45 km S Pomeroy, Garfield Co., 19-VII-1991, Pinus contorta, M. M. Furaiss and J. B. Johnson (2 WFBM). Infesting a 23-cm-diameter standing tree with red foliage. Egg galleiy deeply etched xylem, its branches aligned more or less witli wood grain but cui-v- ing somewhat and irregular due to several turning niches. Base with moist, sour bark. Also present were Trypodendron lineatum (Olivier), Dendroctonus valens LeConte, Oi-tho- tomicus caelatus (Eichhoff), and Pityophthonis confertus Swaine. Bremerton, Kitsap Co., 21- IV-1974, Pinus contoi-ta, M. A. Deyrup (ABS). In a standing, dead tree. Ips pciiurbatus (Eichhoff) Biology. — Polygynous. Breeds abundantly in Picea glauca logging slash and in tops of trees killed by Dendroctonus beetles. Parental galleries have a tuning fork pattern with mod- erately long larval mines. One generation annually but two sets of egg galleries may be constructed by females in one season (Furniss and Carolin 1977). Distribution and notes. — Canada: Alta., B.C., Man., N.B., NWT, Ont., Que., Sask., Yukon; USA: Alas., Me., Mich., Minn., Mont.; Washington: Montesano, Grays Harbor Co., 8-IV-1973, Picea sitchensis, M. A. Deyrup (ABS). Ips plastographus plastograpJnis (LeConte) Bioloc;y. — Polygynous. Usually infests upper side of fallen Pinus contoi'ta, rarely Pinus pon- derosa. Two or three longitudinal egg galleries radiate from each nuptial chamber. Mature larvae and young adults may bore 1 cm into wood prior to emerging (Wood 1982). Distribution and notes. — Canada: B.C.; USA: Calif, Ida., Ore., Mont., Wyo.; Washing- ton: Kettle Falls, Stevens Co., IX-5-1968, Hopkins U.S. no. 54222, Pinus ponderosa, E W. Honing and J. E. Dewey (FS-Rl). Ips woodi Thatcher Biology. — Polygynous. Infests large limbs and boles of unthrifty or felled 5-needle Pinus spp. Egg galleries parallel, resembling a nar- row tuning fork (Wood 1982). Distribution and notes. — Canada: Alta.; USA: Ariz., Ida., Mont., Nev., N.M., Ut., Wyo.; Washington: Tieton Ranger Station, Yakima Co., Pinus alhicaulis (new host), 21-IX-55 to 12-VII-1956, P W. Orr (26 PNW, 3 WFBM). Trypodendron betulae Swaine Biology. — Monogynous. Tunnels are con- structed by females radially through bark into sapwood of Betula spp, rarely Alnus sp. The main tunnel branches at close intei"vals, left or right, in the same plane. Eggs are laid in nich- es oriented above and below the gallery. Larvae excavate short cradles in which they develop and feed on ambrosia fungus. Males are active in keeping the tunnels clean and aerated (Wood 1982).^ Distribution and notes. — Canada: Alta., B.C., Man., N.B., N.S., NWT, Ont, Que.; USA: Ida., Me., Mass., Minn., Mont, N.H., N.J., N.Y, S.D., Wise; Washington: Metaline Rills, Pend Oreille Co., 31-V-1930, Betula occidentalis, Hopkins no. 19839 (PNW). Trypophloeus striatulus (Mannerheim) Biology. — Monogynous. Unstudied, infests stems of Salix scouleriana, Salix spp., Alnus crispa, and A. rugosa. Distribution and notes. — Canada: Newf , N.S., Que., Yukon; USA: Alas., Colo., Ida., 1995] New Records of Washington Scolytidae 339 Minn., Ut.; WASHINGTON: King Co., 20-VI- 1976, Populus trichocarpa, M. A. Deyrup (ABS). In branch. Procryphahis miicronatus (LeConte) Biology. — Monogamous. Infests smooth, outer bark of stems of larger, dying, standing Populus tremuloides. Ovei-Avinter as lai"vae and adults; one and one -half to two generations per year (Petty 1977). Distribution and notes. — Canada: Alta., B.C.; USA: Alas., Colo., Ida., Mont., Nev., N.M., Ore., Ut.; Washington: Kamiak Butte, Whitman Co., 18-VI-1944, Populus tremuloides, M. M. Furniss and Jianlin Zhou (4 WFBM). Infesting a 30-cm-diameter recently dead tree that had no foliage. The bark was necrotic and had an almond odor. New attacks at a density of nine per dm occurred at 10-m-height, 11- cm-diameter. Galleries contained one to two parent beetles, eggs and first instar larvae. Procrijphahis utahensis Hopkins Biology. — Monogynous. Unstudied, infests stems of willows, particularly Salix scouleriana. Distribution and notes. — Canada: B.C., Que.; USA: Alas., Calif, Colo., Ida., Ore., S.D., Ut.; Washington: Bremerton, Kitsap Co., 26- VII-1975, Salix scouleriana, M. A. Deyrup (ABS). Pseudopityophthonis pubipennis (LeConte) Biology. — Monogynous. Infests bole and branches of Que reus spp. that are felled or recently dead. Galleries aligned horizontally across grain, averaging 5 cm long, closely spaced. Lai^val mines are mainly hidden in the phloem and oriented longitudinally. Distribution and notes. — Canada: Soutli- ern B.C. (Bright 1976); USA: Calif., Ore.; Washington: Carson, Skamania Co., 18-VI- 1991, Quereus garrayana, M. M. Furniss and J. B. Johnson (6 WFBM). Infesting a broken, 20-cm-diameter branch on ground. Pityophthorus alpinensis G. Hopping Biology. — Folygynous. Infests broken branches and twigs of Larix lyallii, apparently one generation annually. Distribution and notes. — Canada: Alta.; USA: Ida., Mont.; WASHINGTON: Harts Pass, Okanogan Co., ll-VII-1991, Larix lyallii, M. M. Furniss and J. B. Johnson (3 WFBM). Cadavers collected from old galleries in dead branches 0.5-2.5-cm-diameter. Galleries were branched and variable in shape, each branch containing few (9-11) egg niches; larval mines short, broad, restricted to phloem; adult brood had scored the sapwood as if by feeding. Pityophthorus grandis Blackman Biology. — Polygynous, unstudied. Infests shaded-out branches and young, standing Pin us ponderosa (Wood 1982). Distribution and notes. — Canada: B.C.; USA: Aiiz., Calif, Colo., Nebr, N.M., S.D., Tex., Ut.; Washington: Trout Lake, Klickitat Co., 17-VII-1991, Pinus ponderosa, M. M. Furniss and J. B. Johnson (4 WFBM). Infesting 4-cm- diameter standing tree with straw-color foliage. Umatilla National Forest, 53 km S Pomeroy, Garfield Co., 19-VII-1991, Pinus ponderosa, M. M. Furniss and J. B. Johnson (1 WFBM). Reared from stem of a small, felled tree. WASHINGTON SCOLYTIDAE Hylesininae Hylastini Scients annectens LeConte Scierus puhescens Swaine Hyhirgops porosiis (LeConte) Hijhirgops rcticiilatits Wood Hijliirgops rugipcnnis rugipennis (Mannerheim) Hyhtrgops suhcostulatus subcostulafiis (Mannerheim) Hylastes gracilis LeConte Hylastes longicollis Swaine Hylastes macer LeConte Hylastes nigrinus (Mannerheim) Hylastes ruber Swaine Hylesinini Hylastiniis obscuriis (Marsham) Hylesinus califoniiciis (Swaine) Alniphagiis aspericollis (LeConte) Alniphagus hirsutus Schedl Tomicini Psetidohylesinus dispar pullatiis Blackman Pseiidohylesiniis granulatus (LeConte) Pseiidohylesiniis nebulosiis nebiilosus (LeConte) Pseiidohylesinus nobilis Swaine Pseitdohylesinus pint Wood Pseiidohylesiniis sericeus (Mannerheim) Pseiidohylesiniis sitchensis Swaine Pseiidohylesiniis tsiigae Swaine Xylechiniis nwntaniis Blackman Dendroctonus brevicomis LeConte 340 Great Basin Naturalist [Volume 55 Di'udroi tonus ponderosac Hopkins Dendroctoniis pseudotsu' traits led to similar vulnerabilities. The following 10 montane bird species were categorized as most vulnerable to extirpation from the Great Basin, listed as most to least vulnerable: Olive-sided Flycatcher {Contopiis borealis). Painted Redstart (Mijiohorus pictiis), Hammond's Flycatcher (Empidonax luiiiiinondii), Veery (Cathanis fuscescens). Whip-poor-will {Capriiniil^iii.s vociferiis), Lincoln's Sparrow {Melospiza lincolnii). Black- backed Woodpecker {Picoides arcticus). Three-toed Woodpecker {P. thdactylus), Himalayan Snowcock (TetraogaUus liimalayensis), and Nashville Warbler {yennivora ritficapilla). Species of similar vulnerability scores often were dissimilar in threats related to their vulnerability. No ta.xonomic patterns in vulnerability were found. This type of analysis should be used proactively to identify vulnerable species or populations and to set priorities for research and management. Key words: vulnerability, conservation ]morities, avian diversity. Great Basin, montane islands. Extinction of species worldwide is occur- ring at a high rate (Stanley 1985). For the most part, species disappear following habitat loss (Ehrlich 1988) or after stochastic events elimi- nate relatively small or isolated populations (Mac-Arthur and Wilson 1967, Shaffer 1981, Gilpin and Soule 1986, Rabinowitz et al. 1986, Reed 1990). Because time, money, and other resources for species preservation are in short supply, it is imperative to identify the relative susceptibility to extinction, or extirpation, among species to aid in setting conservation and management priorities. Extremely vulnerable species often are easy to identify because of their scarcity, although sometimes they might be difficult to verify as extant (Solow 1993). Slightly more common species, however, often are difficult to classify by their relative susceptibility to extirpation even if it varies greatly among species (Rabino- uitz 1981, Rabinowitz et al. 1986, Reed 1992). Methods that discriminate among species' sus- ceptibility to extirpation would be valuable for setting management priorities. Such methods exist for selecting geographic areas for conser- vation based on the number or variety of species present (e.g., Kirkpatrick 1983, Margules and Usher 1984, Miller et al. 1987, Scott et al. 1991), but these methods are not applicable to priori- tizing conservation efforts among species. Economic methods can be used to priori- tize conservation efforts (Bishop 1978, Hyde 1989), but they do not accommodate non- monetary appraisals of wildlife conservation goals (Sagoff 1988). The triage method (Myers 1979), whereby species are divided into three categories based on likely success of conserva- tion efforts, might not protect the species that are biologically or anthropocentrically the most important. In the present analysis, I used biological traits to determine the relative sus- ceptibility among species to extiipation. I analyzed susceptibility to extirpation (local extinction) of bird species breeding in tlie semi-isolated montane habitats of the Great Basin. This is a classic island-biogeographic system that has been used to test ideas about extinction and colonization processes (e.g.. Brown 1971, 1978, Johnson 1975, 1978, Behle 1978, Wilcox et al. 1986, Britton et al. 1994). Although there are no endemic bird species in the Great Basin, loss of species from diese mon- tane communities reduces biodiversity and could be indicative of region-wide problems. 'Biological Hcsoiirces Research C:enter, and Deparliiient of Enviroiiniental and Resource Sciences, University of Nevada, Ren NV 89512. lOOU \alle\ Road, Reno, 342 1995] Montane Bird Vulnerability 343 Furthermore, the naturally fragmented habitat of the Great Basin montane forest can act as a model for human-caused fragmentation occur- ring throughout the world. The 74 species con- sidered here differ greatly in their life histo- ries, abilities to colonize, and susceptibility to extirpation. My goal was to rank species by bio- logical characteristics related to their vulner- ability to extirpation, in the anticipation that the information would be useful for setting priorities for research, conservation, and man- agement. Assessing susceptibility to extirpation in- volves some type of decision analysis (sensu Maguire et al. 1987). There are many methods available for assessing susceptibility to extiipa- tion, and they vaiy in complexity from simple classifications to complex multivariate analyses (Table 1). More importantly, classification meth- ods differ in their data requirements. Some sys- tems, such as the lUCN classification scheme (Mace and Lande 1991), are data intensive, while others require far less data (Table 1). The more data available for decision making, the more certain the results, but it is impor- tant to chose a method that makes proper use of the available data. Biological data are rela- tively scarce for birds in the Great Basin. In this analysis, I used a method with intermedi- ate data needs to look at vulnerability to extir- pation of 74 montane breeding bird species. Methods I combined the methods of Burke and Humphrey (1987), Millsap et al. (1990), and Rabinowitz et al. (1986) to develop an analysis appropriate for the species and available data. This analysis involved assessment using seven biological characteristics related to persis- tence ability. Values for each characteristic ranged from 0 to 1, with higher values associ- ated with higher susceptibility to extii-pation. Values for each character were summed to arrive at a final score of susceptibility to extir- pation from the Great Basin. All variables had the same range so that no single character contributed disproportionately to the suscepti- bility score (Given and Norton 1993). Himalayan Snowcock and Ruffed Grouse (sci- entific names are given later) are introduced species in the Great Basin (Alcorn 1988). They were included in the analysis because they are established in the Great Basin avifauna. Variable descriptions used in scoring vulnera- bility to loss from the Great Basin follow. Geographic range. — Species distributions were taken from a subset of 20 montane sites from the Great Basin (Johnson 1975). The con- tribution of this variable to the vulnerability score was calculated as 20 minus the number of ranges on which the species occurs, divided by 20. This results in a value ranging from 0 to 1.0, with higher values associated with fewer ranges occupied by the target species, i.e., greater vulnerability. Mountain ranges here and in Table 2 are numbered the same as in Johnson (1975): 1-Warner, 2-Pine Forest, 3-Santa Rosa, 4-Jarbidge, 5-Raft River, 6-Desatoya, 7-Toiyabe- Shoshone, 8-Ruby, 9-Spruce-S. Pequop, 10-Deep Cr.-Kern, 11-Snake, 12-White-Inyo, 13-Plametto, 14-Grapevlne, 15-Panamint, 16-Spring, 17-Sheep, 18-Mt. Irish, 19-Quinn Canyon- Grant, and 20-Highland. Distributional data were supplemented from Behle (1978), Herron et al. (1985), Ryser (1985), Alcorn Table L Methods for assessing susceptibility to extirpation and for scoring conservation priorities. Data AnaKsis Method intensit\' complexib,' Citations Anthropocentric low veiy low the history of the world Decision analysis: contingency low low Rabinowitz 1981, Rabinowitz et al. 1986, Kattan 1992, Reed 1992 ordinal variable low Burke and Humphre> 1987, Millsap etal. 1990,' this study classical variable medium Maguire et al. 1987 multivariate variable high Given and Norton 1993 Economic variable variable Bishop 1978, Hyde 1989 Viability analysis high high Kinnaird and OBrien 1991, Boyce 1992 lUCN \ en- high high Mace and Lande 1991 344 Gkeat Basin NAXURyVLisT [Volume 55 Table 2. Additions to Johnson's (1975) original l)ii(l dis- tributions. Site numbers are the same as those used b\ Johnson (1975) and are listed in Methods. Scientifie names are listed in Talile 3. Species Sites added American Wigeon 39 g; Dunning 1993) were given a value of 0.5. Species that are small and not known to reject eggs were assigned a score of 1. Data came from Fried- man (1971), Rothstein (1975), Airola (1986), Marvil and Cruz (1989), and Briskie et al. (1992). Migr\tory STATUS. — There is some contro- versy regarding relative costs of migration ver- sus residency in birds. However, because migrants are dependent on habitats in more than one geographic area, I consider them more vulnerable than nonmigrants. I scored migra- tory status as no latitudinal migration = 0 (lowest risk), migrates primarily to U.S. = .25, migrates primariK' to Middle or South America, winters in nonforest = .50, winters in sec- ondan' forest = .75, winters in mature forest = 1.0.' Reproductive potential. — I considered reproductive potential to be the anticipated ability to recover from a population crash and based it on the first age of reproduction, clutch size, and number of broods within a year (data from Ehrlich et. al. 1988). I classified repro- ductive potential based on an index. The index was the mean clutch size times the number of 1995] Montane Bird Vulnerability 345 broods in a year, divided by the age of first reproduction. With this index, a species that breeds repeatedly, at an early age, and with large clutches will have a low score. When no data were available for number of broods, one brood was assumed. Age at first breeding was assumed to be one for small birds, unless data from the literature indicated otherwise. The relationships between the index, reproductive potential, and risk value were made arbitrarily and are presented in Table 3. Data and refer- ences associated with this calculation for each species can be obtained from the author. Diet specialization. — Information on diet breadth came from Ehrlich et al. (1988), and species were classified as generalists (score = 0), moderate specialists (0.5), or specialists (1.0) based on diet described there. This assessment was subjective, based on number of food types typically in the diet and foraging method used. With this system, vulnerability scores could range from 0 to 7, with 7 being the greatest probability of extirpation from the Great Basin. One variable not included in the analy- sis that is important in biological risk to extir- pation was local population trends. Local pop- ulation trends were omitted because they are generally unknown for nongame birds in the Great Basin. Local endemism should be con- sidered in scoring as well, but the Great Basin has no endemic bird species. Another variable that has been suggested as a risk to sunaval is ground nesting. Traditional thought places ground nesters at higher risk to predation than off-ground nesters (e.g., Ricklefs 1969, Slagsvold 1982, Collias and Collias 1984). However, in a reanalysis of the data, Martin (1993) found that ground nesters were not dis- proportionately susceptible to depredation. Given this important ambiguity, nest location was omitted from the analysis. Results and Discussion There were 41 additions of various mountain ranges to breeding bird distributions (Tiible 2). The 74 breeding bird species used in this analysis, their associated scores for each life- histoiy trait, and their vulnerability scores are listed in Table 4. Taxonomy follows the con- vention of the American Ornithologists' Union (1983). Vulnerability scores ranged from 0.60 for the American Robin (scientific names are Table 3. Reproductive potential and its relationship to risk score. The index is mean clutch size times the numher of broods in a year, divided by the age of first reproduction. Index Reproductive Risk value potential score 11.9 high 0 found in Table 4) to 5.70 for the Olive-sided Flycatcher and Painted Redstart. None of the variables alone was sufficient to assess vulner- ability to extirpation. This has been seen by others (e.g., Burke and Humphrey 1987) and is due to other life-history factors affecting susceptibility to extir^Dation (Arita et al. 1990). Therefore, range and density estimates alone cannot be used to assess vulnerability to extir- pation. Another problem with using range and density as the only criteria for extiipation risk is that slice-in-time assessments of rarity can give misleading results due to natural fluctua- tions in distribution and population size (Hanski 1985). Species ranges expand and contract, and population densities can undergo large fluctu- ations annually, even in long-lived species such as birds. Therefore, being uncommon does not, de facto, make a species vailnerable to extir- pation; in contrast, being common does not assure continued presence (e.g., the Passenger Pigeon [Ectopistes migratorhis]; Bucher 1992). Passerines tended to rank as more suscepti- ble to extirpation than other orders, primarily because one threat, vulnerability to cowbird parasitism, did not impact non-passerines. Unlike some earlier studies of birds (Terborgh and Winter 1980, Kattan 1992), I found no tax- onomic pattern in susceptibility to extiq^ation. The 10 species with the highest vulnerability score come from seven families in four orders. There are several likely explanations for this. The first is that no inherent patterns exist. Alternatively, a true taxonomic pattern in extirpation proneness might exist for Great Basin birds but was missed because of incom- plete data, because of a subsampling effect (not enough of tlie Great Basin surveyed), or be- cause tlie anah'sis considers only cuiTcnt species (implying that extirpation-prone species are gone). Many species with similar or identical vul- nerabilit\' scores were vulnerable for different 346 Great Basin Naturalist [Volume 55 Table 4. Data used in analyses and \ulncial)ilit\' scorinj^s; variable definitions given in text. Higher values indieate higher susceptibility to extiipation from the (Ireat Basin. Vulner- (Criteria Some- Habitat Diet Species ability where special- Cowbird ,\ligratoi"y Reproductive special- score Range large? ization? problem? status potential ization Canada Goose (Branta canadensis) L9() .90 0 0 0 .25 .75 0 Green-winged Teal (Anas crecca) 2.90 .90 1 .5 0 .25 • .25 0 American Wigeon (A. americana) 2.90 .90 1 .5 0 .25 .25 0 Canvasback (Aythya valisineria) 2.90 .90 1 .5 0 .25 .25 0 Sharp-shinned Hawk (Accipiter striatus) 4.00 ..50 1 .5 0 .75 .75 .5 Northern Goshawk (A. gentilis) .3.10 .60 1 .5 0 .25 .75 0 Hiniala\an Snowcock (TetraogaUits hiinahiycnsis) 4.20 .95 1 0 0 .75 .5 Blue Grouse {Dendragapus ohscunis) 2.75 ..50 0 0 0 .25 1 Ruffed Grouse (Bonusa umbellus) 3.20 .95 1 0 0 .25 0 Mountain Quail (Oreortyx pictus) L90 .65 0 0 0 .25 0 Common Snipe (GalUnago gallinago) 2.70 .70 1 0 0 .50 .50 0 Flammulated Owl {Otus flammeolus) 3.05 .55 0 1 0 .50 .50 .5 Northern Pygmy-owl [Glaucidium gnoma) 3.30 .80 1 1 0 0 .50 0 Short-eared Owl [Asia flammeus) 2.55 .80 1 0 0 .25 .50 0 Northern Saw-whet Owl {Aegolius acadicus) 2.85 .60 0 1 0 .25 .50 .5 Common Nighthawk {Chordeiles minor) 2.10 .35 0 0 0 .50 .75 .5 Whip-poor-will (Caprimidgus vociferus) 4.70 .95 1 0 0 1 .75 1 Calliope Hummingbird {Stellula calliope) 3.15 .65 0 .5 0 .75 .75 .5 Broad-tailed Hummingbird (Salaspliorus platycercus) 2..30 .05 0 .5 0 ..50 .75 .5 Lewis' Woodpecker {Melanerfjes lewis) 1.90 .90 0 .5 0 .25 .25 0 Yellow-bellied Sapsucker {Sphyrapicus varius) 2.55 .30 0 .5 0 .75 .50 .5 Red-breasted Sapsucker (S. ruber) 3.15 .85 0 .5 0 .75 .50 .5 Williamson's Sapsucker (S. thyroideus) 3.35 .55 0 1 0 .75 .50 .5 Downy Woodpecker (Picoides pubescens) 2.10 .60 0 .5 0 0 .50 .5 Hairy Woodpecker [E villosus) 2.00 0 0 .5 0 0 .50 1 White-headed Woodpecker (P. albolarvatus) 3.45 .95 1 0 0 .50 0 Black-backed Woodpecker {P. arcticus) 4.45 .95 1 0 0 .50 1 Three-toed Woodpecker {P. tridactylus) 4.45 .95 1 0 0 .50 1 Olive-sided Flycatcher (Contopus borealis) 5.70 .45 1 1" .75 .50 1 Hammond's Flycatcher {Empidonax haintnondii) 5.45 .70 1 la .75 .50 .5 Dusky Flycatcher (£. oberholseri) 3.30 .05 0 .5 .75 .50 .5 Western Flycatcher (£. difficilis) 3.95 .45 0 .5 1 .50 .5 Horned Lark [Eremophih alpestris) 2.60 .85 0 0 .25 .50 0 1995] Montane Bird Vulnerability 347 Table 4. Continued. Vulner- Criteria Some- Habitat Diet Species ability where special- Cowbird Migratoiy Reproductive special- score Range large? ization? problem? status potential ization \'iolet-gieen Sw;illo\\ (Taclujcineta thalassina) 3.00 0 0 1 0 ..50 .50 1 Gray Jay (Perisoreus canadensis) 2.95 .95 1 .5 0 0 .50 0 Steller's Jay (Cyanocitta stelleri) 2.10 .60 0 1 0 0 ..50 0 Clark's Nutcracker {Nucifraga cohnnbiana) 1.65 .15 0 1 0 0 ..50 0 Mountain Chickadee {Panis gambeli) 1.50 0 0 1 0 0 0 .5 Red-breasted Nuthatch (Sitta canadensis) 2.15 .40 0 .5 0 .25 .50 .5 White-breasted Nuthatch (S. carolinensis) 2.10 .10 1 .5 0 0 0 .5 Pygmy Nuthatch (S. pygmaea) 1.95 .70 0 .5 0 0 .25 .5 Bro\\Ti Creeper (Ceiihia americana) 1.65 .40 0 .5 0 .25 .50 0 American Dipper (Cinclus mexicamis) 3.25 .50 1 1 0 0 .25 .5 Golden-crowned Kinglet {Regidus satrapa) 2.40 .65 0 .5 1 .25 0 0 Ruby-crowned Kinglet {R. calendula) 2.65 .15 0 .5 1^ .75 .25 0 Western Bluebird (Sialia niexicana) 2.55 .80 1 0 0 .25 .50 0 Mountain Bluebird (S. curntcoides) 1.60 ,10 0 .5 0 .25 .25 .5 Townsend's Solitaire (Myadestes townsendi) 3.00 .25 0 1 .25 .50 0 V'eeiy (Catharus fuscescens) 4.90 .90 1 .5 la .50 .50 .5 Swainson's Thrush (C. ustidatus) 3.60 .60 0 .5 1» .50 .50 .5 Hemiit Thrush (C. guttatus) 2.55 .05 0 .5 1^ .75 .25 0 American Robin {Turdtis migratorius) 0.60 .10 0 0 0 .25 .25 0 Water Pipit (Anthus spinoletta) 3.65 .90 0 1 1" .25 .50 0 Solitary' Vireo (Vireo solitarius) 3.55 .30 0 .5 1 .25 .5 Orange-crowned Warbler {Vennivora celata) 2.60 .35 0 0 .75 .50 0 Nashville Warbler (V nificapilla) 4.15 .90 0 .5 1" .75 .50 .5 N'irginia's Warbler {V. virginiae) 3.25 .25 0 .5 1" .75 .25 .5 Vellow-rumped Warbler {Dendroica coronata) 2..30 .05 0 .5 .50 .25 0 Grace's Warbler (D. graciae) 4.05 .80 0 1 1» .50 .25 .5 MacGillivray's Warbler {Oporomis tohniei) 3.35 .35 0 .5 1" .50 .50 .5 Wilson's Warbler (Wihunia pusilla) 3.85 .85 0 .5 .50 .50 .5 Painted Redstart (Myioboms pictus) 5.70 .95 1 1 1» .75 .50 .5 Western Tanager (Piranga hidoviciana) 3.15 .15 0 .5 P 1 .50 0 Green-tailed Towhee (Pipilo chlorunis) 1.75 0 0 0 1" .50 .25 0 Fox Sparrow (Passerella iliaca) 1.95 .45 0 0 .25 .25 0 Lincoln's Sparrow {Melospiza lincolnii) 4.60 .85 1 1 1" .50 .25 0 348 Great Basin Naturalist [Volume 55 Table 4. ContiiuR'd. Vulner- Criteria Some- Habitat Diet Species ability where special- ('owbird .\1 iterator) Repr oductive special- score Range largc-'^ ization? lirobleni:' status po teulial ization White-crowned Sparrow {Zonotrichia Icucoph njs ) 2.00 ..50 0 1 .25 .25 0 Dark-eyed Jiiiico (Junco hyemalis) 2.05 .05 .5 1 .25 .25 0 Gray-crowned Rosy Finch {Leucosticte tephrocotis) .3.70 .95 1 1'' .25 .50 0 Black Rosy Finch (L. atrata) .3.50 .75 1 1^ .25 .50 0 Cassin's Finch (Caq)oilacus (•(i.s.sinii} 2..50 1) 1 1" .25 .25 0 Red Crossbill (Loxiii cunimstra } 3.15 .40 .5 ^ .25 .50 .5 Pine Siskin {Carduelis pinus) 2.40 .40 0 .5 P .25 .25 0 Evening Grosbeak {Coccothraustes vespcrtUais] 2.35 .85 0 ..5 ..5-' .25 .25 0 ''Assumed to not eject Browii-lieaded Coubird eggs suites of threats to persistence. That is, some equal scores were made up of low values for one or more characteristic and corresponding- ly high values for other traits, which balanced in the ranking. This observation is consistent with Rabinowitz's (1981, Rabinowitz et al. 1986) observations of plant species' rarity in Great Britain. It should be noted tliat this analy- sis refers to species loss in the Great Basin and does not reflect species-wide vulnerability. This type of analysis is sensitive to the num- lier of variables included. Adding or deleting characters from the analysis would change scores. For example, if ground nesting were decisively shown to increase vulnerability, it could be added to tlie analysis and would change relative scores. Results also would be altered if the characteristics were weighted differently. I did not weight any variable as more impor- tant than another because of the lack of data that demonstrates the validity of weighting particular traits over others. Arbitrarily assign- ing different weights in the absence of inde- pendent data supporting the weighting would result in unwarranted bias in the vulnerability scores. The results presented are not absolute rank- ings for susceptibility to extirpation because data are incomplete and more threats might become apparent, which would have to be added to the analysis. Validity of these results depends entirely on reliability of the data used and how representative the 20 mountain ranges are of the rest of the Great Basin. There is a dearth of distributional and life-historv data on many Great Basin birds. Therefore, my results should be taken as a guide for detailed local studies of species and their surrounding communities. Results of these studies can then be used to develop proactive manage- ment plans. Vulnerability Ranks and Management Vulnerability to extirpation and manage- ment priorities are not equal. Scores based strictly on biological variables ignore homo- centric values, such as hunting or local tradi- tional uses. For example, the top 10 vulnerable species in this analysis include only one hunt- ed species (an introduced one at that), though others were scored. In addition, how a given rank comes about can affect management pri- orities. There are four ways a species can have a high score, and they should be interpreted differently for management. (a) High score occurs when the Great Basin is within the greater bounds of a species' dis- tribution and local declines have reduced a species range and population sizes in the Great Basin. These species are probably declining because of local problems, and in this analysis might include Mountain Quail and Northern Goshawk. Specific management plans should be enacted to increase population numbers, sizes, and distributions. (b) High score occurs when the Great Basin is within the greater bounds of a species distri- bution, and the species is declining through- out its range. Problems could be occurring on the breeding grounds, wintering grounds, or 1995] Montane Bird Vulnerability 349 migratory routes. If the cause of decline is known and can be improved through local management, then this should be done. If the cause of the decline is known, but occurs out- side the Great Basin, then I would recom- mend monitoring populations but not making any management efforts. If the cause of the decline is not known, as for many Neotropical migrants, gather information to determine whether or not local management could improve local or region-wide population con- ditions. If management efforts are suspected to work, implement them with proper controls and follow-up work. If no effect is found, dis- continue management. (c) High score occurs partly because the Great Basin is at the edge of a species distri- bution, thus limiting its local distribution and population sizes. Of the top 10 scored species in this analysis, five have Nevada as part of their distributional boundaiy This is possibly tlie trickiest categoiy for management. Species' ranges fluctuate, and population declines might be range retractions having nothing to do with local conditions. These species should be monitored because range retraction might be an early indicator of a species-wide decline (e.g., Laymon and Halterman 1987). However, it can also indicate local problems that require local management solutions. These species need further investigation. (d) High score occurs when species has de- clined severely (thus reducing its range and commonness) but is recovering. Continue exist- ing management efforts, if any, and monitor populations to make sure recoveiy continues. If it does not, these species belong in one of the other three sub-categories. In all instances involving management plans, efforts should be made to set up proper stud- ies or experiments to ascertain the limiting factor(s) and the coiTcct method(s) for counter- acting the problem (MacNab 1983, Gavin 1989, 1991, Muiphy and Noon 1992). This includes monitoring suitable control sites. Without using adequate experimental design, it will not be possible to ascertain the effectiveness of management efforts. Low-score species should still be monitored and management plans developed. Low-score species are those that are closest to recovery or those not threat- ened and thus have potential for the quickest success from management. Acknowledgments I thank J. A. R. Alberico, E E Bmssard, D. A. Delehanty, C. Elphick, N. Johnson, B. Maurer, and one anonymous reviewer for commenting on this manuscript, and K. Reed and S. Dunham for help summarizing the data. I also thank G. Henon, R. Hamlin, M. Elpers, T. Baron, and P. Zenone for discussions regarding threat variables. This work was supported by NSE grant DEB-9322733, the Biological Resources Research Center at the University of Nevada, the U.S. Forest Sei-vice, the Center for Conser- vation Biology at Stanford University, and a donation from the Wells Family Foundation. This is Contribution No. 004 of the Nevada Biodiversity Initiative. Literature Cited AlROLA, D. A. 1986. Brown-headed Covvbird parasitism and habitat disturbance in tlie Siena Nevada. Journal of VVildhfe Management 50: 571-575. Alcorn, J. R. 1988. The birds of Nevada. Fairview West Pubhshing, Fallon, NV 418 pp. American Ornithologists' Union. 1983. Check-list of North American birds. 6th edition. Allen Press, Lawrence, KS. 877 pp. Arita, H. T, J. G. Robinson, and K. H. Redford. 1990. Rarity in Neotropical forest mammals and its ecolog- ical correlates. Consei^vation Biology 4: 181-192. Behle, VV. H. 1978. Avian biogeography of the Great Basin and Intermountain Region. Great Basin Naturalist Memoirs 2: .55-80. Bishop, R. C. 1978. Endangered species and uncertainty: the economics of the safe minimum standard. American Journal of Agricultural Economics 60: 10-18. Bovce, M. S. 1992. Population viability analysis. Annual Review of Ecology and Systematics 23: 481-506. Briskie, J. v., S. G. Sealy, and K. A. Hobson. 1992. Behavioral defenses against avian brood parasitism in sympatric and allopatric host populations. E\'olu- tion 46: 334-340. BRirriNGHA.vi, M. C., and S. A. Temple. 1983. Have cow- birds caused forest songbirds to decline? BioScience 33: 31-35. Britton, H. B., P E Brussard, D. D. Murph\; .and G. T. Austin. 1994. Colony isolation and isozyme variabil- ity of the western seep fritillar>', Speijeria nokomis apacheana (Nymphalidae) in the western Great Basin. Great Basin Naturalist .54: 97-105. Brown, J. H. 1971. Mammals on mountaintops: nonequi- librium insular biogeography. American Naturalist 105: 467-478. . 1978. The theory of insular biogeography and the distribution of boreal birds and mammals. Great Basin Naturalist Memoirs 2: 209-227. Brown, J. H., and B. A. Maurer. 1987. Evolution of species assemblages: effects of energetic constraints and species dynamics on the diversification of the 350 Great Basin Naturalist [Volume 55 North American avilaima. Anicrican Naturalist 130: 1-17. BUCHER, E. H. 1992. The causes oT e.xtinctiou of the Passenger Pigeon. Current Ornitholog)' 9: 1-36. Burke, R. L., and S. R. Humphrey. 19(S7. Rarity as a cri- terion for endangennenf in Florida s fauna. ()r\x 21: 97-102. CoLMAS, N. E., AND E. C. CoLLUS. 1984. Nest building and bird behavior Princeton Universit\' Press, Prince- ton, NJ. Dunning, J. B., Jr. 1993. CRC: handbook of" avian body masses. CRC Press, Boca Raton, FL. Ehrlich, P R. 1988. The loss of diversit)': causes and con- sequences. Pages 21-27 in E. O. Wilson, editor. Biodiversity. National Academy Press, Washington, DC. ' Ehrlich, P R., D. S. Dobkin, and D. Wheye. 1988. The birder s handbook: a field guide to the natural histo- ly of North American birds. Simon & Schuster, New York. 785 pp. Fleischer, R. C., and S. I. Rothstein. 1988. Known sec- ondary contact and rapid gene flow among sub- species and dialects in the Brown-headed Cowbird. Evolution 42: 1146-1158. Friedman, H. 1971. Fnrther information on the host rela- tions of the parasitic cowbirds. Auk 88: 239-255. Gavin, T. A. 1989. What's wrong with the questions we ask in wildlife research? Wildlife Society Bulletin 17: 345-350. . 1991. Why ask "Why": the importance of evolu- tionaiy biology in wildlife science. Journal of Wild- life Management 55: 760-766. Gilpin, M. E., and M. E. Soule. 1986. Minimum viable populations: processes of species extinction. Pages 19-34 in M. E. Soule, editor, Conservation biology: the science of scarcib,' and diversity. Sinauer, Sunder- land, MA. Given, D. R., and D. A. Norton. 1993. A multivariate approach to assessing threat and for priority setting in threatened species consei-vation. Biological Conser- vation 64: 57-66. Hanski, I. 1985. Single-species spatial dynamics may con- tribute to long-term rarity and commonness. Ecology 66: 335-343. Herron, G. B., C. A. Mortimore, and M. S. Ravvlings. 1985. Nevada raptors: their biology and manage- ment. Biological Bulletin No. 8. Nevada Department Wildlife, Reno. 14 pp. Hyde, W F 1989. Marginal costs of managing endangered species: the case of the Red-cockaded Woodpecker Journal of Agricultural and Economic Research 41: 12-19. Johnson, N. K. 1975. Controls of the number of bird species on montane islands in the Great Basin. Evolu- tion 29; 545-567. . 1978. Patterns of avian biogeography and specia- tion in the Intermountain Region. Great Basin Naturalist Memoirs 2: 137-160. Kattan, G. H. 1992. Rarit>' and \'ulnerability: the birds of the Codillera Central of Colombia. Conservation Biology 6: 64-70. Kinnaird, M. E, and T. G. O'Brien. 1991. Viable popula- tions for an endangered forest primate, the Tana River crested mangabey (Cercocehus galeritiis fooler if us). Consei-vation Biology 5: 203-213. Kirkpatrick, J. B. 1983. An iterative method for establish- ing priorities for the selection of nature reserves: an example from Tasmania. Biological Consenation 25: 127-134. L.\Y.MON, S. A., and M. D. Halterman. 1987. Can the western subspecies of the Yellow-billed Cuckoo be saved from extinction'? Western Birds 18: 19-25. Mac;Arthur, R. H., and E. O. Wilson. 1967. The theorv' of island biogeography. Monographs in Population Biology No. 1. Princeton Universit)' Press, Princeton, NJ. 203 pp. M.^CE, G. M., AND R. Lande. 1991. As.sessing extinction threats: toward a reevaluation of lUCN threatened species categories. Consei-vation Biology 5: 148-157. MacNab, J. 1983. Wildlife management as scientific ex- perimentation. Wildlife Society Bulletin 11: .397-401. Ma(;uire, L. a., U. S. Seal, and R F Brus,sard. 1987. Man- aging critically endangered species: the Sumatran rhino as a case study. Pages 141-158 in M. E. Soule, editor. Viable populations for conservation. Cam- bridge Universitv' Press, United Kingdom. Margules, C. R., and M. B. Usher. 1984. Conservation evaluation in practice. I. Sites of different habitats in north-east Yorkshire, Great Britain. Journal of Envi- ronmental Management 18: 153-168. Martin, T E. 1993. Nest predation among \egetation lay- ers and habitat types: revising the dogmas. American Naturalist 141: 897-913. Marvil, R. E., and a. Cruz. 1989. Impact of Brown- headed Cowbird parasitism on the reproductive suc- cess of the Solitaiy Vireo. Auk 106: 476-480. Mayfield, H. F 1977. Brown-headed Cowbird: agent of extermination? American Birds 31: 107-113. Miller, R. I., S. P Gratton, and P S. White. 1987. A regional strategy for reserve design and placement based on an analysis of rare and endangered species' distribution patterns. Biological Conservation 39: 255-268. MiLLSAP, B. A., J. A. Gore, D. R. Runde, and S. I. Cerulean. 1990. Setting priorities for the consei-va- tion of fish and wildlife species in Florida. Wildlife Monograph No. 111. Murphy, D. D., and B. R. Noon. 1992. Integrating scien- tific methods with habitat conservation planning: reserve design for Northern Spotted Owls. Ecolog- ical Applications 2: 3-17. Myers, N. 1979. The sinking ark. Pergamon Press, Oxford. Rabinowitz, D. 1981. Seven forms of rarity. Pages 205-217 in H. Synge, editor, The biological aspects of rare plant consei-vation. Wiley, Chichester Rabinowitz, D., S. Cairns, and T. Dillon. 1986. Seven forms of rarity and their frequency in the flora of the British Isles. Pages 182-204 in M. E. Soule, editor. Conservation biology: the science of scarcity and diversity. Sinauer Associates, Sunderland, MA. Reed, J. M. 1990. The dynamics of Red-cockaded Woodpecker rarity and conservation. Pages 37-56 in A. Carlsson and G. Aulen, editors, Conservation and management of woodpecker populations. Swedish University of Agricultural Science Report 7, Uppsala. . 1992. A system for ranking conservation priorities for Neotropical migrant lairds based on relative sus- ceptibility to extinction. Pages 524-536 in J. M. Hagan and D. W. Johnston, editors, Ecology and conservation of Neotropical migrant landbirds. Smithsonian Institution Press, Washington, DC. RicKLEFS, R. E. 1969. An analysis of nesting mortality in birds. Smithsonian Contributions to Zoolog)' 9: 1—48. 1995] Montane Bird Vulnerability 351 RoTHSTEIN, S. I. 1975. An experimental and teleonomic investigation of avian hrood parasitism. Condor 77: 250-27 L RoTHSTEiN, S. I., J. Verner, .■\.\d E. Stevens. 1984. Radio- tracking confirms a unique diiunal pattern of spatial occurrence in the parasitic Brown-headed Cowhird. EcologN' 65: 77-88. Ryser, E a., Jr. 1985. Birds of the Great Basin; a natural histor\'. University of Nevada Press, Reno. S.\GOFF, M. 1988. Some problems with environmental economics. Environmental Ethics 10: 55-74. Scott, J. M., B. Csuti, and S. Caicco. 1991. GAP analy- sis: assessing protection needs. Pages 15-26 in W. Hudson, editor. Landscape linkages and biological diversit\': a strategy' for sundval. Island Press, Covello, CA. Shaffer, M. L. 1981. Minimum viable populations for species consenation. BioScience 31: 131-134. Slagsvold, T. 1982. Clutch size variation in passerine birds: the nest predation hvpothesis. Oecologia 54: 159-169. SoLOW, A. R. 1993. Inferring extinction from sighting data. Ecology 74: 962-964. Stanley, S. M. 1985. Extinction as part of the natural evo- lutionary process: a paleobiological perspective. Pages 31-46 in R. J. Hoage, editor. Animal extinc- tions: what everyone should know. Smithsonian Institution Press, Washington, DC. Terborgh, J., AND B. Winter. 1980. Some causes of extinction. Pages 119-133 in M. E. Soule and B. A. Wilcox, editors. Conservation biolog}': an evolutionary- ecological perspective. Sinauer, Sunderland, MA. Wilcox, B. A., D. D. Murphy, P R. Ehrlich, and G. T. Austin. 1986. Insular biogeography of the montane butterfly faunas in the Great Basin: comparison with birds and manmials. Oecologia 69: 188-194. Received 27 October 1994 Accepted 19 May 1995 Great Basin Naturalist 55(4), © 1995. pp. 352-358 GRASSHOPPER DENSITIES ON GRAZED AND UNGRAZED RANGELAND UNDER DROUGHT CONDITIONS IN SOUTHERN IDAHO Denni.s |. Melcliii^' and Merlyii A. Brusven'- AUSTKACT. — Low-dt'nsit\ grasshopper populations were sampled at 15 pairs of rangeland sites in south central Idaho. One site of each pair had not been grazed by livestock for at least 10 \ ears. Grazed sites were managed under normal grazing regimes established by the Bureau of Land Management. Mean grasshopper density was higher on ungrazed sites than on grazed sites. Proportions u{ ML'lano})his san^uinipes were higher on ungrazed sites than on grazed sites and were higher on annual grasslands than on otiier \egetation types. Effects of grazing appeared to be independent of vegetation type. Proportions of Gomphocerinae, a subfamily of grasshoppers that feeds almost exclusively on grasses, were affected by \'egetation t\pe, but not grazing. Crested wheatgrass seedings supported the highest proportions of Gomphocerinae. Proj^ortions of Oedipodinae were affected by grazing and vegetation type. Higher proportions of Oedipodinae were found on grazed sites than on ungrazed sites, and on sagebrush/grass sites than on annual grasslands. Results indicate that li\ estock grazing during drought conditions tends to reduce grasshopper populations on southern Idaho rangeland. Keij words: Oiihoptcm, Acrididac. Melanoplus sanguinipes, liccstock grazing, drought, population density, range management. Grasshoppers are frequently the most abun- dant arthropods, in terms of biomass, in the intermountain sagebrush ecoregion of the western United States. As primaiy consumers they may be important in energy and nutrient cychng, and, at outbreak densities, they compete with Hvestock and wildhfe for forage. Because of their ecologic and economic importance, the potential effects of range management practices on grasshoppers are a concern to those interested in the health of rangeland ecosystems. Several studies have addressed the role of livestock grazing on grasshopper populations (Coyner 1938, Smith 1940, Campbell et al. 1974, Holmes el al. 1979, Capinera and Sechrist 1982, Jepson-Innes and Bock 1989, Quinn and Walgenbach 1990, Miller and Onsager 1991). Onsager (1987) sug- gested that there is probably geographic varia- tion in grasshopper responses to grazing among rangeland types and their constituent grasshopper communities. To date no studies have investigated the relationship between livestock grazing and grasshopper densities on rangelands in the intermountain region. This study, conducted during years of below- normal precipitation and low grasshopper densities, examined differences in grasshopper densities between rangeland under normal livestock grazing regimes administered by the Bureau of Land Management and rangeland that had not been grazed for at least 10 years. Study Area The study area is located southeast of Sho- shone, ID, within the Bureau of Land Manage- ment's (BLM) Shoshone District, between longitude 114°30' and 114°00' W and latitude 42°37.5' and 43°00' N. This area receives an average of about 26 cm of precipitation annu- ally, most of it between October and May. Average annual temperature is about 9.0 °C. The intermountain sagebrush ecoregion was subjected to heavy grazing pressure in the late 19th and early 20th centuries, frequent fires, and subsequent invasion by cheatgrass and other e.xotic plant species (Pickford 1932, Stewart and Hull 1949, Mack 1981, Yensen 1982). As a result, stands of grazing-intolerant native grasses were greatly diminished over much of the region, and cheatgrass has become the dominant species on more than 40 million ha of the Intermountain West (Mack 1981, Fellant and Ilall 1994). The present vegetation widiin die stud\' area consists primarih' of cheat- grass, Broniis tectorum L., with sagebrush, Artemisia tridentata wyomingensis Beetle & ' Dcpartnitiil i>t Ph.iit. Soil ciml Knliiniolomcal ScicncfS, University ol Iclulio. Moscow, ID 83843. -.\utlior Ic) wliniii coi rfspondrucc sliould lie acklivssed. 352 1995] Grasshoppers and Livestock Grazing 353 Young and A. tridentata tridentata (Rydb.) Beetle, where it has not burned reeently. As of 1988, about 23% (ca 40,000 ha) of the study area consisted of crested wheatgrass, Agropijron cristatum (L.) Gaertn., plantings (USDI-BLM 1984, 1990). Materials and Methods Ungrazed sites were selected on the basis of grazing history (not grazed for at least 10 years), size (at least 16 ha), and shape (at least 100 m across the narrowest dimension). Fifteen rangeland sites were found within the study area that met these criteria. Most sites were isolated tracts fenced to exclude livestock and to provide habitat for upland game birds. Grazing by wildlife within the ungrazed tracts was negligible. Black-tailed jackrabbits were not abundant during the years in which sam- pling took place, and populations of prong- horn antelope, the only other large vertebrate herbivore present in the summer, are quite low and widely dispersed across the study area, ca 300 individuals over 180,000 ha (J.' Russell, USDI-BLM, personal communication). A grazed site was selected to match each ungrazed site for a total of 30 sites. In most cases grazed sites were adjacent to, and shared a boundary with, ungrazed sites. For six sites adjacent matched pairs were not pos- sible; consequently, grazed sites were chosen within 2 km. All grazed sites matched the ungrazed sites in soil type, topography, vege- tation, slope, and aspect. All grazed sites were located within BLM -administered grazing allotments. Stocking rates for the grazed sites varied from 1.9 to 2.8 ha/AUM (USDI-BLM 1990). Not all grazed sites were grazed each year, as prescribed by rest-rotation grazing management plans. Elevation of the sites ranged from 1180 to 1320 m. Five pairs of sites were located on areas replanted with crested wheatgrass, four pairs were on annual grassland sites having lit- tle or no sagebrush, and six pairs were located on sagebrush-grass sites. Grasshopper and vegetation sampling. — Grasshoppers were sampled on 19 July-7 August 1990 (adult stage), 21-28 June 1991 (primarily nymphal stage), and 2-13 August 1991 (adult stage). Corresponding sites of a grazed and ungrazed pair were always sam- pled on the same day. Grasshopper density on a site was estimated by counting the number of grasshoppers flushed from 50, O.l-m^ rings 5 m apart in a circular transect (Richards and Waloflf 1954, Onsager and Heniy 1977). Species composition on a site was determined by a "flush-capture method (Capinera and Sechrist 1982). Thirty to 100 specimens were captured and identified at each site on each sampling date by slowly walking in a circular transect and, to avoid bias toward more conspicuous species, counting only those grasshoppers en- countered directly in the path of the obsei-ver Vegetation was sampled on the same dates as the grasshopppers by visually estimating the percentage ground cover in 5% incre- ments by plant species in each of 40, 0.1-m^ square quadrats in a circular transect. Plant species unidentifiable in the field were col- lected and identified later. The percent cover of cryptogams, cattle dung, and bare ground was also estimated. Vegetation data from the three sampling dates were combined for sub- sequent analyses. Data analysis. — We classified the sites into three vegetation tyi3es based on dominant vege- tation on a site. Sites that had been seeded to crested wheatgrass were categorized as re- planted. Sites with sagebrush as the dominant plant species were placed in the sagebrush category, and the remaining sites, dominated by cheatgrass without significant sagebrush cover, were categorized as annual grasslands. Differences in percentage ground cover among vegetation types were confirmed with a Kruskal-Wallis non-parametric one-way analysis of variance (Zar 1984). Non-paramet- ric statistical tests were used with the ground cover data because of the large number of zero values involved. Gomparisons between vege- tation types were made with a non-parametric analog of Tukey's test (Dunn 1964, Zar 1984). Because paired sites were in close proximi- ty and of similar vegetation, we used Wilcoxon paired-sample tests to identify differences in percentage ground cover between grazed and ungrazed sites. Grasshopper densities were too low to con- duct meaningful statistical comparisons sepa- rately for all species. Accordingly, analyses were conducted on densities of total grasshoppers, on proportions of Melanoplus sanguinipes (the major pest species in the region), and on proportions of the \hree subfamilies of Acrididae within the region, Melanoplinae, Gomphocerinae, and 354 Great Basin Naturalist [Volume 55 Oedipodinae. Grasshopper densities were transformed by log^,(x + 1) to nornuili/.e the data. The aresin transformation was apphed to the proportions (Zar 1984). Three-way analysis of varianee (PROG GLM, SAS Inst.) was used to determine the significance of sampling date, vegetation type, and grazing treatment effects on grasshopper density and proportions. For the ANOVA, sites were not blocked by location; i.e., pairing was ignored. Gomparisons among vegetation types were made with least significant differ- ence mean separation tests (PROG GLM, SAS Inst.). Paired-sample t tests were used for comparisons between grazing treatments. Trends between habitat characteristics and grasshopper densities/proportions were evalu- ated by Spearman rank correlations (PROG GORR, SAS Inst.). Mean values from the three sampling dates were used for the correlation analyses. Results The replanted vegetation type was domi- nated by crested wheatgrass and also had the greatest amount of bare ground (Table 1). Annual grasslands were dominated by annual plant species, primarily cheatgrass (Table 1). Sagebrush sites had the greatest shrub cover, although other vegetation types had small amounts of sagebrush (Table 1). Annual grass- lands and sagebrush sites had little perennial grass cover, other than Poa sandbergii. Grazing treatment did not greatly affect most ground cover variables (Table 2), but sites grazed by livestock had more bare ground and cattle dung and less total vegetative cover and perennial grass cover than the permanently ungrazed sites. Twenty-three species of grasshoppers were indentified from the 30 sites. Melanoplus san- guinipes was found at all 30 sites and repre- sented 36% of all grasshoppers on the study sites. Other common species included Oeda- leonotus enigma (Scudder), Ageneotettix deonun (Scudder), Aulocara eUiotti (Thomas), Conozoa sidcifrons (Scudder), and TracJiyrachys kiowa (Thomas). No species other than M. san- guinipes comprised more than 10% of all grasshoppers from all sites. Total density of grasshoppers on the three sampling dates ranged from <().2 to 2.6 per m^. Table L Median (s.d.j percentage ground cover by veg- etation t\pe. Vegetation t) pe Ground c()\'er variables Replanted (.V = 10) Sagebrush (.V = 12) Annual grasslands (A' = 8) Annual grasses 1.2bi (3.7) 7.8b (8.1) 2().0a (9.0) Annual Forbs 0.5a (1.6) 1.2b (3.9) 3.1b (2.8) A^ropynm ciisfafiDn 16.8a (7.2) O.Ob (0.02) O.Ob (0.03) Poa sandbergii 6.9a (4.9) .5.0a (1-7) 12.0a (5.4) Other perennial 0.1a 0.6a 1.1a grasses (0.4) (2.9) (2.8) Sagebrush O.Ob (2.8) 13.0a (5.0) O.Ob (2.4) Total vegetation 26.1a 32.9a 41.8b (8.6) (5.7) (8.9) Cn'ptoganis 1.8a 7.0b 3. lab (3.2) (5.9) (4.3) Bare ground 40.,5a 24.()ab IS.Ob (13.7) (6.5) (S.3) 'Meiisures within rows followed by different letters are significantly different, P < .05, non-parametric analog of Tukey's test (Dunn 1964, Zar 1984). Total density of grasshoppers was affected by sampling date and grazing treatment (Table 3). No differences in density were detected among vegetation types (LSD mean separa- tion, P = .05; Table 3, Fig. 1). Proportions of M. sangidnipes and all species within the subfamily Melanoplinae were affected by sampling date, vegetation type, and grazing treatment (Table 3). Annual grass- lands had the highest proportions of A/, san- guinipes and of all species within the subfami- ly Melanoplinae (LSD mean separation, F = .05; Fig. 1). The proportion of grasshoppers within the subfamily Gomphocerinae was strongly affected by sampling date and vegeta- tion type, but not grazing (Table 3). Replanted (crested wheatgrass) sites had the highest pro- portions of Gomphocerinae (LSD mean separa- tion, P = .05; Fig. 1). Proportions of grasshop- pers within the subfamily Oedipodinae were significantly affected by sampling date, vegeta- tion i\npe, and grazing, although F-values were not as great as for proportions of the other subfamilies (Table 3). The mean proportion of 1995] Gr.\sshoppers and Livestock Gfl\zing 355 Table 2. Comparison of median (s.d.) ground cover between grazed and ungrazed sites. Grazing tre atment Grazed Ungrazed Annual grasses 10,0 (S.5) 7.0 (12.3) Annual forbs 1.0 (1.9) 1.5 (3.8) AJI perennial grasses* 3.0 (6.4) 4.3 (11.3) Poa sandhergii 8.7 (5.6) 7.1 (4.9) Sagebrush 3.0 (7.1) 3.0 (6.8) Total vegetation* 32.5 (7.6) 39.9 (6.3) Cattle dungi 0.7 (0.5) 0.0 (0.0) Cryptogams 4.6 (5.7) 6.2 (6.3) Bare ground** 31.0 (13.4) 23.0 (10.2) ^Because no cattle dung was recorded on tlie ungrazed sites, no statistical test of significance was perfomied. *Measures are signiflcantlv different (F < .05, Wilcoxon paired-sample test, N = 15), **Measures are signiflcantK- different (P < .(II, Wilcoxon paired-sample test, IV = 15). Total grasshopper density was not conelated with any ground cover variables. Proportion of M. sanguinipes was negatively correlated with percentage bare ground and cover of perenni- al grasses {r^ = -.59 and -.62, respectively, N = 30, P < .001), and was positively correlated with percentage ground cover of cheatgrass and annual forbs (r^ = .41 and .42, respective- ly, N = 30, P < .05). Proportion of all Melanoplinae combined was correlated posi- tively with cheatgrass (r^ = .52, N = 30, P < .01) and negatively with perennial grasses and percentage bare ground (-.70 and -.64, respectively N = 30, P < .001). As proportions of Melanoplinae declined with increasing cover of perennial grasses and bare ground, proportions of other species in- creased. Gomphocerinae showed trends oppo- site those of the Melanoplinae. Proportions of Gomphocerinae were correlated positively with perennial grasses and bare ground (.66 and .46, respectively, iV = 30, P < .01) and negatively with cheatgrass and annual forbs (-.52 and -.42, respectively N = 30, P < .05). Proportions of Oedipodinae were not signifi- cantly correlated (P > .05) with any of the ground cover variables. Discussion Oedipodinae was greater on sagebrush sites than on annual grassland sites (LSD mean sep- aration, P = .05; Fig. 1). The effect of grazing treatment was consis- tent across vegetation types and sampling dates for all grasshopper variables (Table 3). Because no significant interactions between date and grazing, or vegetation type and grazing, were detected (Table 3), comparisons of grazing treatments were made across all vegetation types and sampling dates. Overall density of grasshoppers was greater on ungrazed than on grazed plots (paired-sam- ple t test, P < .001; Fig. 2). Proportions of Af. sanguinipes and of all species of Melanoplinae combined were higher on the ungrazed sites (paired-sample t test, P < .005; Fig. 2). Oedi- podinae showed a trend opposite that of the Melanoplinae, being found in greater propor- tions on the grazed sites (paired-sample / test, P < .001; Fig. 2). Proportions of Gomphocerinae were not affected by grazing (paired- sample t test, P> 0.10; Fig. 2). Grazing influenced both total density and species composition of grasshoppers. Members of the subfamily Melanoplinae accounted for most of the increase in total density on un- grazed sites. Although vegetation did not affect density in this study, it strongly influenced species composition. Relative abundance of Gomphocerinae increased, and Melanoplinae decreased, with increasing coverage of peren- nial grasses and bare ground, while total num- bers of grasshoppers remained the same. Proportions of Af. sanguinipes, the primaiy pest species in the region, were negatively associ- ated with grazing, perennial grasses (primarily crested wheatgrass), and percentage bare ground. Habitat preferences of A/, sanguinipes, or any organism, represent an integrated response to many stimuli. Short-term changes in habitat due to grazing may include reduced quantities of food, less escape space, increased amounts of bare ground, altered host plant quality, and changes in microhabitat temperature and 356 Grkat Basin Natuiulist [Volume 55 T.-VBLE 3. Summary of Type III F-values (and sijiiiifieaiice knt'ls) from three-way ANOVA for densities of total grasshoppers, Melanoplus sanguinipes, Melanoplinae other tluin M. saii^iiinipcs, Gomphocerinae, and Oedipodinae. SoiMce d.f CJrasshopper (lensit\7m- Percentage Melanoplus sanguinipes Percentage Melanoplinae Percentage Gomphocerinae Percentage Oedipodinae Date (D) 2 12.9 (<.0 1) 18.7 (<.01) 11.9(<.01j 21.3 (.01) 6.1 (<.01) Vegetation t>'pe (V) 2 1.2 (.30) 19.3(<.01) 29.3 (<.01) 22.2 (<.01) 5.6(<.01) Cirazing (G) 1 5.6 (.02) 13.4 (<.01) 11.7(<.0]) 0.5 (.50) 8.8(<.01) V X D 4 2.1 (.09) 2.2 (.07) 2.8 (.03) 1.9 (.13) 1.2 (.34) G X D 2 0.6 (.58) 0.3 (.77) 0.1 (.88) 2.1 (.13) 0.7 (.48) V X G 2 0.7 (.49) 0.1 (.89) 0.1 (.87) 0.8 (.43) 0.9 (.43) D X V X G 4 0.3 (.91) 0.6 (.69) 0.5 (.76) 0.2 (.96) 0.2 (.9.3) Replanted Sagebrush Annuals Oedipodinae M. sanguinipes ^^ Gomphocerinae li-:s»il Other Melanoplinae Fig. 1. Mean density of Oedipodinae, Gomphocerinae, Melanoplinae other than Melanoplus sanguinipes, and M. san- guinipes by vegetation type. humidiW. Differences in plant species compo- sition between grazing treatments were mini- mal, indicating that long-term alteration of the plant community composition was probably not a factor. Results of this study are consistent with some previous studies. Proportions of M. san- guinipes were negatively coirelated with crested wheatgrass in this study. Fielding and Brusven (1992) demonstrated that crested wheatgrass is not a prefeired host plant for M. sanguinipes in southern Idaho. Perennial grasses are favored food plants for Aiiloeara eUiotti (Pfadt 1949, Fielding and Brusven 1992), the most common gomphocerine in the study area. Proportions of M. sanguinipes were also negatively corre- lated with percentage bare ground. Nerney and Hamilton (1969) and Kemp and Sanchez (1987) reported that M. sanguinipes avoids ovi- position in bare soil, whereas A. elliotti prefers to oviposit in bare ground (Kemp and Sanchez 1987, Fisher 1992). High percentages of bare 1995] Gr.\sshoppers and Livestock Gr.\zing 357 Grazed Ungrazed Oedipodinae M. sanguinipes ^^ Gomphocerinae I I Other Melanoplinae Fig. 2. Mean density' of Oedipodinae, Gomphocerinae, Melanoplinae other than Melanoplus sanguinipes, and M. san- guinipes b>' grazing histoiy ground were associated with both grazing and crested wheatgrass seedings in the present study (Tables 1, 2). Previous studies that examined grazing effects on grasshoppers reported results similar in some respects to those reported here. On the short-grass prairie of Colorado, Capinera and Sechrist (1982) reported that Oedipodinae were most abundant on the most heavily grazed pas- tures, while lightly grazed pastures supported the highest total grasshopper densities. Quinn and Walgenbach (1990) found Melanoplinae, particularly Melanoplus sanguinipes, to be dom- inant on ungrazed sites on mixed-grass prairies of South Dakota, even though total grasshop- per abundance was less than on grazed sites. However, Miller and Onsager (1991) were unable to detect any effect of different grazing regimes on adult grasshopper populations, including M. sanguinipes, in a crested wheat- grass pasture in Montana. Our obsen'ations were made under condi- tions of low grasshopper density and drought in southern Idaho. Fielding and Biaisven (1990) showed that grasshopper population density in southern Idaho was positively correlated with precipitation. Results of this study cannot be extrapolated to predict how grasshopper populations will respond to livestock grazing during more favorable years when rangeland productivity is high and grasshopper popula- tions are rapidly expanding. Rangeland grasshoppers have traditionally been viewed solely as destructive rangeland pests. However, in an ecosystem context they may have net beneficial worth during most years as an important food source for at least a pait of the life cycle of many species of mam- mals, birds, and reptiles. Results presented here may serve as a cautionary note regarding range- land ecosystem management under drought conditions. Grazing during years of drought and low grasshopper populations could con- ceivably add to the stress experienced by insectivorous animals by reducing available food resources (i.e., grasshopper populations), especially if other arthropods serving as alter- nate foods are also at low densities. Acknowledgments We thank the staff of the ELM Shoshone District office for their support. J. A. Onsager, M. A. Quinn, and L. P Kish reviewed earlier versions of the manuscript. William Price, sta- tistical research associate, Universitv of Idaho, 358 Great Basin NATuii\LiST [Volume 55 advised on statistical matters. Research was flmded in part by Bureau of Land Management as Cooperative Agreement No. 919-CA7-05 and published with the approval of the director of the Idaho Agricultural Experiment Station as Paper No. 92731. Lite MTU RE Cited Campuell, J. B., W. H. Arnett, J. D. Lambley, O. K. Jantz, AND II. Knltson. 1974. Grasshoppers (Acrididae) of the Flint Hills native tall grass prairie in Kansas. Kansas State University Agricnltural Experiment Station Research Paper 19. 147 pp. Capinera, J. L., AND T. S. Sechrist. 1982. Grasshopper (Acrididae)-host plant associations: response of grasshopper popnlations to cattle grazing intensity. Canadian Entomolgist 114: 1055-1062. Coyner, W. R. 19.38. A report of the effect of overgrazing on the Acrididae. Proceedings of the Oklahoma Academy of Science IS: 83-85. Dunn, O. J. 1964. Multiple contrasts using rank sums. Technometrics 6: 241-252. Fielding, D. J., and M. A. Brusven. 1990. Historical analysis of grasshopper (Orthoptera: Acrididae) population responses to climate in southern Idaho, 1950-1980. Environmental Entomology 19: 1786-1791. . 1992. Food and habitat preferences oiMelanoplus sanguinipes and Aulocara elliotti (Orthoptera: Acrididae) on disturbed rangeland in southern Idaho. Journal of Economic Entomology 85: 783-788. Fisher, J. R. 1992. Location of egg pods of Aidocara elliotti (Orthoptera: Acrididae) in a field of crested wheatgrass in Montana. Journal of Kansas Ento- mology Societ)' 65: 416-420. Holmes, N. D., D. S. Smith, and A. Johnston. 1979. Effect of grazing by cattle on the abundance of grasshoppers on fescue grassland. Journal of Range Management 32: 310-311. Jepson-Innes, K., and C. E. Bock. 1989. Response of grasshoppers (Orthoptera: Acrididae) to livestock grazing in southeastern Arizona: differences be- tween seasons and subfamilies. Oecologia 78: 430-431. Kemp, W. R, and N. E. Sanchez. 1987. Differences in post-diapause thermal requirements for eggs of two rangeland grasshoppers. Canadian Entomologist 119: 653-661. Mack, R. N. 1981. Invasion of Broinus tectorum L. into western North America: an ecological chronicle. Agro-ecosystems 7: 145-165. Miller, R. H., and J. A. Onsager. 1991. Grasshopper (Orthoptera: Acrididae) and plant relationships imder different grazing intensities. Environmental Entomology 20: 807-814. Nerney, N. j., and a. G. Hamilton. 1969. Effects of rain- fall on range forage and populations of grasshoppers, San Carlos Apache Indian Reservation, Arizona. Journal of Economic Entomology 62: 329-.333. Onsager, J. A. 1987. Current tactics for suppression of grasshoppers on range. Pages 60-66 in J. A. Onsager, editor. Integrated pest management on rangeland: state of the art in the sagebrush ecosvstem. USDA- ARS, ARS-50. Onsager, J. A., and J. E. Henry. 1977. A method for esti- mating the density of rangeland grasshoppers (Orthoptera: Acrididae) in experimental plots. Acrida 6: 231-237. Pellant, M., and C. Hall. 1994. Distribution of two exotic grasses on intermountain rangelands: status in 1992. In: Proceedings — Symposium on ecology, manage- ment, and restoration of intermountain annual range- lands. USDA-FS, Intermountain Research Station, General Technical Report INT-GTR-313. Pfadt, R. E. 1949. Food-plants, distribution, and abun- dance of the big-headed grasshopper, Aulocara elliotti (Thos.). JouiTial of the Kansas Entomological Societ\' 22: 69-74. PiCKFORD, G. D. 1932. The influence of continued heav> grazing and of promiscuous burning of spring-fall ranges in Utah. Ecology 13: 159-171. QuiNN, M. a., and D. D. Walgenbach. 1990. Influence of grazing history on the community structure of grasshoppers of mixed-grass prairie. Environmental Entomology 19: 1756-1766. Richards, O. W, and N. Waloff. 1954. Studies on the biology and population dynamics of British grass- hoppers. Anti-locust Bulletin 17. London. 182 pp. Smith, C. C. 1940. The effects of overgrazing and erosion upon the biota of the mixed grass prairie of Okla- homa. Ecology' 21: 381-397. Stewart, G., and A. C. Hull. 1949. Cheatgrass in south- ern Idaho. Ecologv' 30: 58-74. USDI-BLM. 1984. Proposed monument resource man- agement plan and final environmental impact state- ment. USDI-Bureau of Land Mangement, Wash- ington, DC. . 1990. Monument rangeland program smnmaiy Progress report. USDI-Bureau of Land Manage- ment Shoshone District Office, Shoshone, ID. Yensen, D. 1982. A grazing histoiy of southwestern Idaho with emphasis on the Birds of Prey Study Area. Research project report. United States Department of the Interior, Bureau of Land Management, Boise, ID. 82 pp. Zar, j. H. 1984. Biostatistical analysis. 2nd edition. Prentice-Hall, Inc., Englewood Cliffs, NJ. Received 30 March 1994 Accepted 10 April 1995 Great Basin Naturalist 55(4), © 1995, pp. 359-362 PLANT NOVELTIES IN LEPIDIUM (CRUCIFERAE) AND ARTEMISIA (COMPOSITAE) FROM THE UINTA BASIN, UTAH Stanley L. Welsh* and Sherel Goodrich^ Abstr.'VCT. — Named as new tiixa are Lepidiwn hiiberi Welsh & Goodrich, sp. nov., and Artemisia nova A. Nels. var. duchesnicola Welsh & Goodrich, var no\'. The taxa are provided with diagnoses and descriptions, and their relation- ships, provenance, and hahitats are discussed. Key icords: Lepidium huberi, Artemisia nova var. duchesnicola, new faxa, Utah, Uinta Basin. Noted historic pioneer Utah botanist Marcus Eugene Jones (1852-1934) has been quoted, perhaps apocryphally, as saying that he felt sorry for all future generations of botanists because so few plants remained for them to describe and name. Whedier die quote is tnae or not, the generations beyond Jones' time have not suffered from a shortage of areas of botanical inquiiy, including the discoveiy and naming of scores of plants new to science, and there are indications that future generations of botanists beyond the 1990s will continue to find and describe novelties. The flora is not yet fully understood. The Uinta Basin harbors numerous narrow endemics in many genera of plants, due in some part to the availabilit>' of unique geologi- cal substrates. Geomorphological processes have, through time, exposed geological strata of vaiying and diverse composition around the periphery of the basin and onto the slopes of mountains and plateaus that form its borders. The basin proper is the result of uplift during and following the Laramide Revolution, which resulted in a topographically low area south of the Uinta Mountains and north of the Tavaputs Plateau. The exposed strata vaiy in age fiom the present into the remote Precambrian epoch. Revealed are mud and siltstones, shales, sand- stones, limestones, and quartzites of enormous total thickness, each displayed in sequence like pages from a book. Some of the strata, especially the shales and mud and siltstones, weather into fine-textured, salt-laden sub- strates, others into sand and gravel, and still others into platy shales. Each of the substrates presents a different array of texture, salinity. trace elements, and other features important to plant growth. Plants have become adapted to the peculiarities of salt content or its lack, to textural differences, and to the peculiarities of water relationships. During the past several millions of years formations have been exposed and cut by erosional processes, and during that same period floras have developed. Varying attributes of the resulting erosional surfaces have allowed the evolution of present floras of the basin. Some Uinta Basin plant endemics are directly correlated to geological formations and are aligned along the strike of formations as though planted mechanically by some gigan- tic drill. In odiers die coiTclation is more subtle, but most endemics show some affinity to par- ticular formations. The present paper deals with two more narrowly restricted Uinta Basin endemics. Bodi of them have been known in collections since the 1980s. Lepidium huheri Welsh & Goodrich, sp. nov. Similis Lepidio inontano var. ohjssioides in habitu generali amplitudine sed foliis praecipue caulinis (foliis basalibus nullis vel evolutis debiliter) in basim lignosam et in siliculam amplitudinam difiPert. Plants subshrubs, the stems woody at the base, ashy or brownish, 1-2.5 dm long; branches puberulent throughout, green, 15-53 cm long; leaves all cauline, the eophylls reduced, prin- cipal lower leaves 2-3.5 cm long, 8-20 mm wide, pinnatifid, 5- to 7-lobed, the lobes often again lobed or dentate, smaller and entire upwards; panicles 3-10 (14) cm long, branches ^Departnient of Botany and Life Science Museum. Brigliam Young Universih; Pro\o, UT 84602. ^U.S. Forest Service, Ashlev National Forest, .3.5.5 N. Vernal Avenue. Vernal, UT 84078. 359 360 Great Basin Naturalist [Volume 55 corymboscly arranged; pedicels 2-4.5 nnii long, puberulent; sepals glabrous, 1.4-1.9 mm long, oval, green, the margin white; petals white, 2.3-2.6 mm long, 1.8-2.2 mm wide, shallowlv incised, the stvlc 0.4-0.8 mm long (Fig. 1). ' Suffrutices, caules lignei ad basin, cinerei vel brunneis, 1-2.5 dm long; rami puberulenti onuiino, virides, 15-53 cm longi; folia totus caulina, eoph\ His reductis, principalibus infer- nis foliis 2-3.5 cm longis et 8-20 nun latis, pinnatifidis, 5- to 7-lobatis, lobi plerumque lobati vel dentati, parvascens et integra sur- sum; paniculae 3-10 (14) cm longae, ramis corymbose dispositis; pedicelli 2-4.5 mm longi, puberulenti; sepala glabra, 1.4-1.9 mm longa, ovales, virides, marginibus latis albis; petala alba 2.3-2.6 mm long, unquibus 1 mm longis; staminalis filamenti glabra; siliculae glabrae, ovatae vel ovales, ca 2.3-2.6 mm lon- gae, 1.8-2.2 mm latae, incisura vadosa, styli 0.4-0.8 mm longi. Type:— USA: Utah: Uintah County, grow- ing under ledges of Park City (Phosphoria) Formation, above Weber Sandstone, T2S, R21E, S15 NEl/4, Big Brush Creek Gorge, Uinta Mountains, adjacent to black sage- brush/grass community, west exposure, at 2179 m elev., A. Huber 2400, 18 August 1994 (Holotype BRY; isotypes to be distributed). Additional collections: USA: Utah: Uintah County, 8 km N of Maeser, at base of Taylor Mountain, S. Goodrich, 1548, 13 August 1973; do, TIS, R20E, S30, Ashley Creek, drainage N of Sims Peak, 30 km N of Vernal, at 2959 m, D. Atwood 9128a, 30 July 1982; do, T2S, R21E, S34, SW/SW, N side of Red Mts., ca 16 km N of Vernal, at ca 2320 m, J. Tuhy 2693, 31 July 1986; T2S, R21E, S14 NW/SW, Big Brush Creek Gorge, A. Huber 858, 13 June 1994; do, A. Huber 880, 14 June 1994; do, T3S, R21E, S34, SW/SW, N slope of Red Mountain, ca 6.5 km NW of Steinaker Resei-voir, A. Huber & S. Goodrich 2390, 18 Aug. 1994; do, T3S, R21E, S3 NEl/4, Red Mt., ca 6 km NW of Steinaker Reservoir, A. Huber & S. Goodrich, 2392, 18 Aug. 1994; do, T2S, R21E, S34 SW/SE, Red Mt., ca 6 km NW of Steinaker Reservoir, A. Huber & S. Goodrich 2393, 18 August 1994 (all BRY, with numerous duplicates to be dis- tributed). There is a collection, apparently of this, taken from Moffat Countv, CO (R. C. & K. W. Rollins 8387, off countv roads 13 and 789, S of a dm Fig. 1. Habit sketch (a) and silick- (b) oi Lcpidiu)n huhch Welsh & Goodrich. Hamilton), at BRY. Its main difference is the merely toothed unlobed leaves. Pinnately lobed leaves are featured prominently in the material from the range of the species in Uintah Count)'. The following key will serve to distinguish L. hiiheri from other members of the L. )non- tanuiu complex. 1995] Lepidium AND Artemisia, Uinta Basin 361 1. Plants slightK if at all woocK' above the base, biennial to perennial herbs; silicles 2.8-4.1 mm long, 2.1-2.5 mm wide L. inontanum sens. lat. — Plants wood}' well abo\'e the base, long-lived perennial siibshrubs; silicles various 2 2(1). Silicles 4.5-7.5 mm long, 5.2-6.5, obovate; plants of the Moha\'e desert region of SW Utah and southward L. fronontii — Silicles 2.3-2.6 mm long, 1.8-2.2 mm wide; plants montane, in Uintah County, Uttih L. hiiheri This taxon, a definite subshrub, differs from L. montanum Nutt. sens. lat. in about the same degree and manner that the Mohavaean desert L. frcinontii Wats, differs from that species complex, i.e., in degree of woodiness and in size of the silicles, which in L. freviontii are on the large size for that complex and in L. hiiberi are smaller. Members of the montanum complex are widely distributed in the American West and occur in an array of mor- phological races, many of which are geograph- ically or edaphically correlated. Hitchcock (1936) treated 13 infraspecific taxa, some of which are now regarded at specific rank. The phase of the L. montanum complex that is apparently most closely allied to montane L. huheri is the extralimital van ahjssoides (Gray) Jones, to which early collections of this novel- ty were assigned. That variety, which ranges widely from Colorado to New Mexico, Arizona, and Texas, sometimes has a branching subligneus caudex, but is seldom if ever sub- shrubby, and lacks the other morphological features of L. huheri. The spatially and eleva- tionally isolated var. spathulatum (Robinson) C. L. Hitchc, also an ally, is rather common in Uintah County and elsewhere in eastern Utah. It is a tall plant, apparently biennial or short- lived perennial, with a single stem from the base, the caudex not woody or much branched. It is most common at low elevations along drainages, growing with sagebrush. Most phases within the montanum complex have been regarded at specific rank in the past, and there is more than marginal justification for so treating them in the future. Justification for regarding L. huheri at specific rank involves its combination of morphological characters, i.e., long-lived perennial habit, ligneus base, deeply lobed lower cauline leaves, and small silicles. Lepidium huheri grows in sand or silty sands derived from formations of various age from the Shinarump Member of the Chinle, Park City, and Weber Sandstone, all on the south-plunging flank of the Uinta Mountains. It occurs in black sagebrush, mountain brush, ponderosa pine, lodgepole pine, and spruce- fir communities at 2225 to 2960 m elevation. Artemisia nova A. Nels. var. dtichesnicola Welsh & Goodrich, var nov. Persimilis Artemisia nova A. Nels. in mag- nitudine et habitu sed in folius pilis albis dense non-glanduliferis et in floribus gener- aliter 5 (raro 4) et bracteis plus numerosis (10-20, nee 8-12) differt (Fig. 2). Shrubs, 1-3 (5) dm tall, main branches spreading, vegetative stems 1-3 dm long (rarely more); flowering stems mainly 1.5-3 (4) dm long; leaves dimorphic, 0.3-2 cm long, those of old stems shallowly to deeply 3- to 5- lobed or -toothed, lobes or teeth rounded, cuneate basally, appressed white canescent and not punctate; inflorescence narrowly pan- iculate, seldom more than 3 cm wide; involu- cres 3.1-5.8 mm long, 1.4-3.4 mm wide, cylin- dric to narrowly campanulate; bracts 10-20, mm Fig. 2. Drawing of floral head of Artemisia nova A. Nels. \'ai-. dtichesnicola Welsh &c Goodrich. 362 Great Basin Naturalist [Volume 55 white canescent, the margin hyahne; flowers 5 (rare!)' 4), all perfect; receptacle glabrous; ach- enes glabrous. Type.— USA: Utah: Uintah County, T5S R20E S5 NEl/4, 16 km W of Vemal, 1710 m elevation, desert shrub community, on heavy, reddish clay of the Duchesne Ri\ (m- Formation, S. Goodrich 23215, 17 Sept. 1990 (holotype BRY; isotypes to be distributed). Additional specimens. — USA: Utah: Uintah Count>', T5S R19E NW 1/4 S2, along Hw^ 121, 3 km E of Lapoint, 1740 m eleva- tion, Neese et al. 11013, 19 Sept. 1981; do, T5S R19E S2, 3 km NE of Lapoint, along Hwy 121, 1665 m, on red silty clay of the Duchesne River Formation, S. Goodrich 22225, 5 Sept. 1986; do, TIN RIE S26 SEl/4 USiM, 0.6 km SE of Tridell, 1720 m, heavy clay of the Duchesne River Formation, S. Goodrich 23212, 17 Sept. 1990; do, T5S R19E S2 NVVl/4, 3 km E of Lapoint, 1720 m eleva- tion, on heavy, reddish clay of the Duchesne River Formation, S. Goodrich 23214, 17 Sept. 1990; do, T3S R19E S35 El/2 SLM, about 1.4 km N of Hwy 121 between Lapoint and Maeser, red clays of Duchesne River Fomiation, 1800 m elevation, S. Goodrich 23255, 27 Sept. 1990 (all BRY, with numerous duplicates to be distributed). This taxon differs from typical A. nova A. Nels. in the densely white pubescent outer in- volucral bracts and generally denser pubescence of leaves and flowering stalks, and in the lack of conspicuous glandular dots on leaves. Leaves are not the green to lead-gray color typical of most populations of van nova, most of which also have glandular dots. There are, however, a few known populations of var nova that lack glandular dots, l>ut they possess the lead-gray to green color. In var. duchesnicola, the dense white, or silveiy, pubescence of leaves that lack glands is diagnostic. Additionally, mature involucres of var. duchesnicola are less lustrous, the number of involucral bracts is greater on the average (8-12 in var. nova, 10-20 in var. chichesnicola), and the flower number is almost uniformly 5 (not 3-8 as in var. nova). Practically all other features of the variety proposed herein are similar to var. nova. The proposed new variety would key in Welsh et al. (1993) to A. arhuscula Nutt. From that species var. duschesnicola can be distin- guished by its relatively shorter flowering stems, uniformly three-lobed vegetative leaves, much larger number of involucral bracts (10-20, not 4-8), and uniformly 5-flow- ered heads (not 4-9). The following key, modified from Welsh et al. (1993) will aid in identification of this taxon and its near allies. 1. Inflorescence open-paniculate, commonly more than 2 cm wide; plants often more than 5 dm tall A. tridentata var wijo)ningensis — Inflorescence narrowly paniculate, commonly less than 2 cm wide; plants usually less than 5 dm tall 2 2(1). Plants commonly 3-5 dm tall (sometimes taller); involucral bracts 4-8 A. arhuscula — Plants commonly 3 dm tall or less; involucral bracts averaging more than 8 3 3(2). Vesture of plants silverx' white; involucral bracts 10-20 A. nova var duchesnicola — Vesture of plants mainly lead-gi^ay; involucral bracts 8-12 A. nova var nova The var duchesnicola is the dominant plant, often in association with other desert shrubs, on reddish clay soils of the Duchesne River Formation, for which the variety is named, from about 15 km west of Vernal to Tridell. It occurs from about 1700 to 1800 m elevation on low clay uplands in a position ecologically between A. tridentata var wyorningensis (Beetle & Young) Welsh of desert drainages and A. nova var. nova, which grows in rocky sub- strates formed by ancient stream pediments. Suggested as the origin of this entity is poten- tial hybridization of A. nova and A. tridentata var wyomingensis. Although differing only in minor ways, the plants are continuous and uniform over rather large expanses of the Duchesne River Formation, and they are wor- thy of taxonomic recognition. References Hitchcock, C. L. 1936. The genus Lepidhnn in the United States. Madrono 3: 265-320. Welsh, S. L., N. D. Atwood, S. Goodrich, and L. C. HicciNS. 1993. A Utali flora. 2nd edition. Life Science Museum, Brigham Young Unixersitv; Provo, UT. 986 pp. Received 22 March 1995 Accepted 26 June 1995 Great Basin Naturalist 55(4). © 1995, pp. 363-367 PREY CHOICES AND FORAGING EFFICIENCY OF RECENTLY FLEDGED CALIFORNIA GULLS AT MONO LAKE, CALIFORNIA Chris S. Elphick^ and Margaret A. Ruhega^ Abstr.'\ct. — We studied the foraging hiologv' of recenth- fledged California Culls {Lams califoniiciis) at Mono Lake during August-September 1991. We made behavioral observations to collect information on the relative proportions of different prey types in the diet of these birds and took plankton tows to determine the relative abundance of each prey in the water column. These data show that alkali flies {Ephijdra hians) were the primaiy constituent of the diet and that they were eaten at a much higher rate than one would expect based on their abundance. We also detennined the num- ber of feeding attempts and successful captures made during each behavioral observation. From these, we calculated the birds' feeding efficiencies on emergent adult alkali flies and on all other pre>' t\'pes combined. We found that foraging efficiencies on emergent flies were ver\' high and significantly greater than those obtained on other prey types. These results suggest that flies were actively sought in preference to the alternative prey type, brine shrimp {Artcmia monica), presumably because they are easier to capture and of greater nutritional value. Key words: California Gull, Larus californicus, diet, foraging ejficiency. Mono Lake. California Gulls {Larus californicus) breed widely in die arid West, widi the largest con- centrations at two saline lakes: Great Salt Lake in Utah and Mono Lake in east central Cali- fornia (Conover 1983). Various factors may in- fluence the size and reproductive success of the California Gull colony at Mono Lake: pre- dation, food supply, weather, parasitism, nest- ing habitat, and access to freshwater (Winkler 1983, Winkler cited in Botkin et al. 1988). Of these, increased risk of predation caused by the exposure of a "land-bridge" between the mainland and islands on which the birds breed has received most attention (Patten et al. 1987, Botkin et al. 1988). The role of food abundance has received relatively little discussion, primarily because in- fonnation on the diet of California Gulls at Mono Lake is limited. Brine shrimp {Artemia monica) and alkali flies {Ephijdra hians) are the main sources of food available to gulls, although other items (e.g., cicadas, fish, and garbage) are occasionally taken (Patten et al. 1987). Previous reports have focused on the food brought to chicks at the nest. Some of these studies show chick diets to be dominated by brine shrimp (Grinnell and Storer 1924, Winkler et al. 1977, Jehl and Mahoney 1983), while others found high proportions of alkali flies (Nichols 1938, Young 1952, Mason 1967). Diet data for other age classes of gulls are not wide- ly available. Young (1952) dissected two indi- viduals and found their guts to be full of alkali fly pupae, and Jehl and Mahoney (1983) found high proportions (>90% by volume) of shrimp in a sample of free -swimming gulls (18 adults, 20 fledglings). These studies show that both brine shrimp and alkali flies are used by California Gulls at Mono Lake under certain circumstances. The factors that determine which of the two prey species, or which life stages of alkali flies, are taken are not known. Do the patterns simply reflect variation in rel- ative abundances of prey species, or is one species preferred but not always available? During three summers of fieldwork we noticed that over the latter part of summer Cali- fornia Gulls feed extensively on alkali flies, particularly recently emerged adults. Flies of this age class are immotile and presumably easier to catch than either brine shrimp or fly lai-vae (though not necessarily fly pupae). We therefore hypothesized that they would be a preferred prey source when available. In this paper we quantify the incidence of alkali flies in the diet of recently fledged California Gulls. 'EcologN. E\()lution and Conservation Biologx', Universit>- of Nevada, Reno. 1000 \'alle\- Road, Reno, NV 89.512. ."Vuthor to whom correspondence shonld be addressed. ^Department of Ecologv- and Evolutionary Biology, Universit>' of California, Inine, CA 92717. Present address: Ecolog\-, Evolution and Conservation Biology, University- of Nevada, Reno, 1000 Valley Road, jleno, NV 89512. 363 364 Great Basin Naturalist [Volume 55 We restiicted our study to juvenile gulls because inexperienced birds are t\'picall\ the least pro- ficient foragers (Porter and Sealey 1982, Burger 1987, Wunderle 1991) and hence most likely to benefit from the availability of easily cap- tured prey. We demonstrate that under certain circumstances alkali flies (1) constitute a major proportion of the diet and (2) are not eaten in direct proportion to their abundance. As a potential explanation for the birds apparent preference for alkali flies when available, we also test the hypothesis that fledgling gulls are able to achieve greater feeding efficiencies when eating emergent adult flies than when foraging on alternative prey. Methods Data were collected on five days during August and September 1991 from waters just off the northeastern shore of Mono Lake, where feeding gulls were numerous. Feeding observations. — We obtained feeding data by videotaping foraging birds with a Sony 8 mm HandyCam video recorder with an 8X zoom lens (n = 50) or by direct observations {n = 20). In all cases the focal bird was within 10 m of the observer, and for- aging behavior was scored over a 1-min feeding trial. No more than 10 birds were obsei"ved at any site to reduce the chance of obtaining repeated samples of the same individual. Feeding trials were scored for the number of feeding attempts and successful captures, which were divided by one minute to give attempt and success rates. When possible, prey items were identified. An attempt was defined as any occasion on which the bird's bill entered the water or the bird lunged for a prey item on the water's surface. Attempts were deemed successful if (1) the gull was seen "head-throwing" (i.e., inertial feeding; Gans 1961) and swallowing after the attempt, (2) the prey item was observed in the bird's mandibles and not dropped, or (3) the prey item was visible on the water surface before the capture attempt and was picked off by the gull. Filmed trials were scored at half-speed to improve accuracy. Data from the one day when both methods were used were compared to assess the relative accuracies of videotaping and direct obsen'ation. Diet. — We used two measures to determine the incidence of alkali flies in the diet of juve- nile gulls. First, we used the number of attempts directed at flies (all life stages), divided by the total number of attempts, as a measure of the proportion of foraging effort directed at alkali flies. Second, we calculated the mini- miun proportion of the birds diet that consti- tuted flies: fl\' captures attempts on all prey minus known failures. Prey abundance. — Prey abundance was determined from horizontal plankton tows taken at the site of, and immediately after, a series of feeding trials. Tows were made with a 0.5-^tm mesh plankton net, 1 m in diameter, and supported at the surface by floats. The tows sampled approximately 6 ni'^ of water, down to a maximum depth of about 60 cm. Samples were sorted and individuals of each alkali fly life stage counted. Because shrimp were too numerous to count, their numbers were calcu- lated from a previously deteniiined wet weight to number relationship (Rubega unpublished data): Weight (g) = 0.002207*Number (r- = .96, n = 10). Feeding efficiency. — We calculated feed- ing efficiency of juvenile gulls by dividing the number of successful prey captures by the number of attempts for both emergent adult alkali flies and all other prey types combined. These values were compared using a paired / test in which the two efficiency measures for each individual were paired. Feeding efficiency could not be calculated individually for other prey types because, unlike adult flies, they occurred below the water's surface and often could not be seen unless they were captured. Hence, usually we were unable to determine the object of the foraging attempt unless the attempt was directed at an adult fl\. All esti- mates are given in means (± standard error). Results Table 1 compares the minimum propor- tions of the total diet for each prey type with the relative abundances of each prey in plank- ton tows. Alkali fly adults and pupae both were eaten in much higher numbers than expected if prey were taken in proportion to their abundance. The mininumi proportion of 1995] California Gull foraging ecology 365 Table L Mean proportions (± SEM) of different prey b.'pes in the diet of fledged California Gulls {n = 70) and in plankton tows taken where birds were feeding (n = 21). No diet data are axailahle tor fl\' lai"vae or shrimp because they could not be distinguished in our feeding trials. Abundance Abundance in in diet plankton tows Prey type (% by number) (% by number) Alkali fix adults > 22.59 ± 0.35 0.01 ±0.003 Alkali fl\' pupae > 18.20 ± 0.39 0.67 ± 0.40 Alkali fl\ larvae — 0.05 ± 0.0007 Alkali flies (all life stages) > 40.79 ± 0.36 0.74 ± 0.04 Brine shrimp — 99.25 ± 0.04 foraging attempts directed at flies (all life stages) and the minimum proportion of the diet comprised of flies were 41.7 ± 3.0% and 40.8 ± 3.0%, respectively {n = 70). In compar- ison, only 0.7 ± 0.8% (n = 22) of prey items sampled in the water column were alkali flies; the remainder were all brine shrimp. These data indicate that alkali flies were favored over brine shrimp. The two sampling methods are compared in Table 2a. Attempt and success rates for all prey types combined did not differ significant- ly between the videotaped feeding trials and those obtained by direct observation (^33 = -0.1, P = .933 and ^33 = 1.56, P = .128, respectively). Proportions of different prey types recorded did differ, however, with videotaped trials, underrecording the number of pupae captured by an average of 79.7% on the one day for which a comparison was possi- ble. Similar numbers of adult flies were detected by the two methods. This discrepancy was probably because, unlike adult flies, pupae do not float on top of the water surface and are difficult to see on film due to reflection. Values given above for the incidence of alkali flies in the diet are therefore underestimates. Mean foraging efficiency for recently fledged gulls feeding on emergent alkali flies was very high and significantly greater than mean effi- ciency on all other prey (Table 2b; paired ^45 = 10.8, P < .0001). In addition, a comparison of the two measures for each individual showed that in all but one case a bird's effi- ciency was greater when feeding on emergent flies. Although our foraging efficienc)' data for alkali fly pupae are limited because we did not always know what prey type an attempt was directed at, they do indicate that pupae were caught as easily as adult flies (Table 2a). Discussion The large difference between alkali fly use and abundance strongly suggests that flies were actively sought in preference to brine shrimp and that flies were an important component in the diet of the birds we obsei-ved. It is likely that our prey sampling regime underestimated the availability of alkali flies because (1) we sampled deeper in the water column than gulls forage and (2) emergent flies are most abun- dant at the surface. It is unlikely, however, that this could account for the 60-fold difference between observed and expected values for fly abundance in the birds' diet. Two factors may contribute to the apparent preference for flies over shrimp. First, we have shown that 27% higher foraging efficiencies can be attained when feeding on emergent alkali flies than on alternative prey types combined. Second, Herbst (1986) reported that alkali flies are larger and have a greater nutritional value than the alternative food, brine shrimp. Both factors mean that there is an increase in food intake per unit effort when feeding on emergent flies. Although we have no quantitati\'e data for adult gulls, observations made during the course of this study suggest that they also fed predominantly on alkali flies. A supply of easi- ly caught prey, however, would be expected to benefit juveniles more that adults because the former lack foraging experience and are more likely to have difficulty feeding on more motile prey. Conclusions that can be drawn from these results are obviously limited. Our sampling was restricted to a few dates in one year and one portion of Mono Lake. Our anecdotal observa- tions from two additional years and surveys conducted across the entire lake suggest that these findings are not atypical for late summer, when emergent flies and dislodged pupae are common at the water surface. We have no data for other time periods; however, chick diet data collected earlier in the summer suggest that flies were eaten throughout the post-hatching period in 1991 (D. Shuford personal commu- nication). Jehl and Mahoney's (1983) data clearly show that under some circumstances brine shrimp make up a major portion of the diet of fledgling California Gulls. The differ- ence between their result and ours mirrors the variation seen in the diet of chicks (Grinnell and Storer 1924, Nichols 1938, Young 1952, 366 Great Basin Naturalist [Volume 55 Table 2. Mean feeding pcifonnancc v;ilues (± SEM). Sample sizes given in parentlieses. (a) Coniparati\'e \aliies for tlie two obsenation methods from tlie one da> on which hoth wiMe nsed. (li) Values for tlie two prey classifications for which accurate data could be collected from all studv (lavs. Prey type Attempt/min Success/min Efficiencv (%) (a) Comparison of observation methods Alkali tK adults (\idco trials) Alkali tl\ adults (direct observation) Alkali fly pupae (video trials) Alkali fly pupae (direct obsen'ation) All pre\' (\'ideo trials) All pre\' (direct observation) (b) Comparison of prey types Alkali fly adults All prey except adult flies 0.53 ±0.06 (15) 0.65 ±0.05 (20) 2.40 ±0.31 (15)* 9.50 ±0.22 (20)* 16.67 ±0.43 (15) 16.85 ±0.31 (20) 8..30±0.15 (70) 17.20 ±0.99 (70) 0.53 ±0.06 (15) 0.65 ±0.05 (20) 1.93 ±0.286 (15) 9..50±0.22 (20) 13.07 ±0.40 (15) 10.15 ±0.25 (20) 7.79 ±0.14 (70) 11.99 ±0.09 (70) 100 ±0 (5) 100 ±0 (8) 81.75 ±0.03 (8)* 100 ±0 (20)* 78.25 ±0.85 (15) 59.00 ±0.90 (20) 95.77 ±1.0 (70) 68.40 ±0.23 (70) *These data should be viewed with cautinn as attempt rates are Mason 1967, Winkler et al. 1977, Jehl and Mahoney 1983). Alkali fly abundance varies seasonally with an increase during May and June, peak numbers between July and Sep- tember, and a gradual decline thereafter (Herbst 1986). Research by Point Reyes Bird Observatory shows that the relative propor- tions of flies and shrimp in food brought to chicks differ considerably between samples collected during the day and night, and between years (D. Shuford personal communi- ation). These observations not only suggest diat relative availability of the two prey is quite variable at daily, seasonal, and annual time scales but also help explain the discrepancies between studies. Previous diet studies did not present data on relative prey abundances in areas where birds were foraging. In demon- strating a higher than expected abundance of alkali flies in the diet of fledgling gulls and the high foraging efficiencies that can be attained when feeding on them, our study suggests that flies are the preferred prey when they are available. In light of recent research on Red-necked Phalaropes {Phalaropus lohatus), which are physiologically imable to survive on a diet of pure brine shrimp (Rubega and Inouye 1994), our data lead us to speculate that brine fly production may be an important factor in determining fledgling survival rates (currently unknown) for the Mono Lake gull colony. California Gulls clearly eat brine shrimp on a regular basis and apparently are not as depen- dent on alkali flies as Red-necked Phalaropes. However, it is not clear whether the prey supply is limiting the gull population size. Experiments needed to address that issue have yet to be per- formed. In addition, it is possible that gull predation plays an important role in determin- ing alkali fly recruitment rates. The extent to which these issues are important can only be established through further study of the inter- actions between flies and gulls, both at Mono Lake and elsewhere. Acknowledgments We thank D. Elphick, D. Dawson, staff at the Sierra Nevada Aquatic Research Labora- tory, the High Sierra Shrimp Plant, and W Hamner for their contributions to the comple- tion of fieldwork. D. Shuford kindly gave us access to unpublished data collected by Point Reyes Bird Obsei-vatoiy. E. Beedy, L. Oring, M. Reed, D. Shuford, R. Whitmore, and three anonymous reviewers made useful comments on earlier versions of this paper. Fieldwork was supported by a grant fi-om the Universit\' of California, Irvine Foundation to MAR, made possible by a donation from Jones & Stokes Associates, consultants to the Los Angeles Department of Water and Power and the California State Water Resources Control Board. CSE received travel funds fi-om the Uni- versity of East Anglia's Expedition Committee and the Sir Phillip Reckitt Educational Trust. Literature Cited BoTKiN, D. B., W. S. Broecker, L. G. Everett, J. Shapiro, AND J. A. WiENS. 1988. The future of Mono Lake: report of the Community and Organization Research Institute (CORI) "Blue Ribbon Panel." Universit>' of California, Water Resources Center Report 68. 1995] California Gull foraging egology 367 Burger, J. 1987. Foraging efficiency in gulls: a congeneric comparison of age differences in efficiency and age of maturity. Studies in Avian Biolog\' 10: 83-90. CoNOVER, M. R. 1983. Recent changes in Ring-billed and California Gull populations in the western United States. Wilson Bulletin 95: 362-.383. Cans, C. 1961. The feeding mechanism of snakes and its possible evolution. American Zoologist 1; 217-227. Grinnell, J., AND T. I. Storer. 1924. Animal life in the Yosemite. University' of California Press. Herbst, D. B. 1986. Comparative studies of the popula- tion ecology and life histon patterns of an alkaline salt lake insect: Ephydra (Hijdropyrtts) hians (Diptera: Ephydridae). Unpublished doctoral dissertation, Oregon State University, Conallis. Jehl, J. R., Jr., and S. A. Mahoney. 1983. Possible se.xual differences in foraging patterns in California Culls and their implications for studies of feeding ecolog\'. Colonial Waterbirds 6: 218-220. Mason, D. T. 1967. Limnology of Mono Lake, California. University' of California Publications in Zoology 83. Nichols, W. E 1938. Some notes fi-om Negit Island, Mono Lake, California. Condor 40: 262. Patten, D. T, et al. 1987. The Mono Basin ecosystem: effects of changing lake level. National Academy Press, Washington DC. Porter, J. M., and S. C. Sealey. 1982. Dynamics of sea- bird multispecies feeding flocks: age-related feeding behaviour. Behaviour 81: 91-109. RuBEGA, M., AND C. Inouye. 1994. Prey switching in Red-necked Phalaropes {Phalaropus lobatus): feed- ing limitations, tlie flinctional response and water man- agement at Mono Lake, California, USA. Biological Consei-vation 70: 205-210. Winkler, D. W. 1983. Ecological and behavioral determi- nants of clutch size: the California Cull (Lams cali- fornicits) in the Great Basin. Unpublished doctoral dissertation. University of California, Berkeley. Winkler, D. W, C. R Weigen, E B. Engstrom, and S. E. Birc:h. 1977. Ornithology. Pages 88-113 in D. W Winkler, editor. An ecological study of Mono Lake, California. Institute of Ecology Publication No. 12. University of California, Davis. WUNDERLE, J. M., Jr. 1991. Age-specific foraging profi- ciency in birds. Current Ornithology 8: 273-324. Young, R. T 1952. Status of the California Gull colony at Mono Lake, California. Condor 54: 206-207. Received 1 7 October 1994 Accepted 20 June 1995 Great Basin Naturalist 55(4), © 1995, pp. 368-371 HYBRIDIZATION BETWEEN BUFO WOODHOUSII AND BUFO PUNCTATUS FROM THE GRAND CANYON REGION OF ARIZONA Keith Malmos', Holiert Reed', and Bnan Starrett^ Key words: hijhridizatidii, Bulo woodhoiisii, BiiFo punctatus, toads, Anura, dislrihiifion. Natural hybridization between toads of the genus Biifo is eommon; most accounts involve representatives from the same species group (Sullivan 1986). Species groups within the genus Biifo are hypodiesized to be monophyletic groups, based on data that include osteology, lab hybridization studies, advertisement calls, and release calls (summarized in Blair 1972a). Intergroup hybrid adults are expected to be relatively more rare in nature because of the low proportion that develop completely (Blair 1972b). Here we report intergroup hybridiza- tion between Biifo woodhoiisii {ainericanus group) and Biifo punctatus (jninctatus group; Blair 1972c). Hybrid B. punctatus x B. wood- hoiisii previously reported from Colorado near Grand Junction were described as "sterile males with atrophied testes" (McCoy et al. 1967). We present evidence that B. wood- hoiisii and B. punctatus have hybridized at two new localities in Arizona, Coconino Co., and that atrophied testes are not universal in these hybrids. The localities are approximately 3 km upstream from the Colorado River, near Powell Canyon in the Little Colorado River Gorge, and approximately 8 km downstream of the confluence of the Little Colorado River and Colorado River where Lava Creek emp- ties into the Colorado River. We also analyzed specimens collected by S. W. Aitchison in 1973 from Choal Canyon, Coconino Co., approximately 22.5 km NNE of Kaibito; these specimens include putative hybrid B. puncta- tus X B. woodhoiisii. Hybrids from this series are likely the toads that support the comment by Miller et al. (1982) that hybridization between B. punctatus and B. woodhoiisii occurs in Grand Canyon National Paik. Toads were identified and analyzed mor- phologically using methods similar to those of Ferguson and Lowe (1955) and McCoy et al. (1967). Each toad was dissected to determine sex and condition of testes of putative hybrids. Twelve specimens from the Little Colorado River (LCR) site [3 B. punctatus (ASU28935- 28937), 8 B. woodhoiisii (ASU28939-28946), and the hybrid (ASU28938)], and 15 of the 17 specimens fi'om Choal Canyon (CC) [8 B. piinc- tatm (MNA Z6.529-536), 5 B. woodhoiisii (MNA Z6.522-526), and 2 hybrids (MNA Z6.527- 528)] were analyzed. The tv\'o toads fi-om Choal Canyon excluded from the analysis were too small to evaluate reliably since ontogenetic changes in cranial crest and parotoid gland morphology occur in some toads (Sullivan 1986). Measurements were taken from pre- served male toads that were all the size of reproductively mature individuals. A Helios vernier caliper precise to 0.05 mm was used. Body size and parotoid gland variation among species of toads are diagnostic for many species. For the toads we examined, B. wood- hoiisii is larger and has more elongate parotoid glands than B. punctatus, which is a smaller toad with small, round parotoid glands. We measured snout-vent length (SVL) and paro- toid gland length (PL) and width (PW). A ratio of parotoid gland dimensions (PL/PW) was formed to evaluate gland shape. All toads from the LCR collection have developed gonads and secondary sexual char- acteristics. The three B. punctatus and four B. woodhoiisii males exhibit darkened vocal sacs, well-developed thumb pads, and testes typical for the species. The other four B. woodhoiisii appear to be spent females containing ovaries ^Arizona State Universit>, Department of Zoology, Tenipe. AZ 85281-1501. 2plioeiiix Zoo, 455 North Galviii Park-wa\, Phoenix, AZ 85008. 368 1995] Notes 369 with undeveloped eggs. The hybrid male pos- sesses one typical looking testis and one great- ly enlarged testis, approximately 10 times nor- mal size. Morphological analysis supports identifica- tion of ASU28398 as a hybrid. Values present- ed are the mean ± SD. The hybrid was 58.80 mm SVL, larger than B. piinctatus (42.17 ± 1.48 mm) but similar in size to B. woodhoiisii (59.69 ± 6.03). Shape of the parotoid gland, PL/PW, was intermediate for the hybrid, 1.43, relative to B. piinctatus (1.017 ± 0.053) and B. woodhoiisii (2.161 ± 0.330; Fig. 1). Both PL/PW and SVL are different between the species with at least 95% confidence because the means ± 2STD do not overlap. Although no specimens were retained from the Lava Creek site (LC), photographs taken in April 1993 provide clear evidence of hybrid- ization between B. piinctatus and B. wood- housii at this second site in Grand Canyon National Park (Fig. 2). Body size, parotoid gland moiphology, and coloration of the adult male hybrid are intermediate. Biifo ivood- housii is larger, has much more elongate paro- toid glands, and lacks the spinose red warts seen in Biifo piinctatus. We submit this photographic evidence and morphological analysis of toads as support for the suggestion by Stevens (1983) that B. piinc- tatus X B. woodhoiisii hybrids occur in the Grand Canyon region of Arizona. We also sug- gest that, based on specimens not from the Grand Canyon region, but from specimens collected associated with Glen Canyon, Miller et al. (1982) reported that B. piinctatus X B. woodhoiisii hybrids occur in the Grand Canyon. Whether B. piinctatus x B. wood- housii hybrids from LCR and LC could repro- duce would require histological analysis and additional sampling to determine if hybrids have viable sperm. We are, however, unaware of other reports of enlarged testes in hybrid toads. Three toads from Choal Canyon, MNA Z6.527-528 and MNA Z6.496, may be hybrids based on intermediate values of PL/PW (1.42 ± 0.04). As in the LCR series, SVLs of hybrids (58.01 ± 3.37) are greater than B. punctatus (48.84 ± 6.17), but similar to B. woodhoiisii (56.02 ± 7.87; Fig. 1). The means ± 2STD for SVL and PL/PW overlap for the CC sample; 45 50 55 60 65 snout-vent length (mm) Fig. 1. Comparison of relative sizes and parotoid gland dimensions of specimens from the Little Colorado River locality (closed symbols) and the Choal Canyon locality (open symbols). Circles are Bufo woodhoiisii, squares are Biifo punctatus, and triangles are hybrids. therefore, significant statistical differences do not exist. A small sample size is likely influen- tial. Gonadal development in some CC hybrids is unusual; MNA Z6.496 could not be sexed by its gonads or secondary sexual characters. The other two hybrids, MNA Z6.527 and 528, have darkened thumb pads and vocal sacs. Both testes of MNA Z6.527 appear normal, but MNA Z6.528 has one enlarged testis and the other absent or greatly reduced. Again, whether hybrid males of this cross are reproductively functional is unknown. Field observations suggest that hybridiza- tion at LCR may be relatively common. When the LCR collection was obtained, 13-14 May 1993, advertisement calls typical of B. wood- hoiisii and B. punctatus were both heard at night, as well as calls that sounded aberrant, approximately intermediate in duration, pulse rate, and pitch of each species. No other species of toads were obsened during spring months 370 Great Basin Naturalist [Volume 55 Fig. 2. Photographs of toads from the Lava Creek locah- ty: (a) Biifo woodhoiisii, (b) hybrid, (c) Biifo piincfatiis. at the LCR site for two years. Advertisement calls produced by hybrid toads often have characteristics intermediate to their parental forms (Blair 1956, Zweifel 1968, Sullivan 1986, 1990). Calls of suspected hybrids were not heard at Lava Creek, but both species cho- rused together there in April 1993. Habitat disturbance and environmental change associated with Glen Canyon Dam may contribute to hybridization between these taxa in die Grand Canyon region. Other hybrid zones between toads are associated with river regu- lation projects or human impacted areas (Sulli- van 1986 and examples cited therein). Altera- tions to the Colorado River have reduced sea- sonal peak flows, created large daily fluctua- tions in flow, and dramatically lowered the temperature of the water. Tributaries such as the Little Colorado River and Lava Creek are relatively less affected. Perhaps departure from historic conditions contributes to the likelihood of contact and hybridization between B. woodhousii and B. punctatus in the Grand Canyon. Other possible explana- tions for hybridization include natural pertur- bations that disrupt ecological separation. Also, natural cycles in population size and species range are hypothesized to account for many hybrid zones (Hewitt 1989). Acknowledgments We thank M. E. Douglas at Arizona State University and M. Morales and D. Hill at the Museum of Northern Arizona for use of speci- mens. We thank B. K. Sullivan for suggestions on the manuscript. We also thank the Navajo Fish and Wildlife Branch of the Navajo Nation for providing a collecting permit to RNR (#930709-058). Aiizona State University-West provided funds to KBM for some costs associ- ated with this project. Literature Cited Bu\iR, W. F. 1956. The mating calls of h\l)rid toads. Texas Journal of Science 8: 350-355. . 1972a. Evolution in the genus Biifo. Universit>' ot Texas Press, Austin. 459 pp. . 1972b. Exadence from lixbridization. Pages 196-232 in W. F Blair, editor. Evolution in the genus Biifo. University of Texas Press, Austin. . 1972c. Bufo of North and Central America. Pages 93-101 in W. E Blair, editor, Evolution in the genus Bufo. University of Texas Press, Austin. FERGU.SON, J. H., AND C. H. LoWE. 1969. The evolution- ary relationships in the Bufo punctatus group. American Midland Naturalist 81; 435—466. Hewitt, G. M. 1989. The subdivision of species by hybrid zones. Pages 85-110 in D. Otte and J. A. Endler, edi- tors, Speciation and its consequences. Sinauer Associates Inc., Sunderland, MA. McCoy, C. J., H. M. Smith, and J. A. Tihen. 1967. Natural hybrid toads, Bufo punctatus X Bufo wood- housei, from Colorado. Southwestern Naturalist 12: 45-54. Miller, D. M., R. A. Young, T. W. Gatlin, and J. A. Richardson. 1982. Amphibians and reptiles of the Grand Canyon National Park. Grand Can\on Natural Histoiy Association, Monograph No. 4. 1995] Notes 371 Stevens, L. 1983. The Colorado River in Grand Canyon; a guide. Red Lake Books, Fkigstaff, AZ. 110 pp. Sullivan, B. K. 1986. Hybridization between the toad Bufo microscaphiis and Biifo woodhoiisei in Arizona: morphological variation. Journal of HerfDetology 20: 11-21. . 1990. Natural hybrid between the Great Plains toad {Bufo cognatus) and red-spotted toad [Bufo piinc- tatiis) from central Arizona. Great Basin Naturalist 50: 371-372. ZWEIFEL, R. G. 1968. Effects of temperature, body size, and hybridization on mating calls of toads, Bufo a. aynericanus and Bufo woodhousii fowleri. Copeia 1968(2): 269-285. Received 14 October 1994 Accepted 20 March 1995 Great Basin Naturalist 55(4), © 1995, pp. 372-373 REPRODUCTION IN THE BANDED SAND SNAKE, CHILOMENISCUS CINCTUS (COLUBRIDAE), FROM ARIZONA StepluMi H. Goldhergl Key words: Chilonicnisciis ciiictus, handed sand snake. CoJuhridae. reprodneiion. Arizona. The banded sand snake, Chilomeniscus cinc- tiis Cope, 1861, ranges fi-om central Arizona to extreme southern Sonora, and throughout all but the northern part of Baja California (Stebbins 1985). Anecdotal comments on the reproduction of this species have been pub- lished in Stebbins (1954), Wright and Wright (1957), and Behler and King (1979), and in this report I provide data on reproduction in C. cinctus from Arizona. I examined 38 Chilomeniscus cinctus (24 males, 14 females) from Arizona in the heipe- tology collections of Arizona State University (ASU), Tempe; Natural Histoiy Museum of Uos Angeles County (UACM), Uos Angeles; San Diego Natural History Museum (SDSNH), San Diego; and the University of Aiizona (UAZ), Tucson. Museum numbers of specimens exam- ined are given in Appendix 1. All Arizona C. cinctus in the above collections were exam- ined; however, some had been damaged (road- kills) or had not been preserved promptly enough to avoid autolysis. These were not used and are not in Appendix 1. Counts were made of oviductal eggs or enlarged follicles. The left gonad was removed for histological examination, embedded in paraffin, and cut into histological sections at 5 /xm. Slides were stained with Harris' hematoxylin followed by eosin counterstain. Testes slides were exam- ined to determine the stage of the male cycle; ovary slides were examined for the presence of yolk deposition. Data on the male C. cinctus seasonal testic- ular cycle are presented in Table 1. Testicular histology was similar to that reported in Goldberg and Parker (1975) for two other North American colubrid snakes, Masticophis taeniatus and Pituophis nielanoleucus. In the regressed testes, seminiferous tubules con- tained spermatogonia and Sertoli cells. In recrudescence, there was renewal of spermato- genic cells characterized by spermatogonial divisions; primary and secondary spermato- cytes, and spermatids, may have been present. In spermiogenesis, metamorphosing sper- matids and mature sperm were present. Small sample sizes from all months except May-Iune (Table 1) prevented a definitive de- scription of the male cycle. However, since all 10 May males and 5 lune males were under- going spenniogenesis, it is likeK' that C. cinctus breeds during these months. Epididymides from 2 May and 1 lune males contained sperm. The smallest spermiogenic male (sperm pres- ent) measured 151 mm in snout-vent length (SVU). Data on the C. cinctus seasonal ovarian cycle are presented in Table 2. I recorded two clutch sizes: 6 lune, 3 enlarged follicles (3—4 mm diameter), 188 mm in SVU; 4 Inly, 2 ovi- ductal eggs (6 mm diameter), 192 mm in SVU. Yolk deposition (vitellogenic granules) was found on histological examination of ovarian Table 1. MonthK distribution of conditions in seasonal testicular cycle oi Chilomeniscus cinctus. Values shown are the numbers of males exhibiting each of the three con- ditions. Recru- Spermio- Montli X Regressed descence genesis |anuar\ 1 0 0 1 Pebruar\ 1 0 1 0 March 2 0 1 1 April 2 0 0 2 Mav 10 0 0 10 June 5 0 0 5 lulv I 1 0 0 September 1 1 0 0 December 1 0 1 0 ^Department of Biology, VVhitticr College, VVIiiltier, CA 90608. 372 1995] Notes 373 Table 2. Monthly distribution of conditions in seasonal ovarian cycle of Chilo)neni.$cus cinctus. Values shown are the number of females exhibiting each of the four conditions. Month N Inactive Yolk deposition Enlarged follicles Ov iductal eggs Februan 1 0 0 0 March 1 0 0 0 April 2 0 0 0 lime 4 2 1 0 Julv 2 0 0 1 August 1 0 0 0 September 1 0 0 0 October 1 0 0 0 November 1 0 0 0 tissue from two June females (173 mm and 198 mm in SVL). No yolk deposition was seen in the remainder of the female sample. The lack of vitellogenesis in some adult females during the reproductive season may indicate that not all C. cinctus females breed each year. Breeding by only part of the adult female pop- ulation has been reported for other North American temperate zone snake species (see Aldridge 1979). The smallest reproductively active female (yolk deposition in progress) measured 173 mm in SVL. The biology of C. cinctus is poorly known. A few reports on its food habits reveal that it eats centipedes and insects (Vorhies 1926, Stebbins 1954, 1985, Behler and King 1979). According to Lowe et al. (1986), C. cinctus has grooved rear teeth; it is not known whether it has toxic gland secretions. The small numbers of C. cinctus in the two major Arizona her- petology collections (ASU, UAZ) reflect the secretive nature of this snake. Intensive study will be rec^uired before the biology of C. cinc- tus is known. Acknowledgments I thank Charles H. Lowe (University of Arizona), Robert L. Bezy (Natural History Museum of Los Angeles County), Michael E. Douglas (Aiizona State University), and Sally Y. Shelton (San Diego Natural History Museum) for permission to examine snakes in the her- petology collections of their respective institu- tions. Jorge Martinez assisted with histology. Literature Cited Behler, J. L., and E VV. King. 1979. The Audubon Society field guide to North AiTieric;m reptiles and amphibians. Alfred A. Knopf, New York. 743 pp. Goldberg, S. R., and W. S. Parker. 1975. Seasonal tes- ticular histology of the colubrid snakes, Masticophis tacniatus and Pituophis melanoleucus. Heipetologica 31:317-322. Lowe, C. H., C. R. Schvvalbe, and T. B. Johnson. 1986. The venomous reptiles of Arizona. Arizona Game and Fish Department, Phoenix. 115 pp. Stebbins, R. C. 1954. Amphibians and reptiles of western North America. McGraw-Hill, New \brk. 536 pp. . 1985. A field guide to western reptiles and am- phibians. Houghton Mifflin Company, Boston. 366 pp. Vorhies, G. T. 1926. Notes on some uncommon snakes of southern Arizona. Copeia 1926:158-160. Wright, A. H., and A. A. Wright. 1957. Handbook of snakes of the United States and Canada. Volume I. Comstock Publishing Associates, Ithaca, NY. 564 pp. Received 29 November 1994 Accepted 7 February 1995 Appendix 1 Specimens examined by count>' from herpetology col- lections at Arizona State Universit\' (ASU), Natural History Museum of Los Angeles County (LACM), San Diego Natural History Museum (SDSNH), and University of Arizona (UAZ). Maricopa: ASU 04669, 09161, 13903, 26367-26368. LACM 112460. UAZ 24104, 35645, 35795, 35818. Pima: SDSNH 33383. ASU 01231, 15391, 28401. LACM 34918. UAZ 24087, 24089, 24092, 24095-24096, 24103, 24107- 24108, 30241, 33815, 34411, 34680-34681, 35166, 36108, 37819, 37821, 42197. Pinal: ASU 15376, 23.573. 26411, 26413. UAZ 24097. Aldridge, R. D. 1979. Female reproductive cycles of the snakes Arizona elegans and Crotalus viridis. Heipe- tologica 35: 256-261. Great Basin Naturalist 55(4), © 1995, pp. 374-376 NO ACOUSTIC BENEFIT TO SUBTERRANEAN CALLING IN THE CICADA OKANAGANA PALLIDULA DAVIS (HOMOPTERA: TIBICINIDAE) Allen E Sanborni .^ik-i poUy K. Phillips- Key words: cicada, acomfic hehavioi; calling, .sound pressure level, predator avoidance, Okanagana pallidiila. Most male cicadas produce a loud calling song to attract their mates. Sound pulses are produced when specialized muscles buckle the rib-strengthened chitinous membranes, the timbals, located on the dorsolateral surface of the first abdominal segment. Sound pulses are then modified by several body components (Pringle 1954, Bennet-Clark and Young 1992) before being radiated through the tympana (Young 1990). Male cicadas generally use an accessible perch fi-om which they advertise their presence to conspecific females. We came across an ex- ception to this behavior south of Lone Pine, Inyo County, CA, on 15 July 1994. We en- countered the cicada species Okanagana pal- lidiila Davis singing in a scrub habitat. As we began collecting, we noticed that one individ- ual continued to sing as we approached and was very difficult to locate on the plant. By circling the plant we found that the sound was actually coming from the ground near the base of the plant and not from on the plant itself After clearing some grass we could see a hole about 1 cm in diameter from which the sound emanated. Within the hole we could see the head of a cicada that was calling from this sub- terranean site. We measured intensity levels from males calling from burrows and from plants to deter- mine if there is an acoustic benefit for the cicadas calling in burrows. Peak sound pres- sure levels (SPL) were recorded with a Briiel & Kjaer 2235 SPL meter, a Type 4155 1/2" prepolarized condenser microphone, and a UA 0237 wind screen. The system had been calibrated with a Briiel & Kjaer 4230 portable sound pressure calibrator. The SPL meter was used in the linear frequency mode. The peak setting has a time constant of less than 100 ms and was used to ensure that rapid sound tran- sients were measured. Measurements were made peipendicular to the long body axis with the apparatus oriented medially along the dor- sal surface of a singing cicada at the thorax- abdomen junction or directly above the hole in which a cicada was singing. This procedure minimized any inconsistencies between readings due to possible asymmetries in the sound field produced by cicadas (Aidley 1969, MacNally and Young 1981). Each intensity measurement was made at a distance of 50 cm. The distance was kept constant by placing a 1/4" (6.5 mm) dowel, attached to the SPL meter, near a call- ing cicada. If the cicada was disturbed by placement of the SPL meter, the reading was made only after the normal calling pattern had been reestablished. All intensity measure- ments are relative to 1 X IQ-^^ W/cm^. Power output was determined using the following equation: Q = 47Cr2(I) where Q = sound power (W), r = distance from source in cm (= 50 cm), and I = intensi- ty reading for the individual (dB). Since inten- sity is measured on a logarithmic scale, all intensity measurements (dB readings) were converted to pressure levels (W/cm^) prior to calculating the statistics. Mean power output was then used to calculate mean sound inten- sity at 50 cm for each species. Intensity measurements are summarized in Table 1. SPL values recorded for cicadas call- ing from within a burrow are lower than val- ues recorded when the animals were calling from a plant; however, the values are not sig- nificandy different {t = 1.49, d.f = 3, F = .1159). A greater number of trials ma\' provide 'Bari-y University, School of Natural and lli'altli Scicnt'i's, 11300 N.E^ Second A\enue, Miami Shores, FL 33101-6695. 2Mianii-Dade Coninuinity College North C;anipns, Biology Department, 11380 N.W. 27th .\venne, Miami, FL .33167-3495. 374 1995] Notes 375 Table 1. Intensity of Okan(i' of America 69: 299-.306. Walker, T. J. 1964. Experimental demonstration of a cat locating orthopteran prey by the prey s calling song. Florida Entomologist 47: 163-165. Walker, T. J., and D. E. Figg. 1990. Song and acoustic burrow of the prairie mole cricket, Gnjllotalpa major (Orthoptera: Gryllidae). Journal of the Kansas Ento- mological Society 63: 237-242. Young, D. 1990. Do cicadas radiate sound through their ear-drums? Journal of Experimental Biology 151: 41-56. Received 24 March 1995 Accepted 28 June 1995 Great Basin Naturalist 55(4), © 1995, pp. 377-.378 BOOK REVIEW Natural History of the Colorado Plateau and Great Basin. K. T. Haiper, L. L. St. Claii; K. H. Thorne, and W. M. Hess, editors. University Press of Colorado, Niwot, CO. 1994. 294 pp. $24.95 hardbaek. Natural History of the Colorado Plateau and Great Basin, a multi-authored volume, is an introduction to the spectacular arid and remote North American landscape known as the Colo- rado Plateau and the Great Basin. The high, windswept plateau country is interrupted by numerous rocky canyons and arid valleys, and the Great Basin is a huge arid depression with no external drainages. According to the editors, this region is within the boundaries of Nevada, Utah, and Colorado (Fig. 1.2). They indicate that the intended audience of the volume includes students and managers of the region's natural resources. The basic objective of the major eleven chapters is to provide a "ready reference to the best of recent studies that are relevant to the region. " Additionally, the editors hope this volume will stimulate more research, especially on the Colorado Plateau, which is more "biodiverse and perhaps more fragile ecologically than the Great Basin. The map in Chapter 1 of the Great Basin and Colorado Plateau indicates a smaller region than maps in Chapters 2, 5, and 9, excluding areas as far noilli as Oregon and as far south as California, Arizona, and New Mexico. The boundaries of the Great Basin and Colorado Plateau therefore appear mildly confusing. A consensus map or better textual description (as presented in Chapter 5) could have been included in Chapter 1. Chapter 2 presents a rather concise and useful review of the geologic history of the Great Basin and Colorado Plateau. Chapter 3 attempts to summarize the complex climatic weather patterns in the broad context of the western United States and the globe. Anyone who has spent time in the Great Basin or Colorado Plateau knows well the unpredict- able and often extreme weather patterns that have helped form the regional geomorphology. Literature citations of this chapter are very useful. Chapter 4 reviews the extinct late Pleisto- cene mammals of the Great Basin. This region is rich in late Pleistocene vertebrate fossils, and the author provides a discussion of the possible causes of extinction and implications concerning present faunas. Western Great Basin archaeology in the context of regional cultural/environmental models is presented in Chapter 5. Wilde describes various prehistoric ebbs and flows of peoples for the past 12,000 years. Chapter 6 touches on the current politi- cally controversial subject of the changes in plant communities caused by domestic live- stock grazing, the most widespread land-man- agement practice in western North America. Seventy percent of the western United States is grazed, and ecological costs have been great. The author seems to concentrate on deleterious effects of the introduction of alien plants species such as Russian thistle and cheatgrass on the Great Basin plant communi- ties. He predicts that with continued removal of cattle, the "predators" of these plants, the ecosystem structure of the Great Basin may dramatically change in the near future. In Chapter 7 Jackson presents an enjoyable analysis of the unique factors that have influ- enced modem human development of resources in the Great Basin. He traces the cultural his- tory of the region, from the Dominguez and Escalante expeditions of the 1700s to the Mormon farmers who shaped the modern human geography of tlie Great Basin. As Jackson pointed out, the enduring legacy of the Great Basin is the "strange juxtaposition of religion and vice, destruction and recreation.' The authors in Chapter 8 use macrofossil data from packrat {Neotoma) middens to reconstruct tlie evolutionary history of eight modern conifer species. These conifer species now occupy the montane islands of the Great Basin, and the current distribution of these trees is related to past paleoclimatic changes. 377 378 Great Basin Naturalist [Volume 55 Sigler and Sigler in Chaper 9 present a ver\' comprehensive review of the fishes of the Great Basin and the Colorado Plateau. Excellent dis- cussions are presented for each species. However, there appear to be some errors; for example, the Big Spring spinedace is a native to the Colorado River Basin, not Lahontan, and the razorback sucker is a federalh' endangered species as of 1991. Additionally, if the map pre- sented in this chapter is inclusixe, then perhaps several other species could be added: Moapa dace, Moapa speckled dace. Meadow Valley speckled dace, Preston speckled dace, White River sucker, and Sonora sucker Also, I cannot construe the meaning of the last sentence in their chapter, "that many of the species, both native and exotic, have survived in spite of [human] modifications." The fact is, at least for the Colorado River Basin, most native fishes are in serious jeopardy of extinction; they have survived, but with a veiy precarious hold. Chapter 10 by Nelson attempts to cover a daunting subject, the insects of the Great Basin and Colorado Plateau. An estimated 14,000-26,000 species may occur within these boundaries. He discusses several of the better regional known taxonomic groups (stoneflies, butterflies, robber flies, and ants) to answer broad questions, such as, "What range patterns are seen in the Great Basin and Colorado Plateau?" and "Did these groups evolve in the Great Basin?" Many of the insects of this region have a widespread distribution throughout the West, and the insect fauna of the Colorado Plateau have strong affinities with the Rocky Mountains physiographic province. Warren and Harper in Chapter 11 briefly discuss ele- vational patterns of insects in the Great Basin and Colorado Plateau. Most of their examples, however, are higher elevational patterns of the Rocky Mountains and elsewhere, and the dis- cussion is limited to adaptations of insects to harsh environments. Their literature review is excellent. In Chapter 12, Mead and Bell describe the heipetofauna of the Great Basin and Colorado Plateau in the late Pleistocene and Holocene (i.e., during the past two million years, or Quaternaiy Period). Their comparison of mod- em fauna with the Pleistocene-Holocene indi- cates that 61% of the modern fauna is repre- sented in the fossil record, an interesting observation considering the climatic and envi- ronmental change in association with such events as ice ages. In Chapter 13 the editors provide recom- mendations for future directions of research, emphasizing the need for descriptive work. They also state evolutionary and ecological questions about the biodiversity of the Great Basin and Colorado Plateau that need urgent attention. This little book packs in much useful infor- mation, and with its reasonable price it should appeal to all students who work or visit the Intermountain West. The editors have suc- ceeded in presenting a good introduction to many important and conspicuous aspects of the natural history of the Great Basin and Colorado Plateau. B. C. Kondratiefif Colorado State University Department of Entomology Fort Collins, CO 80523 H E GREAT BASIN NATUKALIST INDEX VOLUME 55 — 1995 BRIGHAM YOUNG UNIVERSITY Great Basin Naturalist 55(4), © 1995, pp. 380-386 INDEX Volume 55—1995 Author Index Anderson, Loran C, 84 Anstin, Dennis D., 267 Baker, William L., 287 Bartholomew, Breck, 282 Baumann, R. W, 124 Belk, Mark C, 183 Bodie, Walt, 181 Bowlin, W. R., 19 Brusven, Merlyn A., 352 Callahan, J. R., 89 Carter, Bernard, 169 Cates, Rex G., 29 Cieminski, Karen L., 105 Cifelli, Riehard L., 304 Clements, Charlie, 188 Compton, Stephen B., 89 Crawford, John A., 284 Crompton, Clifford, 322 Czaplewski, Nicholas J., 304 DeBolt, Ann Marie, 237 DeWalt, R. Edward, 1 Dobkin, David S., 315 Ehlerin^er, James R., 135 Elphick, Chris S., 363 Fielding, Dennis J., 352 Flake, Lester D., 105 Flinders, Jerran T, 29 Friedman, Jonathan M., 58 Furniss, Malcolm M., 335 Geer, S. M., 19 Gerdes, Michael G., 315 Gettinger, Ronald D., 315 Goldberg, Stephen R., 372 Goodrich, Sherel, 359 Griswold, T. L., 19 Hansen, E. Matthew, 158 Haiper, Kimball T. (rev.), 286 Heckmann, Richard A., 258 Hubert, Wayne A., 169 Ischinger, Lee S., 58 Johnson, James B., 335 Johnson, Jerald B., 183 KondratiefT, B. C. (rev.), 377 Kucera, James R., 92 Lesica, Peter, 142 Longland, William S., 188 Lytle, C. Mel, 164 Malmos, Keith, 368 McArthnr, E. Durant, 151 McCoy, Matthew, 181 McCune, Bruce, 237 Meinke, Robert J., 249 Miller, Richard F, 37 Minshall, G. Wayne, 193 Munger, James C, 74 Muth, Robert T, 95 Nohavec, Robert D., 282 Owen, Wayne R., 117 Pelren, Eric C, 284 Phillips, Polly K., 192, 374 Qi, Ying, 258 Ratliff, Ra\inond D., 46 Reed, J. Michael, 342 Reed, Robert, 368 Renkin, Roy A., 201 Robinson, Christopher T, 193 Rose, JeffeiT A., 37 380 1994] Index 381 Rose, Kenneth D., 304 Royer, Todd V, 193 Rubega, Margaret A., 363 Rushforth, Samuel R., 193 Sanborn, Allen E, 192, 374 Sandquist, Darren R., 135 Shafroth, Patrick B., 58 Shepard, W. D., 124 Shiozawa, Dennis K., 183, 213 Singer, Francis J., 201 Slichter, Todd A., 74 Smith, Bruce N., 164 Snyder, Darrel E., 95 Starrett, Bryan, 368 Stewart, Kenneth W, 1 Storz, Jay F, 78 Stricklan, Dave, 29 Taye, Alan C, 225 Taylor, El Roy 181 Tepedino, V. J., 19 Urness, Philip J., 267 Vicker>', Robert K., Jr., 174, 177 Walford, Gillian M., 287 Welsh, Stanley L., 66, 271, 322, 359 Winward, Alma H., 151 Wolz, Eric R., 213 Key Word Index Taxa described as new to science in this volume appear in boldface t>'pe in this index. acoustic behavior, 374 Acrididae, 352 age, 183 allopolyploid, 151 alpine, 117 vascular flora, 225 vegetation, 225 analysis elasticity, 142 multivariate, 287 aneuploidy, 174 Anura, 368 Apiosorna campamdatum, 258 aquatic invertebrates, 105 Arizona, 372 Artemisia arhuscula ssp. longicaulis, 151 nova van duchesnicola, 359 Astragalus, 117, 142 Atriplex types, 322 avian diversity, 342 avoidance predator, 374 back-waters, 95 banded sand snake, 372 behavior, 282 acoustic, 374 feeding, 192 benthic habitat, 193 benthic macroinvertebrates, 213 benthos, 213 big sagebrush browsing, 210 bioaccumulation metal, 164 Blue Grouse, 284 browse (lirowsing), 267 big sagebrush, 210 Bufo woodhousii, 368 punctatus, 368 bumblebees, 177 bundle sheath leakiness, 135 California bighorn sheep, 181 Gull, 363 Mono Lake, 363 White Mountains, 117 Calileuctra, 124 calling, 374 candidate species, 315 carbon isotope ratio, 135 caves, 315 Celtis reticulata, 237 Ceratopogonidae, 213 Cervus elaphus, 201 chemotaxonomy, 151 Chenopodiaceae, 322 Chilomeniscus cinctus, 372 Chironomidae, 213 cicada, 374 coevolution, 188 Colorado, 287 Green River, 95 Plateau, 225 Colubridae, 372 competition, 117 conservation priorities, 342 Cottus bairdi, 258 Crotahis scutulatus, 282 Cijprinclla lutrensis, 95 382 Great Basin Naturalist [Volume 54 death canius, 188 defoliators, 158 demography, 142 Dendragapiis ohscunis, 284 density population, 352 description, 124 desert, 169 ecology, 135 diatoms, 193 diet(s), 95, 363 overlap, 95 distribution, 124, 169, 368 disturbance, 193 diversity avian, 342 Douglas-fir, 158 Douglas rabbitbrush, 267 Dragon local fauna, 304 drought, 352 ecological aspects, 258 ecolog\', 237 elasticit\' analysis, 142 elk northern Yellowstone, 201 Erythranthe, 177 escape saltation, 282 evolution, 174 fauna Dragon local, 304 Wagon Road local, 304 faunal list, 335 fecundity, 117 feeding behavior, 192 fish(es), 169 nonnative, 95 flower color mutations, 177 food source, 192 foraging, 192 efficiency, 363 movements, 315 forest insects, 158 Gila copci, 183 Great Basin, 249, 342 Green River, [Colorado], 95 Green River, [Utah], 95, 213 growth, 183, 237 habitat, 169 benthic, 193 halophyte, 135 herbivory, 142 hibernacula, 315 horses, 267 host-symbiont relationship, 258 hummingbirds, 177 hybrid, 151 h\bridizati()n, 368 Idaho, 181, 237 National Engineering Laboratoiy, 105 industrial wastewater, 105 Insecta. 124 interception trap, 284 invertebrates aquatic, 105 Lams culijurnicus, 363 lava flows, 315 leatherside chub, 183 Leguminosae, 271 Lepidium huberi, 359 Leuctridae, 124 life histoiy, 183 livestock grazing, 142, 237, 352 longevity, 237 macroinvertebrates, 193 benthic, 213 Mammalia, 304 matrix projection models, 142 Melanophis san', Jr 1 74 Speciation in Mimulus, or. Can a simple flower color mutant lead to species divergence? Robert K. Vickeiy Jr 1 77 Fall lamb production by a California bighorn sheep Matthew McCoy, Walt Bodie, and ElRoy Taylor 181 Age, growth, and reproduction of leatherside chub [Gila copei) . . Jerald B. Johnson, Mark C. Belk, and Dennis K. Shiozawa 1 83 Consumption of a toxic plant [Zlgadenns paniculatus) by mule deer William S. Longland and Charlie Clements 1 88 Use of an unusual food source by Rock Wrens (Troglodytidae) Polly K. Phillips and Allen E Sanborn 1 92 No. 3— July 1995 Articles Benthic community structure in two adjacent streams in Yellowstone National Park five years after the 1988 wildfires G. Wayne Minshall, Christopher T. Robinson, Todd V. Royer and Samuel R. Rushforth 1 93 Effects of browsing by native ungulates on the shrubs in big sagebrush communities in Yellow- stone National Park Francis J. Singer and Roy A. Renkin 201 Soft sediment benthic macroinvertebrate communities of the Green River at the Ouray National Wildlife Refuge, Uintah County, Utah Eric R. Wolz and Dennis K. Shiozawa 213 Alpine vascular flora of the Tushar Mountains, Utah Alan C. Taye 225 Ecology oiCeltis reticulata in Idaho Ann Marie DeBolt and Bruce McCune 237 Mimulus evanescens (Scrophulariaceae): a new annual species from the northern Great Basin Robert J. Meinke 249 Moiphological and host-symbiont studies of Trichodina teiiuifonnis and Apiosoma campanulatum infesting motded sculpin {Cottus bairdi) fi-om Provo River, Utali Ying Qi and Richard A. Heckmann 258 Effects of horse grazing in spring on sunival, recruitment, and winter injury damage of shrubs . . . Dennis D. Austin and Philip J. Urness 267 North American types of Oxytropis DC. (Leguminosae) at The Natural Histoiy Museum and Royal Botanic Garden, England, with nomenclatural comments and a new variety S. L. Welsh 271 Notes Saltation in snakes with a note on escape saltation in a Crotalus scutulatus Breck Bartholomew and Robert D. Nohavec 282 A trap for Blue Grouse Eric C. Pelren and John A. Crawford 284 ,386 Great Basin Naturalist [Volume 54 Book Review Mountains and plains: tlie ecology of Wyoniin^ landscapes Dennis H. Knif^ht Kimhall T. Harper 286 No. 4— October 1995 Articles classification of the riparian vegetation along a 6-km reach of the Animas River, southwestern Colorado Gillian M. Walford and Willian L. Baker 287 Additions to knowledge of Paleocene manmials from the North Horn Formation, central Utah .... Richard L. Cifelli, Nicholas J. Czaplewski, and Kenneth D. Rose 304 Springtime movements, roost use, and foraging activity of Townsend's big-eared bat {Plecotus toioisendii) in central Oregon David S. Dobkin, Ronald D. Gettinger, and M ichael G. Gerdes 3 1 5 Names and types in perennial AtripJcx Linnaeus (Chenopodiaceae) in North America selectively exclusive of Me.xico Stanley L. Welsh and Clifford Crompton 322 New records of ScoKtidae from Washington state Malcolm M. Furniss and James B. Johnson 335 Relative vulnerability to extiipation of montane breeding birds in tlie Great Basin ... J. Michael Reed 342 Grasshopper densities on grazed and ungrazed rangeland under drought conditions in southern Idaho Dennis J. Fielding and Merlyn A. Brusven 352 Plant novelties in Lepidiwn (Cruciferae) and Arte^nisia (Compositae) from the Uinta Basin, Utah . . . Stanley L. Welsh and Sherel Goodrich 359 Prey choices and foraging efficiency of recently fledged California Gulls at Mono Lake, California Chris S. Elphick and Margaret A. Rubega 363 Notes Hybridization between Bitfo woodhousii and Biifo pitnctatiis from the Grand Canyon region of Arizona Keith Malmos, Robert Reed, and Bryan Starrett 368 Reproduction in the banded sand snake, Chilonieniscus cinctits (Colubridae), from Arizona Stephen R. Goldberg 372 No acoustic benefit to subterranean calling in the cicada Okanogana poUiduIa Davis (Homoptera: Tibicinidae) Allen F Sanborn and Polly K. Phillips 374 Bool< Review Natural histoiy of the Colorado Plateau and Great Basin K. T. Harper, L. L. St. Clair K. H. Thome, and W. M. Hess B. C. Kondratieff 377 !i76 t5 6 INFORMATION FOR AUTHORS The Great Basin Naturalist welcomes previously unpublished manuscripts pertaining to the biologi- cal natural history of western North America. Preference will be given to concise manuscripts of up to 12,000 words. Simple species lists are dis- couraged. 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 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, 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 followdng 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 Naturalist for changes and refer to the CBE Style Manual, 5th edition (Council of Biology Editors, One Illinois Center, Suite 200, 111 East Wacker Drive, Chicago, IL 60601-4298 USA, phone: (312) 616-0800, fax: (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, appendices, 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 bv 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 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 number, and E-mail address • 3 copies of the manuscript and WordPerfect diskette • Confonnity with instructions • Photocopies of illustrations (ISSN 001 7-3614) GREAT BASIN NATURALIST Vol 55 no 4 October 1995 CONTENTS Articles classification of the riparian vegetation along a 6-km reach of the Animas River, southwestern Colorado Gillian M. Walford and Willian L. Baker 287 Additions to knowledge of Paleocene mammals from the North Horn Formation, central Utah Richard L. Cifelli, Nicholas J. Czaplewski, and Kenneth D. Rose 304 Springtime movements, roost use, and foraging activity of Townsend's big-eared bat {Plecotus townsendii) in central Oregon David S. Dobkin, Ronald D. Gettinger, and Michael G. Gerdes 315 Names and types in perennial Atriplex Linnaeus (Chenopodiaceae) in North America selectively exclusive of Mexico Stanley L. Welsh and Clifford Crompton 322 New records of Scolytidae from Washington state Malcolm M. Fumiss and James B. Johnson 335 Relative vulnerability to extirpation of montane breeding birds in the Great Basin J. Michael Reed 342 Grasshopper densities on grazed and ungrazed rangeland under drought condi- tions in southern Idaho Dennis J. Fielding and Merlyn A. Brusven 352 Plant novelties in Lepidium (Cruciferae) and Artemisia (Compositae) from the Uinta Basin, Utah Stanley L. Welsh and Sherel Goodrich 359 Prey choices and foraging efficiency of recently fledged California Gulls at Mono Lake, California Chris S. Elphick and Margaret A. Rubega 363 Notes Hybridization between Bufo woodhotisii and Bufo punctatus from the Grand Canyon region of Arizona Keith Malmos, Robert Reed, and Bryan Starrett 368 Reproduction in the banded sand snake, Chilomeniscus cinctus (Colubridae), from Arizona Stephen R. Goldberg 372 No acoustic benefit to subterranean calling in the cicada Okanagana paUidida Davis (Homoptera: Tibicinidae) Allen E Sanborn and Polly K. Phillips 374 Book Review Natural history of the Colorado Plateau and Great Basin K. T. Harper, L. L. St. Clair, K. H. Thome, and W. M. Hess B. C. Kondratieff 377 Index to Volume 55 379 3 2044 072 231 152