Vol, 37, No.~ November 1984 ARTICLES 483 Forage Response of a Mesquite-Buffalograss Community Following Range Reha- Bilitation by Donald J

TABLE OF CONTENTS: Vol, 37, No.~ November 1984 ARTICLES 483 Forage Response of a Mesquite-Buffalograss Community Following Range Reha- bilitation by Donald J. Bedunah and Ronald E. Sosebee 488 Low. Rates of Tebuthiuron for Control of Sand Shinnery Oak by V.E. Jones and R.D. Pettit 491 Short-term Vegetation Responses to Fire in the Upper Sonoran Desert by George H. Cave and Duncan T. Patten 4% Callie Bermudagrass Yield and Nutrient Uptake with Liquid and Soil N-P-K Fertilizers by Galen D. Mooso, Von D. Jolley, Sheldon D. Nelson, and Bruce L. Webb 501 Establishment of Diffuse and Spotted Knapweed from Seed on Disturbed Ground in British Columbia, Canada by L.D. Roze, B.D. Frazer, and A. McLean 503 Phenological Development and Water Relations in Plains Silver Sagebrush by Richard S. White and Pat 0. Currie 507 Germination ProtIles of Introduced Lovegrasses at Six Constant Temperatures by Martha H. Martin and Jerry R. Cox 509 Seed Pretreatments and Their Effects on Field Establishment of Spring-Seeded ’ Gardner Saltbush by R. James Ansley and Rollin H. Abernethy 514 Leaf Area, Nonstructural Carbohydrates, and Root Growth Characteristics of BlueGrama Seedlings by A.M. Wilson 517 Copper and Molybdenum Uptake by Forages Grown on Coal Mine Soils by Dennis R. Neuman and Frank F. Munshower 521 Natural Establishment of Aspen from Seed on a Phosphate Mine Dump by Bryan D. Williams and Robert s. Johnston 523 Variability of Infiltration within Large Runoff Plots on Rangelands Micheline Devaurs and Gerald F. Gifford 529 Evaluating Soil Water Models on Western Rangelands by Keith R. Cooley and David C. Robertson 534 Characteristics of Oak Mottes, Edwards Plateau, Texas by R.W. Knight, W.H. Blackburn, and L.B. Merrill 538 Soil, Vegetation, and Hydrologic Responses to Grazing Management at Fort Stanton, New Mexico by N. Dedjir Gamougoun, Roger P. Smith, M. Karl Wood, and Rex D. Pieper 542 Forage Preferences of Livestock in the Arid Lands of Northern Kenya by W.J. Lusigi, E.R. Nkurunziza, and S. Masheti 549 Cattle Distribution on Mountain Rangeland in Northeastern Oregon by R.L. Gillen, W.C. Krueger, and R.F. Miller 554 Dietary Selection and Nutrition of Spanish Goats as Influenced by Brush Man- _ 9, agement by Expedito A. Lopes and Jerry W. Stuth Published bimonthly-January, March, May, July, 560 Estimating Seasonal Diet Quality of Pronghom Antelope from Fecal Analysis by September, November B.H. Koerth, L.J. Krysl, B.F. Sowell, and F.C. Bryant Copyright 1984 by the Society for Range Manage- ment TECHNICAL NOTES 565 Technique to Separate Grazing Cattle into Groups for Feeding by J.F. Karn and R.L. Lorenz

INDIVIDUAL SUBSCRIPTION is by membership in BOOK REVIEWS the Society for Range Management. 567 The Genesis and Classification of Cold Soils by Samuel Rieger; Domestication, LIBRARY or other INSTITUTIONAL SUBSCRIP- Conservation and Use of Animal Resources Edited by L. Peel and D.E. Tribe. TIONS on a calendar year basis are $56.00 for the United States postpaid and $66.00 for other coun- 568 Index tries, postpaid. Payment from outside the United States should be remitted in US dollars by interna- 574 Table of Contents tional money order or draft on a New York bank. BUSINESSCORRESPONDENCE, concerning sub- scriptions, advertising, reprints, back issues, and related matters, should be addressed to the Manag- ing Editor, 2760 West Fifth Avenue, Denver, Colo. 80204. EDlTORIALCORRESPONDENCE,concerningmanu- scriptsorothereditorial matters, should beaddressed to the Editor, 2760 West Fifth Avenue, Denver, Colo. 80204. INSTRUCTIONS FOR AUTHORS appear on the inside back cover of each issue. A Style Manual is also available from the Society for Range Manage- mentattheaboveaddress@$1.25forsinglecopies; $1 .OOeach for 2 or more. THE JOURNAL OF RANGE MANAGEMENT (ISSN 0022-409X) is published six times yearly for $56.00 per year by the Society for Range Management, 2760 West Fifth Avenue, Denver, Colo. 80204. SECOND CLASS POSTAGE paid at Denver, Colo. Managlng Editor POSTMASTER: Return entire journal wlth address PETER V. JACKSON Ill change-RETURN POSTAGE GUARANTEED-to 2760 West Fifth Avenue Society for Range Management, 2760 West Fifth Denver, Cola. 80204 Avenue, Denver, Colo. 80204. Editor PATRICIA G. SMITH Society for Range Management 2760 West Fifth Avenue Denver, CO 80204 The Journal of Range Management serves as a forum for the presentahon and dtscus- Book Review Editor sion of facts, ideas, and philosophies pertain- RICHARD E. FRANCIS ing to the study, management, and use of Rocky Mountain Forest and Range Experiment Station rangelandsand their several resources. Accord- 2205 Columbia S.E. rngly. all material published herein IS signed Albuquerque, New Mexico 87106 and reflects the indrvrdual views of the authors and is not necessarily an official position of ASSOCIATE EDITORS the Society. Manuscripts from any source- E. TOM BARTLETT KIETH SEVERSON nonmembers as well as members-are wel- Dept. of Range Science Forest Hydrology come and will be grven every consideration Colorado State University Laboratory Fort Collins, CO 80523 Arizona State University by the editors Submissions need not be of a Tempe. AZ 85281 technical nature, but should be germane to GARY FRASIER the broad field of range management. Edrtor- 780 West Cool Drive DARREL UECKERT ial comment by an indivrdual’is also welcome Tucson, AZ 85704 Texas Agricultural Experiment Station and subject to acceptance by the editor. WIII 7887 N. Highway 87 be pubkhed as a “Viewpornt.” G. FRED GIFFORD San Angelo, TX 76901 Dept. of Range Wildlife, and Forestry University of Nevada BRUCE WELCH Reno, Nev. 89506 Shrub Science Laboratory 735 N 500 E W.K. LAUENROTH Provo, UT 84701 Department of Range Science Colorado State University LARRY M. WHITE Fort Collins, CO 80523 USDA ARS S. Plains Range Research Station LYMAN MCDONALD 200 18th St. Statistics Department Woodward, OK 73801 College of Commerce and Industry University of Wyoming KARL WOOD Laramie, WY 82071 Dept. of Animial and Range Sciences ROBERT MURRAY Box 3-l US Sheep Experiment Station Las Cruces, NM 88001 Dubois, ID 83423 JAMES YOUNG USDA ARS Renewable Resource Center Reno, NV 89512

Forage Response of a Mesquite-Buffalograss Community Following Range Rehabilitation

DONALD J. BEDUNAH AND RONALD E. SOSEBEE

Abstract The influence of different range rehabilitation methods on The soil series of the study area was a Sagerton clay loam which honey mesquite control, herbage production, and grazing capacity is in the fine mixed thermic family of Typic Paleustolls. The were evaluated on a depleted clay loam range site in west Texas. Sagerton series consists of deep, welldrained, moderately slowly Mesquite control by foliar application of 2,4,5-T + picloram, permeable soils that formed in calcareous clays, and loamy sedi- shredding, mechanical grubbing, mechanical grubbing and seeding ments on nearly level to gently sloping uplands. to kleingrass, and mechanical grubbing and vibratilling increased The study area was on a clay loam range site. Climax vegetation herbage production and grazing capacity. Shredding increased soil of this site is primarily a short grass community with a few mid- cover by adding plant litter, but significantly controlled mesquite grasses intermingled (USDA 1965). Climax decreasers include competition for only 2 years. Seeding to kleingrass resulted in a blue grama (Bouteloua gracilis (H.B.K.) Griffiths), side-oats productive stand with a high estimated grazing capacity. Foliar grama (B. curtipendufa (Michx.) Torr.), vine-mesquite (Panicum spraying doubled grass production compared to no treatment and obtusum (H.B.K.)), and western wheatgrass (Agropyron smithii resulted in 76% mesquite mortality 3 years after treatment. Defer- Rydb.). Important increasers of the climax vegetation include ment from grazing was important in increasing herbage produc- buffalograss, silver bluestem (Borhriochloa saccharoides (SW) tion during the study period; however, for maximum grazing Rydb.), tobosagrass (Hiliaria mutica (Buckl.) Nash), white tridens capacity both mesquite control and proper grazing would be (Tridens muricus (Torr.) Nash), and Texas wintergrass (Stipa leu- necessary. corricha Trin. & Rupr.). Common invaders included perennial three-awns (Aristida L. sp.), sand dropseed (Sporobolus cryptan- In much of west Texas, overgrazing by domestic livestock and drus (Torr.) Gray), hairy tridens (Erioneuron pilosum (Buckl.) increasing density of honey mesquite (Prosopis glandulosa Torr. Nash), Texas grama (Bouteloua rigidiseta (Steud.) Hitchc.), tum- var. glandulosa) have resulted in depleted ranges with low forage ble grass (Schedonnarduspaniculatus (Nutt.) Trel.), prickley pear production. Smith and Rechenthin ( 1964) considered mesquite the (Opuntia polyacantha Haw.), cholla (Opunria imbricata (Haw.) most common and widespread noxious plant in Texas. Mesquite Engelm.), mesquite, and lotebush (Ziziphus obtusifolia (T. & G.) competes with valuable range plants for water; thereby, reducing Gray) (USDA 1965). forage production and increasing the aridity of the site. Without At the initiation of the study, the site was in low fair range range improvements many of these areas will continue to decrease condition and was in a downward trend. Mesquite and buffalo- in productivity reducing the possibility of maintaining successful grass were the major overstory and understory dominants, respec- and long-term ranching operations. tively. Mesquite averaged 939 trees/ ha. The area historically had Jacoby et al. (1982) reported that the most dramatic forage been grazed by cattle year long. responses following brush control have occurred on arid to semiarid ranges where there was critical competition between brush and Methods forage plants for water. Studies on semiarid ranges in Arizona The study area was fenced in August, 1977, and protected from (Cable and Tschirley 1961) and Texas (Dahl et al. 1978, Jacoby et grazing by large herbivores for the duration of the study. Three al. 1982) have reported that grass production significantly increased rows of six 0.4-ha plots were located in a completely randomized following mesquite control by aerial application of herbicides. design with 3 replications/ treatment. The treatments. or types of However, few replicated experiments have been conducted on the vegetation manipulation, were: (1) shredding mesquite, (2) foliar influence of different mesquite control techniques on forage pro- spraying mesquite, (3) mechanically grubbing mesquite, (4) mechan- duction of deteriorated semiarid west Texas range sites. Therefore, ically grubbing between mesquite trees, (5) vibratilling, (6) seeding the purpose of this study was to evaluate changes in vegetation to kleingrass (Panicurn coloratum Walt.), and (7) a check or no following several brush control techniques on a deterior- treatment. ated range site with high mesquite density. All treatments had been applied by June 1, 1978, except for the vibratilling and seeding, which were not completed until May, Study Area 1979, because of problems in employing a contractor. A mesquite-buffalograss (Buchloe dacryloides Nutt. Engelm.) Treatments dominated area on the Post-Montgomery Estate Ranch located 7 Shredding km north of Post, Texas, (Garza County) was chosen for the study. Mesquite was top removed on May 18, 1978, using a rotary The area is a semiarid transition zone from the southern short grass shredder and farm-type tractor. Shredding was accomplished at a plains of the Llano Estacado to the Red Rolling Plains of Texas. relatively slow travel rate and no attempt was made to reshred large Average growing season is 216 days. High velocity winds are a debris. No other treatment was applied either simultaneously or critical factor in increasing evapotranspiration which averages subsequently to shredding. 264.5 cm/yr (USDA 1965). Foliar Spray Authors are graduate research assistant and professor Department of Range and Mesquite foliage was sprayed with a I:1 mixture of 2,4,5- Wildlife Management Texas Tech University, Lubbock 79409. Dr. Bedunah’s current Trichlorophenoxy-acetic acid (2,4,5-T) + 4amino-3,5,6-pico- address is assistant professor, Range Resource Management, School of Forestry, linic acid (picloram)at 0.6 kg a.i./ ha. The herbicide was applied on University of Montana, Missoula 59812. The authors would like to thank the Post-Monteomerv Estate Ranch and Mr. Tom May 31, 1978, using a John Bean sprayer equipped with a hand Copeland for their support during this study. - . sprayer. Individual mesquite trees were sprayed until the foliage This article is Contribution No. T-9-353 of the College of Agricultural Sciences. Texas Tech University. was completely wet. Spraying was delayed until soil temperatures Manuscript accepted February 29, 1984.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 463 surpassed the minimum threshold soil temperature of 24OC (Dahl the study site. Precipitation was measured with a tipping bucket et al. 1971). Soil temperatures were measured with standard labor- rain gauge equipped with an event recorder capable of detecting atory thermometers and averaged 25.YC at a 45-cm depth. changes in precipitation events every 5 min. Air temperature and Grubbing Trees relative humidity were measured and recorded with a Skyline Mesquite was removed mechanically by grubbing on May 30 hygrothermograph. Evaporation was measured from a standard and 3 I,1978 using a rear-mounted grubber on a farm-type tractor. Weather Bureau Class A free surface pan. The trees were removed below the basal crown, leaving a pit where Grazing Capacity the tree was removed. Grazing capacity was estimated from herbage production data, Grubbing between Trees similarly to methods used by McDaniel et al. (1982). Grazing Grubbing between trees was used as a treatment to evaluate if capacity was determined from the proper use factor (PUF) and the herbage response was a result of the method of removal (grub- production according to the following equation: bing and possibly impounding water) or from the removal of mesquite competition. Grubbing between trees was done on June PUF X Species dry weight/ha = ha/ AUY 1, 1978, using the same equipment and procedures as used for 4967 kg grubbing the trees (including size of pit). An attempt was also made to simulate the number of pits per plot created by the grubbing tree Desirable plants, decreasers and the more palatable increasers, treatment. were assigned a 50% PUF (Table 1). Intermediate plants, increas- Vibratill ers, and palatable invaders, were given a PUF between 30 and 40%. Mesquite was removed mechanically by grubbing and raking. A vibratiller (large chisel with an oscillating unit, driven by a power Table 1. Palatability rating and proper use factor (PUP) of plants used for takeoff that causes the tynes to fracture subsurface soil simultane- deteminging grazing capacity. ously with ripping) with the tynes set for a 76-cm row spacing and a 60-cm depth was pulled across the prevailing slope. The vibratill Proper use factor disturbed the soil surface and fractured the claypan, but left much Palatability rating Plant species or grouping (%) of the vegetation intact. High Bouteloua gracilis 50 Seeded Bothriochloa saccharoides 50 Mesquite was removed mechanically by grubbing. The plots Digitaria caltyornica 50 were then plowed with a vibratiller, disked, and kleingrass was Panicum coloralum 50 drilled-seeded at 1.4 kg/ha (PLS) on May 10, 1979. Kleingrass- Panicum obtusum 50 seeded plots were never fertilized nor irrigated. Setaria macrostachya 50 Perennial forbs 45 Mesquite Mortality Mesquite mortality (%) was measured by counting living trees in Moderate Buchloe dactyloides 40 each plot before treatment and 3-years post-treatment. Mesquite Chloris sp. 30 trees showing any resprouting 3 years post-treatment were consi- L.eptoloma cognatum 30 30 dered to be living. Mesquite mortality was considered to be impor- Sporobolus cryptandrus tant in assessing the potential longevity of the treatment. Low Aristida sp. 20 Hilaria mutica 20 Standing Crop and Soil Cover Muhlenbergia arenicola 20 Herbage data (standing crop) were collected after each growing Panicum hallii 20 season (approximately October 1). Herbage was determined by Erioneuron pilosum 20 clipping 21 randomly located 0.45mr quadrats/treatment at I-cm Xanthocephalum sp. 0 stubble height. Herbage was separated by grass species, broom- Annual forbs 0 weed species (Xanthocephalum dracunculoides (D.C.) Shinners Annual grasses 0 and X. sarothrae (Pursh) Shinners), or by grouping all other forbs. Woody, herbaceous, and standing litter were also collected for Invader plants were assigned a PUF of 20 to 30%. Annual and each quadrat after removing the current year’s growth. The her- perennial broomweed and annual grasses were not included in bage was oven dried at 50°C for at least 7 days and then weighed. grazing capacity determinations. Intake for an animal unit (AU) Weights were converted to kilograms of oven-dried material per was considered to be 13.6 kg/day (Bell 1973). hectare. Herbage was classified by 3 groups. The first group was total Statistical Analysis grass production, which was a sum of standing crop (kg/ha) of The Statistical Analysis System (SAS) package programs were threeawns (Aristida fongiseta Steud. and A. purpurea Nutt.), buf- used (Helwig and Council 1979). Analysis of variance was used to falograss, sand dropseed, blue grama, hairy tridens, windmill test for differences in treatment means at the 0.05 level of probabil- grasses (Chloris cucullata Bisch. and C. verticillata Nutt.), sand ity. If the analysis of variance tests showed a significant treatment muhly (Muhlenbergia arenicola Buckl.), feather fingergrass (Chlo- effect, means were separated using Duncan’s new multiple range ris virgata Swartz), silver bluestem, Arizona cottontop (Digitaria test (Steel and Torrie 1960). californica (Benth.) Henr.), vine-mesquite, plains bristlegrass (Setariamacrostachya H.B.K.), tobosagrass,and kleingrass. Total Results and Discussion forb production or the sum production of broomweeds and other Mesquite Control forbs constituted the second group. The third group was the sum of At the initiation of the study, mesquite canopy cover and density standing crop (total production) of grasses and forbs. averaged 22% and 939 trees/ ha, respectively. All mesquite control Ground cover was estimated ocularly for each quadrat by spe- techniques had the immediate effect of eliminating live mesquite cies (foliar cover) and for litter before clipping. Total ground cover canopy cover. was determined as the sum of litter and the canopy cover of living Foliar application with 2,4,5-T + picloram resulted in 78% root herbaceous vegetation. kill 3 years post-treatment. However, the reduction in mesquite Climatological Data canopy cover and transpiration surface was estimated to be 98%. A climatological station for measurement of precipitation, air Mechanically grubbing mesquite resulted in top removal of all temperature, relative humidity, and evaporation was located on mesquite trees and 90% root-kill. Mesquite grubbing followed by

484 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 r 1978 1979 1980

Fig. 1. Influence of range rehabiliration treatments on roraiforb, totalgrass and rotalproducrion (sum of the standing crop of totalforbs and totalgrass) for 1978, 1979 and 1980. Means within the same year with a similar superscript are not significantly different (P

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 485 Grazing Capacity significant species composition change for nonseeded areas during The estimated grazing capacity was significantly increased by the study, except for an increase in annual broomweed in 1979. The all range rehabilitation practices where mesquite was controlled high production of annual broomweed in 1979, compared to 1978 (Table 2). Seeding to kleingrass resulted in the greatest estimated and 1980, was probably caused by the higher amount of precipitation during June, July, and August (Fig. 3). Table 2. Influence of range rehabilitation treatments on grazing capacity (ha/AU/yr) during 1978,1979 and 1980’.

Year Rehabilitation treatment 1978 I979 I980

Foliar spray 27.2b (x)2 10.9bc(y) 7.6bc(z) Shred 20.3b (x) 10.3cd(y) 10.2ab(y) Check 53.9a (x) 20.8a (y) 15.0a (z) Grub between trees 30.0ab(x) 15.2ab(y) I 1.9ab(z) Grub trees 19.4b (x) 13.4c (xy) 9.0bc(y) Kleingrass -3 5.7d (x) 5.4c (x) Vibratill 12.0bc(x) 8.3bc(y)

‘It was assumed that 9934 kg of total forage (dry weight) are required to support an animal unit (AU) per year. ‘Means followed by a similar letter within each column or in parenthesis wthm each row are not significantly different at the 0.05 level of probability. ‘No data were available in 1978 for the vibratill or kleingrass treatments. grazing capacity for 1979. However, in 1980, grazing capacities for the foliar spray and vibratill treatments were similar to the seeded treatment. Mean estimated grazing capacity across treatments established in 1978 showed an increase in grazing capacity for each year. In 1978 mean grazing capacity was estimated at 28 ha/ AUY compared to 13 ha/AUY in 1979 and IO ha/AUY in 1980. Grazing capacity is a function of the amount and kind of forage available. Most of the increase in grazing capacity was a result of an increase in grass production each year (Fig. 2). Grazing capacity

m FORES ‘;;i c 0 GRASS \ F-_l TOTAL ul 1600 J

n P 1978 1979 1980

Fig. 3. Precipilarion andfreepan evaporation for 1978, 1979 and 1980.

Much of the increase in grass production for this site was a result of mesquite control but grass showed an increase even for the check b treatment. McDaniel et al. (1982) stated that a dormant season grazing regime following honey mesquite control should be carried Y out for one or more years, depending upon the range condition of 0 treated pastures and the management goals. Therefore, we believe that some of the increased grass production was from an increase in vigor of the perennial plants associated with the rest from grazing a and an increase in plant cover for protection of the soil surface. The

0 range trend was up, but 3 years was not long enough to detect a A_ change in range condition. 1978 1979

Fig. 2. Mean total forb, rota1 grass, and roral production (sum of the Summary and Conclusions standing crop of toralforbs and [oral grass) for the foliar spray, shred, Mesquite removal by all range rehabilitation methods resulted check. and grub treatments combinedfor 1978. 1979 and 1980. Means in increased herbage production and grazing capacity. Each with a similar superscript are not significantly different (lYO.05). method influenced site conditions in a particular manner. The best was more related to total grass production than total herbage range rehabilitation method for range sites in west Texas will production because of the high production of annual broomweed depend on initial site conditions, management concerns and during 1979. Buffalograss and sand dropseed were the most impor- expected economic returns. tant species, averaging 38% and 19%, respectively, of the total For very depleted sites, with few valuable forage plants, seeding herbage production for nonseeded treatments. Brock et al. (1978) improved grasses would result in a rapid increase in grazing capac- and McDaniel et al. (1982) reported greater production of decreas- ity. Mechanical grubbing alone or followed by vibratilling, decreased ers within the mesquite canopy zone. For this site, decreaser species surface runoff and would result in long term control of mesquite. comprised less than 2% of the total herbage production. Few Shredding mesquite resulted in only a short term (2 years) con- decreasers were present even under mesquite trees. There was no trol and increase in herbage production. Shredding influenced site

486 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 conditions by increasing plant litter, returning nutrients to the soil, Dahl, B.E., R.E. Sosebee, J.P. Goen, and C.S. Brumley. 1978. Will mes- and increasing grass production the year of the treatment. Shred- quite control with 2,4,5-T enhance grass production. J. Range Manage. ding mesquite, in combination with another treatment, may offer a 31:129-131. valuable range rehabilitation alternative for sites with poor her- Dahl, B.E., R.W. Wadley, M.R. George, and J.L. Talbot. 1971. Influence of site on mesquite mortality from 2,4,5-T. J. Range Manage. 24:210-215. baceous plant cover, but still having some valuable forage plants. Helwig, J.T., and K.A. Councils (eds.). 1978. SAS User’s Guide. SAS Foliar spraying with 2,4,5-T + picloram was the most feasible Institute, Inc., Raleigh, N. C. control method for this site. The foliar spray resulted in satisfac- Jacoby, P.W., C.H. Meadors, M.A. Foster, and F.S. Hartmann. 1982. tory mesquite control, provided high grazing capacity and cost Honey mesquite control and forage response in Crane County, Texas. J. would be much lower than mechanical rehabilitation methods. Range Manage. 35:424-426. For deteriorated range sites a deferment from grazing would be Klett, W.A. 1969. An evaluation of equipment used to treat depleted important to improve the vigor of the forage plants. Nevertheless, shortnrass ranges. M.S. Thesis. Texas Tech Univ I.ubbock. for maintenance of maximum grazing capacity both mesquite Langley, B.C., aid C.E. Fisher. 1939. Some effects of contour listing on control and proper grazing would be necessary. native grass pasture. J. Amer. Agron. 3 I: I I. McDaniel, K.C., J.H. Brock, and R.H. Haas. 1982. Changes in vegetation Literature Cited and grazing capacity following honey mesquite control. J. Range Man- age. 35:551-557. Bedunah, D.J. 1982. Influence of some vegetation manipulation practices Scifres, C.J., and G.O. Hoffman. 1974. Mesquite: growth and develop- on the biohydrological state of a depleted deep hardland range site. ment, management, economics, control, uses-a brief. Texas Agr. Exp. Ph.D. Diss., Texas Tech Univ., Lubbock. Sta. Res. Monogr. I. Texas A&M Univ., College Station. Bell, H.M. 1973. Rangeland management for livestock production. Uni- Smith, H.N., and C.A. Rechentbin. 1964. Grassland restoration: I. The versity of Oklahoma Press, Norman. Texas brush problem. USDA , Soil Conservation Service, Temple, Brock, J.H., R.H. Haas, and J.C. Shaver. 1978. Zonation of herbaceous Texas. vegetation associated with honey mesquite in northcentral Texas. P. Steel, R.G.D., and J.H. Torrie. 1960. Principles and Procedures of Statis- 187-189. In: Proc., 1st Internat. Range Cong., Denver, Colo. tics. McGraw-Hill Book Co., New York. Cable, D.R., and F.H. Tschirley. 1961. Responses of native and introduced U.S. Department of Agriculture. 1965. Texas Soil Survey, Garza County. grasses following aerial spraying of velvet mesquite in southern Arizona. U.S. Government Printing Office, Washington, D.C. J. Range Manage. 14: 155-l 59.

Director and Associate Dean POSITION AVAILABLE The director and associate dean is responsible for coordinat- Position: Research Associate, Texas Agricultural Experiment ing the mission of the New Mexico Agricultural Experiment Station. Station and acceptsother appropriately delegated responsibilities assigned by the dean of the College of Agriculture and Location: Texas Experimental Ranch, Throckmorton County. Home Economics. Applicants should have earned doctorate and administrative experience is desirable. Candidates should Minimum Qualifications: MS. degree in Range Science or have substantial experience in research plus knowledge of the closely related field. Research experience in range nutrition teaching, research and extension organization unique to land- preferred. grant universities. Application deadline is Januaryl,1985. Send resume plus names, addresses and phone numbers of five (5) Salary: Competitive with other States and consistent with references to (;.M. Southward, Experimental Statistics, BOX experience. 3730, New Mexico State University, Las Cruces, NM 88003 (505-646-2936). New Mexico State University is an equal oppor- Closing Date for Applications: April 15 or until position is tunity/affirmative action employer. filled.

Duties and Responsibilities: Individual will supervise and coordinate routine ranching operations at the 2900 ha Texas Experimental Ranch. Other responsibilities will include the maintenance of detailed cow/calf performance records, coor- dination of all field research projects, and direct supervision of !23 ranch foreman, ranch office secretary, and part-time workers. Significant opportunity for individual to maintain an active research program.

To Apply: Send resume, official transcripts and three letters of recommendation to: Dr. RodHeitschmidt, TexasA&M Research & Extension Center, P.O. Box 7658, Vernon, TX 76384.

An Equal Opportunity and Affirmative Action Employer

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 487 Low Rates of Tebuthiuron for Control of Sand Shinnery Oak

V.E. JONES AND R.D. PETTIT

Abstract Tebuthiuron [N-(5-(l,l-dimethylethyl)-l,3,4-thiadiazol-2-yl)N, 3 to 5 years (Scifres 1972). N’-dimethylurea] pellets (20% ai) were broadcast at 0.2 kg incre- Forage increased up to 387% where pelleted herbicides were ments to 1.0 kg/ha onto a sand shinnery oak (Quercus havardii) used to control oak in the Texas Rolling Plains (Jones et al. 1978). community in west Texas (33” 23’52”N and 102°46’38”W). Treat- On the more arid Southern High Plains, grass yields tripled where ments 50.4 kg/ha reduced oak canopy 98% and standing crop at oak yield was reduced 25% (Pettit 1979). Greater oak control gave least 90%. Grass yield was unaffected by herbicide treatments a 4-to 9-fold increase in grass. during the first year. Thereafter, yield on treated areas increased Data reported here reflect the effectiveness of low rates of tebu- from 420 to 690 kg/ha as contrasted to 140 kg/ha on the control. thiuron for control of sand shinnery oak. The associated yield of Where oak was untreated, grasses became quiescent, due to grass is also addressed. drought, up to 6 weeks earlier than on treated areas. Materials and Methods Sand shinnery oak on 1.2 million ha of (Quercus havardii)grows The study was on the Southern High Plains of Texas (33O 23’ 52” rangeland in Texas, two-thirds having a canopy of 20% or more N and 102O 46’ 38” W), about 25 km S and 2.5 km E of Lehman, (Deering and Pettit 1971). This plant infests an additional 0.4 Texas. Soils of the study area were Brownfield, Circleback (tenta- million ha in Oklahoma (Mcllvain 1956) and 1.1million ha in New tive series), Patricia and Tivoli fine sands. Excluding the Tivoli, Mexico (Garrison and McDaniel 1982). Along the precipitation these soils differ primarily in the depth of fine sand over a sandy gradient across this species’ range, density and stature of oak clay loam subsoil. The first 3 soils are classed as Alfisols (Paleus- increases with precipitation and locally with depth of surface talfs) while the Tivoli is an Entisol (Ustipsamment). sands. A warm-temperate, semiarid climate typifies the area, though This oak is toxic to livestock and competes with better forages. rapid temperature fluctuations, especially during the winter, are Presence of it increases livestock production costs because of (1) common. Precipitation averages 4 1cm and is variable. About 80% increased death loss, (2) reduced conception rates and weight gains of the precipitation is received from May through October. Fre- caused by chronic poisoning, (3) required supplemental or alterna- quent winds, high temperatures, and low relative humidity enhance tive feeds during the most toxic period, and (4) increased manage- evaporation. The growing season averages 200 days. rial costs. A motor-driven “Cyclone” seeder behind a tractor was used to Repeated applications of foliar herbicides or land conversion by broadcast tebuthiuron pellets (20% ai) onto 2-ha plots (100 by 200 deep-plowing have accounted for most control efforts. Risk asso- m) at 0.2,0.4, 0.6, 0.8, and 1.0 kg/ha during May 1978. A func- ciated with these techniques and increases in their cost have stimu- tional swath of 16.7 m, offset slightly to the right, was used so that lated research to develop better controls. Use of root-absorbed overlap between swaths would help equalize ,herbicide distribu- herbicides has been suggested as a means of improving the reliabil- tion. Treatments plus a control were assigned in a completely ity, efficiency, and safety of treatment. These herbicides, applied as randomized design with 3 replicates. Buffers of 20 m were left to pellets or granules, reduce the risk of off-target movement; and prevent overlap during application. since they are root absorbed, plant condition at application is not Four transects, perpendicular to treatment swaths, were chosen believed as critical as for foliar applied compounds. at random across each plot. Along each, IO random points were Pettit (1975) and Jones et al. (1978) reported that oak could be permanently marked. Density and canopy cover by species were controlled with pelleted tebuthiuron [N-(5-( I,l-dimethylethyl)-l,3, measured within square 0.25-m* quadrats at each reference point 4-thiadiazol-2-yl)-N,N’-dimethylurea], a substituted urea. Tebu- during the spring, summer, and fall from 1978 through 1980. thiuron at 1.0 kg/ ha or more killed all oak (Pettit 1979) while 2 Standing crop was estimated at 10 randomly selected reference years after treatment with 0.6-kg/ha stem density decreased over points in each plot during the spring, summer, and fall from May 80% (Jones et al. 1978). Other oaks are killed or suppressed by this 1978 through June 1981. Square 0.5-m2 quadrats were placed at herbicide (Meyer et al. 1978, Meyer and Bovey 1980, Scifres et al. preselected distance and direction from each point to prevent 1981). resampling the same area. Herbage was clipped at a 1 cm height Oak-dominated rangelands in Oklahoma produce twice as much and separated into current year’s growth by major species or group grass following single, foliar applications of phenoxy herbicides (i.e., forbs, shrubs, other grasses) before bagging. Samples were (Mcllvain and Armstrong 1959, 1963; Greer et al. 1968). Consecu- dried at 44 C to a constant water content and weighed. tive yearly treatments provide better oak control and result in a Density and cover data were subjected to either square-root or tripling of grass production. Production, however, soon declines arcsin transformation before being analyzed (Steel and Torrie and stabilizes at twice that of untreated areas (McIlvain and Arm- 1960). Data were analyzed using standard analysis of variance or strong 1963). In Texas, grass production increases up to 6-fold the analysis of covariation techniques. Means separation (X0.05) second year after a single foliar herbicide treatment; but since oak utilized either Duncan’s multiple range test or the least significant regrowth is not suppressed, retreatment is deemed necessary within difference (LSD) test (Steel and Torrie 1960). The authors, at the time of research, were graduate research assistant and associate professor, Department of Range and Wildlife Management, Texas Tech University, Results and Discussion Lubbock 79409. This article is Texas Tech University College of Agricultural Sciences Publication Oak Control No._. T-Q-776. _ _._. Manuscript accepted March 19, 1984. In 1978, oak defoliated cyclically with defoliation and regrowth

488 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 associated with precipitation events. Little regrowth occurred after treated areas yields were 600 to 1,500 kg/ ha greater in 1979. leaf drop unless either the soils were wet or an appreciable amount Early June 1980 grass yields in treated areas were almost 3.5 of precipitation was received. New leaves then developed rapidly, times (417 vs 120 kg/ha) greater than in the control and nearly 1.2 but before they were fully expanded, phytotoxic symptoms times greater than at the same period in 1979. Later grass growth appeared. Maximum herbicidal effects on oak occurred when soils was suppressed by the drought that prevailed throughout the were wet and when climatic conditions favored high transpiration summer of 1980. Fail rains came too late to stimulate additional rates. Plant injury developed more slowly in swath overlap areas growth. and in deeper sands. Drought-induced quiescence of grasses during 1980 occurred up Pretreatment canopy cover of oak, through only 17%, accounted to 6 weeks earlier in control than in treated areas. From early June for 60% of the total canopy cover. Oak frequency was 93% with a through July, grass yield in untreated areas increased only 5 kg/ ha density of 12,000 old and 1,600 new (sprouts) shoots/ ha. Standing as compared to increases of 34, 170, 193, 130, and 193 kg/ha in crop was likewise dominated by oak. On 6 June 1980 standing crop each successively higher herbicide rate. The reduction in competi- (1,309 kg/ ha) in untreated areas was 78% oak and 10,9, and 2% for tive use of available soil water by oak diminished the effect of other shrubs, grasses, and forbs, respectively. drought on herbage species in treated areas. By May 1979 density and canopy cover of old oak shoots were Management Concerns reduced in treated areas. In 1980 the mean density of old shoots in We have observed oak ranges treated with tebuthiuron since the 0.2-kg/ ha and LO.Ckg/ ha treatments was reduced 8 1and 96% 1974. Perennial brush species typically respond to treatment by with a reduction in canopy of 84 and 98%, respectively. undergoing cyclic defoliations. Up to 3 years may be required As oak topgrowth died, rhizome buds broke dormancy; by fall before control stabilizes. No reinfestation has been observed on 1978 the density of new shoots was greater than in the control. areas where treatment provided total oak kill. However, a few sand Herbicide induced mortality reversed this trend by mid 1979. There- sagebrush (Artemisia filifolia) have reestablished, and may be a after the density and canopy of new shoots continued to be lower in management problem in the future. Soapweed (Yucca angustifo- areas receiving at least 0.4 kg/ ha of herbicide. lia) was not killed by tebuthiuron to 1.0 kg/ha and also may Current year’s standing crop of oak was collected as an alterna- become a problem. tive assessment of control. From July 1979 through June 1981 On the area studied, tebuthiuron did not kill all the oak. Grow- tebuthiuron treatments a.4 kg/ ha reduced oak standing crop ing on Tivoli soils, it defoliated repeatedly, yet lived. Variability in 90% or more (Table 1). Oak yield was also reduced by the 0.2 kg/ ha sand depth, even on level topgraphy, caused a mottling effect of treatment but due to unequal distribution of herbicide, control in total oak kill. However, the patches of oak that remain may benefit individual plots ranged from negligible to near total in a pattern of wildlife habitat, especially for lesser prairie chickens (Tympanu- alternating bands. thus pallidicinctus). This oak produces primarily short shoots on which leaf expan- We were able to convert a sand shinnery oak ecosystem to a sion is rapid. When not damaged by frost or insects, leaves are fully mid-grass prairie; but this prairie has developed on easilyerodable expanded by late May. During the summer drought in 1980, soils. More than 90,000 ha of sand shinnery oak have been treated untreated oak shed 36% of its leaves as compared to 24% in 1978. with tebuthiuron. The resulting prairie may require different man- No effects of the drought were noted in 1979. Leafout in 1981 was agement strategies than the native oak-grass type. normal and foliage was comparable to that of the previous years. Literature Cited Grass Response After removing treatment effect by covariate analysis, grass Deering, D., and R. Pettit. 1971. Sand shinnery oak acreage survey. Res. yield was rarely influenced by the pretreatment canopy or density Highlights, Texas Tech Univ., Lubbock. 2: 14. of grass. Thus all analyses of grass yields use data unadjusted for Garrison, G.I., and K.C. McDaniel. 1982. New Mexico brush inventory. pretreatment population variables. New Mexico Dep. of Agr. Rep. No. I. Mean grass yields throughout the total study period were greater Greer, H.A.I., E.H. McIlvain, and C.G. Armstrong. 1968. Controlling shinnery oak in western Oklahoma. Oklahoma State Univ. Ext. Facts in treated than untreated areas (Figure l-a). Maximum yield of 571 No. 2765. kg/ ha (nearly 4 times that of the control) occurred in the 0.8 kg/ ha Jones, V.E., C.H. Meadors, and P.W. Jacoby. 1978. Pelleted herbicides for treatment. Yield in the 0.2 kg/ha treatment was 2.5 times greater control of sand shinnery oak (Quercus havardii). Thirty First Meeting than in the control. Sot. Range Manage., Proc. 3 159. (Abstr.). In 1978, grass yield (Figure I-b) did not differ among treatments. Mcllvain, E.H. 1956. Shinnery oak can be controlled. Southern Weed Sci. In more recent samples (Figures l-c, d, e), yields were consistently Sot., Proc. 9:95-98. greater in treated areas. By late May 1979 grass yields on treated McIlvain, E.H.,and C.G. Armstrong. 1959. Shinnery oak control produces areas had already approached or exceeded the maximum yields more grass. Southern Weed Sci. Sot., Proc. 12: I34- 137. recorded during the previous year. Fall 1978 and 1979 grass yields Mcllvain, E.H., and C.G. Armstrong. 1963. Progress in shinnery oak and sand sage control at Woodward. U.S. Southern Great Plains Field Sta. on control areas were comparable (239 vs. 23 1 kg/ ha) while on Prog. Rep. 630 I.

Table 1. Mean yield (kg/ha) of sand shinnery oak at study site in Co&ran County, Texas, during 1978,1979,1980 and 1981 after tebuthiuron treatment in May 1978.

Treatment 1978 1979 1980 1981 (kg/ha) Spring Summer Fall Spring Summer Fall Spring Summer Fall Spring

0.0 604 a 1 1041 a 906 a 1027 a 781 a 655 a II49 a 0.2 323 b z2 458 _ 186b 166 b 237 b 293 b 136b 346 b 0.4 462 ab 53 c 41 c 51 c 78 c 23 bc 46 c 0.6 436 ab 25 c Ic 39 c 24 c I8 bc oc 0.8 350 b - 2c oc l5c oc IC oc I.0 332 b oc oc 2c oc oc oc ‘Means followed by a similar letter within the same column do not differ at the 0.05 level of probability. ‘Data not available.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 469 0.6 0.6 1.0

(k6hl

4 1978 (b)

i 0.0 QL 04 0.6 0.1 1.0

Tr rotm*fll I kg/ha)

c2 04 06 06 I3 00 a2 0’6 db o:a I.0

Trcotmem (kq/bo Treolment ( kg/ha 1

Fig. 1. Mean yield ofgrass both acrossdates (a)and bysamplingdote (b-e)ot study site in Cochron County, Tex., following treatment with tebuthiuron in May 1978. Means on the some sampling dote designed with the same letter are similar (X0.05).

Meyer, R.E., and R.W. Bovey. 1980. Control of live oak (Quercus virgini- Pettit, R.D. 1979. Effects of picloram and tebuthiuron pellets on sand ono) and understory vegetation with soil-applied herbicides. Weed Sci. shinnery oak communities. J. Range Manage. 32:196-200. 28:51-58. Scifres, C.J. 1972. Herbicide interactions in control of sand shinnery oak. J. Meyer, R.E., B.W. Bovey,end J.R. Bauer. 1978. Control of an oak (Quer- Range Manage. 25:386-389. cus) complex with herbicide granules. Weed Sci. 26444-453. Scifres,C.J., J.W.Stuth,andR.W. Bovey. 1981.Controlofoaks(Qu~cus Pettit, R. 1975. Comparative effects of two pelleted herbicides on a shin oak spp.) and associated woody species on rangeland with tebuthiuron. community. Res. Highlights, Texas Tech Univ., Lubbock. 6~43. Weed Sci. 29:270-215. Steel, R.G.D., and J.H. Torrie. 1960. Principles and procedures of statistics. McGraw-Hill Book Co., Inc., New York.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Short-term Vegetation Responses to Fire in the Upper Sonoran Desert

GEORGE H. CAVE AND DUNCAN T. PATTEN

Abstract Annual and perennial plant vegetation was sampled following a (Wright 1980), the western Colorado Desert (O’Leary and Minnich controlled burn (1981) and a wildflre (1980) in the Upper Sonoran 1981), the Chihuahuan Desert (Ahlstrand 1982), and the Great Desert near Phoenix, Ariz. Perennial plant composition 1 year Basin Desert (Rogers 1980). All these efforts reflect contemporary after controlled burning included 32% shoot survivors, 30% interests in re-evaluating and contributing to knowledge of fire sprouters, and 38% seeders, mostly brittle bush (Enceliafarinosd). ecology in North American deserts. To the extent that large por- Several invader species, stickweed (Stephanomeria exigua) and tions of the Upper Sonoran Desert are used as rangeland, the four o’clock (Mirabilis bigelovii) were important seeders, indicat- present study is of particular interest because of the potential ing that there may be postfire successional communities in the stimulatory effect ftre has on rangeland productivity. Upper Sonoran Desert. Most cacti were fire killed or died eventu- This study was designed to examine short-term effects of fire on ally from fire damage. Total annual plant density decreased (69%) both annual and perennial plant communities in the Upper Sono- while biomass increased significantly (131%) on burned areas. Red ran Desert through the use of controlled burning and study of an brome (Bromus rubens) was essentially eliminated 1 year after fire adjacent wildfire area. Specific objectives were: (1) to characterize while schismus (Schismus arabicus) and Indian wheat (Plantago l- and 2-year postfire, herbaceous plant communities (annuals) spp.) increased in both density and biomass. Fire appears to with respect to changes in density and biomass in both open/shrub enhance rangeland productivity in the Upper Sonoran Desert. (small shrubs plus interspaces) and shaded (small trees) microhabitats; and (2) to examine the density and survival/ recovery strate- Deserts of the southwestern United States have been increas- gies of l- and 2-year postfire tree, shrub, and cactus plant ingly impacted by man and as a result may be rapidly deteriorating. communities. One such impact is an increase in fire occurrence on desert or semidesert recreation and rangeland areas. Desert fire ecology Methods research, however, is lacking compared to available data from The study site was located in Bulldog Canyon, a desert canyon other ecosystems (Wells et al. 1979 and Lotan et al. 1981). This near Phoenix, Ariz., in the Tonto National Forest at 33O15 ’N and deficiency may be attributed to Shreve’s (1925 and 1951) state- I1 l”33’W with an elevation of 450 m. Three fire treatment sites ments characterizing the desert responses to fire disturbance as a were studied: (1) a wildfire area in which 84 ha burned on 26 May direct and relatively rapid recovery back to the climax community. 1980; (2) I unburned hectare used for the controlled burn site and Muller (1940) and Whittaker (1975) have presented similar hypo- located adjacent to the wildfire site; and, (3) another adjacent, theses for general disturbances in the desert. In addition, Humph- unburned hectare selected as a no fire control site. The controlled rey (1963 and 1974) stated that fires have never been a factor of burn site was burned on I2 June 1981 by fire crews from Tonto much importance and that their occurrence in the Upper Sonoran National Forest. Desert is rare. All of these statements have undoubtedly discour- Vegetation in the canyon was typical of the Upper Sonoran aged researchers from studying desert fires. Fire can, however, Desert in central Arizona and is characterized by the palo verde- occur relatively frequently in the Upper Sonoran Desert during dry cactus (mostly Opuntid and Carnegiea gigunreu)-shrub (mostly seasons that follow moist winters (USDA 1980). Ambrosia deltoideu) association (Shreve 1951). These perennial Fires in southern Arizona desert grassland near Tucson, where plants occupy about one-third of total ground cover. Perennial fires can occur more frequently than in most desert areas, have grasses are rare in this portion of the Upland Desert, but herbace- been studied by Humphrey (1949), Humphrey and Everson (195 I), ous annual forbs and grasses are abundant after winter and heavy Reynolds and Bohning (1956), Humphrey (1963), Cable (1967), summer rains. White (1969) and Martin (1983). These studies reported vegeta- Precipitation data were from Stewart Mountain Weather Sta- tion responses to fire and investigated the use of fire as a tool for tion operated by the National Oceanic and Atmospheric Adminis- controlling undesirable species on grazing land. tration and located approximately 5 km (33’34’N and 11 l”32’W) Fire ecology in the Upper Sonoran Desert of central Arizona has from the study site. only been recently studied (Whysong and Heisler 1978, Rogers and Steele 1980, McLaughlin and Bowers 1982, Patten and Cave 1984). Herbaceous Plants This desert is floristically different from the southern Arizona Herbaceous plant data were collected during 2-5 April 198 1and desert grasslands, and fire recovery data are not comparable. Other 13-17 March 1982, during each year’s peak annual plant growth recent studies have been conducted in southern Arizona semidesert period. Plants were sampled on the no fire, controlled burn, and Authorsaregraduateassistant, Department of Botany/ Microbiology,anddirector wildfire sites using randomly located 20 X 20 cm plots within both of The Center for Environmental Studies and professor of botany, Arizona State shaded and open/shrub areas. This segregation was designed to University, Tempe 85287. The authors are indebted to the staff at Tonto National Forest, in particular to permit separate comparisons for annual plant species that grow James Kimble and James Brown. In addition, W.D. Clark, M. Knox, S. Link, R. primarily in shaded microhabitats (Tiedemann et al. 1971 and Maw. D. Robinson. G. Ruffner. T. Thomas. S. Workman. and esoeciallv Stan Smith Patten 1978), and to distinguish fire recovery effects in shaded de&e special recognition for their various contributionsand a&stan& We would also like to thank Jack Dieterich and Leonard DeBano from the U.S. Forest Service. areas where annual plant growth can frequently be much greater Rocky Mountain Forest and Range Experiment Station, Arizona State University; than in partially shaded or unshaded areas created by low shrubs Tempe. Manuscript accepted March 7, 1984. ‘Nomenclature follows Kearney and Peebles (1960) and Lchr (1978).

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 491 and open interspaces. Twenty plots were located in each area and above-ground biomass were calculated, the latter by harvest- except for open/ shrub areas on the no fire site which had 14 plots ing above-ground growth and drying for 48 hrs at 60°C before during 1981 sampling. For each species within the plots, density weighing.

Table 1. Mean berbaceous (annual) plant density (no. plants/m’) and biomass (pm/m’) in 1981 and 1982 for two microhabitats (open/shrub and shade) within three fire treatment areas (no fire, wildfire 1980, and control burn 1981). Means followed by the same letter (a orb for 1981 comparisons among the three treatments and x or y for 1982 comparisons) are not significantly different (JYO.05) according to Dunn’s multiple comparison test. Where no significance is given, data were absent for comparison or were grouped (e.g., others).

Wildfire Controlled burn No tire 1980 1981 Species/microhabitat Variable 1981 1982 1981 1982 1981’ 1982 Indian wheat Open/shrub Density 232a 542xy l26a 79oY 336a 149x Biomass 2.6a 21.4x 16.0a 84.7~ 7.la 19.4x Shade Density 38a 36x 61a 258x 25a 60x Biomass l.6a 1.9x 5.4a 31.8~ 0.9a 18.Oxy Red brome Open/ shrub Density 396a 120x 20b 2OY I84ab l lY Biomass l2.4a 5.6x 3.8b 2.5~ I I .Oab 2.4~ Shade Density 824a 258x Ilb 105xy lOl6a 4Y Biomass 26.3a 29.0x l.2b 34.7xy l8.la I.Oy Six-weeks fescue Open/shrub Density l2la 123x 2Oa 34x 88a 153x Biomass 0.5a 1.5xy l.2a 0.7x 0.6a 9.7y Shade Density 84a 8x 40a 35x 99a 23x Biomass 0.3a 0.3x 0.5a I .0x 0.3a 1.5x Schismus grass Open/shrub Density 27a 26x 66a l86y 34a 99x Biomass 0.3a 0.4x 13.8b 32.8~ 0.4a 10.6~ Shade Density 20a 48x 98b 284~ 56a 125x Biomass 0.7a I .9x Il.lb 63.2~ 0.2a 31.2~ Filaree Open/shrub Density 29a 49x IOa 26x 41a 33x Biomass 0.6a 3.9x 4.la 1.8x 3.4a 5.5x Shade Density la 61x IOa 18x 4a IIX Biomass 0.3a 4.7x l.2a 8.8x O.la 7.2x Goldfields Open/shrub Density Ila 266x l9a 85xy 9a 53Y Biomass O.la 5.9x 0.3a 4.3x 0.2a 6.3x Shade Density 4a 193x 26a 48x 8a 51x Biomass 0.03a 4.4x 0.5a 1.6x 0.04a 5.8x Filago Open/shrub Density 4a 451x 3a 7Y la 145xy Biomass O.la 5.8x O.la 5.3x O&la IO.Ix Shade Density - 151x 9 56xy - 25~ Biomass - 2.6x 0.1 2.9x - 1.3x Comb bur Open/shrub Density 20a 208x 4a llY 40a 39xy Biomass l.3a 10.8x 0.4a 0.9y 0.7a 2.9~~ Shade Density 9a 55x 4a lY 8a llxy Biomass 0.2a 2.6x O.ly O.ly 0.2a 0.9xy Red maids Open/ shrub Density - 6 4 3 - 4 Biomass - 0.2 2.9 I.0 - I.8 Shade Density l8a 46x 33a 33x 4a 76x Biomass 0.3a 4.0x 2.7a 6.1~~ O.la 31.4y Others Open/ shrub Density 39 I41 83 199 I90 230 Biomass 2.4 8.7 10.8 2.1 3.2 66.3 Shade Density 41 252 66 I28 36 210 I.6 28.5 9.4 34.0 5.9 51.6 TOTALS Open/shrub Density 879a 1914x 355b 1361~~ 923a 916~ Biomass 20.3a 64.2x 53.4b 136.1~ 26.6ab l38.Oy Shade Densitv l045a 1108x 358b 966x 1256a 596~ 31.3a 79.9x 32.2a 184.2~ 25.8a 155.94,

*Preburn measurements.

492 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Shaded area herbaceous plant plots were located under ran- Herbnceous Plants domly selected palo Verde (Cercidium microphyllum) and iron- Mean density and biomass values for the 9 abundant annual wood (Olneya tesota) trees at 4 aspects (N, S, E, and W) from the plant species are presented in Table 1. Species listed as “Others” stem at one-half canopy radius. Open/shrub area plots were include plants from the following genera: Plagiobothrys, Amsinc- located at random points not associated with tree canopies but kia, Cryptantha, Astragalus, Poa. Eriophyllum, Calandrinia, including both open interspaces or low shrubs. Density and bio- Lotus, Lupinus, Dichelostemma, Eschsholzia, Platystemon, Eu- mass values for the 9 most abundant annual plant species were phorbia, Silene, Pholistoma, Bowlesia, Orthocarpus, Androsace, analyzed among the 3 collection sites at each sampling time using Daucus, Phacelia, Parietaria, Eucrypta, Chenopodium, Rajines- the Kruskal-Wallis procedure (Dixon 1977) followed by Dunn’s quia, Linathus, and Lepidium. test (Hollander and Wolfe 1973). Separate comparisons were per- Before the controlled burn in 198 I, all herbaceous plant means formed for shaded and open/ shrub area plant data. Values for all calculated for the no fire and controlled burn (198 l* - Table I) species were then totaled and analyzed in the same manner. sites were not significantly different. This supports the assumption that comparisons between the 1982 no fire and controlled burn Woody Perennials and Cacti sites will accurately reflect fire-induced changes. Since the 1980 Tree, shrub and cactus plant data were collected on permanent 4 wildfire site was located adjacent to the other 2 sites, its prefire X 8-m quadrats (Cox 1974) for both the controlled burn (n q 23) status was presumed to have been sufficiently similar to allow for and wildfire (n = 10) sites. Plant densities werecensused 3 times on statistical comparisons among all sites. No comparisons were cal- the controlled burn site (27 May 1981 for preburn values, and 16 culated between 198 1and 1982data because differing temperature June 1981 and 17 March 1982 for postfire values), and 2 times on and rainfall patterns between years affects each desert growing I the wildfire site (23 June 1981 for year after fire and 17 March season’s annual plant productivity and species composition (Pat- 1982 for 2 years after fire). The 5 most important species (based on ten 1978, Webb et al. 1978). importance values) on the controlled burn site were analyzed using In 1981, of the 9 species listed in Table 1, only red brome Kruskal-Wallis procedure and Dunn’s test in order to test for (Bromus rubens) and schismus (Schismus arabicus) grasses res- significant differences among sampling times. In addition, for all ponded significantly to the 1980 wildfire. In open/ shrub areas, red postfire measurements, the survival and/or recovery strategy of brome density on the wildfire site was reduced by 95 and 89% while each plant was classified and recorded as a seedling, sprout, or biomass declined by 69 and 65% when compared to the no fire and surviving adult. controlled burn (preburn conditions) site. Similar trends were The use of the word significant in statistical analysis and discus- recorded for red brome in shaded areas. In contrast to red brome, sion of data indicates a probability level <0.05, unless otherwise schismus grass density increased significantly in both open/ shrub stated. and shaded areas I year after the wildfire. Red brome continued to show a reduction for the second year Results and Discussion after the 1980 wildfire and repeated the first-year reduction pat- Precipitation at Stewart Mountain for the winter growing sea- terns on the controlled burn. After 2 years, intermediate values for son (December through March) was 275 mm in 1979-80,90 mm in brome in the shade indicated a gradual recovery. Horton and 1980-81, and 154 mm in 1981-82. These represent 216, 71, and Kraebel(1955) also found brome grasses in southern California to 12 1% of the 195 I- 1980 30-year annual average of I27 mm for the be greatly reduced in the first postfire year. Studies by Keeley et al. 4-month period. As a result of above-normal precipitation, prefire (1981) suggest that reductions may be caused by low seed survival standing crop of herbaceous annual plants prior to the 1980 wild- during fire or low soil seed reserves at the time of the fire. fire was lush. The wildfire was probably more intense than the 198 1 Horton and Kraebel(1955) postulated that peak populations of controlled burn; however, the 2 sites appeared similar enough after brome grasses require several postfire years for either seed disper- burning to justify intersite comparisons. Lower than normal 198 1 sal into burned areas and/or on site build-up from pioneering precipitation did not produce good winter and spring annual plant postfire plants. Since red brome remained significantly reduced in growth for ground cover, but this biomass was supplemented by open/shrub areas 2 years after burning in the present study, it remaining litter on unburned areas from the 1979-80 wet winter. appears that at least 3 to 4 years will be necessary for full red brome Good germination and growth conditions for most desert species recovery in the desert. However, in the more mesic, shaded areas did occur in 1982 following above-normal winter rainfall. where seeds may have survived better because of light surface burning and low fire temperatures below the thick insulation (Pat

Table 2. Mean tree, shrub, and cactus densities (no. plants/ha) at three sampling times on the 12 June 1981 controlled burn site (n = 23) and at two sampling times on the adjacent 26 May 1980 wildfire site (n = 10). Means for the controlled bum site followed by the same letter (a or b) are not significantly different (KO.05) according to Dunn’s multiple comparison test. Only the five most important species were analyzed. There were no significant differences between 1981 and 1982 data on the wildfire site.

Controlled burn 1981 Wildfire 1980 Soecies 27 Mav 1981; I6 June 1981 I7 March 1982 23 June 1981 I7 March 1982 Bursage 6275a ll4lb ll4lb 3590 3844 Cholla cactus 450a l9lb l77b 280 157 Foothill palo Verde 203a l50a 55a Jojoba l43a l42a 68a 31 31 Brittle bush 82a l4a 707a 63a 81 lronwood 32 27 27 31 31 Desert mallow 27 I4 31 31 Barrel cactus I4 I4 I4 Saguaro I4 Four o’clock 109 Stickweed 95 310 313 Others 149 177 69 31 31 TOTALS 7607 1856 2476 4367 4516

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 493 ten and Cave l984), red brome density and biomass attained after fire, while density decreased, the average biomass per plant intermediate values within 2 years after burning. must have also increased after fire. This demonstrates the ability of In 1982, schismus continued to respond to fire with higher desert annuals to respond positively to fire, initiating a solid recov- densities in open/shrub areas on both the controlled burn and ery with increased biomass within 2 years. wildfire sites (Table 1). Schismus biomass production also signifi- The reduction in total herbaceous plant density was primarily a cantly increased manyfold on the 2 burn sites with increases in the result of low red brome density on burned sites, while increases shade greater than in the open. Apparently schismus grass seeds from schismus, Indian wheat, and red maids contributed to total survive fire well, and subsequent germination, which appears to be increases in biomass. These results for herbaceous plant responses stimulated by fire, allows this species to achieve higher density and parallel findings for southern Arizona desert grasslands (Reynolds biomass in the postfire herbaceous plant community. and Bohning 1956,and Cable 1967), the Chihuahuan Desert (Ahl- Of the remaining 7 genera and species in Table 1, 3-goldfields strand 1982) North American annual plant grasslands (Dauben- (Lasthenia californica); filago (Filago arizonica and F. californica; mire 1968) and California chaparral (Christensen and Muller and comb bur (Pectocarya recurvata and P. playtcarpa)-were 1975). reduced by fire. One-filaree (Erodium cicutarium)---showed no Woody Perennials and Cacti significant response; and 3-red maids(Calandrinia ciliata), six- Tree, shrub and cactus responses are presented in Table 2. weeks fescue (Vulpia octoflora),and Indian wheat (Plantago insu- Density of the most abundant species, bursage, was reduced by laris and P. purshii)-were stimulated by fire. Goldfields was 82% after the controlled burn. However, numerous seedlings had reduced in density primarily in the open/shrub areas. Filago, established on the adjacent wildfire burn site (ca. 78% of all bur- which normally grows best in intermediate shade locations, was sage individuals) producing intermediate density values. McLaugh- reduced in all conditions; however, it recovered more rapidly in the lin and Bowers (1982) also recorded very high mortality rates for shade. The absence of filago in the shade in 198 1 was probably a bursage after fire in the Sonoran Desert in south-central Arizona. result of year-to-year variations in temperature and precipitation Cholla cactus (Opuntia acanthocarpa, 95%; 0. fulgida, 3%; and because it had a uniform presence in 1982. Comb bur had lower 0. bigelovii, 2%) densities were reduced by about 58% immediately densities and biomass in 1982 on the wildfire site than on the after controlled burning. In addition to initial reductions in den- controlled burn site, indicating that the probable higher intensity sity, these cacti declined further between the 2 postfire sampling of the 1980 wildfire reduced these species more than the controlled times on both sites (Table 2) indicating that some cacti, not killed burn. at the time of or shortly after fire, die eventually from latent heat Of the herbaceous species that showed a positive response to fire, damage to epidermal and mesophyll tissue. Visual observations of red maids increased in biomass primarily in the shade, but these cacti in later sampling dates verified this result. No seedlings or increases diminished by the second year. The density of red maids sprouts were recorded for any cactus species; postfire cacti consist was not significantly affected by fire. Indian wheat, a characteristic entirely of adults that survive fire with little or no injury. Barrel annual in open and semishaded areas in the desert, had signifi- cactus (Ferocactusacanthodes)was unaffected by controlled burn- cantly greater biomass (ca. 300%) on the wildfire site than on the ing while all saguaro cacti (Carnegiea gigantea) found within sam- controlled burn or no fire sites in 1982. Its response in shade was pling quadrats were killed by fire. However, other larger saguaros, high biomass values on the 2-year-old wildfire, intermediate values that were not sampled on the site, managed to survive, although on the l-year-old controlled burn, and lowest values on the no fire scorched epidermis may permit invasion of fungi or other lethal site, indicating a continued increase for a few years following fire. vectors. Observations of burned sites indicate that small saguaros Six-weeks fescue increased dramatically in biomass (547%) in the less than about 2-4 m in height do not survive when large fuel open/shrub areas 1 year after fire as measured on the controlled quantities are present near the base, while larger individuals with burn and then showed a reduction after 2 years measured on the extensive corky bases usually survive with limited damage. On the wildfire site. Densities of this species showed little change. There wildfire site, all cacti except chollas had died within a year after were no significant changes in density or biomass of six-weeks fire. As noted earlier, this site probably burned more intensely fescue in the shade, although the pattern of first-year increase with thereby consuming more of the perennial plant community than on a subsequent decline occurred. the controlled burn site. Results for cacti presented in this study In open/ shrub areas during 198 1 there were significant reduc- differ slightly from McLaughlin and Bowers (1982) as they did not tions in total herbaceous plant density (ca. 60%) and increases record any cactus mortality subsequent to initial postfire mea- (163%) in biomass on the wildfire site when compared to the no fire surements. Rogers and Steele (1980), however, recorded observa- site (Table I). In 1982, herbaceous plant density in open/shrub tions in the Upper Sonoran Desert similar to this study as did areas was also reduced 1 year after controlled burning, but the Bunting et al. (1980) for cacti in Texas mixed prairie. 2-year wildfire site had intermediate values that were significantly The 2 tree species, palo Verde and ironwood, had different different from either the no fire or controlled burn sites. In con- responses to fire. Density of ironwood was virtually unaffected by trast, biomass was significantly greater on both fire treatment sites fire while palo Verde underwent an initial 26% reduction followed when compared to the no fire site. by an overall 73% reduction 9 months after fire (Table 2 -controlled Similar trends in total herbaceous plant density and biomass burn site). These results are supported by high ironwood densities were recorded in shaded areas (Table 1). In this microhabitat, on the wildfire site where palo Verde was essentially eliminated however, increased biomass (ca. 13%) in 1981 on the wildfire site after fire. The decline in palo Verde density (Table 2) indicates that was not significant, but the reduction in density (ca. 69%) was. In this species also suffered from heat damage (probably to the cam- 1982, a 46% reduction in herbaceous plant density on the con- bium) during the fire, taking several months to eventually kill trees trolled burn site was observed when compared to the no fire site. or portions of their crowns. In comparison, ironwood with thicker The 1980 wildfire site density was intermediate again. Biomass was bark is rather resistant to fire and/or heat damage. Twenty-two greater on both fire treatment sites (controlled burn 95% and percent of palo verdes that survived had basal or aerial sprouting wildfire 131%) when compared to the no fire site. These data within 9 months after fire; ironwood was present only as surviving indicate significant reductions in herbaceous plant density 1 year adults. These results parallel findings by Rogers and Steele ( l980), after fire. During the second postfire year herbaceous plant density and McLaughlin and Bowers (1982). achieved intermediate status between prefire and 1 year after fire. Density of the economically important plant, jojoba (Simmond- Aboveground biomass, on the other hand, increased significantly sia chinensis), was substantially reduced within 9 months after the after fire. This increase was attenuated into the second postfire year controlled burn (Table 2). On the wildfire site, jojoba was less for both shaded and open/shrub areas. Since biomass increased abundant than on the controlled burn site, but its density remained

494 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 unchanged between sampling times. Jojoba is an active postfire hypothesized by Shreve (1925 and 195 I) and Humphrey (1963 and sprouter in the desert in that 60% of postfire jojoba plants on the 1974). Desert fire ecology therefore requires additional study in controlled burn site and 100% on the wildfire site were sprouting. order to determine the full extent of the effects of fire on the desert Brittle bush (Enceliu furinosu) underwent an initial 83% reduc- ecosystem. tion in density, but within 9 months it had increased to 762% above preburn status on the controlled burn site (Table 2). This was caused by very successful seed germination and subsequent seed- Literature Cited ling establishment on the controlled burn site. Similarly, desert Ahlstrand, G.M. 1982. Response of Chihuahuan Desert mountain scrub mallow (Sphaeralcea ambigua) appears to be greatly reduced or vegetation to burning. J. Range Manage. 35:62-65. eliminated by fire and shows recovery or reinvasion within a year. Bunting, SC., H.A. Wright, and L.F. Neuenschwander. 1980. Long-term Two species that were not recorded in the prefire vegetation effects of fire on cactus in the southern mixed prairie of Texas. J. Range appeared after fire: four o’clock (Mirabilis bigelovii)and stickweed Manage. 33:85-88. (Stephanomeriu exiguu) (Table 2). Densities of these species were Cable, D.R. 1967. Fire effects on semidesert grasses and shrubs. J. Range large enough to indicate that they may be initiating an invasion of Manage. 20: I70- 176. the site. Both species were present only as seedlings and neither Christensen, N.L., and C.H. Muller. 1975. Effects of fire on factors control- species was recorded by Rogers and Steele (1980), or McLaughlin ling plant growth in Adenostoma chaparral. Ecol. Monogr. 4529-55. Cox, G.W. 1974. Laboratory manual of general ecology. Wm. C. Brown and Bowers (1982) on other Upper Sonoran Desert burn locations. Co., Dubuque, la. Rogers and Steele (1980), however, recorded two genera (Cussiu Daubemnire, R. 1968. Ecology of fire in grasslands. In: Advan. Ecol. Res. and Castilfeja) not normally found in the climax community grow- 5:209-266. ing at their study area. Dixon, W.J. 1977. BMDP - Biomedical Computer Programs P-Series. Total tree, shrub and cactus density was reduced by 76% imme- Univ. California Press, Berkeley. diately after the controlled burn (Table 2). Nine months after the Hollander, M., and D.A. Wolfe. 1973. Nonparametric statistical methods. fire, density for these growth forms had increased slightly because John Wiley and Sons, New York. of extensive brittle bush seedling establishment. Averages from Horton, J.S., and C.J. Kraebel. 1955. Development of vegetation after fire recovery and/or survival strategy data indicate that on 16 June in the chamise chaparral of southern California. Ecology 36:244-262. Humphrey, R.R. 1949. Fire as a means of controlling velvet mesquite, 1981, 98% of all perennials were adults surviving the controlled burroweed, and cholla on southern Arizona ranges. J. Range Manage. burn while on I7 March 1982,9 months after burning, the same site 2:175-182. was composed of 38% seedlings, 30% sprouts, and 32% adults. At Humphrey, R.R. 1963. The role of fire in the desert and semidesert grass- the latter sampling time, the wildfire site consisted of 48% seed- land areas of Arizona. Proc. Tall Timbers Fire Conference, Tallahassee, lings, 13% sprouts, and 52% adults indicating that seedling estab- Fla. lishment is more important than sprouting for postfire desert Humphrey, R.R. 1974. Fire in the deserts and desert grassland areas of recovery as noted by Rogers and Steele (1980). North America. p. 365-401. fn:T.T. Kozlowski and C.E. Ahlgren (eds.), Fire and Ecosystems, Acad. Press, New York. Conclusions Humphrey, R.R., and A.C. Everson. 1951. Effect of fire on mixed grass shrub range in southern Arizona. J. Range Manage. 4:264-266. The herbaceous plant community responds rapidly to fire with Keamey, T.H., and R.H. Peebles. 1960. Arizona Flora. Univ. California an overall reduction in plant density and increase in above-ground Press, Berkeley. biomass. Soil fertility increases commonly observed after burning Keeley, SC., J.E. Keeley, S.M. Hutchinson, and A.W. Johnson. 1981. in other ecosystems also occur in the desert (Whysong and Heisler Postfire succession of the herbaceous flora in southern California chap- 1978) and at least partially account for the reported increases in arral. Ecology 62: 1608-1621. above-ground biomass after fire. Lehr, H.J. 1978. A catalogue of the flora of Arizona. Desert Botanical Although red brome seeds (with large awns) apparently do not Garden, Phoenix, Ariz. survive fire well in the desert, schismus seeds (without awns) do; Lotan, J.E., M.E. Alexander, S.F. Arno, R.E. Freneh, O.G. Langdon, subsequently, postfire schismus growth is extensive. If herbaceous R.M. Loomis, R.A. Norum, R.C. Rothermel, W.C. Schmidt, and J.V. Wagtendonk. 1981. Effects of fire on flora. USDA Forest Serv. Gen. plant biomass remains high for several more postfire years, it is Tech. Rep. WO-16. Forest Serv. National Fire Effects Workshop, reasonable to conclude that burning may stimulate desert range- Denver, Cola. land productivity. Where grazing allotments are involved, this Martin, S.C. 1983. Responses of semidesert grasses and shrubs to fall suggestion has very important ramifications for desert rangeland burning. J. Range Manage. 36604-610. management, especially in light of current trends to convert desert McLaughlin, S.P., and J.W. Bowers. 1982. Effects of wildfire on a Sonoran land into more profitable uses. Therefore, any efforts to use pres- Desert plant community. Ecology 63:246-248. cribed burning should remain experimental until a long-term data Muller, C.H. 1940. Plant succession in the Lurreu-Flourensiu climax. base exists on which to make more reliable predictions. Ecology 21:206-212. Desert perennials also respond quickly to fire with large O’Leary, J.F., and R.A. Minnich. 1981. Postfire recovery of creosote bush scrub vegetation in the western Colorado Desert. Madrono 2:6l-66. amounts of seed germination and aerial or basal sprouting. Due to Patten, D.T. 1978. Productivity and production efficiency of an Upper the short-term nature of this study, data presented primarily record Sonoran Desert ephemeral community. Amer. J. Bot. 65:891-895. immediate fire loss and the initial recovery phase. Patten, D.T., and G.H. Cave. 1984. Fire temperatures and physical charac- The appearance of two nonclimax species (four o’clock and teristics of a controlled burn in the Upper. . Sonoran Desert. J. Range stickweed) suggest that in the Upper Sonoran Desert ecosystem, Manage. 37~277-280. specific postfire successional plant communities may exist, or at Reynolds, H.G., and J.W. Bohning. 1956. Effects of burning of a desert the very least, specific postfire successional species occur. Results mass shrub ranae in southern Arizona. Ecolonv 37:769-777. from Rogers and Steele (1980) support this hypothesis as they also RGgers, C.F. 1986 Photographic documentationof vegetation change in recorded nonclimax postfire species. They postulated that it may the Great Basin Desert. Ph.D. Diss., Univ. Utah, Salt Lake City. take 20 years for plant density and decades for species composition Rogers, G.F., and J. Steele. 1980. Sonoran Desert fire ecology. USDA Forest Serv. Cen. Tech. Rep. RM-81. to return to prefire status. These hypotheses seem tenable in light of Shreve, F. 1925. Ecological aspects of the deserts of California. Ecology the marked vegetation changes recorded here during the first two 6:93-103. postfire years, especially the reduction in such long-lived species as Shreve, F. 1951. Vegetation and flora of the Sonoran Desert. Vol. 1. saguaro. Vegetation. Carnegie Inst., Washington, D.C. Publ. 591:1-192. If partial postfire successional plant communities do exist in the Tiedemann, A.R., J.O. Klemmedson, and P.R. Ogden. 1971. Response of desert, then it is reasonable to infer that present plant distribution four perennial southwestern grasses to shade. J. Range Manage. patterns may have been more influenced by fire than originally 24442-447.

JOURNAL OF RANGE MANAGEMENT 37t6). November 1964 495 U.S. Department of Agriculture. 1980.Tonto National Forest Annual Fire Whittaker, R.H. 1975,Communities and Ecosystems. 2nd. ed. MacMillan, Reprt. Phoenix, Ariz. New York. Webb, W., S. Szarek, W. Lauenrotb, R. Kinenon, and M. Smith. 1978. Wbysong, G.L., and M.H. Heisler. 1978. Nitrogen levels of soil and vegeta- Primary production and water use in native forest, grassland and desert tion in the Upper Sonoran Desert as affected by fire. Proc. 1st Internat. ecosystems. Ecology 59: 1239-1247. Range Cong., Sot. Range Manage., Denver, Colo. Wells, C.G., R.E. Campbell, L.F. DeBano, C.E. Lewis, R.L. Fredrikson, Wright, H.A. 1980. The role and use of fire in the semidesert grass-shrub E.C. Franklin, R.C. Froelicb, and P.H. Dunn. 1979. Effects of fire on type. USDA Forest Serv. Gen. Tech. Rep. INT-85, lntermountain soil. USDA Forest Serv. Gen. Tech. Rep. WO-7. Forest and Range Exp. Sta., Ogden, Utah. White, L.D. 1969. Effects of wild fire on several desert grassland shrub species. J. Range Manage. 22:284-285. Callie Bermudagrass Yield and Nutrient Up- take with Liquid and Solid N-P-K Fertilizers

GALEN D. MOOSO, VON D. JOLLEY, SHELDON D. NELSON, AND BRUCE L. WEBB

Abstract A 2-year study to compare the effect of liquid and solid N-P-K N yielded 9.6 Mg ha-’ of forage in 1975. At another location (9:1:4)fertilizers on ‘Callie’bermudagrass (Cynodon dactylon var. (Monson and Burton 1982), Callie produced 27.4and 15.3 Mg ha-’ aridus Harlan et de Wet) production and nutrient uptake was of forage from the annual application of 336 kg ha-’ N for 2 conducted in Central Florida. There was a positive linear relation- consecutive years. Yields from the unfertilized controls were not ship between yield and amount of N-P-K fertilizer applied from reported in either study. In the latter investigation, the forage N both sources. Forage N and K concentrations were positively concentrations averaged I .6 and I .8% with applications of 336 and affected and P levels were unaffected by increased fertility levels. 672 kg ha“ N, respectively, when harvested at 8-week intervals Solid fertilizer increased dry matter production and resulted in (Monson and Burton 1982). These concentrations parallel sim- higher relative uptake efficiencies of the applied N, P, and K than ilarly treated coastal bermudagrass (Ashley et al. 1965, Burton et the liquid source. It also maintained higher N concentrations in the al. 1969). In a 2-year study, N uptake efficiency was 12 I and 80% of forage in some cuttings than the liquid, but neither P nor K the applied N the first year and 79 and 52% the second year for concentrations were affected by the fertilizer source. Ammonia annual rates of 336 and 672 kg ha-‘, respectively (Monson and volatilization of the urea in the liquid source was probably the Burton 1982). These excellent uptake efficiencies at high levels of major reason for the lower yield, N concentration, and N uptake application suggest a great potential for Callie in intensely- efficiency with that source. The trend for lower P and K uptake managed forage systems. efficiencies by the liquid-treated forage appears to be associated The effect of phosphorus and potassium applications on nut- with the lower yields obtained with this source. rient concentration in Callie bermudagrass has not been reported. ‘Callie’bermudagrass (Cynodon dncrylon var. aridus Harlan et Crespo et al. (1976), studying the effect of P and K application on de Wet) is a vigorous, robust, warm-season perennial grass selected coastal bermudagrass, reported forage P levels increased from 0.12 in 1967 in Mississippi as an aberrant plant. Cytological and elec- to 0.16% with rates of 0 to 68 kg ha-’ P. Forage K contents also trophoretic studies suggest that it is a natural hybrid of a Cynodon increased from I. I6 to I .2470 with the application of 0 to I32 kg accession introduced in 1964 from South Africa (PI 290814) and a ha-’ K. tetraploid form of C. dactylon var. aridus (Burson and Tischler, Although the application of liquid nitrogen and liquid N-P-K 1980). ‘Callie’ has proven to be taller, wider-leaved, and more sources on coastal bermudagrass has been studied (Burton and vigorous than other bermudagrass used for forage and has conse- Jackson 1962, Walker et al. 1979), liquid and solid N-P-K applica- quently elicited considerable interest in the past several years. tion to the more productive and robust bermudagrass hybrids has Despite recent interest in’Callie’, only limited information on its yet to be investigated. The objective of this research was to com- response to fertilization exists (Monson and Burton 1982, Ruelke pare the effect of liquid and solid N-P-K fertilizer sources on the 1976). In both studies, Callie produced more forage per unit of production and nutrient uptake of Callie bermudagrass. fertilizer applied than coastal bermudagrass (C. ductylon (L.) Materials and Methods Pers.). Ruelke ( 1976) reported that the application of 336 kg ha-’ of The experiment was established in March 1980 in a relatively Authors are former research associate at Brigham Young University and current pure stand of Callie bermudagrass on a Myakka fine sand (sandy predoctoral research associate at Iowa State University; assistant professor of agronomy, associate professor of agronomy, and director of Soil and Plant Analysis siliceous hyperthermic, Aeric Haplaquods) on the Deseret Ranches Laboratory. respectively, Brigham Young University. Provo, Utah. Florida, Inc. in Osceola County, Florida. A summary of the them Theauthors thank Deseret Ranches of Florida for funding the project. Appreciation is also given to the Brigham Young University Soiland Plant Analysis Laboratory, for ical characteristics of the soil is given in Table I. The experiment chemical analysis and 10 the Ezra T. Benson Food and Agriculture Institute for its consisted of 4 replications of a split-plot design. The main plots and support. This article is a contribution of Brigham Young University, 275 WIDB, Provo. subplots were fertilizer level and fertilizer source, respectively. Five Utah 84602. levels of liquid (L) and solid (S) fertilizers were applied in a 9: I:4 Manuscript accepted March 7, 1984.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 delayed the initial application sufficiently to require that treat- Table 1. Soiltest levels of Myakka fine sand. ments be applied in 3 equal applications. Applications were made approximately 7 weeks before each harvest. A 1.5 X 10.6-m area Measurement Soil test level from each plot was mechanically clipped to a height of 5 cm 4 times in 1980 ( 14 June, 8 August, 9 October, and 1 December, designated CEC, meq/ 100 g 7.1 cutting 1,2,3, and 4, respectively) and 3 times in 198 1 (5 August, 25 PH 5.8 N % 0.074 September, and 7 December, designated cutting 2,3, and 4, respec- O.M., % 2.83 tively) with a Hege 211 B research plot harvester. A 200- to 500-g P, mg kg-’ 7.6 subsample was taken from each plot for dry-matter determination K. mg kg-’ 20.8 and for chemical analysis. The samples were oven dried at 65’C Ca, mg kg“ 950.6 and ground to pass a l-mm screen. Plant samples were dry ashed Mg, mg kg-’ 55.6 (Chaplin 1970) and analyzed for K by atomic absorption spectros- Cu, mg kg-’ 0.18 copy. Digestion of samples in preparation for semiautomated Fe, mg kg-’ 4. I calorimetric analysis of N and P was according to a procedure Mn, mg kg-’ 1.78 described by Horwitz (1980). Composite soil samples were ob- Zn, mg kg“ I .4o tained from each plot in 1980 and 198 1 to 20-cm depth, air dried, and sieved(2 mm) in preparation for chemical analyses. Soil ratio (N:P:K) as follows: (1) O-O-O,(2) 45-5-18, (3) 90-10-37, (4) chemical analyses consisted of cation exchange capacity (CEC), 179-19-73, and (5) 358-38-146 kg ha‘ year“ of N-P-K. The liquid pH, organic matter (O.M.), total nitrogen (N), and double-acid fertilizer was composed of urea (86% of the N); ammonium pbly- extractable calcium (Ca), copper (Cu), iron (Fe), magnesium (Mg), phosphate (11% of the N and 100% of the P); potassium nitrate (3% manganese (Mn), phosphorus (P), potassium (K), and zinc (Zn) of the N and 17% of the K); and potassium chloride (83% of the K). (Bremner 1965, Graham 1948, Mitchell and Rhue 1979, Peech The solid fertilizer was composed of ammonium nitrate, triple 1965). The relative uptake of N, P, and K was calculated by superphosphate, and potassium chloride. It is recognized that the multiplying the oven-dry yield by the nutrient concentration for liquid and solid fertilizers differ in chemical composition. How- each treatment, subtracting the corresponding value for the con- ever, the solid fertilizer materials represent the long-established trol (level I), dividing the difference by the total amount of N, P, or standards and those most commonly applied to pastures. The K applied during the year, and multiplying by 100. Growing season liquid source is composed of materials most often available in (April l-December 3 1) precipitation was 750 and 803 mm for 1980 solutions. The total annual fertilizer treatments were made in 4 and 1981, respectively, which was below normal for both years. equal applications in 1980. In 1981 an early summer drought Precipitation for cuttings 1,2, and 3 ( 1980) and 3 and 4 (198 I) was

1980 1981

n 1st cutting q 2nd cutting 12 q 3rd cutting 11 q 4th cutting 10

1 2L 2s 3L3S 4L4S SLSS 1 2L2S 3L 3s 4L4S 5L 5s

TREATMENT LEVEL

Fig. 1. Cumulative dry matter production (Mg ha-‘) of Collie bermudagrass from liquid (L) and solid (S) N-P-Kfertilizer applications (I = O-O-O, 2 =

4.5-5-18. 3 : 90-10-37, 4 q 179-19-73, 5 = 358-38-146 kg ha-’ of N-P-K).

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 497 near or above normal and for cuttings 4 (1980) and 2 (1981) was cutting. Significant differences in forage production were observed below normal. A severe drought from October 1980 to June 1981 among cuttings within a given year, indicating that environmental prevented both fertilization and harvest for the period correspond- factors influenced the response of Callie to applied fertility. ing to cutting 1 in 198 1. Statistical significance is reported at the (Yq Fertilizer level significantly affected the amount of dry matter 0.05 level, unless stated otherwise. produced in each cutting period. In general, Callie responded linearly to increasing N-P-K levels. In only 3 cases (the liquid and Results and Discussion solid treatments for cutting 4 in 1980 and the liquid treatment for Yield cutting 2 in 1981) was it necessary to include quadratic and/or The cumulative dry matter production for the various cuttings of cubic components in the regression models to increase the fit. In all Callie bermudagrass at the 5 levels of liquid and solid N-P-K other cases the r* values ranged from 0.90 to 0.99 from a linear fertilizers is given in Figure 1. An application of 45 kg ha-’ N was regression analysis. inadvertently made to the entire experimental area in March 1980, Significant differences in production between liquid and solid before the experiment was initiated. This may have resulted in N-P-K sources were observed in the majority of the 7 cuttings and higher than normal yields for the first cutting for 1980. Signifi- for the total production in both years. The liquid source produced cantly lower cumulative yields were observed in 1981 compared to from 8 to 12% less forage in 1980 and from 8 to 19% less forage in 1980, partially in response to the above-mentioned application as 198 1 than the solid for the various rates, and averaged 10 and 15% well as to an early summer drought in 198 1 which prevented a first less production for 1980 and 1981, respectively. A similar study on

Table 2. Average nitrogen, phosphorus, and potassium concentrations (%) in Callie bermudagrass from liquid and solid fertilizer sources applied at five N-P-K levels.

1980 Cutting 1981 Treatment level source1 I 2 3 4 2 3 4 ---NT %- 1.55 1.39 1.28 1.92 1.59 1.53 1.95 2L I .49 1.29 1.27 2.02 1.58 1.54 2.07 2s 1.50 1.31 1.27 I .97 1.44 1.35 2.12 3L 1.55 1.31 1.33 1.89 1.62 1.37 2.10 3s 1.64 1.29 1.22 2.22 1.57 1.37 2.39 4L 1.54 1.25 I.16 2.18 1.58 1.38 2.27 4s 1.75 1.40 1.32 2.35 1.60 1.66 2.68 5L 1.78 1.31 1.27 2.29 1.52 I .46 2.54 5s 1.97 1.51 1.77 2.75 1.58 1.83 2.64 Mean2 1.63 1.34 1.32 2.18 1.56 1.50 2.31 LSD,& 0.17 NS 0.20 0.38 NS 0.09 0.46 LSD.& 0.17 0.16 0.17 NS NS NS NS

--p, %- 0.21 0.20 0.19 0.36 0.29 0.31 0.25 2L 0.21 0.20 0.19 0.40 0.28 0.30 0.24 2s 0.19 0.20 0.17 0.35 0.26 0.27 0.25 3L 0.16 0.22 0.19 0.39 0.27 0.26 0.27 3s 0.18 0.22 0.17 0.39 0.26 0.27 0.29 4L 0.22 0.19 0.19 0.40 0.30 0.27 0.28 4s 0.20 0.19 0.18 0.37 0.28 0.28 0.33 5L 0.22 0.20 0.21 0.43 0.24 0.27 0.30 5s 0.21 0.21 0.20 0.42 0.25 0.29 0.31

Mean2 0.20 0.20 0.19 0.39 0.27 0.28 0.28

LSD.05’ NS NS NS NS 0.03 0.02 0.04 I33.05~ NS NS 0.02 NS NS NS 0.03 --K, %-- 1.20 1.27 0.84 I.21 1.00 1.08 0.39 2L 1.20 1.30 0.97 1.44 1.07 1.28 0.40 2s 1.27 1.24 I .oo I .36 1.16 1.28 0.45 3L 1.41 1.39 1.09 I .50 1.32 1.41 0.54 3s 1.24 1.16 1.04 I .49 1.12 1.30 0.50 4L 1.45 I .40 I .03 1.45 1.23 1.48 0.45 4s 1.44 1.51 1.16 1.49 1.30 1.57 0.62 5L 1.47 1.24 1.17 1.62 1.26 1.63 0.54 5s I .49 1.57 1.33 1.73 1.36 1.76 0.61 Mean* 1.35 1.34 I .07 I .48 1.20 1.42 0.50 LSD.OE? 0.21 NS 0.16 NS 0.18 0.17 NS LSD.o# NS NS NS NS NS NS NS

‘Symbols refer to: I = O-O-O,2 =45-5-l& 3 = 90-10-37.4 = 179-19-73, 5 = 358-38-146 kg ha-’ of N-P-K. L = Liquid, S = Solid. *LSDos values to compare differences among cutting means for 1980 and 1981 for N are 0.32 and 0.31, for P are 0.048 and NS. and for K are 0.27 and 0.20, respectively. ‘LSDos value to compare differences among treatment means within a cutting. ‘LSD.05 value to compare differences between liquid and solid sources for a given treatment level.

498 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 the application of liquid urea, ammonium polyphosphate, and slightly over-estimate the efficiency of the applied nutrients. potassium chloride to coastal bermudagrass reported that the liq- Relative N uptake (Fig. 2) ranged from 12 to 44% in 1980,22 to uid source produced significantly less forage than the solid coun- 42% in 198 1, and averaged 28 and 30% for 1980 and 198 I, respec- terpart (Walker et al. 1979). Burton and Jackson (1962) found tively. The solid-treated forage achieved a higher relative uptake liquid urea to be only 80% as efficient for forage production as ammonium nitrate. Lower yields observed with the use of liquid n 1st cutting N-P-K on Callie are less pronounced than those reported for coastal bermudagrass. The lower yields from the liquid fertilizer application are most likely a result of ammonia volatilization of the urea (Volk 1966). Concentration Nitrogen Treatment level significantly increased forage N concentrations 50 in 5 of 7 cuttings (Table 2). Only in the second cutting in each year t did treatment level not significantly affect N concentration. The 40 general effect observed was for the highest N-P-K level (X8-38- 146) to produce higher concentrations than the control. In the 30 fourth cutting 198 I, level 4 (179- 19-73) was also positively affected. Thus, at least 90 kg ha-’ N was necessary to positively influence N 20 concentration. Some differences between the N concentrations of the liquid-and 10 solid-treated forage were observed, but these were limited to the first 3 cuttings in 1980. These differences usually occurred at the 0 highest N-P-K level (358-38-146). Significant differences in forage 1 N concentrations were observed among cuttings within each 1981 year-the forage fourth-cutting in both years contained signifi- 50 cantly higher levels than relatively consistent earlier cuttings. 40-l Phosphorus In 2 of the 7 cuttings (cuttings 2 and 3 in 198 l), increasing levels 30 of N-P-K reduced the concentration of P in the forage (Table 2); but in cutting 4 in 1981, forage P concentrations increased with increasing levels of N-P-K. The trend for the other 4 cuttings 20 i (cuttings I, 2, 3, and 4 in 1980), although not statistically significant, was for depressed P concentrations at the low and interme- 10 _:_~- diate N-P-K levels (90-10-37 and 179-19-73) and for elevated concentrations at the highest N-P-K level (358-38-146). 0 IL Although differences in P concentrations between liquid- and 2L 2s 3L 3s 4L 4s 5L 5s solid-treated forage were observed in 2 cuttings, the effects of the 2 sources in these cases were contradictory. In one (cutting 3, 1980), TREATMENT LEVEL the P content was higher with liquid than solid fertilizer, but in the other (cutting 4, 198 I), it was lower. Because of this contradiction, Fig. 2. Relative nitrogen uprake by Cailie bermudagrass. Fertilizer applied the limited number of affected observations, and the trend for P to as liquid (L)and solid (S)ar 4 N-P-K levels (2~ 45-5-l&3= 90-10-37.4 = be unaffected by source in most cuttings, it was concluded that 179-19-73.5 = 358-38-146 kg ha“ of N- P-K). there was no effect of fertilizer source on P concentration. efficiency than the liquid-treated forage, averaging an additional Significant differences in forage P concentration were observed 14.3 and 13.6% of the applied N for 1980 and 1981, respectively. among cuttings in 1980, but not in 198 1. In 1980, the fourth cutting These differences between sources were statistically significant was higher in P concentration than the 3 earlier cuttings. This only in 198 I. Treatment level did not significantly affect the effi- could be attributed to the relatively low yields of that cutting. ciency of N uptake in most cases nor were differences in N uptake Potassium efficiency among cuttings observed in either year. There was a trend in all cuttings for forage K content to increase Relative uptake of P (Fig. 3) ranged from 25 to 52% for 1980, with increasing N-P-K levels (Table 2). The trend was significant in from 25 to 56% for 1981, and averaged 35 and 40% for 1980 and 4 of the 7 cuttings, and generally treatment levels 3, 4, and 5 1981, respectively. Solid-treated forage averaged an additional 8.3 (90-10-37, 179-19-73, and 358-38-146) produced higher forage K and 17.7% uptake of the applied P above the liquid-treated forage concentrations than the control. in I980 and 198 1, respectively. This trend for more efficient uptake Forage K concentrations were not affected by the source of with the solid P source was significant for the second and fourth fertilizer. Differences in forage K concentration were observed cuttings and for total P uptake in I98 1. It appeared to be associated among the cuttings within each year, but no consistent trend for with the higher yields of the solid source. Generally the level of any given cutting to be higher or lower than another was found. N-P-K application had no effect on the efficiency of P uptake of The depressed forage K concentration observed in the fourth cut- Callie nor did it vary among cuttings in either year. ting in 1981 was probably related to a killing frost before harvest. The relative uptake of K (Fig. 4) ranged from 41 to 98% of the applied K for 1980, from 36 to 76% for 198 I, and averaged 71 and Effkiency of Applied N, P and K 62% for 1980 and 1981, respectively. The solid-treated forage An estimate of the efficiency of uptake of the applied N, P and K averaged an additional uptake of 23 and 10% more of the applied K fertilizer, expressed as relative uptake, is given in figures 2, 3, and than the liquid-treated forage in 1980 and 1981, respectively, but 4, respectively. This technique of assessing uptake has been used these differences were not statistically significant. On an individual extensively for this purpose in field experiments (Jansson and cutting basis, relative uptake efficiency with the solid treatments Persson 1982), even though it has been suggested that it may was significantly higher in 3 of the 7 cuttings (cutting 3, 1980, and

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 499 n 1st cutting q 2nd cutting W 1st cutting 1980 q q 2nd cutting 3rd cutting 1980 q 3rd cutting 60 q 4th cutting

40 60

30 60

0 1981 60

50 1981 60 40 6C 30

0 2L 2s 3L 3s 4L4S 5LSS 2L 2s 3L 3s 4L 4s 5L 5s TREATMENT LEVEL TREATMENT LEVEL Fig. 3. Relative phosphorus uptake by Callie bermudagrass. Ferrilizer

applied as liquid (L) and solid (S) at 4 N-P-K levels (2 = U-5-18, 3 q 90-10-37, 4 = 179-19-73. 5 = 358-38-146 kg ha-’ of N-P-K). Fig. 4. Relativepofassium uptake by Callie bermudagrass. Fertilizer app- liedas liquid (L)andsolid (S)ar 4 N-P-K levels (2 = 45-5-18,3= 90-10-37, cuttings 2 and 4, 198 1). Treatment level had no significant effect on 4 q 179-19-73, 5 q 358-38-146 kg ha-’ of N- P- K). the relative uptake efficiency of K in most cases. The large differences in relative K uptake among cuttings within each year were related to conditions unfavorable for growth resulting in lower Literature Cited relative uptake efficiencies. Ashley, D.A., O.L. Bennett, B.D. Doss, and C.E. Scarsbrook. 1965. Effect of nitrogen rate and irrigation on yield and residual nitrogen recovery by Conclusions warm- season grasses. Agron. J. 57:370-372. The results of this study indicate that the liquid N-P-K fertilizer Bremner, J.M. 1965. Total nitrogen. In: C.A. Black (ed.) Methods of soil was consistently less efficient than the solid fertilizer for Callie analysis. Chemical and microbial methods. Part 2. Agronomy 9: I 149- bermudagrass production. The lower efficiency, which was likely a 1178. Amer. Sot. of Agron. Madison, Wis. Burson, B.L., and C.R. Tischler. 1980. Cytological and electrophoretic result of ammonia volatilization of the urea in the liquid source, investigations of the origin of'Callie' Bermudagrass. Crop Sci. 20409-410. was reflected in lower yield and in lower relative uptake efficiencies Burton, G.W., and J.E. Jackson. 1962. Effect of rate and frequency of of N, P, and K. Although uptake efficiencies of N, P, and K were applying six nitrogen sources on Coastal Bermudagrass. Agron. J. not affected by source in 1980, higher relative uptake efficiency of 54:40-43. N and P for 198 1 was associated with the solid N-P-K source. The Burton, G.W., W.S. Wilkinson, and R.L. Carter. 1969. Effects of nitrogen, physical form of the fertilizer had no consistent effect on the phosphorous, and potassium levels and clipping frequency on the forage concentrations of P and K in the forage, but the solid source did vield and nrotein, carotene, and xanthophyll of Coastal Bermudaerass. increase N concentration in some cuttings in 1980. kgron. J.‘61:60-63. Response of Callie to increased levels of N-P-K was generally Chaplin, M.H. 1970. Use of atomic absorption spectrophotometer and chemical methods for plant analysis. Horticulture Deoartment. Oreaon linear. Forage N and K concentrations were elevated and P con- State University, Corvallis. centrations were often depressed by increasing levels of N, P, and Crespo, G., J.J. Paretas, and D. Pupo. 1976. Coastal Bermudagrass K. Treatment level had no significant effect on the relative uptake response to PK fertilization. Cuban J. Agr. Sci. 10:91-97. efficiency of applied nutrients. Differences in yield, N, P, and K Graham, E.R. 1948. Determination of soil organic matter by means of a concentrations, and relative efficiency of K uptake were observed photoelectric calorimeter. Soil Sci. 65: 181-183. among cuttings within each year indicating that the environmental Horwitz, W. 1980. Official methods of analysis of the Association of conditions had significant effects on yields and nutrient utilization Official Analytical Chemists. p. 127-128. Association of Official Analyti- of Callie bermudagrass. cal Chemists. Washington, D.C.

500 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Jansson, S.L.,and J. Petsson. 1982. Mineralization and immobilization of Peech, M. 1965. Hydrogen-ion activity. In: C.A. Black (ed.). Methods of soil nitrogen. In: F.J. Stevenson. Nitrogen in agricultural soils. Agron- soil analysis. Chemical and microbial methods. Part 2. Agronomy 9:914- omy 22:229-248. 926. Amer. Sot. of Agron., Madison, Wis. Mitchell, C.C., Jr., and R.D. Rhue. 1979. Procedures used by the Univer- Ruelke, O.C. 1976. Bermudagrass variety evaluation in Florida. University sity of Florida Soil Testing and Analytical Research Laboratories. Soil of Florida, Institute of Food and Agricultural Sciences. Agron. Res. Science Research Rep. 79-1, Inst. of Food and Agr. Sci., Univ. of Fla., Rep. AY77-I. Gainesville. Volk, G.M. 1966. Efficiency of fertilizer urea as affected by method of Monson, W.G., and G.W. Button. 1982. Harvest frequency and fertilizer application, soil moisture, and lime. Agron. J. 38:249-252. effect on yield, quality, and persistence of eight bermudagrasses. Agron. Walker, M.E., T.C. Keisling, and W.H. Matchant. 1979. A comparison of J. 74137 l-374. solid and liquid fertilizer for Coastal Bermudagrass hay production. Soil Sci. Sot. Amer. J. 43597-601 Establishment of Diffuse and Spotted Knap- weed from Seed on Disturbed Ground in Brit- ish Columbia, Canada

L.D. ROZE, B.D. FRAZER, AND A. MCLEAN

Abstract The rangeland weeds diffuse and spotted knapweed (Centaureu mine if the density of seeds sown affects the chance of knapweed diffusa L. and C. maculosa L.) were sown at densities of 208 to establishment on disturbed ground. 1,504 seeds/m2 on disturbed rangeland in Westwold, British Materials and Methods Columbia, in 25 X 25cm plots. Both species established well to the rosettes stage at the lowest sowing densities, but only 5% of the This study took place on rangeland near Westwold, B.C. at an diffuse knapweed rosettes bolted in the second year compared to elevation of 646 m where spotted knapweed predominates. The site 45% of the spotted knapweed rosettes. Intraspecific competition was a Douglas fir (Pseudotsuga menziesii)and pinegrass (Calama- appeared to decrease the number of spotted knapweed rosettes grostis rubescens)community that was logged 40 years earlier and bolting at the higher sowing densities. now consists of trees interspersed with large clearings. A I5 X I5 -m area, free of knapweed, was cultivated and raked smooth. Plots Diffuse and spotted knapweed (Centaurea diffuusu Lam. and C. were marked by delineating their borders with baler twine. Knap- maculosa Lam.) threaten 10.7 million ha in western Canada (Har- weed seed was collected in August, 1976-diffuse, from Pritchard, ris and Cranston 1979). In British Columbia, spotted knapweed and spotted, from Westwold. The seeds were stored for 2 months at prefers the more mesic areas of the Interior, growing at elevations room temperature and then handsown in late October, 1976. Seeds to 1,200 m. Diffuse knapweed, occupying 7 times the area of were sprinkled evenly on the soil and pressed down firmly, a spotted knapweed, grows only up to an elevation of 900 m and is method found by Watson (1972) to give optimal seed germination. more successful in the hotter, drier habitats (Watson and Renney The sowing density for each plot was chosen by drawing random 1974). Diffuse knapweed is a biennial, producing a rosette in its numbers. first season and bolting in the second, when it bears many spiny Diffuse knapweed was sown at densities of 368,944, and 1,504 seed heads. Spotted knapweed also produces a rosette in the first seeds/m* and spotted knapweed at 208,528, and 832 seeds/ m*. The year, but it is a shortlived perennial, producing new rosettes at the actual numbers sown were 23, 59, and 94 for diffuse, and 13, 33, bases of bolted stalks each fall for up to 3 years. Diffuse and and 52 for spotted knapweed seeds because the plots were 25 X 25 spotted knapweed seed yield has been recorded as high as 28,000 cm. Plots were 75 cm apart. Fewer spotted knapweed seeds were and 6,000 seeds/m*, respectively on rangeland (Roze 198 I). sown because they are heavier ( I .7 f .I mg) than diffuse knapweed Seed-reducing gall flies, Urophora affinis Frauenfeld and U. seeds (1.05 f .03 mg; Roze 198 I) and could have a better chance of quadrtfasciata Meigen (Diptera:Tephritidae) are two biological establishment. Each density was replicated 15 times. No seeds were control agents released against knapweed. Gall flies oviposit in added to I5 plots. The site was checked in September, 1977, to see if immature seed heads, and the larvae feed on ovariole tissue. These any rosettes had bolted. In June and September, 1978, numbers of flies have reduced diffuse and spotted knapweed seed yields by 80% rosettes were counted, and in September, 1978, the bolted plants at Pritchard and Chase, B.C.; but in spite of this, 2,000 seeds/m* and numbers of seed heads on each bolted plant were counted. were produced on both sites (Roze I98 I). Can this amount of seed reduction decrease the rate of spread of knapweed, or decrease Results knapweed density in existing populations? In this study we deter- No plants were found on control plots in September, 1977, and Authors are research scientists, Westwold. B.C., Agriculture Canada, Research one of the rosettes had bolted. In June, 1978 average numbers of Station. Vancouver, B.C., and range ecologist, Agriculture Canada, Research Sta- tion. Kamloops. B.C.. respectively. This project was funded in part by a grant from the diffuse knapweed rosettes did not vary significantly between sow- B.C. Forest Service and an Extramural Research Grant EMR 7808 from Agriculture ing densities of 368 and 944 seeds/m* but did increase significantly Canada. Manuscript accepted March 12. 1984. between sowing densities of 944 and 1,504 seeds/m* (Table I).

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 501 ulations would have to be greatly reduced. Schirman (1981) pre- Table 1. Average number of rosettes in June, 1978. dicted that even if the seed yield of diffuse and spotted knapweed were reduced to below 0. I % of their potential yield, and plant density declined, knapweed would spread to adjoining, non- Knapweed Sowing density Number of rosettes’ species seeds/m* per plot infested land. In this study some rosettes and bolted plants appeared on plots with the lowest sowing densities, indicating that Diffuse 368 3fla knapweed will establish on disturbed rangeland from very sparsely 944 6fla scattered seed. Once establishment occurs, seeds from the mature 1504 l3f2b plants form a seed bank in the soil, and seedlings emerge for years Spotted 208 2fla to come. The gall flies were in the vicinity of this study site, and 528 6flb although they did find the new plant population, seed production 832 8flc was not prevented entirely. ‘Means followed by the same letter do not differ significantly @ = .05) according to There was a decrease in the number of spotted knapweed Duncan’s New Multiple Range Test (LeClergeet al. 1962). Plot size is 25 cm X 25 cm. rosettes per plot with a decrease in numbers of seeds sown, but there appeared to be a plastic effect, where individual plant size, Spotted knapweed rosettes increased significantly with the sowing i.e., the average number of seed heads per plant, increased with density. Only 5% (0.3 + 0.1 per plot) of the diffuse knapweed decreasing rosette density. Thus, even if rosettes were thinned by rosettes present in the spring of 1978 bolted that season and there biological control agents, seed yield per unit area may remain the were no significant differences in numbers of bolted plants between same. Intraspecific competition among spotted knapweed rosettes sowing densities. In contrast, 45% of the spotted knapweed rosettes accounted for the decrease in the number of bolted plants from the bolted in 1978 and the number per plot increased significantly in sowing density of 528 to 832 seeds/m* because rosette numbers plots with 208 to 528 seeds/m2 (Table 2). However, the number of were significantly greater at 832 than 528 seeds/ m2. A reduction in rosette density by biological control agents may result in a higher Table 2. Average number of bolted spotted knrpweed plants per plot in percentage of plants bolting in the second year. September, 1978. So few diffuse knapweed rosettes bolted that the effect of rosette density on mature plant size could not be measured. A study similar to this was made in Kamloops, where diffuse knapweed is Sowing densitv (seedslmzj Mean’ the predominate species, but the probability of bolting in the 208 1.2 f 3 a second year was not greater than in Westwold (Roze, unpublished 528 5.5 f .I b data). Therefore, diffuse knapweed appears to remain vegetative 832 2.4 f .5 c for at least two years on rangeland before it bolts. This species can ‘Means followed by the same letter do not differ significantly @ = .05) according to become very well established in the rosette stage before biological Duncan’s New Multiple Range Test (LeClerge et al. 1962). Plot size is 25 cm X 25 cm. control agents that feed on the mature plants can attack it. bolted plants per plot decreased significantly from the sowing Literature Cited density of 528 to 832 seeds/m*. The average numbers of seed heads per plant for diffuse (72 f Harris, P., and R. Cranston. 1979. An economic evaluation of control IO) and spotted ( I6 f I) knapweed were not affected by the sowing methods of diffuse and spotted knapweed in western Canada. Can. J. density. There was a trend, however for spotted knapweed: the Plant Sci. 59:375-382. number of seed heads per plant decreased with increasing sowing Le Clerge, E.L., W.H. Leonard, and A.G. Clark. 1962. Field Plot Tech- density (20 + 3, I6 f 2, and I3 f 2 seed heads per plant for the nique. Burgess Publishing Company, Minneapolis. 2nd Edition. sowing densities of 208, 528, and 832 seeds/m*, respectively). Dis- Rose, L.D. 1981. The biological control of Cenfaureu diffuse Lam. and C. section of undehisced seed heads revealed that Urophora affinis maculosa Lam. by Urophora affinis Frauenfeld and U. quadrifasciata and (1. quadrifasciata had attacked both diffuse and spotted Meigen (Diptera:Tephritidae). Ph.D. Thesis The University of British Columbia. knapweed, but most seed heads contained some seeds. Schirman, R. 1981. Seed production and spring seedling establishment of Discussion diffuse and sootted knapweed. J. Range Management. 3414547. Watson, A.K. i972. The biology and c&trol ofkenraurea diffusa Lam. The density of seeds sown affects the chance of knapweed estab- and Cenraureu maculosu Lam. M.Sc. Thesis, The University of British lishment on new ground, with fewer diffuse and spotted knapweed Columbia. rosettes produced by fewer seeds. Therefore seed-reducing biologi- Watson, A.K. and Renney, A.J. 1974. The Biology of Canadian Weeds. 6. cal control agents could theoretically lower the chance of knap- Centaurea dif/usa and C. maculosa. Can. J. Plant Sci. 54:687-701. weed establishment in new locations. But, existing knapweed pop-

502 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Phenological Development and Water Rela- tions in Plains Silver Sagebrush

RICHARD S. WHITE AND PAT 0. CURRIE

Abstract Detailed measurements were made on silver sagebrush plants to quantify phenologicsl development, plant water potential, and soil water status. Measurements were made at bi-weekly intervals from early April to late October. A pbenological scoring system was employed and the data used in linear regression equations with calendar date or plant water potential as independent variables. Both variables were used successfully in predicting pbenologicnl development in silver sagebrush. However, calendar date bad less variability around the regression line, and it would therefore probably have greater direct application. The results should prove of value in future autecological studies of the species and will provide important information to manage plant communities that contain silver sagebrush. Species ofsagebrush (Artemisia spp.) have long been recognized as havingan important ecological role on rangelands overmuch of the West. In areas where it is abundant, sagebrush often has considerable economic impact on livestock operations, and it is frequently regarded as a plant which should be reduced or eliminated through brush control practices. In other circumstances, sagebrush species may provide desirable forage or cover for either livestock or wildlife. The degree ofdesirability of sagebrush generally depends upon the species and in some cases the variety that is present. The complexity of this relationship is exemplified from the vast literature that has been compiled. Harniss et al. (19X1), for example, cite almost 1,500 references dealing primarily with the section Tridentatae of Artemisia. The majority of published material cm sagebrush species has been directed toward a better understanding of big sagebrush (Anemisia tridentam). Since big sagebrush occupies more acreage than other Artemisio species, this is not surprising. In comparison, very little research work has dealt directly with silver sagebrush (Arfemisiacana). However, silver sagebrush has beenestimated to occupy 13 million ha in the I I western states, and it ranks second in importance to big sagebrush (Beetle 1960). Therefore, it merits much more intensive research attention. Additional information is particularly needed with respect to how to manage plant cammuni- ties in which silver sagebrush is a major component. Knowledge of phenological development in plants is basic to an understanding of how they react to different management practices. Hyderetal. (1962), for example, found substantial differences in herbicide efficacy on big sagebrush and green rabbitbrush (Chrysothamnus viscid~@wus) in relation to environmental, phenological, and physiological conditions. With silver sagebrush, White and Curie (1983) reported wide differences in response to fire that could have been caused by differences in phenological development of the plant. Lack of more precise information on shrub phenology, however, has too often limited interpretation of research results. Because a more concise description of phenologi- agement options, we decided to establish a more comprehensive cal events would be of considerable benefit to understanding man- description of phenology in Plains silver sagebrush (Arremisia cam cam). We also believed that it was important to define more concisely water relationships that existed during the period when the plant was physiologically most active. Shrub phenology has received relatively little attention in comparison to phenological development in grass species. West and

JOURNAL OF RANGE MANAGEMENT 37(6),‘November 1964 503 mixed prairie in the Northern Great Plains (Daubenmire 1978, Weaver and Clements 1938). Major plant species included western Table 1. Numeric scores that were assigned to phenological stages of wheatgrass (Agropyron smirhii), green needlegrass (Stipa vir- development in silver sagebrush. Principle anatomical characteristicsare id&), annual bromes (Bromus spp.), threadleaf sedge (Curex indicated in the morphological description. filifoliu), and Plains silver sagebrush. Long-term precipitation for the area averages about 350 mm/year, but it commonly ranges Numeric from 250 to 450 mm/year. During the year this study was con- Phenological ducted, precipitation was 272 mm. Temperatures during the study Score Morphological Description were near normal throughout the growing season, and hence they probably exerted a normal influence on phenological development. 0.0 Winter dormant, foliage dull gray. Soils on the study sites were Borollic Camborthids and belonged 1.0 Buds swelling, initial signs of leaf extension present but to the Yamac and Kobar series. These series are characterized as no obvious elongation, foliage green. deep, well-drained soils that form from alluvial sediments. The 2.0 New leaves emerging from buds, showing rapid elonga- primary difference between series is in texture. The Kobar series is tion along with shedding of old leaves, no distinct stem a silty clay loam, while the Yamac series is a loam. Both series are elongation. relatively fertile and represent highly productive range sites. Topo- 3.0 New leaves near full size, initial signs of stem growth graphy was nearly level and study sites were typical of plant com- present but no extensive elongation. munities in which silver sagebrush often becomes abundant. 4.0 Undergoing rapid stem elongation of primary stems. Selection of individual study areas was based primarily on sta- 5.0 Secondary stem growth evident in axes of leaves of pri- ture and density of silver sagebrush. An effort was made to select mary stems but without much stem elongation in areas that were representative of plant communities throughout secondary tissue. the Northern Great Plains where silver sagebrush can become 6.0 Secondary stems undergoing rapid elongation. abundant. Three degrees of sagebrush infestation (light, moderate, 7.0 Floral branches differentiating. heavy) were represented by the different study areas. The lightly infested area was characterized by sagebrush plants that were 8.0 Floral buds swelling. generally under I m in height. Plants were scattered in a relatively 9.0 Anthesis. open stand with distances between individuals commonly 5 to 10 10.0 Post anthesis, pre-seed development, yellow floral parts. m. The heavily infested area had sagebrush plants that commonly Il.0 Brown and dry floral parts, seeds developing moist and exceeded 1 m in height, and intervals between individuals were semi-succulent achenes. usually less than 2 m. The moderately infested area was interme- 12.0 Seeds dry and undergoing dissemination. diate. Figure 1 shows the range of infestation that was encountered 13.0 Leaves on floral branches shed. across the 3 study areas. Five 100 m long transect lines were established in each of the 3 study areas. Three plants were selected at random for detailed analysis from each transect at bi-weekly intervals between the first Wein (197 1) suggested a numeric scoring system that could be of April and the end of October. This interval covered that period adapted to a variety of shrub species. DePuit and Caldwell (1973) of the year when plants showed first morphological signs of physio- subsequently modified this system to relate phenological events in logical activity and continued until seeds were dispersed and plants big sagebrush to photosynthesis and respiration, and Campbell were becoming dormant. The 45 plants that were sampled on each and Harris (1977) used the same system to study water relations in date were evaluated for phenological development and assigned a big sagebrush. An acceptable numeric scoring system, however, numeric score based on morphological categories and the approach has not been devised for silver sagebrush. Moreover, the sequence suggested by West and Wein (1971) (Table 1). of phenological events has not even been described although it Moisture conditions were evaluated by means of a neutron seems likely that phenological development could account for scattering depth gauge for measuring soil moisture and a Scho- some of the response differences that have been observed under lander pressure bomb for determining plant moisture stress (War- different management regimes. Our work has been designed to ing and Cleary 1967). Soil water content was derived from a total of obtain an increased understanding of phenological development in 12 neutron access tubes located within the 3 study areas. Readings silver sagebrush. Specific objectives were: (1) to define readily were taken every 2 weeks at the same time phenological observa- recognizable phenological stages of development in silver sage- tions were made. Sample depths measured were 30,9 1, 122, and brush that would be of value in subsequent autecological and 152 cm. Plant water potential was determined on each date for each management related studies, (2) to relate phenological response to of the 45 plants that were phenologically scored. Readings were calendar date with respect to ontogenetic development within a obtained between 8 and 11 AM. This interval was selected because single season, and (3) to examine the extent to which soil moisture diurnal variation with individual plants at this time showed little and plant moisture stress’ were related to phenological develop- deviation from values obtained at sunrise. Three twigs were ment. sampled from each plant. Pressure bomb values were then Methods obtained for each twig, and the plant moisture stress of each plant was calculated as the average of 3 readings. Data for this work were obtained in 198 1 from 3 study areas on the Livestock and Range Research Station west of Miles City, Results were analyzed statistically by regarding individual Mont. Individual study areas were 0.56 ha in size and measured plants within each study area as replications. Mean value compari- 61.0 by 91.5 m. Vegetation of the study areas was typical of the sons between areas were made within each sampling date by using Student’s t-test. Regression relationships were derived by pooling ‘As pointed out by Taylor (1968).considerable discrepancy exists when using termi- observations of plants from all 3 areas. Dates and plant moisture nology dealing with plant water relations. We have chosen to regard “plant moisture stresses were used as independent variables in separate linear equa- stress” and “plant water potential” as homologous and have used them interchange- ably. Both terms have been used extensively by researchers so that there seems to be tions to predict phenological stage of development. little differentiation between the two. Results and Discussion Bud swelling in early April was the first readily observable sign of the end to winter dormancy in silver sagebrush plants. This was

504 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 followed by leaf expansion, twig growth, floral differentiation, and simple linear expression: seed dispersal with a normal progression of phenological events. ?=a+bX Comparison of our results with general observations reported by where ? is the predicted numeric growth stage as described in Table I Harvey (198 1) over a 3-year period indicated that phenological X is the calendar date expressed as day of the year (92 to 301) development during our study was characteristic of normal mor- a and b arc regression coefficients phogenesis in the species. It is likely, therefore, that our results are This relationship is illustrated by our data in Figure 2. The estab- typical of those usually encountered. However, it should be noted lished relation between date and plant phenology is important that year-to-year weather differences in the same area can result in different phenological patterns with respect to date or morphological development. Comparison of phenological development among the 3 study areas showed small differences throughout the duration of the study with respect to phenological scoring units (PSU). Differen- ces between high and low PSU values on individual sampling dates were less than 0.2 on 10 occasions, between 0.3 and 0.5 on 4 occasions, and greater than that only once. Plants from sites that were lightly or moderately infested had comparable PSU values. On 3 of the 16 sampling dates, lightly infested areas had significantly (KO.05) higher values than moderately infested areas. On 4 sampling dates this situation was reversed, and on 9 sampling dates there were no significant differences between light and moderate Y= -4.22 + .06x infestations. Plants from the heavily infested site tended to be more R’ = .QQ advanced phenologically than those from the other 2 sites. This was most readily seen by examining the relative rank of individual study areas with respect to PSU values (Table 2). The heavily . . Table 2. Relative rank of individual study areas expressed as /

frequency of occurrence (70) across all sampling dates. 1 412 4130 s/27 6/24 r/22 at19 0116 10114

Numeric rank of area CALENDAR DATE (MO/DAY) Characteristic Infestation low middle high Fig. 2. Phenological development of silver sagebrush as observed Phenology light 53 28 19 moderate 47 22 31 from early April to late October. Each point represents a mean of heavy 12 38 50 45plants. Criteria used to establish numericphenologicalscores Plant moisture stress light 88 6 6 are presented in Table 1. moderate 0 66 34 heavy 13 28 59 because it demonstrates that changes in morphological attributes are readily described in silver sagebrush by a linear equation. The numeric values that were assigned to individual morphological infested site had the highest value on 50% of the sample dates and stagesofdevelopment were well suited toquantitativeexpression. Moreover, the lowest value on only 12% deviations from the regression line were small on individual sam- Plant water potential values were more related to the degree of pling dates. As a result, the criteria that we established to describe sagebrush infestation than was phenological development. Plants plant phenology should have direct application in mathematical from the lightly infested site were consistently under less moisture modeling in this or other similar species. In addition, discrete stress than plants from the heavily infested site (Table 2). Mean development stages are identified that can be examined in subse- differences between these 2 sites were less than 3 bars on 8 of the 16 quent studies that address management of plant communities with sample dates, between 3 and 6 bars on 3 dates, and between 6 and 9 silver sagebrush. bars on the 5 remaining dates. Plants from the moderately infested Evaluation of neutron probe data showed that seasonal shifts in area usually had intermediate water potential values. Similar results have been reported by Wambolt (1973) in conifer species SOIL WATER CONTENT ( lb ) and apparently reflect intraspecific competition for available mois- 20: 203 ture. Soil moisture content showed very consistent relationships - among the 3 study areas. It was highest in the heavily infested site and lowest in the moderately infested area. It appeared, therefore, to have no easily identified cause-and-effect association with shrub density. In addition, soil moisture measurements seem to be in conflict with plant water potential observations. That is, high plant i 7 moisture stress was associated with higher soil moisture content. 5127 6124 Similar results have been reported by Branson et al. (1976) during the spring period and by Haas and Dodd (1972) for mesquite in Texas. The most plausible explanation of this situation would appear to be site differences in osmotic potential of the soil and/ or the plant. However, other factors such as vertical root distribution may also have contributed. i i When phenological development was examined among all 1 7122 I015 I3119 I9130 10114 plants as a function of calendar date, it was apparent that a direct relationship existed which could be mathematically described as a Fig. 3. Soil moisrure profiles obtained by neutron attenuation from I.2 access tubes at bi-weekly intervals from April to October. Individual sampling dates (molday) are indicated beneath each curve.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 505 moisture content were primarily restricted to the upper 60 cm of the soil profile (Fig. 3). Moisture content of this layer was relatively high until the end of June. As a result, silver sagebrush and other 12 plant species were able to sustain vigorous growth during this period (Phenological stages 4.0 to 6.0). From early July until late September, soil moisture content was relatively low throughout the soil profile with little change from one sampling date to 10 another. During this time period, silver sagebrush ceased stem elongation. Differentiation of reproductive parts, however, continued on a regular basis as flowers matured, were pollinated and z developed seed (Phenological stages 7.0 to 11.O). Rainfall toward 08 the end of summer and in early fall provided some recharge of the surface horizons, but soil moisture content did not attain as high a 6 level as that observed during spring and early summer. During this 3 interval, silver sagebrush seeds matured and seed dispersal began. Z6 Changes in plant water potential primarily reflected changes in soil moisture content. These results were consistent with those reported for big sagebrush by Branson and Shown (1975). As the season advanced, there was an overall decrease in plant water 54 potential (Fig. 4). During April and May, plant water potential was d

0 0 10 20 30 40 PLANT MOISTURE STRESS ( - BARS 1

Fig. 5. Relationship between phenological development and plant moisture stress in silver sagebrush. Plant moisture stress data were obtained with a Scholanderpressure bomb, andphenologicalscores were assigned according to criteria in Table I. E -- 4;2 4i30 6i27 ai 7i22 sita Q;I~ joi

CALENDAR DATE ( MO/DAY 1 stages exist in silver sagebrush that can be quantitatively character- izedand described by linear regression. In doing this, a mathematical expression of phenological development was obtained that Changes in plant water potential of silver sagebrush that were Fig. 4. utilized either calendar date or plant water potential as independ- observedfrom early April to late October. Individualpoints on each date represent mean values from 45 plants. ent variables. Water relations were determined for the species between early April and late October. This period corresponded to approximately -15 bars. This corresponded to a time of the year the interval that silver sagebrush was physiologically most active. when soil moisture content was favorable and air temperatures Additional data will be needed to increase resolution of the pheno- were relatively cool. In June, soil moisture continued to remain logical events that we observed in reference to year-to-year varia- high, but plant water potential began to decrease. This decline tion, but our work provides preliminary autecological information appeared to be associated with increased air temperatures and on the species. It should thereby furnish a better foundation for vapor pressure deficits, and it can probably be attributed to pro- management of plant communities that contain silver sagebrush. portionately higher rates of transpiration. Similar patterns have been reported in other shrub species (Moore et al. 1972), and Literature Cited Wambolt (1973) has pointed out the importance of these factors in Beetle, A.A. 1960. A study of sagebrush--Section Tridentatae of Artemi- influencing water potential. In July, August, and September, plant sia. Wyom. Agr. Exp. Sta. Bull. 368. water potential continued to decline gradually to an average low of Branson, F.A. and L.M. Shown. 1975. Soil-moisture stress as related to about -35 bars. This trend apparently resulted from continued plant-moisture stress in big sagebrush. J. Range Manage. 28:212-215. depletion of soil water concomitant with lack of rainfall and matu- Brmson, F.A., R.F. Miller, and LS. McQueen. 1976. Moisture relation- ration of reproductive parts in silver sagebrush. Improved soil ships in twelve northern desert shrub communities near Grand Junction, moisture conditions and cooler temperatures in late September Colorado, Ecology 27:1104-I 124. and early October were associated with a slight increase in plant Campbell, G.S.,nnd G.A. Harris. 1977. Water relations and water use water potential. patterns for Artemisia tridentata Nutt. in wet and dry years. Ecology Examination of relationships between plant moisture stress and 58~652-659. phenological development indicated that as stress became greater, Daubenmire, R. 1978. Plant geography. Academic Press. New York. DePuit, E&and M.M. Caldwell. 1973. Seasonal pattern of net photosyn- plant phenology advanced (Fig. 5). It should be noted, however, thesis of Artemisia tridentata. Amer. J. Bot. 60:426-435. that this response may be partially an artifact of statistical depen- Haas, R.H. and J.D. Dodd. 1972. Water-stress patterns in honey mesquite. dencies existing among date, plant water potential, and plant Ecology. 53:674-680. phenology. Since plant water potential decreased at the same time Harniss, R.O., S.J. Harvey, and R.B. Murray. 1981. A computerized that calendar data advanced, these 2 factors are confounded to the bibliography of selected sagebrush species (genus Artemisia)in western extent that it is difficult to ascertain which is most critical to North America. USDA Forest Serv., Gen. Tech. Rep. INT-102. phenological development. Harvey, S.J. 1981. Life history and reproductive strategies in Artemisia. In summary, our work showed that distinct morphological M.S. Thesis, Montana State Univ., Bozeman.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Hyder, D.N., F.A. Sneva, and V.H. Freed. 1962. Susceptibility of big Wambolt, C.L. 1973. Conifer water potential as influenced by stand density sagebrush and green rabbitbrush to 2,4-D as related to certain environ- and environmental factors. Can. J. Bot. 12:2333-2337. mental, phenological and physiological conditions. Weeds 101288-295. Waring, R.H., and B.D. Cleary. 1967. Plant moisture stress: evaluation by Moore, R.T., R.S. White, and M.M. Caldwell. 1972. Transpiration of pressure bomb. Science 155:1248-1254. Arriplex confertifoliu and Eurotic lanaru in relation to soil, plant, and Weaver, J.E.,and F.E. Clements. 1938. Plant Ecology. McGraw-Hill Book atmospheric moisture stresses. Can. J. Bot. 50~241 I-2418. Co.. Inc.. New York. Taylor, S.A. 1968. Terminology in plant and soil water relations. p. 49-72. West,‘N.E.; and R.W. Wein. 1971. A plant phenological index technique. In: T.T. Kozlowski (Ed.). Water deficits and plant growth. Academic Bioscience 2 I : I I6- I 17. Press, New York. White, R.S., and P.O. Currie. 1983. The effects of prescribed burning on silver sagebrush. J. Range Manage. 36:6l I-613. Germination Profiles of Introduced Love- grasses at Six Constant Temperatures

MARTHA H. MARTIN AND JERRY R. COX

Abstract Seeds of A-68 Lehmann lovegrass (Eragrostis lehmanniana lower soil temperatures, it may be possible to seed in fall and rely Nees), cochise lovegrass (Eragrostis Iehmanniana Nees X Eragros- on cool-season moisture rather than on warm-season moisture for tis trichophora Coss & Dur.) and A-84 and Catalina boer love- germination and seedling growth. The purpose of this study was to grasses (Eragrostis curvula var. conferta Nees) accessions were determine the germination characteristics of four lovegrasses at germinated for 14 days at constant temperatures of 15,18,21,24, constant temperatures. 27, and 30” C. Light intensity was 216 p mol m-k-’ and photope- Methods riod was 15 h. Germination of Catalina seeds varied from 87 to 96% between 18 and 300 C after 12 days. Germination of cochise seeds Fifty caryopses of either A-68, A-84, Catalina or cochise love- was optimum between 21 and 270 C after 12 days. Germhtation of grasses were placed on filter paper in separate plastic petri dishes. A-68 seeds was optimum at 27” C and A-84 seeds at 30” C. This Approximately 9 ml of distilled water was added to each dish at the study indicates that Catalina boer lovegrass and cochise lovegrass beginning of the study; and seeds were germinated at either 15, 18, will germinate at relatively low temperatures. A-68 and A-84 love- 2 1,24,27, or 30” C on a thermogradient plate (Larsen 1962) under grasses, in contrast, require higher temperatures for optimum alternating I5 h light and 9 h dark. All experiments were conducted germination. in a light regime of 216 p mol mdsel photosynthetic active solar radiation. Previous laboratory experiments had established that The most commonly recommended grasses for rangeland seed- lovegrass germination was not influenced by light intensity or ing in the arid Southwestern United States and Northern Mexico length of photoperiod, as long as photoperiod was greater than 1 h were either introduced from Southern Africa or developed from (Unpublished data, USDA-ARS, Tucson Ariz.). Germinated genetic lines collected in Southern Africa. The most easily estab- seeds were counted daily for 14days after water was added. Germi- lished and apparently persistent grasses are A-68 Lehmann love- nation was considered complete when the seed radical was 0.5 cm grass (Erugrostis fehmonniuna Nees), cochise lovegrass (Erugrostis long. Dur.) plus lehmanniunu Nees X Erugrostis trichophoru Coss & Dishes were arranged in a stratified randomized block design A-84 and Catalina boer lovegrass (Erugrostis curvulu var. conferru because temperatures were constant at points across the thermo- Nees). A-68 Lehmann and A-84 boer lovegrasses were introduced gradient plate. One petri dish was used for each accession at each in 1930 (Crider 1945). Cochise lovegrass was introduced in 1961 temperature, and the experiment was repeated 6 times. Total ger- (Holzworth 1980) and Catalina boer lovegrass was selected for mination was determined by accumulating the number of germi- seedling drought tolerance (Wright and Jordan 1970) and released nated seed over the 14-day period. Germination values for the in 1969 (Wright 1971). respective accessions were compared at the same temperature with The amount and distribution of summer precipitation in the arid analysis of variance at days 6 and 12. When Fvalues were signifi- Southwest is sporadic, and a successful summer seeding may be cant, a Duncan’s new multiple range test (Steel and Torrie 1960) expected in 1 of 10 years (Cox and Jordan 1983). We have observed was used to separate accession means (EO.05). Catalina and cochise lovegrass seedling emergence following wet sequences in spring, fall, and winter and when seeds were sown prior to a dry summer; Cox and Martin (1984) have determined Results and Discussion that Catalina and cochise seedlings are more drought tolerant than Lovegrasses responded differently to variations in temperature A-68 and A-84 seedlings. If some lovegrasses will germinate at (Fig. 1). Germination of A-84 seeds began 2 days after water was applied and germination of A-68 seeds began 4 days after water The authors are graduate student, School of Renewable Natural Resources, Uni- was applied at 27 and 30” C. Germination of both A-68 and A-84 versity of Arizona, Tucson 85721; and range scientist. USDA ARS, Arid Land Ecosystems Improvement. 2000 E. Allen Road, Tucson, Ark 85719. seeds were similar in that germination began on day 4 at 24’ C, day Manuscript accepted December 21. 1983. 6 at 21° C, day 8 at 18’ C and between days 1I and 12 at 15’ C.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 507 LEHMANN LOVEGRASS A-68 IOO- COCHISE LOVEGRASS ‘OO[- - 30’c 90 - 27% go- _._._ 24.C *0- -.-.-._._ 21-c / ~yz-L_._._._ --- FJ 70. -- - IS’C 5 ----___ ,60C /,/y_._._._._‘-‘-‘-‘--

5 60- / / / E ,._’ j ; 50- ///’ i I- z 40- 1 Y _/ \ ;_._./ ___ “w JO- _-_-- a i’ ;’ /I /----- 20. -._. ,-,* /’ 1’ ,’ IO- ,./’ / ,’ I’ -- l f #’ I I. * I I I I I I I I I I, ’ I 2 3 4 5 6 7 6 9 IO II 12 I3 I4 ’ ; 2 3 4 5 6 7 6 9 IO II I2 13 I4 DAYS DAYS

IOO- BOER LOVEGRASS A-54 CATALINA LOVEGRASS r------90. go- / _-----

60. 80-

g 70- 5 5 60-

E g 50-

I- z 40-

30-

I I. 11 I I I. I1 I I I I ’ I 2 3 4 5 6 7 6 9 IO II I2 13 14 DAYS DAYS

Fig. 1. Germination (%) of 4 lovegrasses over 14 days at 6 constant temperatures. Germination of Catalina seeds began within 1 day at 18 to 30” C tion of A-68 and cochise seeds occurred at 27O C and optimum but was delayed 1 day at 15’ C. Germination of cochise seeds began germination of A-84 seeds occurred at 30° C within 12 days. within 1 day at 2 1 to 30’ C, but was delayed 3 days at 18’ C and 5 Germination of Catalina seeds was significantly (EO.05) greater days at 15’ C. than the remaining accessions from 15 to 24OC after 12 days (Table Germination of the lovegrasses generally increased as tempera- 1). Optimum Catalina germination occurred at 18’ C and this ture increased from 15 to 27’ C after 6 days (Table 1). Germination occurred in all petri dishes. Germination of cochise seeds was of Catalina and cochise seeds occurred at 1So C, but germination of intermediate and unchanged from 2 1 to 27” C, but A-68 germina- Catalina seeds increased 300% at 18’ C while the germination of tion peaked at 27” C while A-84 germination peaked at 30” C. cochise seeds was unchanged. Germination of Catalina seeds was The rapid and consistently high germination of Catalina seeds relatively uniform from 18 to 30° C. Germination of A-68 and A-84 across all temperatures suggests this accession is adapted over a seeds did not occur below 2 lo C within 6 days. Optimum germina-

Table 1. Mead germination (%) of four lovegrasses accessions at six constant temperatures after six and twelve days.

Temperature (” C) Accession Day I5 18 21 24 27 30 -%- A-68 6 3’ 34’ 45’ 33’ Cochise IS” 17b Sb 67b 72b 64b A-84 8’ 28’ 65b 7lb Catalina 19” 63” 81” 81a 84” 87’ A-68 I2 15’ 28’ 66’ 74b 84” 77b Cochise 50b 51b 79b 79b 79ab 68’ A-84 2d 13d 44d 54’ 72b 85” Catalina 72” 96’ 87” 87” 87’ - 88” ‘Each mean is the average of six replications of SO seed. ‘Means within columns at 6 and 12 days are no,t significantly different (8’9.05) when followed by the same superscript.

508 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 greater elevation gradient and can be expected to germinate follow- Cox, J.R., and G.L. Jordan. 1983. Density and production of seeded range ing both warm- and cool-season moisture in the Sonoran Desert. grasses in Southeastern Arizona (1970-1982). J. Range Manage. 36:649- Germination of cochise and A-68 seeds was inhibited at both 652. higher and lower temperatures and this might suggest that cochise Cox, J.R., and M.H. Martin. 1984. Effects of planting depth and soil and A-68 seeds should be planted in either fall or spring. This might texture on the emergence of four lovegrasses. J. Range Manage. 37:204- 205. also explain why A-68 has persisted in the Sonoran Desert when Crider, F.J. 1945. Three introduced lovegrasses for soil conservation. the probability of fall-spring precipitation is greater than in the USDA, Circular No. 730. Chihuahuan Desert. Seeds of A-84 seed should probably be Holzworth, L.K. 1980. Registration of “Cochise” atherstone lovegrass. planted only in summer, and seeding limited to areas where Crop Sci. 20:823-824. summer precipitation exceeds 20 cm (Cox et al. 1982). Larsen, A.L. 1962. Two-way thermogradient plate for seed germination research. USDA-Agricultural Research Service, ARS 51-41. Literature Cited Steel, R.G.D., and J.H. Torrie. 1960. Principles and procedures of statis- Cox, J.R., H.L. Morton, T.N. Johnsen, Jr., G.L. Jordan, SC. Martin, and tics. McGraw Hill, New York. L.C. Fierro. 1982. Vegetation restoration in the Chihuahuan and Sono- Wright, N.L., and G.L. Jordan. 1970. Artificial selection for seedling ran Deserts of North America. USDA-Agricultural Research Service, drought tolerance in Boer lovegrass (Eragrosris curvula Nees). Crop Sci. ARM-W-28. 10:99-102. Wright, N.L. 1971. Registration of Catalina weeping lovegrass. Crop Sci. 11:939. Seed Pretreatments and Their Effects on Field Establishment of Spring-Seeded Gard- ner Saltbush

R. JAMES ANSLEY AND ROLLIN H. ABERNETHY

Abstract Gardner saltbush [Atriplex gurdneri (Moq.) D. Dietr.] seeds regions of the Intermountain West. It is particularly adapted to collected from the Red Desert Basin of Wyoming were subjected to saline, alkaline, and clay soil conditions, extreme temperatures and pretreatments of scarification (Se), washing (W), and stratification aridity (Stubbendieck et al. 1981). It has use on rangeland and (St) to alleviate dormancy. Laboratory germination was evaluated. disturbed lands because it is a soil stabilizer and provides valuable Subsequently, seedling vigor was observed by determining field winter browse. emergence of similarly pretreated seeds spring planted at 1 irri- Direct seeding of Gardner saltbush and other native arid land gated and 2 dryland sites in Wyoming. Effects of l-cm and 3-cm shrub species on disturbed lands has often met with limited success planting depths on emergence were also evaluated. (Bleak et al. 1965, Nord et al. 1971, Sindelar et al. 1974, Plummer Seed was pretreated, then dehydrated with minimal impact on seed 1976, McKell 1979, DePuit and Coenenberg 1980). Reasons for germination. Field emergence was much less than laboratory ger- this can be attributed to seed related factors including low purity mination for all treatments at all sites, indicating that establish- seed sources, low viability, seed dormancy, and poor seedling ment for this species is related to poor seedling vigor as much as to vigor, in addition to inimical environmental conditions. seed dormancy. Moreover, when compared to untreated controls, Much recent research has contributed to the understanding of relative responses to seed pretreatments often differed between dormancy and germination requirements of native arid land shrub laboratory and field trials. In the laboratory SC = W = St provided seeds (McDonough 1970, Wood et al. 1976, Eddleman 1978, Sabo the greatest germination, whereas the best seed pretreatment for et al. 1979, Young et al. 1980, Young et al. 198 I, Young and Evans field establishment was SC + St. Washing had little effect on 1981) including Gardner saltbush (Foiles 1974, Eddleman 1978, enhancing field emergence and appeared to inhibit effects of St in Ansley 1983). Most of this research has been restricted to labora- scarified seed. The most effective planting depth varied with clima- tory germination studies. These studies are necessary because tic/edaphic severity of the site. establishment of a species from direct seeding would be impossible if seed dormancy was not first broken. However, an equally impor- Gardner saltbush (Atriplex gardneri (Moq.) D. Dietr.) is a low tant area of research involves post-germination events, including growing half-shrub (20-50 cm high) which occurs in cold desert seedling emergence from soil and subsequent survival under natu-

Authors are graduate assistant and assistant professor, Plant Science Division, ral conditions. As Osmond et al. (1980) stated: “. . . it is the University of Wyoming, Laramie 82071. Senior author is currently located at Texas establishment phase rather than germination which is more Agricultural Experiment Station,, Vernon 76384. vulnerable under natural conditions, and if this is so, the detailed This article is Wyoming Agncultrual Experiment Station Journal Article No. JA-1264. physiology and biochemistry of germination is of somewhat aca- Research was supported by funds from the University of Wyoming Agricultural demic interest.” Experiment Station and the High Plains Grassland Research Station, through USDA-ARS Coop. Agreement No. 58-9AHZ-9-467. Spring seeding is recommended for Gardner saltbush (Eddle- Manuscript accepted December 13. 1983.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 509 man 1978, Vories 1981). However, this species has a substantial combination treatments (i.e. Sc+W, Sc+St, W+St and Sc+W+St), stratification requirement (Ansley and Abernethy 1982). Young seeds were initially scarified, then washed, air dried 24 hours, and and Evans (1981) noted that once seeds of most species are strati- finally rehydrated and stratified. Nontreated controls were includ- fied they must be planted immediately in a hydrated state for the ed for all comparisons. germination enhancement to be effective and to retain viability. To determine ovendry weight loss due to the scarification Seeding hydrated seeds is impractical for a variety of reasons. treatment, fresh seeds were weighed, scarified, oven dried at 100°C Therefore, fall seeding of species with a stratification requirement for 24 hours, and reweighed. These were then compared on an oven is the common practice on revegetation sites. However, problems dry weight per seed basis to a set of unscarified seeds which had encountered with fall seeding include seed loss from runoff, envi- been weighed, oven dried, and reweighed. To determine oven dry ronmental exposure, loss of viability through time, fall emergence, weight loss from the wash treatment, scarified and unscarified and rodent predation. seeds were weighed while in the fresh condition, then enclosed in Little is known regarding seedling vigor in Gardner saltbush. cheesecloth and exposed to running tap water 24 hours. Following Seeding failures have been attributed to poor seedling vigor in washing, seeds were removed from the cheesecloth, oven dried, and some studies but the distinction between germination failure and again weighed. events relating to post-germination emergence from soil (i.e. seed- Laboratory Germination ling vigor) was not made (McLean 1953, Nord et al. 1971). Pretreated 8 month post-harvest seeds were germinated from The first objective of this study was to determine if artificially April to May, 1982, in a germinator cycled at 24°C-light (16 stratified seeds could be dehydrated immediately after treatment hours)/ 13OC-dark (8 hours). Each treatment was replicated 6 and subsequently spring planted without adversely affecting times. Each replication consisted of a IOXlOX2.5-cm plastic box emergence. and tightly fitting lid containing 100 seeds on one sheet heavy The second objective of this study was to characterize seedling blotter with an initial application of 10 ml distilled water. Germina- vigor in Gardner saltbush by (a) comparing laboratory blotter tion, as indicated by 15-mm radicle extension, was assessed at germination and field emergence of seed pretreated to alleviate 5-day intervals for 35 days. dormancy and (b) determining effects of planting depth on field emergence. Post-germination emergence from the soil and first establish- Field Emergence ment were considered in this study to reflect “seedling vigor.” Concurrent with the laboratory studies, seeds were similarly pretreated and spring planted at 2 seeding depths and at 3 different Methods and Materials sites in Wyoming: Seed Procurement (1) on April 7 at a topsoiled coal mine reclamation site oper- All seeds used in this study were collected within a 4-ha area 10 ated by Bridger Coal Co., 30 km west of Wamsutter, km west of Wamsutter in South Central Wyoming on August 4, (2) on April 28 at a topsoiled bentonite mine reclamation site 1981. The elevation and average annual precipitation are 2,100 m operated by Wyo-Ben Inc., 12 km east of Thermopolis, and and 1S-18 cm, respectively. Average mean monthly temperatures (3) on June 9 at University of Wyoming research plots in range from -2°C to 20°C with extremes ranging from -39OC in Laramie. January to 41°C in July (Becker and Alyea 1964). Vegetation at the Treatments were arranged in a split-plot design with seed plant- site consists of predominantly Gardner saltbush interspersed with ing depths as main plots and seed pretreatments (i.e. SC, W, St) bud sagebrush (Artemisia spinescens D.C. Eaton), bottlebrush arranged in a completely randomized 2X2X2 factorial as subplots. squirreltail (Sirunion hystrix Nutt.), and indian ricegrass (Oryzop- Wash and stratification treatments were conducted in the labora- sis hymenoides Roem. & Schult.). tory and seeds were air dired at 21-24” C for 16 to 24 hours prior to Seeds were collected by hand stripping the flower stalks (Eddle- seeding. Effects of various periods of air drying at 2 1”C following man 1978). Freshly harvested seeds were spread on a canvas tarp to 24 hour wash and 3 week stratification treatments on Gardner air dry (Young et al. 1978). Impurities were removed via screening saltbush germination on blotter paper were concurrently observed. and by a “Dakota”adjustable plexiglass column blower (Young et A single row cone seeder with a knife furrow opener was used for al. 1978). Seeds were stored in paper bags at 20-24“C (Springfield seeding. Two seeding depths, 1 cm and 3 cm, were evaluated. 1970, Foiles 1974). Each treatment replication consisted of 350 seeds planted in a single row 6.4 m long. Treatments were replicated 4 times at Seed Fill and Viability Bridger, 8 at Thermopolis, and 5 at Laramie. Replication number The female Gardner saltbush flower has no perianth, protection was dependent on the size constraints of each site. Seedlings were being provided by 2 bracteoles 3-6 mm long which form a false fruit counted periodically at each site from June to September, 1982. or utricle (Young et ai. 1980). In our study the utricle was consi- Seedlings were considered emerged when the first pair of (non- dered the”seed”(Nord et al. 1971, Younget al. 1980). Seed fill was cotyledon) leaves were observed. determined by slicing 15 replicates of 100 utricles with a razor The reclamation sites were seeded in April to take advantage of blade. Viability was determined by soaking embryos bisected dur- soil moisture accumulated during winter. The sites were not irri- ing the seed fill test in a 0.1% 2,3,5-triphenyl-2H-tetrazolium chlo- gated after seeding. The Thermopolis site was fenced to exclude ride (TZ) solution for 4-8 hours (Grabe 1970, Weber and Wiesner sheep and deer, which are prevalent in the area. The Bridger site 1980). Five replicates, each containing 25 embryos, were used for was unfenced. It was assumed mining activity less than 300 m from the TZ test. the research site would preclude large herbivores from grazing on Seed Pretreatments the site. Rodents and other small herbivores as well as weeds were Seed pretreatments used in this study were previously deter- not controlled at either of the reclamation sites. mined to provide varying levels of increase in laboratory blotter To observe seedling emergence under conditions which might be germination and included: scarification (SC), washing (W), and considered least-limiting, the Laramie site was seeded later in the cold stratification (St), arranged in all combinations. Seeds spring on June 9 to avoid colder soil temperatures. One centimeter occupying a volume of 250 cm3 were scarified for 20 seconds in a of water was sprinkler applied every 5 days for the first 30 days Forsberg scarifierr at 1,725 RPM. Washing was conducted in after seeding. Large herbivores, rodents, and weeds were excluded. flowing tap water (flow rate: 35-40 ml per set; 6’C) for 24 hours. Data from laboratory and field studies were subjected to analy- Washed seeds were then air dried (21OC) for 24 hours. Stratifica- sis of variance procedures. Duncan’s multiple range test or the tion involved maintaining imbibed seeds at 2°C for 3 weeks. For LSD (m.05) was used to test differences between means.

‘Forsbergs, Inc., Thief River Falls, Minn. 56701.

510 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Results and Discussion affect germination enhancement from stratification (St) and stratification + washing (St + W) of scarified seed (Fig. 1). Air drying up !3eedFill and Viability to 240 hours did not adversely affect germination enhancement Average seed fill was 48%. Viability was difficult to determine from washing (W) of scarified seed. In unscarified seed, air drying using the TZ test. Some staining occurred on 94% of the embryos was not detrimental to effects of St, W, and St + W, although tested, which would suggest a maximum pure live seed (PLS) of germination under these treatments was not, in some cases, signifi- (.94 X .48) 45%. However, in most of those embryos, staining was cantly greater than in the untreated control. nonuniform with portions (i.e., radicle, hypotocyl, cotyledons) This laboratory study demonstrated that germination enhance- being either dark red, pink, or unstained. Blotter germination tests ment obtained from stratification and wash treatments was conducted 17 months post-harvest on this seed source provide 90% retained by scarified seed allowed to air dry. This implies that germination of filled seed or (90/94) 96% of PLS (Ansley 1983). pretreatments would not have to be applied immediately prior to Therefore it was assumed that embryos with some staining were planting to retain their benefit. viable. Only those embryos with no staining were classified as nonviable. Pretreatment Effects on Blotter Germination Weber and Wiesner (1980) noted it is often difficult to achieve Laboratory blotter germination data included in Table 1 were adequate TZ staining in embryos of many native forbs and shrubs. transformed so that comparisons between blotter germination and Further research is needed to more accurately evaluate TZ embryo field emergence could be made on a “seedlings per seeds treated” staining in Gardner saltbush. rather than a percentage basis. Scarification alone did not enhance germination (Table 1). However, scarification appeared to facilitate effects of W, St, and SEEDS NOT SCARIFIED W + St. Differences between St and SC + St were not significant at m.05, although trends were apparent. s EDS In the laboratory, stratification alone significantly enhanced SCAE IFIED I germination over the control (Table 1). Moreover, all combination I I LSD*05 treatments which included stratification significantly enhanced germination on blotter paper. Washing alone increased germination, although this difference was not statistically significant at m.05 (Table 1). The combination “scarification-wash-stratification” (SC + W + St) yielded the greatest germination. However, even with this treatment, germination was only 48.8% of the PLS treated, suggesting that the afterripening requirement was not fully met (Ansley 1983). These results agree with other works in showing that stratification and washing (i.e. leaching) increase germination in arid shrub- land Arriplex spp. (Beadle 1952, McLean 1953, Cornelius and Hylton 1969, Eddleman 1978). Scarification alone was not effective, which agrees with results of Springfield (1970) and Graves et 24 48 COI&.OL soTRA~.4 tS?l WASH al. (1974) in scarification tests on fourwing saltbush [A. canescens HOURS AIR DRYING FOLLOWINGTREATMENT (Pursh) Nutt.]. Other authors have found scarification to be effective in some Arriplex spp. (Nerd and Whitacre 1957, Edgar and Fig. 1. Effects of various drying periods (2P C)fol/owing 3 weeks stra@- Springfield 1977). Ansley (1983) found that germination response cation (St), 24 hours washing (W), and stratifiearion + washing (St + W) of Gardner saltbush seeds to scarification was related to seed on blottergermination ofunscarifiedandscarifiedSmon?hspost-harvest source. Little research has been done on interactive effects of Gardner sahbush seeds. stratification, leaching, and scarification in Atriplex spp. Effects 01 Air Drying Following Wash and Stratification Pretreatment Effects on Field Emergence Results indicated that up to 48 hours air drying did not adversely For all treatments, germination on blotter paper was much

Table 1. Effects of seed pretreatments and planting depth on field emergence of Gardner saltbusb at one irrigated site (Lammie) and two dryland reclamation sites (Thermopolis and Bridger). All seed treated was g-10 months post-harvest.

Average Number of Seedlings’ Laboratory Planting depth blotter Laramie Thermopolis Bridger Seed pretreatment2 germination I cm 3 cm I cm 3 cm I cm 3cm I. Control 25 d I d’ 2b 2d Id Oa I ab 2. SC 29 d 3 cd 2b 5 bed 5b la 0.5 b 3. w 38 cd 2d 2b 3 cd 2 cd 0.5 a lb 4. SC+ w 58 b 6 cd 4b 4 cd 4b 0.5 a 2 ab 5. St 50 bc 15 b II a 5 bc 5b 0.5 a 2 ab 6. SC + St 64 ab 28 a II a 9a 8a 0.5 a 2 ab 7. w + St 51 bc I4 b 9a 5 bed 4 bc 0.5 a 2 ab 8. SC + W + St 78 a IO bc 9a 8 ab 5b la 4a

‘Each mean based on turner of seedlings blotter germinated or field emerged per 39 seeds (160 pure live seeds) treated. Seeds were germinated for 5 weeks on germination blotter. Emergence data were obtained August 25-Sept. I at Laramie and Thermopolis, I I and 17 weeks after seeding, respectively. Data were obtained July 27 at Bridger, 16 weeks after the seeds were planted. Treatments: Control = untreated seed, W = 24 hour wash, St = 3 week cold stratification, SC = 20 second scarification. All pretreatments allowed to air dry 24 hr prior to germination or planting. ‘Means within each co/umn having similar letters are not significantly different at EO.05 according to Duncan’s new multiple range test.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 511 greater than emergence at any field site, including Laramie where have reduced water availability to seeds at Bridger (Ansley 1983). growing conditions were least-limiting. The greatest field emer- At Thermopolis, relative responses among seed pretreatments gence for any treatment at any site was 28 of 160 PLS planted generally resembled results at the Laramie site, with greatest emer- (Table I), or 17.5%, at Laramie. This was about one third the gence occurring in the SC + St treated seed at 1cm. Unlike Laramie highest laboratory blotter germination. These results suggest that washing did not inhibit effects of stratification in scarified seed at 1 limited field establishment for this species is related to poor seed- cm. There was some inhibition at 3 cm, however. ling vigor as well as to seed dormancy. Seedling vigor responses of a seed lot can be due both to the genetic makeup of the parent seed plants and to the environment prevalent at the time the seed was 1 CM SEED developing. Seedling vigor has been related to a number of morphological and/or physiological factors in various crop species (McDonald 1980) but little is known about seedling vigor of wildland shrubs. Emergence in this study was much lower than that Nord et al. (1971) reported for Gardner saltbush in California. At the Laramie irrigated site relative responses among pretreatments were similar to laboratory results in that all pretreatments including stratification enhanced emergence for both seed depths (Table I). Moreover, scarification appeared to facilitate stratification effects at l-cm seed depth but had little effect itself on emergence. Response to washing was markedly different between field emergence at Laramie and blotter germination (Table 1). Unlike blotter germination, washing alone or with scarification had no significant effect at the Laramie site at either seed depth. Moreover, washing appeared to inhibit effects of stratification in scarified seed, in contrast to the laboratory, where SC + W + St slightly LARAMIE THEBNOPOLIS BRIDGER increased germination over SC+ St. An explanation for emergence PLANTINGSITE response to the wash treatments at Laramie is not readily apparent. Fig. 2. Effect of planting depth on 8-10 month post-harvest Gardner The amount of oven-dry material lost from both unscarified and solrbush seedling emergence oi 3 locations overoged over all prerreot- scarified seed during the 24-hour wash treatment was greater than merits. Bors represent the LSD.05. 10% of oven-dry seed weight (Table 2). Loss of this material, which Results from Bridger were insufficient to make many conclu- Table 2. Effects of scarification and 24 hour wash treatments on fresh and sions about effects among pretreatments (Table 1). It appeared SC oven dry weight of Gardner saltbusb seeds. + W + St at 3 cm was more effective than other seed pretreatments. Effects of Seeding Depth Weight per Seed Emergence, when averaged over all seed pretreatments, was (mg) significantly different between the 2 seed depths at both Laramie Percent weight and Bridger (Fig. 2). At Laramie, significantly more seedlings Unscarified Scarified loss from Wash treatment seed seed scarification emerged from I cm than from 3 cm. Conversely, at Bridger, significantly more seedlings emerged from 3 cm than 1 cm. Thermopolis -Fresh Weight- had no real difference in emergence between the 2 seed depths. It No Wash 2.59 2.16 16.9 was difficult to prepare a smooth seedbed at Thermopolis due to --Oven-Dry Weight- the silty/clayey topsoil texture (Ansley 1983). The uneven, often No Wash 2.46 2.02 17.9 cloddy surface may have precluded accurate regulation of planting 24-Hour Wash 2.14 1.78 depth at this site. Seedbeds were smooth at both Laramie and Percent Oven Dry Bridger at the time of seeding. Weight Loss Results from this study suggest that Gardner saltbush emergence From Washing 13.2 11.8 is greater from l-cm than 3-cm planting depth when soil moisture is adequate. However, on extremely arid sites, such as Bridger, dry- could be largely perisperm reserves, combined with 17.9% oven- land seeding may be more successful if seeds are planted deeper dry weight loss due to scarification did not inhibit blotter germina- than I cm. These results agree with Nord et al. (1971), who found tion but may have been deleterious to proper field seedling devel- that Gardner saltbush emergence was greater from I .3 cm than 2.5 opment from scarified seed. cm when seeds were planted in February and March. At an April At the dryland reclamation sites emergence was much lower planting date, emergence from 1.3 cm decreased from that than at the Laramie irrigated site for both seed depths (Table I). obtained by the earlier planting date, while emergence from 2.5 cm The greatest emergence for any treatment at Thermopolis was 9 of remained unchanged. Soil moisture availability may have decreased 160 PLS planted, or 5.6%. Greatest emergence at Bridger was 4, or more at 1.3 cm than 2.5 cm soil depth from February to April. 2.5% of PLS planted. This compares to 17.5% emergence at Lara- mie and 48.8% germination on blotter paper. Thus, the effect of Conclusions poor seedling vigor suggested by the Laramie field results was Seed pretreatments used in this study enhanced both laboratory accentuated under progressively more limiting climatic/edaphic blotter germination and field emergence of Gardner saltbush seeds. conditions at Thermopolis and Bridger. However, even under relatively nonlimiting growing conditions, Emergence from both seed depths was greater at Thermopolis field emergence was much lower than laboratory blotter germina- than at Bridger even though both sites were not irrigated following tion, indicating that this species has poor seedling vigor in addition seeding (Table I and Fig. 2). The Thermopolis site received IO cm to seed dormancy. Moreover, when compared to the untreated precipitation during the first 2 months following seeding while control, relative response to seed pretreatments often differed Bridger received 4 cm precipitation. Moreover, topsoil at Bridger between germination on blotter paper and field emergence studies. was coarser textured than topsoil at Thermopolis, which might Spring-seeding artificially stratified Gardner saltbush seeds was

512 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 successful. Although emergence of stratified seeds was low relative MeDonough, W.T. 1970. Germination of 21 species collected from a to total PLS planted, it was substantial relative to untreated con- high-elevation rangeland in Utah. Amer. Midl. Natur. 4:551-554. trols at both Laramie and Thermopolis. McKell, C.M. 1979. Selection, propagation and field establishment of Previously recommended planting depth for Gardner saltbush is native plant species on disturbed arid lands. Utah Agric. Exp. Sta. Bull. 1.3 cm (Vories 198 I). McLean (1953) found emergence of Nuttall 500, Utah State Univ., Logan. McLean, A. 1953. The autecology of Atriplex nuttallii S. Wats. in south- saltbush declined rapidly when seeds were planted below 1.3 cm. western Saskatchewan. MS. thesis. Utah State Agr. Coll., Logan. We found that emergence was significantly greater from 1 cm than Nord. , E.C.. and J.E. Whitrcre. 1957. Germination of fourwine saltbush 3 cm when moisture was adequate. However, deeper planting improved byscarificationand grading. USDA Forest Serv., &if. Forest depths may have merit on extremely arid sites. and Range Exp. Sta. Res. Note. 125. Nord, E.C., P.F. Hartless, and W.D. Nettleton. 1971. Effects of several Literature Cited factors on saltbush establishment in California. J. Range Manage. Ansley, R. James. 1983. Dormancy, germination, emergence and ecology 24216-223. of Gardner saltbush [Afriplexgurdneri (Moq.) D. Dietr.] seeds. Ph.D. Osmond, C.B., 0. Bjorkman, and D.J. Anderson. 1980. Physiological Diss., Univ. Wyoming, Laramie. processes in plant ecology-toward a synthesis with Atriplex. Springer- Ansley, R. James, and R.H. Abernethy. 1982. Seed dormancy and germi- Verlag, New York. nation enhancement of Gardner saltbush (Atriplex gardneri). Abstr. p. Plummer, A. Perry. 1976. Shrubs for the subalpine zone of the Wasatch 23. In: Proc. Sot. Range Manage. 35th Ann. Meeting, Calgary, Alberta. Plateau. p. 33-40. In: Proc. High Altitude Revegetation Workshop No. Beadle, N.C.W. 1952. Studies in halophytes. I. The germination of the seed 2. R.H. Zuckand L.F. Brown(eds.). Colorado State Univ., Fort Collins. and establishment of the seedlings of five species of Arriplexin Australia. Sabo, D.G., G.V. Johnson, W.C. Martin, and E.F. Aldon. 1979. Germina- Ecology 33:49-62. tion requirements of I9 species of arid land plants. USDA Forest Serv. Becker, C.F., and J.D. Alyea. 1964. Temperature probabilities in Wyom- Res. Pap. RM-210. USDA Forest Serv., Rocky Mt. Forest and Range ing. Univ. Wyo. Agr. Sta. Bull. 415. Exp. Sta., Fort Collins, Colo. Bleak, A.T., N.C. Frischknecht, A.P. Plummer, and R.E. Eckert, Jr. 1965. Sindelar, B.W., R. Atkinson, M. Majerus, and K. Proctor. 1974. Surface Problems in artificial and natural revegetation of the arid shadscale mined land reclamation research at Colstrip Montana. Montana Agr. vegetation zone of Utah and Nevada. J. Range Manage. 1859-63. Exp. Sta. Res. Rep. 69. Bozeman. Cornelius, D.R., and L.O. Hylton. 1969. Influence of temuerature and Springfield, H.W. 1970. Germination and establishment of fourwing salt- leachate on germination of kriplexpolycarpa. Agron. J. bl:209-21 I. bush in the Southwest. USDA Forest Serv. Res. Pap. RM-55. Rocky DePuit, E.J., and J.G. Coenenberg. 1980. Establishment of diverse native Mountain Forest Range Exp. Sta. Fort. Collins, Colo. plant communities on coal surface-mined lands in Montana as influ- Stubbendieek, J., S.L. Hatch, and K.J. Kjnr. 1981. North American Range enced by seeding method, mixture and rate. Mont. Agr. Exp. Sta. Res. Plants. Nat. Resources Enterprises, Inc., Lincoln, Neb. Pap. 163. Montana State Univ., Bozeman. Vories, K. 1981. Growing Colorado plants from seed: a state of the art. Vol. Eddleman, L.E. 1978. Survey of viability of indigenous grasses, forbs and I: Shrubs. USDA Forest Serv. Gen. Tech. Rep. INT-103. Intermtn. shrubs. Annu. Prog. Rep. prepared for US Energy Res. and Develop. Forest and Range Exp. Sta., Ogden, Utah. Admin. Weber, G.P., and L.E. Wiesner. 1980. Tetrazolium testing procedures for Edgar, R.L., and H.W. Springfield. 1977. Germination characteristics of native shrubs and forbs. J. Seed Tech. 5:23-24. Broadscale: a possible saline-alkaline site stabilizer. J. Ranae Manaae. Wood, M.K., R.W. Knight, and J.A. Young. 1976. Spiny hopsage germi- 301296-299. - nation. J. Range Manage. 29:53-56. Foiles, M.W. 1974. Atriplex L., Saltbush. p. 240-243. In: Seeds of woody Young, J.A., and R.A. Evans. 1981. Germination of seeds of antelope plants in the U.S. USDA Aar. Handb. 450. Washineton. D.C. bitterbrush, desert bitterbrush, and cliff rose. USDA Science and Educ. C&be, Don F. 1970. TetrazolLm testing handbook fo;agricultural seeds. Admin. ARR W-17. Contrib. No. 29 to Handbook on Seed Testing. Ass. Off. Seed. Anal., Young, J.A., B.L. Kay, H. George, and R.A. Evans. 1980. Germination of Corvallis, Ore. three species of Atriplex. Agron. J. 72:705-709. Graves, W., B.L. Kay, and W.A. Williams. 1974. Seed-treatment studies of Young, J.A.,R.A. Evans, R. Stevens,and R.L. Everett. 1981. Germination seven Mohave Desert shrub species. In: Test of seeds of Mohave Desert of Kochia prosrrata seed. Agron. J. 731957-96I. shrubs. Prog. Rep., Dep. Agron. and Range Sci., Univ. Calif., Davis. Young, J.A., R.A. Evans, B.L. Kay, R.E. Owen, and F.L. Junk. 1978. Mimeo. Collecting, processing, and germinating seeds of western wildland plants. McDonald, M.B. Jr. 1980. Assessment of seed quality. Hon. Sci. USDA Science and Educ. Admin. Agr. Rev. and Manuals. ARM-W-3. I5:784-788.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 513 Leaf Area, Nonstructural Carbohydrates, and Root Growth Characteristics of Blue Grama Seedlings

A.M. WILSON

Abstract Establishment of blue grama [Boutelouu gracilis (H.B.K.) Lag 35OC. Maximum midday _fhotosynthetic photon flux density ex Steud.] seedlings requires extension of adventitious roots into (PPFD) was 1,650 pmole m set-’ in early summer and 850 pmole the soil profile. The objectives of this study were to determine the me2 set-’ in late summer. effects of leaf area and total nonstructural carbohydrates (TNC) on Plastic pots (15 cm diameter by I5 cm deep) were filled with root growth characteristics of blue gmma. Seedlings supported by 1,800 g of sterilized (100°C dry heat for 2 days) sandy loam soil the seminal root only were treated with 3 days of reduced light and (fine-loamy, mixed, mesic Aridic Argiustoll). The soil was surface then with 0, 1,2, and 3 days of full sunlight to alter TNC percentage irrigated with 250 ml of water, and 25 ‘Lovington’blue grama seeds in crowns. Seedlings within each of these treatments were then were planted at a depth of 2 mm in the moist soil. Seeds were then clipped at a height of 3,6,9, and 12 cm, or left unclipped to alter covered with 2.5 cm of air-dry soil. Seedlings emerged through the leaf area. Adventitious root growth was studied during a j-day test. dry soil layer within about 5 days. Path coefficients indicating the effects of leaf area on number of Pots were weighed on alternate days and the amount of water roots per seedling, depth of roofs, and root weight per unit length needed to bring the lower 1,800-g of soil to field capacity was (diameter) were 0.72, 0.47, and 0.77, respectively. The TNC had placed in a petri dish. Water moved into the soil through holes in smaller effects on root growth than did seedling leaf area. Clipping the bottom of the pot. The subirrigation procedure maintained a treatments probably reduced root growth because of a deficiency moist subsoil and a dry soil surface which promoted seedling of photosynthetic products. But, the reduction was explained by an growth but prevented growth of adventitious roots. Thus, seed- adjustment in all components of growth rather than in root depth lings were supported by the seminal root only during this phase of only. Thus, blue grama seedlings maintained a reasonable rate of the study. root elongation even under severe clipping treatments. At 3 weeks after planting, pots were thinned to 8 vigorous, well-spaced seedlings. At 5 weeks, seedlings were exposed to 3 days Establishment of blue grama [Bouteloua grucilis (H.B.K.) Lag. of shade and then to 0, 1,2. or 3 days of sunlight to create various ex Steud.] seedlings requires the initiation and the extension of levels of total nonstructural carbohydrates in seedling crowns. adventitious roots (Wilson and Briske 1979). That process depends Average midday PPFD in shade was 240 pmole m-’ sec.-‘. After on seedling leaf area (Wilson 198 I), total nonstructural carbohy- shade and light treatments, the pots were separated into 5 groups drates (TNC), tolerance of dehydration (Briske and Wilson 1980, and seedlings were clipped at heights of 3,6,9, or 12 cm. The fifth Khan 1980), and favorable environmental conditions (Hyder et al. group of seedlings was left unclipped. Four seedlings in each pot 1971). were randomly sampled to determine the weight of shoot removed The relationships among clipping treatments, TNC, rate of root by clipping and the weight of shoot remaining after clipping. elongation, and total dry weight of roots have been studied by Leaf-blade area remaining after clipping was also determined. The several investigators (Booysen and Nelson 1975, Buwai and Trlica lower 3-cm portion of stem base was dried at 60°C and used for 1977, Crider 1955, Hansen 1978, Smith 1974, Youngnerand Nudge determination of percent TNC (Association of Official Agricultu- 1976). Little information is available, however, on the possible ral Chemists 1965, Heinze and Murneek 1940, Smith et al. 1964). effects of clipping and TNC on other components of root growth Measurements before the root growth test were based on a compo- (Parker and Sampson 1930). The objectives of this study were to site sample of 12 seedlings harvested from 3 pots. determine effects of seedling leaf area (clipping treatments) and Immediately after clipping, the remaining seedlings in each pot crown TNC percentage (light treatments) on 3 components of root were surface irrigated to promote growth of adventitious roots growth in blue grama seedlings: number of roots, root length during a 3-day test. Shade, light, and clipping treatments were (depth), and root weight per unit length (diameter). Information scheduled so that the root growth test could be started on the same on root growth characteristics associated with clipping and envir- day for all treatments. After 3 days with a moist soil surface, onmental stress will aid in seedling establishment and in managing seedlings in each pot were harvested, adventitious roots were blue grama stands. counted, the length of each root (main axis) was measured, and roots were oven-dried and weighed. Root weight per unit length Materials and Methods was estimated from the total length of the main axis of all roots in a The effects of light and clipping treatments on the components sample divided by the dry weight of the sample. There was little or of adventitious root growth were investigated on blue grama seed- no branching of adventitious roots at this stage of development. lings under greenhouse conditions during June through Sep- Leaf blades were removed and leaf area was measured. Measure- tember. Air temperatures in the greenhouse varied from 25 to ments after the root growth test were based on a composite sample of 6 seedlings harvested from 3 pots. The author, now deceased, was plant physiologist, USDA ARS, Crops Research The study was conducted as a randomized complete-block. Fac- Laboratory, Colorado State University, Fort Collins 80523. This study was a cooperative investigation of USDA Agricultural Research Service tor I represented 4 light treatments and factor 2 represented 5 and Colorado State University Experiment Station. Scientific Paper 2862. clipping treatments. The study included 6 replications which The technical assistance of C.J. Carren, D.S. Edyvean, D.A. Nason, and B.L. Oskroba during this study is gratefully acknowledged. represented different dates of growth in the greenhouse. Analysis Manuscript accepted March 7, 1984. of variance was used to test for differences among treatments. Path

514 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Table 1. Effects of days of sunlight after shade and height of clipping on leaf am per seedling (cm’) before and after the 3&y root growth test.

Clipping height (cm) Days of sunlight 3 6 9 12 Unclipped control Mean Before root growth test 0 1.55 4.51 6.86 8.52 10.95 6.49’ 1 1.34 4.15 6.64 8.82 11.58 6.51 2 1.59 4.23 6.61 8.49 11.43 6.46 3 1.31 4.04 6.48 8.72 II.32 6.37 Mean I .4s 4.25 6.65 8.63 11.32 Pooled standard error (si) = 0.59 After root growth test 0 3.32 6.83 10.20 10.95 13.07 8.87’ 1 3.13 6.08 10.17 11.32 12.24 8.71 2 3.74 7.00 9.30 13.06 12.58 9.13 3 3.94 7.57 9.40 12.18 12.24 9.06 Mean 3.692 6.87 9.71 II.88 12.53 Pooled standard error (si) = 0.67

‘Within each sampling period (before or after root growth test), differences among light treatments were not significant (KO.05). 2Within each sampling period (before or after root growth test), differences among clipping treatments were significant (KO.01).

Table 2. Effects of days of sunlight after shade and height of clipping on the average number of adventitious roots and the length of longest adventitious root (cm) produced per seedling during the 3-day test.

Clipping height (cm) Days of sunlight 3 6 9 12 Unclipped control Mean Number of roots per seedling 0 11.3 18.4 21.4 24.0 27.3 20.5’ 1 13.0 18.3 21.2 23.2 25.0 20.1 2 13.9 20.6 22.2 21.4 26.6 22.2 3 14.8 22.2 24.3 21.4 30.2 23.8 Mean 13.22 19.9 22.3 25.5 27.3 Pooled standard error (si) = 1.3 Length of longest root per seedling 0 5.63 7.48 8.97 8.90 9.85 8.17’ I 5.80 7.72 8.83 9.92 9.83 8.42 2 6.67 8.20 9.22 9.62 9.98 8.74 3 6.40 8.53 9.13 9.65 9.87 8.72 Mean 6.122 7.98 9.04 9.52 9.88

Pooled standard error (Q q 0.25

‘Differences among light treatments in number of roots and length of longest root per seedling were significant (cUO.01). zDiffcrences among clipping treatments in number of roots and length of longest root per seeding were significant (KO.01).

Table 3. Effects of days of sunlight after shade and height of clipping on the weight per unit length of adventitious roots (Ccg/cm)and the total weight of adventitious roots (mg/seedllng) produced during the May test.

Clipping height (cm) Days of sunlight 3 6 9 12 Unclipped control Mean Weight per unit length of adventitious roots 0 145 137 160 164 187 159’ I 124 144 159 181 197 161 2 137 149 189 184 216 175 3 136 152 176 172 208 169 Mean 1362 146 171 175 202 Pooled standard error (si) q 7 Total weight of adventitious roots per seedling 0 6.2 12.9 20.5 22.6 34.8 19.4’ 1 6.4 13.6 20.3 28.3 34.6 20.6 2 8.0 17.1 26.5 32.2 40.4 24.8 3 9.1 20.2 21.2 32.8 42.1 26.3 Mean 7.42 16.0 23.6 29.0 38.0 Pooled standard error (si) = 2.2 ‘Differences among light treatments in weight per unit length of adventitious roots and total weight of adventitious roots per seedling were significant (KO.01). ZDifferences among clipping treatments in weight per unit length of adventitious roots and total weight of adventitious roots per seedling were significant (KO.01).

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 515 coefficient analysis was used for evaluating possible cause and roots per seedling (~4s q 0.132) and root weight per unit length @41q effect relationships among variables associated with adventitious 0.212). The effect of TNC on length of longest root was not root growth (Nie et al. 1975). significant (~142 q 0.026). In a related study, TNC varied widely among treatments and was significantly associated with root Results and Discussion length (Wilson, in press). The relative shoot weight removed by clipping at a height of 3,6, There was a positive association between number of roots and 9, and 12 cm was 5 1, 36, 25, and 16%, respectively. The relative length of longest root per seedling @32=0.42 I) and between length shoot weight removed by clipping seedlings in the 0-, I-, 2-, and of roots and root weight per unit length @21 = 0.232). Those 3-day light treatments was 25, 25, 26, and 27%. Before the root associations probably do not have a genetic basis because each growth test, leaf-blade area in the 3-, 6-, 9-, and 12-cm clipping sample represented 6 seedlings. Rather, they are explained by treatments, and in the unclipped treatment, was 1.4, 4.2, 6.6, 8.6, differences in environmental conditions during growth of seedlings and I I .3 cm2 per seedling (Table I). Days of sunlight after shade in replications that represented different dates of growth. Favora- did not affect the leaf-blade area remaining when clipping treat- ble light conditions in the greenhouse early in the summer resulted ments were imposed. The increases in leaf-blade area (during the in seedlings with a high number of roots, rapid elongation of roots, root growth test) in the 3-, 6-, 9-, and 12-cm clipping treatments, and a high root weight per unit length. and in the unclipped treatment, were 2.24,2.62,3.12,3.25, and 1.21 The positive association in this study between length of roots cm2 per seedling. Apparently leaf-blade area of unclipped seedlings and root weight per unit length differs from the results found in a had approached the maximum that could be supported by the previous study in which increasing soil temperatures (10 to 30” C) seminal root. Therefore, leaf area expansion began slowly as new caused a substantial increase in root length but a decrease in root adventitious roots developed. Days of sunlight after shade had weight per unit length (Wilson 198 1). Temperature affected the 2 little or no effect on the increase in leaf area during the root growth components of growth in a different way. The negative effect of test. number of roots @s1 = -0.268) on root weight per unit length is Average crown TNC values in the 0-, I-, 2-, and 3day light consistent with the results of the earlier study. Thus, seedlings that treatments were 11.07, 12.26, 12.75, and 12.98%, respectively. All produced many roots tended to have roots that were small in of the TNC levels were considered favorable for the development diameter. of adventitious roots (Wilson 1984). The possible reasons that TNC exerted smaller effects on root A decrease in leaf-blade area (11.3 vs I .4 cm2) resulted in a growth than did leaf area are as follows: (1) shade treatments did decrease of 52% in number of adventitious roots per seedling not cause TNC to fall below critical levels, (2) amounts of TNC in (Table 2), 38% in length of longest adventitious root per seedling, seedling shoots generally are lower than the amounts of current net 33% in root length per unit length (Table 3), and 81% in total assimilate produced in leaves during the 3day root growth test, adventitious root weight per seedling (KO.0 I). A decrease in days and (3) current assimilate is more readily utilized for root growth of sunlight after shade (3 vs 0) resulted in a decrease of 14% in than is the TNC accumulated in crowns before the root growth test number of roots per seedling (Table 2), 6.3% in length of longest (Wilson 1984). There were similarities in the effects of TNC and root per seedling, 5.9% in root weight per unit length (Table 3), and leaf area on the components of root growth notwithstanding dif- 26% in total adventitious root weight per seedling (KO.01). ferences in the magnitude of the effects. Path coefficient analysis showed possible cause and effect rela- The results suggest that blue grama seedlings possess morpho- tionships among variables associated with root growth (Fig. 1). logical and physiological characteristics which favor survival dur- Leaf area (before the root growth test) affected number of roots per ing stress. Plant survival under grazing and drought stress appar- seedling 01% q 0.724), length of longest root per seedling @52 q ently is favored by the effective allocation of photosynthetic 0.474), and root weight per unit length (PSI = 0.766). Crown TNC products into the various components of root growth. Clipping percentage, to a lesser degree than leaf area, affected number of treatments probably reduced root growth because of a deficiency of photosynthetic products. But the reduction was explained by an P = 0.724** (X,) LEAF AREA 53 , (X3)NUM13ER OF ROOTS adjustment in all components of root growth rather than in root BEFORE TEST PER SEEDLING length or depth only. Thus, blue grama seedlings maintained a reasonable rate of root elongation even under severe clipping treatments. A deep root system may be more critical than diameter of roots or number of roots in the survival of seedlings under long-term drought conditions. It is not known whether mature stands of blue grama make similar adjustments in the components of root growth in response to clipping, grazing, or changes in TNC. Literature Cited

Association of Official Agricultural Chemists. 1965. Official methods of analysis. p. 489-499. 10th Ed. Washington, D.C. PER SEEDLING Briske, D.D.,and A.M. Wilson. 1980. Drought effects on adventitious root development in blue grama seedlings. J. Range Manage. 33:323-327. Booysen, P. de V., and C.J. Nelson. 1975. Leaf area and carbohydrate reserves in regrowth of tall fescue. Crop Sci. 15:262-266. Buwni, M., and M.J. Trlica. 1977. Defoliation effects on root weights and total nonstructural carbohydrates of blue grama and western wheatgrass. Crop Sci. 17:15-17. Crider, F.J. 1955. Root-growth-stoppage resulting from defoliation of grass. USDA Tech. Bull. 1102. Hansen, G.K. 1978. Utilization of photosynthates for growth, respiration, and storage in tops and roots of Ldium mulriflorum. Physiol. Plant. (X,)TNC % - > (X,) ROOT WEIGHT PER 42:5-13. BEFORE TEST UNIT LENGTH P4, = 0.212** Heinze, P.H., and A.E. Murneek. 1940. Comparative accuracy and efficiency in determination of carbohydrates in plant material. Missouri Fig. 1. Path coefficient analysis offactors associated with root growth in Agr. Exp. Sta. Res. Bull. 314. blue grama seedlings. **Path coefficients were significant (FCO.01). Hyder, D.N., A.C. Everson,and R.E. Bement. 1971. Seedling morphology and seeding failures with blue grama. J. Range Manage. 24:287-292.

516 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Khan, S.M. 1980. Nonstructural carbohydrates and tolerance of dehydra- Wilson, A.M. 1981. Air and soil temperature effects on elongation of tion in blue grama. Ph.D. Diss. Colorado State Univ., Fort Collins. adventitious roots in blue grama seedlings. Agron. J. 73:693-697. Nie, N.H., C.H. Hull, J.G. Jenkins, K. Steinbrenner, and D.N. Bent. 1975. Wilson, A.M. 1984. Nonstructural carbohydratesand root development in Path analysis and causal interpretation. p. 383-397. In: Statistical pack- blue grama seedlings. J. Range Manage. 37:28-30 age for the social sciences. McGraw-Hill Book Co., New York. Wilson, A.M., and D.D. Briske. 1979. Seminal and adventitious root Parker, K.W., and A.W. Sampson. 1930. Influence of leafage removal on growth of blue grama seedlings on the Central Plains. J. Range Manage. anatomical structure of roots of Stipapulchra and Bromus hordeuceus. 32:209-2 13. Plant Physiol. 5543-553. Youngner, V.B., and F.J. Nudge. 1976. Soil temperature, air temperature Smith, D. 1974. Growth and development of timothy tillers as influenced and defoliation effects on growth and nonstructural carbohydrates of by level of carbohydrate reserves and leaf area. Ann. Bot. 38:595-606. Kentucky bluegrass. Agron. J. 68:257-260. Smith, D., G.M. Paulsen, and C.A. Rsguse. 1964. Extraction of total available carbohydrates from grass and legume tissue. Plant Physiol. 391960-962. Copper and Molybdenum Uptake by Forages Grown on Coal Mine Soils

DENNIS R. NEUMAN AND FRANK F. MUNSHOWER

mine soils (Munshower and Neuman 1978b and 1980). Abstract At the Big Sky Mine the possibility of elevated plant available Coal mine soils have shown a tendency to produce leguminous MO in spoil material is increased by the high levels of extractable vegetation containing elevated concentrations of molybdenum MO (I to 4 pg MO/g) in the blue-grey shale interburden material (MO). The potential for cattle developing copper (Cu)deticiency by located between the Rosebud and McKay coal seams. The mining grazing vegetated areas is increased at one mine where a shale company has been required to selectively handle the shale and bury interburden material contains elevated MO levels. The purpose of it below 2.4 m of sandy overburden in the reclamation process. The this study was to determine if mixing or dilution of the interburden ultimate goal of this study was to determine if mixing or dilution of with low-M0 sandy overburden would produce vegetation with the shale interburden with sandy overburden would produce vege- undesirably high MO levels or low Cu/Mo ratios. tation with undesirably high MO levels or low Cu/Mo ratios. A Concentrations of Cu, MO, sulfur, and Cu/Mo ratios of several corollary to this goal was the investigation of the impact of topsoil- legumes and one grass species grown on these alkaline coal mine ing on these plant MO concentrations and Cu/ MO ratios. soils suggest that, with the exception of white sweetclover, mixing Leguminous plants are known to be accumulators of MO. This of the MO-bearing interburden material with sandy overburden element is an integral part of the nitrogenase enzyme system resulted in desirable elemental levels and ratios for grazing cattle if responsible for the reduction of atmospheric nitrogen to ammonia the mine soils were covered with an adequate depth (0.6 m) of and hence is important to the symbiotic bacteria in root nodules of suitable topsoil. Vegetation uptake of MO was species and site nitrogen-fixing legumes. specific. Although elevated levels of MO in vegetation are not injurious to the plants themselves, animals are very sensitive to high MO levels. In 1976, an investigation of copper (Cu) and molybdenum (MO) The required dietary level for grazing cattle and sheep is small and levels in 2 leguminous species, yellow and white sweetclover (Meli- concentrations in the total diet as low as 5 pg Ma/g may be toxic lotus officinalis and M. alba). grown on revegetated mine soils of under certain conditions (Webb and Atkinson 1965). Dietary levels the Northern Great Plains was initiated by Erdman and his co- of IO-20 pg/g are nearly always associated with disturbed Cu workers (1978). They reported a Cu/ MO ratio of 0.44 and a MO metabolism because of the antagonistic action of these elements level of 13 pg/gfor white sweetclover collected at the Big Sky Mine (Allaway 1977). The ratio of these 2 elements in vegetation is of in southeastern Montana and concluded that cattle grazing pre- prime concern when discussing the potential for induced Cu defi- dominantly on this vegetation could display subclinical or acute ciency in grazing animals. A Cu/ MO ratio of 2.0 is considered the symptoms of molybdenosis. Legumes growing on revegetated lowest acceptable value in ruminant feeds (Miltmore and Mason mine areas have exhibited a tendency to accumulate higher concen- 1971). trations of MO than those growing on native range or cultivated The sulfur (S) intake of the ruminant animal exerts a significant fields (Munshower and Neuman 1978a). As a complicating factor influence on the Cu-MO interrelationship (Suttle 1975). Ingested the Cu content of vegetation on the Northern Great Plains in MO plus S can either increase or decrease the absorption of Cu, southeastern Montana is generally marginal in terms of required depending on their ingestion relative to Cu (Underwood 1977). cattle nutritional needs. This low Cu status is common to both Suggested dietary intake of S is 0.1% for beef cattle (National native range vegetation and established plants on revegetated coal Research Council 1976) and 0.2% for dairy cattle (National Research Council 197 1). Authors are associate research chemist and director and associate research plant ecologist at the Reclamation and Research Unit, respectively, Montana Agricultural High MO forages that may be toxic to grazing animals in the Experiment Station, Department of Animal and Range Sciences, Montana State spring have been shown to lose their toxicity when cut as hay or University, Bozeman 59717 This contribution was submitted as Montana Agricultural Experiment Station when dry and cured in the field (Ferguson et al. 1943). However, Journal Series No. 1413. This work was supported entirely by the Peabody Coal Barshad (1948) reported large increases in MO concentrations for Company, Denver, Cola. several plants as they aged. Furthermore, a seasonal decrease in the

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 517 Cu content of native grasses has been reported (Munshower and Table 1. Mean vrlua of extractable molybdenum @g/g) and copper&g) Neuman 1978b). in study site soils. Methods and Materials Bare Mixed Segregated Native Site Selection spoils interburden interburden range The Big Sky Mine, owned and operated by Peabody Coal Co., is Soil depth #I #2 #3 #4 located in Rosebud County 5 miles south of Colstrip, Mont. Four study sites were established at this mine. Site I consisted of old Molybdenum Concentrations mine spoils of mixed interburden and overburden devoid of any Surface 53 aAr .I7 bA .I7 bA .23 bA 20 cm 54 aA .I9 bA .I6 bA .22 bAB topsoil. Site 2 was mixed interburden and overburden covered with 40 cm .45 aA .23 bB .24 bA .I4 bC 0.6 m of topsoil. At Site 3 the mixed interburden and overburden 80 cm .32 aA .28 aB .62 bB .I7 aBC was buried below 2.4 m of overburden material and 0.6 m of 120 cm .45 aA .32 aB .46 aB .23 aA topsoil. An undisturbed, native range site was used as a control plot. overall MO means .46 a .23 ab .3l ab .19b Plot Design and Seeded Species Copper Concentrations Yellow sweetcover (Melilotus officinalis Lam.), white sweet- Surface 1.2aA 0.6 CA 0.5 CA 0.8 bAB clover (M. a&a Dew.), alfalfa (Me&ago sutivu L.), white prairie- 20 cm l.4aA 0.8 aA 0.6 aA I.1 aA clover (Petulostemon candidurn Michx.), purple prairieclover (P. 40 cm 1.6 aA I.0 abA 0.5 CA 0.8 CAB purpurem Rydb.), cicer milkvetch (Astrugulus cicer L.) and green 80 cm 1.6aA I.2 bA 0.3 cB 0.5 cc needlegrass (Stipu viridulu Trin.) were seeded in 1.2 X 1.5-m plots 120 cm 1.2 aA 1.9 aA 0.3 bB 0.7 bBC that were replicated 3x at each of the 4 study sites in fall 1980. overall Straw mulch was used as an amendment except on the native range Cu means l.4a 1.0 b 0.5 c 0.8 b site where native vegetation provided a natural mulch. Germina- tion and emergence of most species were complete in spring 198 I. ‘Means followed by same lower-case letter in rows and same capital letter in columns are not different at pSO.05 N=8 for all depths except 120 cm where N4, N-36 for Soil Determinations overall mean values. Soils were collected on 2 occasions, at vegetation seeding (fall 1980) and in fall 1981. Soil cores were augered at the corners of depths which are within the root zone of the legume species and each study site and the surface, 20, 40, 80, and 120 cm fractions below the 60 cm of applied topsoil. were air-dried and gently crushed to pass a 2-mm sieve. Several Table 1 shows the mean extractable Cu concentrations for each physical and chemical determinations were made including pH of profile depth and overall mean values for each study site. The the saturated soil paste, DTPA extractable Cu, and ammonium topsoiled spoils at Site 3 exhibited a significant b50.5) depression oxalate extractable MO. in available Cu at the 80 and 120 cm depths. This is in sharp contrast to the extractable MO levels which were elevated in soils Vegetation Determinations from this site at these depths. Several grams of vegetation were collected from each plot at each site on 4 occasions: after 1growing season in September 198 1, Seasonal Variation of Elemental Levels and Ratios and on June 1, July 14 and September 9, 1982. Short plants were Plant samples were collected throughout the growing season to clipped approximately 5 cm above ground level to avoid soil evaluate elemental concentrations and subsequent Cu/ MO ratios contamination, while only the top 22 cm of taller plants, sweetclov- from early growth through senescence. ers and alfalfa, were collected. The vegetation was oven dried, The single grass species evaluated in this Cu/ MO uptake study ground to pass a 40-mesh screen and analyzed for Cu and MO revealed a relatively constant MO level at each site throughout the concentrations by electrothermal atomic absorption spectroscopy growing season. Grass Cu concentrations decreased at 3 of the 4 (AAS) after acid digestion (Neuman and Munshower 1981). Total sites between spring and fall collections. Alfalfa showed signifi- S concentration was indirectly determined by analysis of excess cantly higher levels of MO in spring samples collected from Sites 2, barium (Ba) by flame AAS after formation of insoluble barium 3, and 4. The other legumes did not demonstrate any change in Cu sulfate in the dissolved ash of plant samples (Horwitz 1975). or MO concentrations throughout the growing season or only Results and Discussion minor fluctuations at one site. The effects of the few seasonal variations in the elemental status of vegetation on the Cu/Mo The question of burial of the MO bearing interburden material or ratios resulted only in a lower ratio for alfalfa collected in the dilution with sandy overburden may be resolved by examining the spring at Site 4 and an elevated ratio for cicer milkvetch in the following: plant available Cuand MO within the root zone; vegeta- summer at Site 1. tion MO, Cu, and S concentrations; and the subsequent Cu/Mo Since seasonal variations appeared to be relatively unimportant, ratios. the data were combined for the 4 sampling periods with mean Copper and Molybdenum in Soils values for Cu, MO, and S concentrations and Cu/ MO ratios exhi- Soil pH is one of the most important single factors affecting the bited in Table 2. For most species the mean values are based on uptake of MO and Cu in plants. In this study the soil pH levels N= 12. The prairieclovers were generally not available for collection ranged from 7.0 to 7.9 with little intersite variation. Molybdenum in the spring from all plots at all the sites. is readily available to vegetation in this pH range and Cu less Molybdenum Status of Seeded Species available. The concentrations of extractable MO in the study site Most of the legumes from Site I demonstrated elevated MO soils are shown in Table 1 with the data arranged by soil profile levels Q710.05) when compared with the same species grown on the depth. The overall mean extractable MO level for each site is also 2 topsoiled mine soil sites or the native range site (Table 2). given. Non-topsoiled mine soils from Site I exhibited the highest Legumes grown on the segregated interburden at Site 3 contained overall mean level of 0.46 pg MO/g. Of the topsoiled study sites, higher MO levels than those grown on the mixed interburden at Site 3 (segregated interburden) and Site 2 (mixed interburden) Site 2, although these differences were not significant. White revealed similar extractable MO concentrations. However, when sweetclover demonstrated the highest mean MO concentration of intersite variation by profile depth was evaluated, Site 3 contained all the species at Sites 1, 2, and 3. higher Q&0.05) extractable MO at the 80cm level than soils from The accumulation of MO was species specific with white sweet- Site 2. Both of these sites exhibited elevated MO at the lower soil clover and white prairieclover accumulating more of the metal

518 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Copper Status of Seeded Species Table 2. Mean clcmentnl concentrations and ratios in seeded species. Table 2 lists Cu concentrations in the 6 legumes and I grass species. The data indicate that vegetation grown at Site 1 contained Bare Mixed Segregated Native the highest levels of Cu with alfalfa and white prairieclover concen- spoils interburden interburden range trations being elevated b10.05). Green needlegrass, cicer milk- Sites #I #2 #3 #4 vetch, purple prairieclover, and both sweetclovers exhibited no variation in Cu levels among the 4 study areas. The Cu status of the Yellow Sweetclover MO’ 3.06 a2 1.12 b 1.51 ab 0.86 b vegetation appeared to be species specific, but not site specific. cu 7.8 a 6.9 a 6.5 a 6.4 a Only some species accumulated higher levels of Cu when grown on s 0.47 a 0.28 b 0.37 ab 0.31 b soil containing more available Cu. Cu/ MO 3.2 a 7.6 b 5.2 ab 7.5 b Levels of this metal in vegetation grown on native range soils White Sweetclover were statistically equivalent to levels of Cu in vegetation from the 2 MO 5.69 a 1.75 b 2.73 b 0.75 b topsoiled sites. Other studies (Kubota et al. 1967, Kubota 1975) cu 8.1 a 6.2 a 6.9 a 6.1 a have also confirmed that plants of the same species with widely s 0.61 a 0.29 b 0.33 b 0.32 b different MO concentrations may have similar levels of Cu. Cu/ MO l.9a 4.5 a 4.0 a 9.9 b Copper:Molybdenum Ratios in Seeded Species Alfalfa Legume species from Site 1 revealed lower Cu/ MO ratios than MO 2.17a 0.96 b 1.13 b 0.67 b the same vegetation grown on the native range study site (Table 2). CU 8.4 a 6.4 ab 6.7 ab 5.9 b White sweetclover from Site 1 had a mean ratio below the critical S 0.50 a 0.27 b 0.31 b 0.29 b dietary ratio for cattle of 2.0. Vegetation grown on the two top- Cu/Mo 4.4 a 7.3 ab 7.3 ab 10.3 b soiled sites exhibited similar Cu/ MO ratios except for white prai- Green Needlegrass rieclover. The mean ratios of all species at Sites 2,3, and 4 were well MO 0.85 ab l.OOa 0.70 ab 0.62 b above the critical dietary value with vegetation grown on the native cu 4.2 a 4.0 a 3.5 a 4.1 a soil displaying the highest ratios. S 0.21 a 0.16 ab 0.15 b 0.17 ab Cu/Mo 5.5 ab 4.0 a 5.4 ab 7.4 b Sulfur Status of Seeded Species High levels of dietary S and MO restrict Cu utilization by White Prairieclover MO 4.32 a 0.68 b 1.27 b 0.62 b depressing its solubility in the digestive tract through the precipita- cu 10.3 a 7.0 ab 5.9 b 6.4 b tion of insoluble copper thiomolybdate, CuMoS4 (Dick et al. S 0.29 a -3 .26a 1975). Suttle (1975) demonstrated that total S rather than inor- Cu/ MO 2.7 a 10.3 b 4.9 a 11.3 b ganic S is the more useful measurement in evaluating the Cu-MO-S Purple Prairieclover interrelationship. MO 3.15 a 0.83 b 1.02 b 0.57 b Most legumes grown on Site 1 exhibited elevated S levels cu 7.9 a 7.7 a 7.2 a 5.9 a (p10.05) compared with vegetation from the other sites (Table 2). S 0.37 a 0.21 b 0.18 b White sweetclover from Site I showed the highest mean S level Cu/Mo 3.0 a 9.3 b 9.6 b ll.4b (0.61%) and green needlegrass from Site 3 exhibited the lowest S Cicer Milkvetch concentration (0.15%). Martin and Matocha (1973) summarized MO 2.23 a 0.76 b 0.92 b 0.83 b the S status of alfalfa and described deficient, critical, adequate, cu 7.2 a 5.5 a 6.2 a 5.8 a and high nutrient concentrations. This species was considered S S 0.47 a 0.26 b 0.27 b 0.27 b critical at <0.20-0.35%, adequate at 0.26-0.50%, and high at levels Cu/Mo 4.3 a 7.4 b 8.1 b 8.9 b greater than 0.5 l.%. In comparison to alfalfa, legumes in this mine ‘MOand Cu concentrations in &g, S in %. soil study appear to be within low-normal limits for adequate +kaas followed by same letter in rows are not different at p50.05. N=I2 for most vegetation S nutrition. means; N varies from 6 to IO for prairieclover mean values. Limits of S concentration associated with disturbed Cu metabo- J(-_) indicates insufficient sample for analysis. lism in animals have not been well defined in the literature. How- than the other legumes. Uptake of MO was also site specific with ever, several investigators have indicated that levels above 0.35% the plants containing elevated levels of MO in soils of higher are generally found in relation to abnormal Cu metabolism (Spais extractable concentration. These findings are in agreement with 1959, Whitehead and Jones 1969, Havre and Dishington 1962). Allaway (1977) who found that different plant species accumulated Mean S concentrations in the legumes from Sites 2, 3, and 4 are substantially different amounts of MO from the same soil or cul- below this level, while those from Site 1 exceed 0.35% (Table 2). ture, with leguminous vegetation containing more of the metal than grasses in the same area.

Table 3. Vegetation showing acceptable molybdenum concentrations and ratios when grown on each study site.

Acceptable MO content Acceptable Cu:Mo ratio Overall acceptability Vegetation species (90% of samples C5.0 pg/g) (90% of samples >2.0) of vegetation Sites Sites Sites Yellow sweetclover 1234 234 234 White sweetclover 234 2 4 2 4 Alfalfa 1234 1234 1234 Green needlegrass 1234 1234 1234 White prairie clover 234 234 234 Cicer milkvetch 1234 1234 1234 Purple prairieclover 234 234 234

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 519 Summary and Conclusions Erdmm, J.A., R.J. Ebens, and A.A. Case. 1978. Molybdenosis: A potential problem in ruminants grazing on coal mine spoils. J. Range Manage. Statistical differences between specific parameters at Site 1 (no 3 1:34-36. topsoil), Site 2 (mixed interburden and overburden) and Site 3 Ferguson, W.S., A.H. Lewis,and S.J. Watson. 1943. The teart pastures of (segregated interburden) were of prime importance. These parame- Somerset: I. The cause and cure of teartness. J. Agr. Sci. 33:45-51. ters included differences in extractable MO within the plant root Havre, G.N., LW. Dishington. 1962. The mineral composition of pasture zones, differences in Cu and MO levels in the seeded species, and as influenced by various types of heavy nitrogen dressings. Acta. Agr. differences in vegetation Cu/ MO ratios. Stand. 12~298-308. Extractable soil MO concentrations in the topsoil applied at Horwitz, W. (Ed.). 1975. Official Methods of Analysis of the Association of Sites 2 and 3 were similar. All species grown on the two topsoiled Official Analytical Chemists. 12th Ed. Method. 3-059. Assoc. Off. Anal. Chem. Washington, D.C. areas exhibited mean MO levels below critical dietary concentra- Kubota, Joe. 1975. Areas of molybdenum toxicity to grazing animals in the tions for grazing cattle of 5 pg/g. western states. J. Range Manage. 28:252-256. The uptake of MO by the legumes was species and site specific. Kubota, Joe, V.A. Lazar, G.H. Simonson, and W.W. Hill. 1%7. The All of the legumes accumulated higher levels of MO when grown on relationship of soils to molybdenum toxicity in grazing animals in soils with increasing extractable MO concentration. Oregon. Soil Sci. Amer. Proc. 3 1:667-671. Table 3 shows which of the seeded species evaluated may be Martin, W.E., and J.E. Matocha. 1973. Plant analysis as an aid in the planted on different sites and meet acceptable MO levels and fertilization of forage crops. In: Walsh, L.M. and J.D. Beaton (Eds.). Cu/Mo ratios for grazing cattle. Acceptable performance was Soil Testing and Plant Analysis. Soil Sci. Sot. Amer. Madison, Wis. based upon 2 factors: 90% of the samples of any species must reveal Miltimore, J.E.and J.L. Mason. 1971. Copperand molybdenum ratio and MO concentrations below 5 pg/g and Cu/Mo ratios must be molybdenum and copper concentrations in ruminant feeds. Can. J. Anim. Sci. 51: 193-200. greater than 2.0. Based on these criteria, 4 of the legumes should Munsbowcr, F.F. and D.R. Neuman. 1978a. Trace element concentrations not be seeded on bare spoils of Site 1. in vegetation from revegetated strip mined land and native range in With the exception of white sweetclover, the legumes grown on southeastern Montana. p. 887-891. In: M.K. Wali (Ed.). Ecology and the 2 topsoiled mine soils exhibited mean Cu, MO, and S levels and Coal Resource Development, Vol. 2, Pergamon Press, New York. Cu/ MO ratios within normal dietary limits for grazing cattle and Munshower, F.F., and D.R. Neuman. 1978b. Elemental concentrations in were not statistically distinct. They could safely be seeded and native range grasses from the Northern Great Plains of Montana. J. consumed by livestock on these topsoiled mine soils. Range Manage. 31:145-148. Since leguminous species would not constitute the sole diet of Munshower, F.F., and D.R. Neuman. 1980. Elemental concentrations in native plant species growing on minesoil and native range. Reclamation grazing animals, the intake of MO in the total diet would not be as Review. 3:4 l-46. high as levels provided by these legumes alone. The subsequent National Research Council, Subcommittee on Beef Cattle Nutrition. 1976. dietary Cu/ MO ratio of free ranging livestock or wildlife would Nutrient requirements of beef cattle. Printing and Publishing Office, probably be higher than values reported in this study. Nat. Acad. Sci., Washington, D.C. Mixing or dilution of the MO bearing shale interburden located National Research Council, Subcommittee on Dairy Cattle Nutrition. between the Rosebud and McKay coal seams at the Big Sky Mine 1971. Nutrient requirements of dairy cattle. Printing and Publishing with sandy overburden in the reclamation process did not result in Office, Nat. Acad. Sci., Washington, D.C. undesirable MO levels or Cu/ MO ratios in the legumes evaluated as Neuman, D.R., and F.F. Munshower. 1981. Rapid determination of long as the site was covered with an adequate depth of suitable molybdenum in botanical material by electrothermal atomic absorption spectrometry. Anal. Chim. Acta. 123:325-328. topsoil. Spais, A.G. 1959. Lx cuivre en pathologie ovine et bovine. Rec. Med. Vet. 135:161-194. Literature Cited Sullivan, J.T. 1969. Chemical composition of forages with reference to the needs of grazing animals. Agr. Res. Serv. Rep. No. 34-107. USDA, Allaway, W.H. 1977. Perspective on molybdenum in soils and plants. p. Washington, D.C. 317-339. In: W.R. Chappell and K.K. Petersen (Eds.), Molybdenum in the Suttle, N.F. 1975. The role of organic sulfur in the copper-molybdenum-S Environment. Marcel Dekker, Inc. N.Y. relationship in ruminant nutrition. Br. J. Nutr. 3441 I-420. Barshad, I. 1948. Molybdenum content of pasture plants in relation to Underwood, E.F. 1977. Molybdenum in animal nutrition. p. 9-32. In: W.R. toxicity to cattle. Soil Sci. 66: 187-195. Chappell and K.K. Petersen (Eds.). Molybdenum in the Environment. Dick, A.T., D.W. Dewey, and J.W. Gawthome. 1975. Thiomolybates and Marcel Dekker, Inc., N.Y. the copper-molybdenum-sulphur interaction in ruminant nutrition. J. Webb, J.S., and W.J. Atkinson. l%S. Regional geochemical reconnaissance Agr. Sci., Camb. 85:567-568. applied to some agricultural problems in Co. Limerick. Eire. Nature 208:1056-1059. Whitehead, D.C., and E.C. Jones. 1969. Nutrient elements in the herbage of white clover, red clover, lucerne and sainfoin. J. Sci. Food Agr. 20:584-59 1.

520 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Natural Establishment of Aspen from Seed on a Phosphate Mine Dump

BRYAN D. WILLIAMS AND ROBERT S. JOHNSTON

Abstract The natural reproduction of aspen (Populusfremuloides Michx.) of the farthest seedlings. Seeds from this source were collected, from seed was discovered on a phosphate mine dumo in southenst- germinated in the greenhouse and showed 96% germination. em Idaho. Aspen seedlings w&e f&d growing on k-as that were The germination and establishment of aspen from seed is a rare essentially bare except for scattered plantings of containerized occurrence over large areas of interior Western United States shrubs and trees. Aspen survival and growth was monitored for 4 under the present climate of warm, dry summers. Discovery of the growing seasons. Seedling density varied from 2 to 10 per ml, seedlings provided an opportunity to monitor survival and growth seedling heights varied from 16 to 81 cm, and survival rate was 73% to determine if aspen can indeed become established naturally. at the end of 4 growing seasons. No changes in the number of Site Description seedlings were noted after the second growing season. The aspen seedlings are growing on a 3-ha north-facing phos- Aspen(Populusrremuloides Michx.)isa highly desirable, major phate mine spoil dump at an elevation of 2,100 m, about 40 km component of the predisturbance vegetation on large areas of northeast of Soda Springs, Ida. The average annual precipitation Western U.S. phosphate minelands. It is also desirable on mine from 1977ro 1981 was 50cm, mostofwhich wassnow. Rainfall for waste spoils because it enhances plant species diversity, has recog- June through September of 1979, the year the seedlings became nized value for wildlife habitat, watershed protection, and aesthe- established, was 7 cm. The dump was constructed as a “head-of- txs. valley”fill with a 3: I slope and a maximum depth of about 30 m. Since the early 1970’s. most revegetation efforts have concen- Thearea wassurfaced witha”mid-waste”shale,a highly weathered trated on grasses and forbs to stabilize the surface of mine dumns. shale material located between alternating beds of phosphate ore. However, once grasses and forbs are established, invasion by Numerous small water seeps are evident in the middle portion of native shrubs and trees and survival of planted stock are greatly the dump, where perched groundwater tables intersect the dump reduced due to shading and competition for water and nutrients face. Most seepareas persist until mid-July and afewareasare wet (Williams, USFS, Logan, Utah, unpublished data). To avoid this through most of the growing season. competition, a mine dump was revegetated using 3-m widecontour Aspen seedling distribution within the shrub strips is concen- strips planted with containerized shrubs and trees, alternating with trated on, but is not limited to, these wet areas. While aspen 10-m strips that were fertilized and seeded with grasses and forbs appears throughout the shrub strips, no aspen seedlings have been (Fig. I). The grass strips were seeded in the fall of 1978 and the found growing in the alternating grass strips.

Methods In the fall of 1979, IO plots, each I m square, were established within four of the shrub strips. The plots were deliberately located so they contained aspen seedlings. Five plots were located on wet areas near water seeps and 5 on drier areas away from any visible signs of surface water. The design served as a means of monitoring survival and growth of the aspen and was not intended to provide any other statistical inference. The number and height of seedlings on each plot was recorded at the end of each growing season for 4 years.

Results The number of aspen seedlings on all plots declined from 73 in 1979 to 53 at the end of the 1982 growing season. This is a 73% survival over that 3 years. Seedling losses occurred between the second and third growing season (Fig. 2). Seedling numbers remained constant during the third and fourth year. The initial number of seedlings was slightly higher on the wet sites than on the contamerized trees and shrubs were planted the following spring. drier sites and this general relationship persisted. The range of Quite unexpectedly in the fall of 1979 naturally regenerated seedling numbers on all individual plots varied from 2 to IO per aspen seedlings were discovered in the grass-free strips. Inspection plot with a slightly narrower range on wetter plots than on drier of the roots of these seedlings by Dr. George A. Schier’ confirmed ones. that the aspen originated from seed and not from sprouting of Plant height was slightly greater on the wet areas than on dry residualaspen roots. Aseed sourceislocated within 350mupwind areas (Fig. 3). The growth rate on the wet sites was markedly greater during the 1980 growing season and declined the next year. Except for this I year, the rate of growth was about the same for both areas. The average height of seedlings after I growing season was I .4 f .36cm. Attheend ofthefourthseason theaverage height was385

JOURNAL OF RANGE MANAGEMENT W(6), November ,984 521 \ W- \ (Dm AREAS) N------20-

15 -

10 -

Ob 1WJ 1051 1052 Ymu Fig. 2. Aspen seedling survival. Fig. 3. Aspen growth rate (average of 5 plots each area).

15 cm. The range in height was I6 cm to 8 1 cm. The mean height of grass strips regardless of observed surface moisture conditions. aspen on the wetter sites was 14 cm greater than on the dry sites at Our observations, while not conclusive, indicate that aspen can the end of 4 seasons. In the fall of 1982 the height of 60 additional be established on selected phosphate mine dumps from seed. Since seedlings growing throughout the shrub strips was measured. this manuscript was written another phosphate mine dump has Average height was 35 f 19 cm. These heights are similar to the been discovered where aspen are growing from seed. The natural heights of seedlings within the sample plots. establishment of aspen on mine sites eliminates costly propagation and planting of containerized trees. An additional advantage may Discussion be realized because the introduction of a large gene pool through Although most reproduction of western aspen is achieved by natural seeding increases the probability of site adaptability and vegetative sprouting, aspen can occasionally become established survival. under favorable seedbed and climatic conditions (Ellison 1943, Larson 1944, Barnes 1966). Aspen are prolific seed producers and Literature Cited seeds germinate rapidly under a wide temperature range. The main deterrents to aspen reproduction from seed are the rigorous Barnes, B.V. 1966. The clonal growth habit of American aspens. Ecology requirements of friable mineral soil, limited competition, and a 47z439-447. continuous supply of soil water (Barnes 1966, McDonough 1979). Ellison, L. 1943. A natural seedling of western aspen. J. Forestry. 41:767-768. Areas within the bare strips in or near ground water seeps fulfill Larson, G.C. 1944. More on seedlings of western aspen. J. Forestry 42:452. these requirements. No aspen seedlings were found in the adjacent McDonough, W.T. 1979. Quaking aspen-seed germination and early seedling growth. USDA, Forest Serv., Res. Pap. INT-234.

Membership in the Society for Range Management. . .

l is open to those engaged in or interested in the study, n offers opportunities for face-to-face exchange of management, or use of range ecosystems and the ideas at local, national, and international meetings intelligent use of all range resources of the Society. l includes research scientists, ranchers, governmental Dues vary according to type of membership and agency administrators and technical personnel, geographical section. For application forms and ad- teachers, students, and people from the business ditional information, contact the: community l provides members with two pclblications-one ori- Society for Range Mana~emer~t 2760 West Fifth Avenue ented t0 research (JO~rr~idOf Ri3ng8 M8nfSgement) Denver, Colorado 80204 and the ottret oriented to practical resource manage- (303) 571-0174 ment (Rangelands)

522 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Variability of Infiltration within Large Runoff Plots on Rangelands

MICHELINE DEVAURS AND GERALD F. GIFFORD

Abstract In this study we investigated the variability of infiltration on These models implicitly transfer point infiltration properties to a native rangeland sites. A rainfall simulator was used to collect data watershed-wide application. They are significant developments, on runoff from small (0.37 m*) plots located within large plot yet they are computer simulations and not adequately field vali- boundaries (32.5 mr). Three range sites were sampled and data dated. In this field study, we evaluated the variability of infiltration were collected from unfenced, fenced, and rototilled conditions on characteristics and soil properties on small (0.61 m X 0.61 m) each site. In addition data were collected on vegetation, antecedent runoff plots within large (3.05 m X 10.67 m) runoff plots on moisture, bulk density, soil texture, and organic matter as possible “homogeneous*’ semiarid rangeland sites. explanations for variations in hydrologic response on small and Site Descriptions large plots. The field study demonstrated large variability in measured infiltration and soil physical properties on relatively uniform Study Area rangeland sites, suggesting that inherent variability patterns need The field study was conducted during summer, 1981, on range- to be examined to provide appropriate confidence intervals for land sites located on the Reynolds Creek Experimental Watershed single parameter values that may be applied to larger areas. No set near Boise, Ida. The watershed soils, geology, vegetation, and land of factors consistently explained the observed variability within use are representative of plateau and foothill grazing areas of the large plots. Northwest (Stephenson 1977). Elevations range from about 1,097 m to 2,225 m. The climate of the watershed ranges from arid to Historically it has been assumed that watersheds are homogene- temperate, with annual precipitation varying from 25 cm at the ous. The only data collected were streamflow measurements, lower elevations to 127 cm at the higher elevations. Nearly 75% of which were assumed to integrate hydrologic processes on an entire the annual runoff is from snowmelt; however, flash runoffs from watershed. Use of rainfall simulators made it possible to assess the smaller areas do occur following intense summer rain storms. impacts of management practices on specific parts of a watershed. Three sites on the Reynolds Creek Experimental Watershed That capability led to an emphasis upon characterizing the spatial (referred to as Flats, Nancy, and Lower Sheep) were selected for variability of hydrologic properties to eventually link point and this study. These sites are described in Table 1. area1 measurements. For simplicity, many hydrologic models still assume that a Plot Preparation watershed is homogeneous. Such models ignore the spatial varia- Rainfall simulator plots were located on 3 sites in unfenced, bility of hydrologic properties and lump watershed characteristics fenced, and tilled conditions. Average large plot slopes were 3, 6, as area1 averages. However, numerous recent investigations of and 9% on the Flats, Nancy, and Lower Sheep sites, respectively hydrologic responses (Achouri 1982, Blackburn 1975, Grah 1983, (with the exception of one tilled plot on the Flats site with a 9% Gifford 1976, Springer and Gifford 1980, Lyford and Qashu 1969, slope); see Table 2. Tilled plots were tilled up and down-slope to Merzougui 1982, Murabayashi and Fok 1979, Sharma et al. 1980, approach a fallow condition approximately one week prior to Rogowski 1980, Tricker 198 I, Vieira et al. 198 1) have illustrated rainfall simulation. They were restored to the original bulk density great spatial variability of hydrologic properties within short (l.l-1.4g/cc) by natural settling and trampling (walking on the distances. plot) before simulator runs. Fenced sites have been protected from To estimate the effect of spatial variability of infiltration on grazing by domestic livestock since 197 1. Unfenced sites are grazed hydrologic models’ assumptions of homogeneous sites or water- both by domestic livestock and wildlife, but exact stocking rates, sheds, area1 loss rates must be contrasted with point infiltration season, and duration of use data are not available. Generalized rates. Three theoretical explanations of this functioning of natural grazing allotment data are available from local Bureau of Land watersheds have been proposed. Hawkins (1981) submits that Management offices, but the applicability of these data to specific watersheds act as a collection of runoff elements, each with inde- locations on the watershed are limited. pendent, uniform hydrologic characteristics. Computer simulation Vegetal Cover has been used by Smith and Hebbert (1979) and Cundy (1982) to The 3 sites are typical of sagebrush-grass communities found on investigate the effects of spatial variability on plot and hillslope the Reynolds Creek Experimental Watershed. The vegetation at performance. Smith and Hebbert developed a model that consi- the Flats site is chiefly shadscale (Atriplex confertifolia) but dered only the spatial variability of rainfall excess. The Cundy includes minor amounts of big sagebrush (Artemisia tridentata), model considers variability in soil properties, initial moisture, and clasping pepperweed (Lepidiumperfoliatum), cheatgrass (Bromus rainfall intensity, and routes rainfall excess to downslope using the tectorum), squirrel tail (Sitanion hystrix), bluegrass (Poa spp.), kinematic wave technique. and moss. At the Nancy site, the dominant species are big sage- Authors are graduate research assistant and professor and chairman, Watershed brush, little-leaf horsebrush (Tetradymia glabrata), squirrel tail, Science Unit, Range Science Department, Utah State University, UMC 52, Logan bluegrass, and moss. Low sagebrush (Artemisia arbuscula) domi- 84322. Gifford’s current address is Dept. of Range, Wildlife,and Forestry, University of Nevada, Reno 895 12. Ms. Devaurs is currently with Los Alamos National Lab. Los nates on the Lower Sheep site. The diverse grass and forb under- Alamos, NM 87545. story includes mainly low pussytoes (Antennaria dimorpha). This project was supported through the Northwest Watershed Research Center, locoweed (Astragalus spp.), bluegrass, and squirrel tail. USDA, ARS, Boise, Ida.,and the Utah Agricultural Experiment Station, Projects 749 and 771. Tech. Pap. 2877, Utah Agricultural Experiment Station, Logan 84322. Manuscript accepted February 7, 1984.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 523 Table 1. Site characteristics on Reynolds Creek watershed near Boise, Idaho.

Elevation Annual precipitation Site fm) (cm) Soil series Soil descriotion Flats 1219 25 Nannyton Fine loamy, mixed mesic typic haplargids Nancy 1402 33 Glasgow Fine, montmorillonitic, mesic xerollic durargid Lower sheep 1646 36 Harmehl Fine, montmorillonitic, frigid pachi argixeroll

Table 2. Large plot characteristics for each site.

Number of small plots where infil- Rainfall rate Total number Number of small tration rate exceeded rainfall Mean final 30- Slope on small plots of small plots where run- rate for duration of 30-minute minute infiltra- Site condition (%) (cm/ hr) plots off occurred run tion rate (cm/hr)t Flats unfenced 3 6.35 20 19 1 2.5 f I.0 Flats fenced 3 6.35 20 15 5 4.0 f I.0 Flats tilled 9 6.35 10 10 0 1.7 f0.2 Flats tilled 3 6.35 IO 10 0 1.3 f 0.3 Flats tilled 3 12.7 10 IO 0 1.3 f 0.6 Nancy unfenced 6 12.7 20 19 1 5.4 f 2.3 Nancy fenced 6 12.7 20 I9 1 7.2 f 3.0 Nancy tilled 6 12.7 10 10 0 2.3 f 0.7 Lower Sheep unfenced 9 12.7 20 I9 1 7.6 f 2.3 Lower Sheep fenced 9 12.7 20 15 5 9.6 f 1.2 Lower Sheep tilled 9 12.7 10 IO 0 2.3 f 0.5 tValues in this column are reported as mean f standard deviation; small plot data (only plots where runoff occurred).

Table 3. Regression models for three sites.

Standard error of Slope regression Site Condition (%) Model (cm/hr)

Flats unfenced 3 Y q 13.08 - .036X8 47.55 0.806 Flats fenced 3 Y q -3.24 + 0.152X6 + 0.106X7 + 0.502X8 -3.99X11 95.0 0.276 Nancy unfenced 6 Y = 1.437 + 1.476X12 58.66 1.570 Nancy fenced 6 Y = -4.104 + 0.108X7+ 0.18X13 68.17 I.858 Lower Sheep unfenced 9 Y = 11.56 - 11.56X1 + 0.468X2 - .072X+ 92.04 0.8676 0.288X4 +.036X5 - .216X6 + 0.180X7 Lower Sheep fenced 9 Y = 0.282 + 0.396X2 + 0.036X1 - 0.072X6 + .324X6 95.79 0.3528

In all of the above equations, Y = final 30-minute infiltration rate (cm/ hr) and X1 = bulk density of 7.62 cm soil sample (g/cm’); X2= organic matter of 7.62 cm soil sample (%); Xs = sand in 7.62 cm soil sample (vc); XC = sand + silt in 7.62 cm soil sample (%); Xs = total live overstory of shrubs, grasses,and forbs (%); Xs 3 bare ground (%); XT = Inter cover (%); Xs = shrub canopy (%); XI1 = bulk density of 2.54 cm soil sample (g/cm)); XI2 = organic matter in 2.54 cm soil sample (%); Xts = sand m 2.54 cm sod sample (%).

Table 4. Necessary sample size to estimate true population mean.

Sample size necessary for X Sample size necessary for 51(final (timeto ponding) within infiltration rate after 30 minutes) two minutes of fi with within one cm/ hr of p with Rainfall confidence confidence Slope rate on small Site Condition (%) (cmihr) .80 .90 .80 .90 Flats unfenced 3 6.35 35 58 8 14 Flats fenced 3 6.35 70 118 8 14 Flats tilled 9 6.35 I 2 I 1 Flats tilled 3 6.35 I I I 2 Flats tilled 3 12.7 I 2 4 6 Nancy unfenced 6 12.7 21 35 38 64 Nancy fenced 6 12.7 29 48 62 105 Nancy tilled 6 12.7 1 1 4 7 Lower Sheep unfenced 9 12.7 84 143 38 64 Lower Sheep fenced 9 12.7 59 100 II 19 Lower Sheep tilled 9 12.7 1 I 2 4

Above values calculated using the formula n = ts G/d* where n = required sample size, t : tabulated value for desired confidence and degrees of freedom of initial sample, d = half width of desired confidence interval, s = standard deviation of a given sample.

524 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Methods All small plot infiltrometer data were collected using a modular drop-forming device (designed after one described by Chow and Harbough (1965) and modified by Meeuwig (1971) and further modified by Malekuti and Gifford (1978)). Rainfall simulation on small plots (0.37 m2) began with soils at field capacity (pre-wet by applying 9,440 cc water to each small plot approximately I2 hours prior to the infiltrometer run), to eliminate confounding effects of any antecedent moisture. Rainfall was applied for 30 minutes to each small plot (wet run). On the Flats site only, one-half hour after the wet run ended, simulated rainfall was again applied for one-half hour. This was called the very wet run. Average rainfall application rates (& 0.4 cm/ hr) were either 6.35 cm/ hr or 12.7 cm/ hr. The median drop diameter of the simulated rainfall was 2.9 mm. Using data from Laws (1941), the kinetic energy associated with this simulator, when run at a height of 183 cm, is about 40% that of natural rainfall. The time to ponding (or time runoff begins) was defined as the time when measurable (i.e., approximately 10 to 15 ml runoff in 15 seconds) runoff occurred from the small plot. Volume of runoff was recorded for 15 or 30 seconds (depending upon the volume) at the following times after runoff began: at I-minute intervals for the first 4 minutes, at 2-minute intervals for the next 6 minutes, and 5-minute intervals thereafter. Small plots were located within large (3.05 m X 10.67 m) rainfall simulator plots, which were part of a concurrent Agricultural Research Service (USDA) study. (Large plot data are not included in this analysis because, due to differences in instrumentation and routing on the large plot, large and small plot infiltration data are not comparable. This has been amply demonstrated by Smith (1979). Runoff from the large plot simply represents the response of a micro-watershed, in which the “rainfall excess” has been routed over the surface.) On large plots in either the fenced or Fig. 1. Maximum, average, andminimum infiltrarion rates versus timefor unfenced condition, 20 small plots were sampled within each large IO small plots on large tilled plots (sire. slope and rainfall intensity are plot. Ten small plots were systematically located within two ran- given). domly identified belt transects on each large plot. (Figure 7 illustrates the sampling design as described.) On large plots in a roto- sity were not included in calculations of the average infiltration tilled condition, 10 small plots, randomly located within each large rate because the actual infiltration rate is not quantifiable for these plot, were sampled. points. This causes a downward bias of the maximum and average After obtaining runoff data from a small plot, surface soil char- curves by an unknown amount. acteristics were sampled. Two samples of soil (with a core diameter Similar graphs (Fig. 2, 3) for the unfenced and fenced sites of 5.4 cm and core depths of 2.54 cm and 7.62 cm) were taken for portray a much wider range of possible values. In some cases, even determination of bulk density (g/cc), particle size distribution at a rainfall intensity of 12.7 cm/hr., the infiltration capacity (hydrometer method-Bouyoucos 1962) and soil organic matter exceeded the application rate for the duration of the 30-minute run (calorimetric analysis-Sims and Haby 1971). (see Table 2). Vegetal cover was measured with a point frame (Levy and Mad- The mean final infiltration rates are given in Table 2. The mean den 1933). Fifty evenly spaced points were sampled on each small final infiltration rates on fenced and unfenced plots are at least an plot. Strikes from the first hit to the ground surface were recorded. order of magnitude greater than those on the tilled plot on the same Cover was classified as bare soil, litter, vegetal basal cover, and site. Unlike on the tilled plot, the minimum curves on the fenced vegetal crown cover (by species). Rock was included in ground and unfenced plots approach values above zero. The minimum cover with small rock (2-6 mm diameter), gravel (6-20 mm diame- curves approximate the mean tilled plot curve, indicating that ter), medium rock (20-50 mm diameter), and large rock (greater conditions giving the lowest infiltration rate on fenced and than 50 mm diameter) recorded separately. unfenced plots are similar to those responsible for average infiltra- For each site, all small plot infiltration curves were used to tion rates on a tilled plot. investigate site variability. The increases and decreases in infiltration rates as shown in average and minimum infiltration curves in Figure 1-3 were possi- Results and Discussion bly due to air counterflow, inhomogeneity of the soil, other unquantified changes in soil characteristics with time, duration of Average, Maximum, and Minimum Infiltration Curves for Small the rainfall simulator runs, and instrument or sampling error. Plots within a Given Large Runoff Plot Figure 1 shows the average, maximum, and minimum infiltra- Individual Small Plot Characteristics: Lower Sheep Site as an tion curves on the tilled plots on the 3 sites. The curves were Example obtained by dividing the 30- minute run into 2-minute intervals, Having examined general small plot curves, it is useful to exam- finding the average, maximum, and minimum infiltration rates ine specific small plot curves, and their variability. The Lower over the given interval, and plotting these rates at the midpoint of Sheep Site is used as an example; the variability patterns on this each time interval. site are representative of the variability pattern found at all of the Times where the infiltration capacity exceeded the rainfall inten- sites. Figures 4, 5, and 6 show the differences between individual small plot responses on rototilled, unfenced and fenced sites. Small

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 525 0 0.090 0.180 0.110 o.am OASO O.MO o.s?a time (hours) 2a FLATS SITE. 3% SLOPE, I=&36 CM/M 16 of 20 SMALL SITES PLOTS RECEIVINQ RUNOFF

2b NANCY SITE. 0% SLOPE. l-12.7 CM/M 10 of 20 SMALL SITES PLOTS RECEIVMQ RUNOFF

2c LOWER SHEEP SITE. 9% SLOPE, i-12.7 CMlHR 3c LOWER SHEEP SITE. 0% SLOPE,,.12.7 CMlnR 15 of 20 SMALL SITES PLOTS RECEIVING RUNOFF 16 Of 20 SMALL PLOTS RECEIVING RW”OFF Fig. 2. Maximum, average, and minimum infiltration rates versus timefor Fig. 3. Maximum, average and minimum infiltration rate versus the smallplots where runoff occurred on each unfenced largeplot (site, time for the small plots where runoff occurred on each fenced slope, and rainfall intensity are given). large plot (site, slope, and rainfall intensity are given). 16.0 I I I I I I 3

12.0

2 Q-0 2 Q-0 f s E E s 6.0 s 8.0 r c

3.0 i

0 I I 1 I I I 0 0.000 0.130 0.270 0.360 0.460 0.540 0.630 0 0.000 0.130 0.270 0.380 0.430 0.340 0.630 time (hours) time (hours)

Fig. 4. Three of IO individual infiltration curves (as examples) showing Fig. 5. Six 20 infiltration curves (as examples) showing variability on the variability on the Lower Sheep tilled, 9% slope, large plot at the average Lower Sheep unfenced, 9% slope, large plot at the average rainfall rainfall intensity of 12.7 cmjhr. intensity of 12.7 cmlhr.

526 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 15.0

12.0 \.., \ %L , i \ ‘*+,m-- .q...... ,, “\ _i_ I\ “J--..a c-s ‘,:‘._...... “._...... _ .._., -1 2 0.0 ..A____ .\._._.A.- 5 -\____ ----____ E s 6.0 Ic

3.0

F”: a a* 0 J H 16 2 0 0.000 0.100 0.270 0.360 0.450 0.540 0.630 c ]t, :016 ‘: LJ. 4 4 01 6 94 time (hours) ‘s :: : ‘4” lb

Fig. 6. Six of 20 infiltration curves (as examples)showing variability on the 0, 6 Lower Sheep fenced, 9%slope, large plot at the average rainfallintensity 5 of 12.7 cm/hr. plots where the infiltration capacity exceeded the rainfall intensity for the duration of the run are not included in Figures 5 and 6. The relationship between small plot position on the slope and final 30-minute infiltration rate and vegetai cover is shown in Figures 7 and 8. These figures show that plot positioning has no impact on 30-minute infiltration rates; vegetai cover and associated soil factors have the most influence on variability of infiltration. Given this large variability in infiltration on a relatively homogeneous site, we looked at the ranges of other parameters that may be contributing to this variability. Figure 9 shows mean values for O:o,.s. ~:t.rb a:arru. L:LU,W u:uoak or :Om@ad E,w....d .I 2 time to ponding and final infiltration rate. These graphs show Fig. S Final 30-minute infiltration rates and vegetal cover relative to plot mean values f standard deviations for unfenced, fenced and tilled position on Lower Sheep fenced site. (Ground (Gr) category includes percent bare ground plus percent of soil surface covered by minute organic particles.) plots on the Lower Sheep site. There was much greater variability of values on the unfenced and fenced plots than on the tilled plot. This same trend is also evident when looking at bulk density, percent organic matter, and soil texture. Figure 10 illustrates this trend for percent organic matter. Differences between tilled and natural surfaces (i.e. unfenced and fenced) indicates that surface condition found on unfenced and fenced sites reflects either (I) the effects of tilling on creating a

b unfenced

c tancod . - .

a tllled - I 1 I I I

1.5 4.0 6.5 0.0 11.5

FINAL lNFlLTRATlON RATE AFTER 30 MINUTES (cm/hr)

a unfenced. .

, fwlced .

b 11lhd -

L I 1 I I 0 4.0 8.0 12.0 16.0

0:0t*a* C:C*m t:*.w. 1:,,,,., “:“.#fi ov :o,.vm, ,,rr..,.cl .s x TIME TO PONDING hid Fig. 7. Final 30-minute infiltration rates and vegetal cover relative to plot position on Lower Sheep unfenced site. (Ground (G) category includes Fig. 9. Mean f standard deviation for times to ponding andfinal infilrra- percent bare ground plus percent of soil su$a.ce covered by minute tion rates for 3 treatments on the Lower Sheep, 9% slope site at the organic particles). rainfall intensity of 12.7 cmf hr.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 527 a unfenced The results of this study indicate a need for further investigation of variability on rangelands. First, the variability in infiltration on a fenced . a homogeneous soil surface must be characterized for use as a baseline. Then studies very similar to the one described herein a tilled - should be done on diverse rangeland sites. Eventually, such studies I I I I I would enable a researcher to characterize the variability associated with an individual measurement or model parameter estimate for a 2.6 4.0 1.1 7.0 8.1 given site. ORGANIC MATTER (%I FOR 2.54 cm DEPTH Literature Cited Achouri, M. 1982. Spatial and seasonal variability of field measured infil- a unfenced l . tration rates on rangelands. M.S. thesis, Utah State Univ., Logan. a fenced . Blackburn, W.H. 1973. Infiltration rate and sediment production of selected plant communities and soils in five rangelands in Nevada. Final Report forcontract No. 14-l l-0001-4632. Agr. Exp. Sta., Univ. Nevada a tilled - Reno in cooperation with U.S. Dep. Interior, BLM. Blackburn, W.H. 1975. Factors influencing infiltration and sediment pro- 2.5 4.0 5.5 7.0 8.5 duction of semiarid rangelands in Nevada. Water Resour. Res. I 1:929-937. Bouyoucos, G.J. 1962. Hydrometer method improved for making particle ORGANIC MATTER (%I FOR 7.62 cm DEPTH size analyses of soil. Agron. J. 54:464-465. Busby, F.E.,and C.F. Ciiord. 1981. Effects of livestock grazing on infiltra- Fig. 10. Mean f standard deviation ofpercent organic marrerfor2.54and tion and erosion rates measured on chained and unchained pinyon- 7.62 CMsoil depthsfor 3 treatments on the Lower Sheep, 9% slope site. juniper sites in southeastern Utah. J. Range Manage. 34:400-405. homogeneous surface or (2) the effect of the diversity of vegetal Chow, V.T., and T.E. Harbough. 1965. Raindrop production for laboratory watershed experimentation. J. Geophys. Res. 70:61 I I-6120. cover (i.e., shrub, interspace, or bare ground), on soil surface Cundy, T.W. 1982. An analysis of the effects of spatial variability of point characteristics. infiltration rates on the comparison of small and large plot rainfall- runoff. Ph.D. Diss., Utah State Univ., Logan. Impact of Soil Properties and Vegetal Cover on Infiltration Rates Infiltration rate and sediment production trends on a In an effort to explain site variability on the unfenced and fenced Gifford, G.F. 1972. plowed big sagebrush site. J. Range Manage. 2553-55. plots, rangeland soil properties and vegetal cover were correlated Gifford, G.F. 1976. Applicability of some infiltration formulae to range- with hydrologic response (e.g., measured as final infiltration rate land infiltrometer data. J. Hydrol. 28:l-1 I. after 30 minutes). As has been found in previous studies (Black- Grrh, O.J. 1983. Spatial and temporal distribution of infiltration rates on a bum 1973, Gifford 1972, Williams et al. 1972, and Busby and small subalpine watershed. M.S. thesis, Utah State Univ., Logan. Gifford 198 l), the correlations showed that no set of factors con- Hawkins, R.H. 1981. Interpretation of source area variability on rainfall- sistently explained small plot variability. Stepwise multiple regres- runoff relationships. Intern. Symp. on Rainfall-Runoff Modeling. Mis- sions gave much the same results. Although R2 values indicate that sissippi State Univ. on a given area it is possible to find a set of factors that accurately Laws, J.O. 1941. Measurements of the fall-velocity of water-drops and raindrops, Trans. Amer. Geophys. Union, 22:709-72 I. predicts the final infiltration rate (see Table 3), relationships Levy, E.G., and E.A. Madden. 1933. The point method of pasture analysis. between rangeland soil properties, vegetal cover, and infiltration New Zealand J. Agr. 46~267-269. change from area to area. Hence it would be difficult to select a set Lyford, F.P., and H.K. Qashu. 1969. Infiltration rates as affected by desert of factors that would consistently explain the observed variability vegetation. Water Resour. Res. 5:1373-1376. from one area to another. Malekuti, A., and G.F. Gifford. 1978. Natural vegetation as a source of diffuse salt within the Colorado River Basin. Water Resour. Bull. 14:195-205. Given the large variability in measured infiltration and soil Meeuwig, R.O. 1971. Infiltration and water repellancy in granitic soils. U.S. Forest Serv. Res. Paper INT-I II. physical properties that we noted on relatively uniform rangeland Merzougui, M. 1982.The effect of instrument type in measuring infiltration sites, it is logical to ask how many samples are necessary to accu- rates on spatial variability patterns. M.S. thesis, Utah State Univ., rately define a site. As experienced in this study, it would take from Logan. 1 to 84 plots in order for the average time to ponding to be within 2 Murabayashi, E.T., and Yu-Si Fok. 1979. Urbanization-induced impacts minutes of the real time to ponding (population value) with 80% on infiltration capacity and on rainfall-runoff in a Hawaiian urban area. confidence. Fewer samples are necessary to accurately define the Tech. Rep. No. 27, Water Resour. Res. Center, Honolulu, Hawaii. average final (after 30 minutes) infiltration rate, as shown in Table Rogowski, A.S. 1980. Hydrologic parameter distribution on a mine spoil. 4. To be within 1 cm/ hr of the actual final infiltration rate, how- p. 764-780. In: Symposium on Watershed Management (Proc.). Amer. ever, 4 times the number of plots shown in Table 4 would be Sot. Civil Engineers, New York. Spatial variability of needed. Due to the variability on any given area, single parameter Sharma, M.L., G.A. Gander, and C.G. Hunt. 1980. infiltration in a watershed. J. Hydrol.. 45:101-122. values exhibit wide confidence intervals. Small-plot studies are Sims, J.R., and V.A. Haby. 1971. Simplified calorimetric determination of useful in defining the confidence interval for a given parameter. soil organic matter. Soil Sci. 112:137-141. The large spatial variability of field-measured infiltration rates Smith, R.E. 1979. Rainfall simulation as a research tool-simulation for and associated soil parameters suggests that, if variability is of infiltration studies. In:Proc. Rainfall Simulator Workshop, Tucson, importance, extensive sampling will be necessary to define inherent Ariz. ARM-W-10:78-84. variability patterns. Just how to incorporate such spatial variabil- Smith, R.E., and R.H.B. Hebbert. 1979. A Monte Carlo a’nalysis of the ity into hydrologic models or into a practical land management hydrologic effects of spatial variability of infiltration. Water Resour. approach is unclear; unless insights are available on this, it may be Res. 15:419-429. Spatial variability of rangeland best to first ignore the problem of spatial variability and lump Springer, E.P., and G.F. Gifford. 1980. infiltration rates. Water Resour. Bull. 16:550-552. watershed characteristics as area1 averages. If the variability of a Stephenson, G.R. 1977. Soil-geology-vegetation inventories for Reynolds given site is characterized, however, when average, effective, or Creek Watershed. Agr. Exp. Sta., Univ. of Idaho. Misc. Series No. 42. fitted values are subsequently used in modeling watershed res- Tricker, A.S. 1981. Spatial and temporal patterns of infiltration. J. ponse, inaccuracies and biases in such assumptions will be better Hvdroloay 49~261-277. defined. On the other hand, lumped parameters may be sufficiently Vie&, S.K.; D.R. Nielsen, and J.W. Bigger. 1981. Spatial variability of accurate in most cases when the costs of additional sampling are field-measured infiltration rate. Soil Sci. Sot. Amer. J. 45: 1040-1048. balanced against expected uses of the model output and the accu- Williams, G., Gifford, G.F., and G.B. Coltharp. 1972. Factors influencing racy required. infiltration and erosion on chained pinyon-juniper sites in Utah. J. Range Manage. 25:201-205.

528 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Evaluating Soil Water Models on Western Rangelands

KEITH R. COOLEY AND DAVID C. ROBERTSON

Abstract The soil profile is an important water storage reservoir within Table 1. Input requirementsfor running SPAW and ERHYM soil water the hydrologic cycle. An understanding of the factors affecting balance models. daily soil water status is necessary to increase or modify vegetation or water yields. Many mathematical simulation models have been developed to assess soil water status, but none were found that SPAW Model ERHYM Model were specifically developed for use on Western rangelands. The Initial soil water (or default Initial soil water purpose of this report was to test soil water models that appeared option) to be sufftciently general for adaption to rangeland conditions, to Daily precipitation Daily precipitation determine if they could provide adequate results, and the level of Daily runoff or SCS curve number SCS curve number sophistication required. The 2 models selected for evaluation were the Ekalaka Rangeland Hydrology and Yield Model (ERHYM) Daily pan evaporation Daily max & min temperature developed for use during the growing season on grasslands of the Monthly pan coefficient Daily solar radiation northern Great Phtms,and the Soil-Plant-Air-Water model (SPAW), Avg. annual pan evaporation which was developed for use with cultivated crops in the Midwest. Layer thickness Layer thickness Results indicated that both models could be adapted to produce A time interval adequate soil water information under rangeland conditions of A soil pressure for tolerance southwestern Idaho. Overall, the somewhat simpler ERHYM parameter model produced results more closely aligned to observed values Soil water evaporation parameter than did SPAW. The lack of a snow accumulation and melt routine Soil freezing dates in SPAW (which could be added) appeared to be the main source Root distribution Soil temperature curves by layer of observed differences. These differences were a function of timing Canopy cover curve Crop coefficient rather than a difference in total soil water at the end of each year, Canopy susceptibility to water where results for the 2 models were very similar. stress Phenology curve Plant growth curve The soil profile is one of the most important water storage Phenology susceptibility to water reservoirs within the hydrologic cycle. In arid regions, available stress soil water seldom exists for more than a few months at a time, Moisture-stress curves (as provided) because it is rapidly extracted by plant transpiration and soil Planting & harvest dates Starting growth date evaporation. Therefore, very little water ever percolates below plant rooting depth. In more humid regions, the magnitude of For more details and units refer to Model User’s Guide. infiltrated water may be more than adequate for plant needs, and excess water may percolate into ground water reservoirs. (1983) and the SPAW (Soil-Plant-Air-Water Model) developed by An understanding of the factors affecting the day-today soil Saxton, Johnson, and Shaw (1974). water status is important when attempting to increase or modify Model Descriptions vegetation or water yields. The time distribution of soil water within the root zone is a complex interaction of many variables The ERHYM model is relatively simple and requires a minimum related to present and historical climate, plants, and parent soil of input information. The SPAW model, on the other hand, materials. attempts to treat in some detail, all of the physical processes and Hildreth ( 1976,1978) indicated that many mathematical simula- interactions involved. As such, it requires more information on tions or computer models have been developed to assess soil water initial conditions and limits. However, both use essentially the status. However, most of these models were developed to satisfy a same hydrologic and meteorological data. particular need (i.e., spring wheat yield predictions in the northern ERHYM Great Plains)and may not represent other crops or locations. None The ERHYM model was developed for use in predicting runoff of the models were specifically developed for use on Western and herbage production for northern Great Plains rangelands rangelands. (mainly grasses). It provides daily runoff, soil water, snow pack, The objective of this study was to determine if existing models evaporation, transpiration, and soil water routing for up to 4 soil with soil water accounting procedures included, representing dif- layers at a range site. The model can be run on a seasonal basis, or ferent levels of complexity and data requirements, could be continuously, using daily precipitation, solar radiation, tempera- adapted to particular Western rangeland conditions. Two models ture, soil characteristics, and a plant growth curve (Table 1). It uses were selected for evaluation based on the following criteria: (1) a slightly modified Soil Conservation Service (SCS) curve number models appeared to be general enough to be adaptable to range method to determine runoff (Smith and Williams 1980; USDA, conditions; (2) models had been tested against field data (even SCS 1972) and a constant times temperature procedure for snow though for cultivated crops); and (3) documentation was readily accumulation and melt (Stewart et al. 1975). available. The models selected were the ERHYM (Ekalaka Range- The model functions as a series of cascading reservoirs (one for land Hydrology and Yield Model) developed by Wight and Neff each layer); that is, all of the water infiltrated is assumed to be stored in the first layer until it reaches its field capacity, at which Authors are hydrologist and hydrologic technician. respectively. Northwest Watershed Research Center, USDA-ARS. 270 South Orchard, Boise, Ida. 83705. time any additional infiltration is assumed to be stored in the Manuscript accepted February 29, 1984.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 529 second layer, and so on. Soil water extraction also proceeds one Potential evapotranspiration (PET) is determined from actual layer at a time beginning at the surface layer. or estimated daily pan evaporation for each day of calculation. The evapotranspiration portion of the model is essentially the Interception water on the plant and soil surfaces is a constant same as that used by Wight and Hanks (198 1). A climatic potential amount for each storm. The PET value is reduced by the amount of evapotranspiration is first calculated using the Jensen and Haise interception evaporation before plant and soil water evapotranspi- (1963) equation, which assumes a full cover of alfalfa with water ration are computed. Soil evaporation is represented by inclusion nonlimiting. The potential evapotranspiration from rangeland is of a separate thin (1.27 cm) upper boundary layer (evaporation then obtained by multiplying the value determined above by a layer) of soil from which water is readily evaporated and limited range crop coefficient. Actual transpiration is estimated by multi- only by PET. plying the potential evapotranspiration for the range site by a site Plant transpiration is a function of percent soil shading with specific transpiration coefficient and a relative growth curve fac- time during the year. Since all of the canopy does not transpire, the tor. Soil evaporation is a function of potential soil evaporation and quantity of canopy (soil shading) and its average biological ability time since the soil surface was last wet. Potential soil evaporation is to transpire, assuming adequate PET demand and water availabil- the difference between potential evapotranspiration and the poten- ity are considered separately. Therefore, a second time distribution tial transpiration. Soil evaporation is limited to water in the top 30 representing a phenological factor is included as a direct modifier cm of the soil profile that is in excess of air-dry soil water content, of the plant canopies’ ability to transpire. which is less than the lower limits of soil water availability (per- Root water abstraction is represented using a “typical” root manent wilting point). More detailed information can be obtained distribution for the plants being represented, with depth and time from the ERHYM User’s Manual (Wight and Neff 1983). through the year, thus providing, for selected dates, the percent of water to be abstracted from each soil layer containing roots. Well- SPAW watered, vigorous plants usually transpire at nearly the rate The SPAW model was developed to provide daily soil water demanded by the atmospheric conditions (PET), but as water profile estimates on cultivated crop lands in the Midwestern Uni- supply becomes limited, physical and biological controls begin to ted States. It computes a daily estimate of runoff, soil water, actual restrict the rate of transpiration. evapotranspiration including soil water evaporation, transpira- A daily estimate of actual evapotranspiration is obtained by tion, interception evaporation, and deep percolation. Principal adding the components of interception evaporation, soil water inputs include daily potential evapotranspiration (PET) and pre- evaporation, and plant transpiration. cipitation, crop descriptions of canopy, phenology, and rooting, A daily infiltration value is determined for each day with precipi- plus soil profile descriptions (Table 1). tation, by simply subtracting measured runoff if data are available, The soil profile is represented by a user selected number of layers or by using the SCS curve number method. Daily infiltration to reflect the average soil profile over the field or watershed being amounts are added to the uppermost soil layers and cascaded to studied. Each layer may be assigned a unique depth and set of lower layers as each layer’s near-saturation capability is reached. water characteristic curves (tension and conductivity) selected by After all daily infiltration has been distributed, further redistribu- soil texture. The computational sequence for the model is shown in tion is determined using a Darcian soil moisture model based on schematic block diagram form in Figure 1. pressure gradients and unsaturated conductivity for the soil types specified. The SPAW model does not contain a mechanism to account for snow accumulation and melt, although an algorithm could be added. More detailed information can be obtained from the SPAW User’s Manual (Saxton et al., In Progress).

Model Evaluation Description of Study Sites The 2 sites selected for use in this evaluation are part of the Reynolds Greek Experimental Watershed in southwestern Idaho (Robins et al. 1965). These watersheds were selected because of data availability, and because they represent “typical” sagebrush ecosystem rangeland sites found in the Northwestern United States. The two sites, called Flats and Lower Sheep, vary somewhat in elevation, size, vegetation, soils, and climate as shown in the summary of site characteristics presented in Table 2. Model Parameters and Data Requirements Each model requires specific types of input parameters and data. Comparison of the 2 models includes a comparison of input factors in addition to comparing final results. Both models require the normal input parameters such as output options, beginning and ending computation dates, etc. Data requirements too, are similar in many respects, in that both models require certain weather data to drive them, in addition to soil characteristic and vegetation data (Table 1). Although most of the parameters can be obtained from the literature or data sets, with some modifications made for different site conditions, some of the parameters require information that is not readily available for rangeland sites. This is especially true for the SPAW model, where 6 parameters or relationships concerning vegetation roots, growth, and stress are needed. Fig. 1. Schematic representation of the expanded Soil-Plant-Air- Water To run the SPAW model, some information from the literature (SPA W) model. (From Saxton et al. 1974) was used for root distribution, realizing that such data were col-

530 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Table 2. Comparisons of watershed characteristics at two Reynolds Creek Table 3. Soil water contained in specific layers, and in the total soil study sites. profile (O-137 cm), at the end of each year (19761981). at the Flats Site (in mm).

Plats Lower Sheen Layer (cm) ERHYM Measured SPAW Characteristic Size .91 Ha 13.36 Ha 1976 Elevation 12OOm 1650 m O-23 23 30 30 avg. slope 5% 16% 23-46 23 30 31 aspect n nw 46-8 I 38 36 41 soils 81-137 16 74 80 subgroup typic haplargids calcic argixerolls Total I60 170 182 family fine loamy mixed, loamy skeletal mixed Difference -10 +12 mesic % Difference 5.9 7.1 series nannyton loam searla gravelly loam 1977 geologic material seimentary basalt O-23 56 56 59 vegetation bottlebrush sandberg bluegrass 23-46 48 50 48 cheatgrass low sagebrush 468 I 36 41 50 shadscale 81-137 74 79 85 squirreltail Total 214 226 242 % vegetative cover 25% 25% Difference -12 +I6 Average precipitation 25 cm 36 cm % Difference 5.3 7.1 SCS hydrologic soil B B 1978 group O-23 46 40 44 (USDA, SCS, 1972) 23-46 23 36 35 46-81 69 43 43 lected from sites with different climatic and soil characteris- 81-137 81 81’ 83 tics. Total 219 200 205 Canopy and phenology curves were developed from vegetation Difference -19 +05 measurements made at Reynolds Creek. These data do not repres- % Difference 9.5 2.5 ent potential unstressed conditions. The canopy and phenology susceptibility curves and the moisture stress curves were used as 1979 presented in the SPAW User’s Manual, which was developed from O-23 41 33 42 studies based on crops in the Midwest. 23-46 28 32 33 ERHYM has 3 factors that are not readily available. Of these, 46-81 41 44 43 the plant growth curve, which is similar to the phenology curve in 81-137 79 85 84 the SPAW model, is the most difficult to make site specific. Data to Total 189 194 202 develop soil temperature curves by layer are also lacking (if the Difference -05 +08 model is to be used for areas with a climate different from that of 70 Difference 2.3 4.1 Ekalaka, Mont., where the soil temperature relations in the model 1980 were derived). Since the temperature of the deeper layers varies O-23 56 50 49 only slightly through the year, these curves can be estimated with 23-46 46 41 37 some confidence. The third factor not readily available for range- 46-8 I 41 45 44 lands is the crop coefficient, which is merely a ratio of transpiring 81-137 84 84 83 canopy to total canopy cover. Little guidance for estimating range- Total 227 220 213 land crop coefficients can be obtained from the literature, unless Difference +07 -07 lysimeter data are available. % Difference 3.2 3.2 1981 Model Calibration O-23 56 Both models were first run using data from the Flats sites for the 61 63 23-46 64 1979 calendar year and parameter values based on previous expe- 60 49 46-8 I 48 rience, site characteristics, or best estimates. After the initial runs, 47 45 81-137 81 some model parameters were adjusted to better match observed 88 80 Total 249 soil moisture trends and totals. Once the differences between 256 237 Difference -07 model determined values and observed values were minimized, the -19 % Difference 2.7 model was assumed to be calibrated for that site. The model was 7.4 then run for the total period of adequate record (1976-1981) at the Flats site. Since this record contained years that received more and radiation or evaporation data. The 4 soil layers used (O-23 cm less precipitation than the year chosen for initial calibration, minor 23-46 cm, 46-76 cm, and 76- 137 cm), were selected on the basis of adjustment of the soil water characteristics were made to improve data availability rather than soil characteristics. As noted, differ- overall performance of the models. ences between model estimated and observed values for the indi- Soil water content produced by the 2 models after final adjust- vidual layers, and years, are generally small. The largest deviations ments were made are presented in Table 3 for individual layers and are found in the upper 2 layers where most of the activity takes for the total profile, for each year. The runs were made for each place. Because of compensating differences between the different year individually, that is, the initial soil moisture values were set layers, and the inherent bookkeeping procedures contained in the equal to the last measured values for the previous year (usually models, individual layer values may be off somewhat, but the total near December 27 or 28), and the model was run for the entire year annual soil water for the entire soil profile is within 10% of the as driven by observed temperature, precipitation, and observed value. The differences are somewhat greater for the

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 531 ERHYM model in this case. by ERHYM. In general both models follow the observed data A better evaluation would be to compare model predicted soil quite well, one being a little better one year or in a certain period, water with observed soil water throughout the year, since conceiv- but the opposite being true for another year or period. ably the year-end totals could be close, but the annual trends might be out of phase. A continuous record as estimated by the models Model Comparison and Discussion and the observed soil water at approximately 2-week intervals is Validation of the models consisted of using the information shown for the 1976 and 1978 years at the Flats site (Fig. 2 and 3). gained during calibration, concerning operation of the models and range of parameter values applicable to rangeland conditions, to test the models at a new site where climate and watershed characteristics differed from the calibration site. In applying the models to the new site, Lower Sheep, only physical parameters related to watershed characteristics were changed, rather than adjusting coefficients and parameters to provide a best fit, as was done in the calibration runs. The growth or phenology curves were delayed 15 days to account for the slightly cooler climate at the higher elevation, but shape and length of growing season remained the same. Results of the model runs at the Lower Sheep site are presented in Table 4. The differences between the observed and calculated ,o_ +_.o_.~ - .r . .. _ . . ...*._ ~ =zo- I Table 4. Soil water contained in specific layers and in the total soil profile g _ IG to lS)ICnl L.yw 100 - (O-137 cm) at the end of each year (1977-1981)at Lower Sheep (in mm).

clo-P *-P ...___. t . -*” *** --e* ? * 0 - & - e* __ (Lo- .o- , , , , Layer (cm) ERHYM Measured SPAW JAN FEE MAR APR tA*‘l JUN JUL AU0 SEP OCT NO” DEC 1976 1977 O-23 I97784 68 67

Fig. 2. Field measured ( @ ) and model-predicted (SPA W q --; 23-46 76 71 67 ERH YM= --)soil water contentfor 4 layers,from the surface to 137cm 46-8 1 74 82 93 depth, with time for 1976 at the Flats site. 81-137 137 136 144 Total 371 357 371 As previously mentioned the major changes in soil moisture Difference +I4 +I4 occur in the upper 2 layers. The only significant change in the 46-76 % Difference 3.9 3.9 cm layer occurred during the first half of 1978 (Fig. 3), where 1978 O-23 76 60 61 23-46 43 57 61 46-81 76 81 89 81-137 I50 148 I52 Total 345 346 363 Difference -01 +I7 % Difference 0.3 4.9 1979 O-23 74 63 61 23-46 46 65 56 46-81 76 79 86 81-137 137 134 I48 Total 333 341 351 Difference -08 +I0 % Difference 2.3 2.9 IOOC (10 ~...*..g..l _...... _A.-- ~ ” ” :w 1980 60 - .o - O-23 84 62 61 23-46 58 57 57 JAN FEG MAR APR MAY J”N J”L AUG SEP OCT NOV OEC 1979 46-8 I 84 75 86 81-137 147 133 147 Fig. 3. Field measured (& ) and model-predicted (SPA W = --; Total 373 327 351 ERHYM= ---- soil wafer content for 4 layers, from the surface to 137 Difference +46 +24 cm depth, with time for 1978 at the Flats site. TO Difference 14.1 7.3 observed soil moisture increased about 16 mm. Both models also 1981 indicated an increase in the 46-76 cm layer during the first half of O-23 84 82 73 1978, suggesting that they properly accounted for the deeper infil- 23-46 84 84 70 tration that occurred then. The SPAW model best matched the 46-8 I 97 99 105 81-137 132 155 162 observed trend in magnitude although it predicted the increase Total 397 420 410 would start before it was observed. The SPAW model does not Difference -23 -10 account for storage in the form of snow which would delay the % Difference 5.5 2.4 entry of the water into the soil. The ERHYM model which contains a snow accumulation and melt algorithim followed the actual timing of the soil water increase better, but differed considerably in soil water values are slightly greater for some years than those the magnitude of the change, predicting it to be about 60 mm rather obtained at the Flats site, but the maximum difference was still than the 16 mm observed. Similar results were obtained for the only 14%. Values produced by ERHYM (Tables 3 and 4) for lowest layer where the magnitude of the observed changes was best individual years differ from observed values more than values predicted by SPAW, but timing of the change was predicted best produced by SPAW, but ERHYM long-term (1976-198l)averages are actually closer to observed averages.

532 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Table 5. Regression equationsand coefficients for model simulated versus observed soil water on measurement days (approximately two-week intervals), in mm.

SPAW ERHYM Site Year Regression Equation R2 Regression Equation R* Flats 1976 Y = 95.499 + 0.4431x .56 Y = 7.046 + 0.9233X .9l 1977 Y = 108.049 + 0.4453X .36 Y = 41.298 + 0.7462X .76 1978 Y = 3.178 + 1.0162X .83 Y = -3.058 + 0.7462X .95 1979 Y = 28.524+ 1.1351X .6l Y q 308.224 + 2.5982X .9l 1980 Y = 7.442 + 0.8960X .83 Y = 76.154 + 0.6501X .74

1976-8 I Y q 50.584 + 0.7065X .65 Y = -29.093 + 1.163X .84 Lower Sheep 1977 Y = 100.292 + 0.6870X .73 Y = -61.239+ 1.2183X .78 1978 Y = 92.819 + 0.7272X .85 Y q -11.971 + 1.0376X .98

1979 Y q 198.984 + 0.4091X .64 Y = -18.341 + 1.0978X .91 1980 Y = 123.269 + 0.6226X .84 Y q 104.160 + 0.7694X .89 1981 Y = 99.515 +0.6583X .84 Y = 32.923 + 0.8899X .95

1977-81 Y = 117.488 + 0.6379X .79 Y q 21.958 + 0.9603X .89

Y q Model simulated soil water X = Observed soil water

Regression analysis between model predicted soil water and should be considered if daily values are important and snow is a measured soil water, on each day of observation provides another factor. way to compare results (Table 5). Presented are: (I) the regression Figures 4 and 5 show the soil water content in the 0 to 137 m cm equations and coefficients for both models and sites for individual profile as predicted by the models, and observed, for 1978 and 1980 years; (2) the regression equations and coefficients for both models at the Lower Sheep site. These figures represent a true validation and sites for all years combined; and (3) the coefficient of determination R2 associated with each equation. ERHYM predicted values correlate better (high R2) with the approximately 26 biweekly observations than SPAW predicted values, in all but one case ( 1980 at the Flats site). This indicates that trends between observed soil water changes and ERHYM predicted changes are similar even though the magnitude of actual values may be different. Precipitation falling as snow may cause the reduced correlations observed for the SPAW model. In other words, the trend between observed conditions and SPAW predicted conditions could be opposite at times because snowfall Fig. 5. Field measured (0 ) and model-predicted (SPA W q -; occurred, but was handled by the SPAW model as rain. This would ERHYM = -) soil water conlenl for the loral soil profile, from the cause the SPAW predicted soil water to show an increase when surface to 137 cm deplh. wilh limefbr 1980 at the Lower Sheep sire. observed soil moisture indicated no change, the water still being stored on the surface as snowpack. Later as the snow melts, period, and again indicate that both models produce results that observed soil moisture would increase while SPAW predicted are similar to the observed trends, with ERHYM being slightly values would remain unchanged. This out-of-phase soil moisture better one year and SPAW being slightly better the other year. accounting would occur in late fall and winter, and can be observed Generally, observed data fall between the 2 model predicted traces. somewhat in the top layer of Figure 3. A snow accumulation and melt routine could be added to Conclusions SPAW to improve results. However, the purpose of this study was to test “existing”models to determine their applicability to western Conclusions reached in this evaluation of 2 soil water balance rangelands. Results indicate that (I) even without accounting for models can be summarized as follows: Both models were adaptable snow the SPAW model produced values closed to observed, and to Western rangeland conditions and produced results which fol- (2) storage of water in a snowpack could cause timing errors, and lowed observed trends adequately for most sites. In the present form of the models, the ERHYM model would probably produce better, or more economical, results if (1) a snowfall occurred during the study period that would cause temporary storage of significant amounts of precipitation on the surface; (2) plant growth and stress data are not available for the species under investigation at a semiarid range site; and (3) computer capacity is limited, or cost of computer operation is significant. (SPAW takes about 4 times longer to run than ERHYM). The daily scheme of modeling used by both models is a significant advancement in soil water determination for use in investigat- ing climatic influences, timing of vegetation manipulation, etc.

Fig. 4. Field measured (a ) and model-predicted (SPA W = ---_; Literature Cited ERH YM = -) soil water content for the lotal soil profile. from the sur$ace to 137 cm deprh. with rime for 1978 at rhe Lower Sheep sire. Hildreth, W.W. 1976. Plant and Crop Growth Simulation Models: A Literature Survey, Technical Memorandum LEC-9174. Lockheed Elec- tronics Company, Inc., Aerospace Systems Division, Houston, Texas.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 533 Hildreth, W.W. 1978. Soil Moisture Modeling Review, Tech. Memo. Smith, R.E., and J.R. Williams. 1989. Simulation of surface water hydrol- LEC-I 1857. Lockheed Electronics Company, Inc., Aerospace Systems ogy. In: Walter C. Knisel (ed.) CREAMS: A Fieldscale Model for Division, Houston, Texas. Chemical, Runoff, and Erosion from Agricultural Management Sys- Jensen, M.E., md H.R. Hnise. 1963. Estimating evapotranspiration from tems, USDA, Conserv. Rep. No. 26, Washington, D.C. solar radiation. Amer. Sot. Civil Eng., Proc., J. Irrigation and Drainage Stewart, B.A., D.A. Woolhiser, W.H. Wischmeier, J.H. Caro,and M.H. Div. 89:15-41. Frere. 1975. Control of water pollution from cropland. Vol. 1,A manual Robins, J.S., L.L. Kelley, and W.R. Hnmon. 1965. Reynolds Creek in for guideline development, USDA-ARS Headquarters, Rep. No. ARS- southwest Idaho: An outdoor hydrologic laboratory. Water Resour. H-5-1, Washington, D.C. (Series discontinued). Res., 1:407-413. U.S. Department of Agriculture, Soil Conservation Service. 1972. National Saxton, K.E., P.F. Brooks,and R. Richmond. Users Manual for SPAW- Engineering Handbook, Hydrology, Section 4. Washington, D.C. Soil-Plant-Air-Model, In progress (To be published as USDA-ARS- Wight, J.R., and R.J. Hanks. 1981. A water-balance, climate model for Western Region Ser. Pub.). range forage production, J. Range Manage. 34:307-311. Saxton, K.E., H.P. Johnson, and R.H. Shaw. 1974. Modelingevapotrans- Wixht. J.R., and E.L. Neff. 1983. Soil-Vegetation-Hydrology’studies: piration and soil moisture. Trans. Amer. Sot. Agr. Engr. 17:673-677. v&me If. A User Manual for ERHYM: The Ekalaka Rangeland Hydrology and Yield Model. USDA-ARS, Agr. Res. Results- ARR-W- 29, Oakland, Calif. Characteristics of Oak Mottes, Edwards Pla- teau, Texas

R.W. KNIGHT, W.H. BLACKBURN, AND L.B. MERRILL

Abstract Infiltration of live oak mottes on watershed properties was 1975). Infiltration rates of honey mesquite (Prosopis glandulosa evaluated in July 1979 at the Sonora Agricultural Research Sta- var. glandulosa) canopy zone in the Rolling Plains of Texas were tion. Infiltration rates of undisturbed live oak mottes and those greater than in adjacent midgrass- or shortgrassdominated areas. with mulch layers removed were greater than rates of adjacent The lowest infiltration rates and greatest sediment loss occurred in grass-dominated areas. However, infiltration rates of oak mottes shortgrassdominated areas. Livestock grazing and brush control with mulch layer and organic layer removed exposing the mineral practices altered infiltration rates and sediment loss of midgrass- were similar to those of adjacent grass-dominated areas. Infiltm- dominated areas the most but had little or no influence on the tion rates of midgrass-dominated sites were greater than those of shrub canopy zone or shortgrass-dominated areas (Wood and shortgrass-dominated sites. Greatest soil loss occurred from oak Blackburn 1981, Brock et al. 1982). motte with mineral soil exposed with little differences between Live oak (@ercus virginiana var. virginiana) mottes generally other treatments. Infiltration rates and sediment production of oak occur on 20 to 50% of the land of the Edwards Plateau region of mottes were most influenced by grass standing crop, mulch and Texas. The mottes provide important goat and wildlife forage and organic layers, soil organic matter, and water stable aggregates. wildlife cover. The objectives of this study were to determine: (1) Organic matter and water stable aggregates in the oak motte were soil hydrologic properties of oak mottes and adjacent grass- similar and significantly greater than in the adjacent grass- dominated areas, and (2) the influence of mulch and organic layers dominated areas. Surface soil bulk density and texture were similar on infiltration rates of oak mottes. for the oak mottes and grass-dominated areas. Grass standing crop Study Area was similar for the oak motte and midgrassdominated areas but significantly greater than for shortgrass-dominated areas. Mulch The study was conducted during July 1979 at the Texas Agricul- accumulation was 6 times greater in oak motte than midgrass areas tural Research Station located about 45 km south of Sonora, and 43 times greater than in shortgrass areas. Texas. The rolling, stoney hill topography that characterizes the station is typical of the Edwards Plateau. The study area was Infiltration, the movement of water into the soil, and soil erosion located on a gentle sloping (4%) Tarrant silty clay soil series. are modified by the amount and kind of vegetation (Blackburn Tarrant soils are members of the clayey-skeletal, montmorillonitic, 1975). Range practices such as brush control and livestock grazing thermic family of Lithic Haplustolls. The station is approximately that alter vegetation also have the potential to alter soil surface 632 m in elevation with an average growing season of 240 days and hydrologic characteristics (Blackburn et al. 1982, Blackburn 1983). mean annual precipitation 554 mm. The Edwards Plateau is Infiltration rates of sagebrush (Arremisiu spp.) canopy zones in second only to the Trans-Pecos region of Texas in length and Nevada were greater than the adjacent interspace areas (Blackburn frequency of drought (Sprott 197 I). The site is characterized by live The authors are presently assistant professor and professor of watershed manage- oak mottes with short and midgrasses dominating the interspace ment, Department of Range Science, Texas A&M University, College Station 77843, between mottes. Dominant herbaceous plants on the study site and former professor of range science, Texas Agricultural Experiment Station, Son- ora 76950. include common curly mesquite (Hiluriu belungeri) the dominant Published with the approval of the Director, Texas Agricultural Experiment Sta- shortgrass, sideoats grama (Bouteloua curtipendula)the dominant tion as TA-18880. midgrass, and threeawn (Aristidu spp.). Dominant woody plants Manuscript accepted December 21, 1983.

534 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 are live oak, honey mesquite, shinoak (Quercus pungens var. Results and Discussion and ash juniper The study area was vaseyana), (Juniperus ashei). Vegetstion and Soils grazed by cattle, sheep, and goats stocked at a moderate rate (8.1 Grass standing crop of the shortgrass and oak mottedominated ha/ AU/ yr), on a continuous yearlong basis. areas were similar and significantly lower than midgrassdomina- Methods ted areas (Table I). Forb standing crop was low but similar on all

Infiltration rates and sediment production were determined Table 1. Mean above ground plant biomass accumulation (kg/ha) for using a drip-type rainfall simulator (Blackburn et al. 1974) on ten undistrubed oak mottes and grass dominated areas, Edwards Plateau, 0.5-m* randomly located runoff plots in each vegetation type and Texas. treatment. Runoff plots in the grass-dominated areas were pre-wet with 120 liters of water, using a sprinkler system, to remove ante- Vegetation type cendent soil-water content differences and covered with plastic to maintain uniform surface water conditions. After the runoff plots Variable Oak motte Midgrass Shortgrass drained to field capacity (approximately 24 hr), simulated rainfall Grass standing crop 332 b’ 682 a 150b was applied at a rate of 20.3 cm/ hr for 0.5 hr. Simulated rainfall at Forb standing crop 18a I1 a 13a a rate of 22.9 cm/ hr was applied to oak mottes with soil antecedent Mulch layer 21,500 a 3,400 b 5ooc moisture condition initially dry and at field capacity (approxi- Surface organic layer 23,000 -2 - mately 24 hr after the first simulated rainfall). Between simulated IMeans of each variable with the same letter are not significantly different (E.05). rainfall events the plots were covered with plastic to maintain ‘No measurement taken. uniform surface water conditions. These application rates approx- ‘Organic layer was included in mulch determinations for grass dominated areas. imate a storm with a I50-year return period and were necessary in order to insure runoff from all plots. Runoff was collected contin- 3 vegetation types. Mulch accumulation was greatest in the oak uously and measured by weight at S-minute intervals. Infiltration mottes and lowest in the shortgrassdominated areas. Mulch rates were determined by calculating the difference in simulated accumulation in the midgrass areas was 6.8 times greater than in rainfall applied and runoff collected from each plot. Runoff col- shortgrass areas, and 6.3 times less than in oak mottes. The decom- lected at the termination of the simulated rainfall event was tho- posed surface organic layer averaged 23,000 kg/ ha in oak mottes, roughly mixed and a l-liter subsample was collected. The sediment but could not be separated from the mulch layer on the grass- from each subsample was collected on #1 Whatman filter paper, dominated areas. Bare ground was not exposed in the oak mottes; dried at 105°C for 24 hr., weighed, converted to sediment yield in however, in the midgrass and shortgrassdominated areas, bare kg/ ha, and used as an index of sheet erosion. ground was exposed on 16 and 2290, respectively (Table 2). Rock The percentage ground cover by midgrass, shortgrass, and forb Table 2. Mean berbaceous and ground cover values (Q) for undisturbed foliage and mulch, rock, and bare ground were determined by oak mattes and grass dominated areas, Edwards Plateau, Texas. ocular estimates on each runoff plot from a gridded sampling quadrat. Grasses, forbs, and standing dead material were clipped to a 2-cm stubble height and mulch was hand-collected from each Vegetation type runoff plot. The herbaceous material was dried at 60“ C for 48 hr Cover component Oak motte Midgrass Shortgrass and weighed. Bare ground 0 b’ 16a Soil bulk density and soil moisture content were measured at 22 a Rock Oa 3a 3a depths of 0 to 3 cm and 5 to 8 cm on areas adjacent to runoff plots Mulch 77 a 46 b 46 b prior to each simulated rainfall event. Soil bulk density was deter- Forb 2a 2a 2a mined by the core method (Black 1965), and soil moisture content Midgrass 21 b 28 a IC by the gravimetric method (Gardner 1965). A soil sample was Shortgrass oc 5b 26 a collected to 3 cm deep within each plot after each simulated rainfall ‘Means of each cover component followed by the same letter are not significantly event and analyzed for organic matter content by the Walkley- different (p1.05). Black method (Walkley and Black 1934), aggregate stability by the wet sieve method (Kemper and Koch 1965), and texture by the cover was low and similar in all 3 vegetation types. The percentage hydrometer method (Bouyoucos 1962). of the study plots covered by mulch was greater in the oak mottes Oak trees were removed from the study plots by hand slashing than in the midgrass or shortgrassdominated areas. Although and the following conditions within the oak mottes were evaluated mulch accumulation on midgrassdominated areas was signifi- with soil moisture initially dry and at field capacity: (1) mulch layer cantly greater than on shortgrass areas, the percentage of the area and organic layer (decomposed mulch) left in place, (2) mulch layer covered by mulch was the same for these 2 vegetation types. Forb removed, and (3) mulch layer and organic layer removed. The cover was low and similar for the 3 vegetation types. Midgrass adjacent midgrass- and shortgrass-dominated areas were evalu- foliar cover was greater on midgrassdominated areas than in oak ated with antecedent soil moisture initially at field capacity. The mottes or on shortgrass areas. Shortgrasses did not occur in oak 7-ha study area was stratified by oak mottes, midgrass, and mottes and only covered 5% of the midgrass areas. shortgrass-dominated areas. Ten plots were randomly located Surface soil organic matter content was similar for the 3 oak within each vegetation type and oak motte mulch and organic layer motte treatments and more than double that of the adjacent grass condition for the 2 antecedent moisture conditions. Sample size dominated areas (Table 3). Soil organic matter content of the was 80 plots. Data normality was determined by tests for skewness midgrass and shortgrass dominated area was similar. No differen- and kurtosis (Snedecor and Cochran 1971). Values for sediment ces occurred in bulk density or texture between the oak mottes and production were highly skewed, requiring a log10 transformation grassland areas. However, there was a trend for greater percentage of the data set. Differences between treatments were determined by of sand size particles in the oak motte soils. analysis of variance, and treatment means were separated by Dun- Infiltration Rate and Sediment Production can’s multiple range test. Within treatment variation (variation Oak motte infiltration rate under dry antecedent moisture con- among subplots) was allocated to residual for testing differences ditions with the mulch layer removed was significantly lower than (~~5.05) among treatments (Steele and Torrie 1980). under undisturbed conditions or with the mulch layer plus the organic layer removed exposing the mineral soil (Fig. 1). However,

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 535 Table 3. Mean infiltration rate (after 30 min), sediment production and surface soil characteristics of oak mattes and adjacent grass dominated areas of the Edwards Plateau, Texas.

Oak mom Mulch and organic Variable Undisturbed Mulch removed Midgrass - layers removed Shortgrass Infiltration rate (cm; hr) Dry 22.0 a 19.9 b 21.3 a Field capacity 22.0 b 21.4 a 19.6 b 19.9 b 19.4 b Sediment production (kg! ha) 5 c I41 b 796 a 69 b 315 ab

Soil organic matter (F) 14.2 a 16.6 a 14.8 a 5.7 b 5.9 b Soil water stable aggregates (q>) 81 a 80 a 78 a 69 b 68 b Bulk density (g; cc) O-3 cm 1.2 a 1.3 a 1.2a 1.2 a 1.3 a 5-8 cm 1.3 a 1.4 a 1.3 a I.4 a l.4a

Texture (R) Clay 43.4 a 43.9 a 44.3 a 50.1 a 51.2a Sand 20.0 a 19.2 a 20.2 a 12.6a 13.3 a

‘Means of each bariable wth the fame letter are nor significantly different (K.05). ‘No measurement taken. infiltration rates under field capacity antecedent moisture condi- I%, were present on the plots with mineral soil exposed. tions with the mulch layer removed were similar to those on the The mulch layer on the undisturbed plots retained water longer, undisturbed oak mottes and significantly greater than with the thus allowing infiltration around the non-wettable areas. The rela- mulch and organic layers removed (Fig. 2). Normally, areas with a tively rapid infiltration rate of the plots with mineral soil exposed dry organic layer would have a greater infiltration rate than areas was due to well-aggregated soils and large macropores. However, with mineral soil exposed to raindrop impact. The greater infiltra- infiltration rate after 30 minutes on the bare mineral soil plots was tion rate of the areas with exposed mineral soil was attributed to 1.7cm/ hr greater with antecedent soil moisture initially dry rather the presence of non-wettable areas in the organic layer. Dry than at field capacity (Table 3). The lower infiltration rate on field organic layers often become difficult to wet. This can occur capacity plots was attributed to the swelling of clays and the sealing because of an irreversible drying of organic matter (Hooghoudt of soil pores from raindrop impact. Infiltration rates of undis- 1950) or because of fungal mycelia in the organic layer (DeBano turbed oak mottes were similar regardless of antecedent soil 1981). A dry water repellent layer strongly resists water penetra- moisture. tion. The initial infiltration rates are slow but generally increase if Infiltration rates of the grass-dominated areas were lower than water remains in contact with the repellent layer. Once the repel- those of undisturbed oak mottes and mottes with mulch layer lent layer is saturated, the infiltration rate will usually be as rapid as removed, but were similar to oak mottes with mineral soil exposed that on a wettable layer (DeBano 1981). The non-wettable areas (Table 3). Infiltration was more rapid in midgrass (bunchgrass) covered 10 to 20% of the undisturbed and mulch-layer removed dominated areas than in shortgrass (sodgrass) dominated areas. plots. Small amounts of non-wettable areas, generally less than Although the difference between these 2 vegetation types was not

23 l-

[I- 231 O 22 %;---___ -, ‘\ ---Cl--__ ----+_-_~ ‘; ‘1 -o--_-* f 4 -\ E ‘n 3 21I- ‘\ -\ -\ -1 s! Undisturbed “ndlsfurbed “4 d M”lCh Layer Removed L\ --- Mulch Layer Removed Mulch and orgmc ‘TA., s -.- Mulch and Organic Layers Removed “..A 0 20,- Layers Removed k E -c

I8 1 I 18I- 10 15 20 25 30 0 5 10 15 20 25 30 O 5

Time (min) Time (min)

Fig. 1. Mean infiltration rates for the oak matte treatments with dry Fig. 2. Mean infiltration rates for the oak motte treatments, antecedent antecedent soilmoisture conditions, Edwards Plateau. Texas. Meansfor soilmoisture atfieldcapacity, Edwards Plateau, Texas. Meansforeach 5 each J-minute interval with rhe same letters are not significantly different minute interval with the same letter are not signtficant1.v different (PS.05). (PS.05).

536 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 significant, more recent long-term studies on this area have shown Blackburn, W.H. 1983. Livestock grazing impacts on watersheds. Range- significantly greater infiltration rates on midgrass area than on lands 5: 123-125. shortgrass areas (Blackburn 1983, M&alla et al. 1984). Blackburn, W.H., R.W. Knight,and M.K. Wood. 1982. Impact of grazing Sediment production was variable and few differences occurred on watersheds: a state of knowledge. Texas Agr. Exp. Sta. Pub. among the various soil cover situations. Soil loss was greatest from MP-1496. Bouyoucos, G.J. 1962. Hydrometer method improved for making particle oak mottes with mineral soil exposed and shortgrass areas and size analysis of soil. Agron. J. 54464-465. lowest from undisturbed oak mottes (Table 3). The greater soil loss Brock, J.H., W.H. Blackburn, and R.H. Haas. 1982. Infiltration and from the areas with mineral soil exposed than the other treatments sediment production on a deep hardland range site in north central was attributed to raindrop impact which destroys soil aggregates Texas. J. Range Manage. 35195-198. and dislodges soil particles from the soil surface. DeBano, L.F. 1981. Water repellent soils: a state-of-the-art. USDA Pacific Infiltration rates and soil loss not only are influenced by the Southwest Forest and Ranae Exn. Sta. Gen. Tech. Rep. PS W-46. amount of vegetation but by the kind of vegetation. Plant foliage Gardner, W.H. 1965. Water c&tent In: C.A. Black (ed.); Methods of soil and mulch provide protection to the soil from raindrop impact, analysis. Amer. Sot. Agron. Series No. 9. Madison, Wis. slow down runoff, and provide organic matter that aggregates soil Hooghoudt, S.B. 1950. Irreversiblydisiccated peat, clayey peat, and peaty clay soils. The determination of the degree of reversibility. Int. Cong. particles. Thus as the amount of vegetation increases, infiltration Soil Sci. Trans. 43 l-34. rates should increase and soil loss decrease. However, some plants Kemper, W.O., and E.J. Koch. 1965. Aggregate stability of soils from the have a greater influence on infiltration rates and sediment produc- western portions of the U.S. and Canada. USDA Tech. Bull. 1355 tion than others. The soils under oak mottes are strongly aggre- McCaIIa II, G.R., W.H. Blackbum, and L.B. Merrill. 1984. Effects of gated and covered with mulch, organic layers, and bunchgrasses, livestock grazing on infiltration rates, Edwards Plateau of Texas. J. which results in greater infiltration rates and lower soil loss than Range Manage. 37~265-269. from adjacent grassland areas. Likewise, midgrasses (bunchgrasses) Snedecor, G.W.,and W.G. Cochrm. 1971. Statistical methods. Iowa State are more effective in modifying adjacent grassland areas than Univ. Press. Ames. shortgrasses (sodgrasses). Thus, greater infiltration rates and lower Sprott, J.M. 1971. Texas droughts during the 20th Century. Chapter 1. Beef cattle management during drought. Texas Agr. Ext. Serv. Pub. soil losses occurred from midgrass-dominated areas than short- B-l 108. grass-dominated areas. Steele, R.G.D., and J.H. Torrie. 1980. Principles and Procedures of Statis- Literature Cited tics. McGraw-Hill Book., Inc. New York, N.Y. Walkley, A., and A.1. Black. 1934. An examination of the Deqtjareff Black, C.A. (ed.). 1965. Methods of soil analysis. Amer. Sot. Agron. Series method fordetermining soil organic matter and a proposed modification No. 9. Madison, Wis. of the chromic acid titration method. Soil Sci. 37:29-38. Blackbum, W.H., R.O. Meeuwig, and C.M. Skau. 1974. A mobile infil- Wood, M.K.,and W.H. Blackburn. 1981. Grazing systems: their influence trometer for use on rangeland. J. Range Manage. 27:322-323. on infiltration rates in the Rolling Plains of Texas. J. Range Manage. Blackbum, W.H. 1975. Factors influencing infiltration and sediment pro- 34:33 l-335. duction of semiarid rangelands in Nevada. Water Resour. Res. 1:929-937.

Announcing the 1985 Annual Meeting of the Society for Range Management: Hotel Utah Salt Lake City, Utah February 11 - 14,1985 You are invited to join the expected 1,500 range professionals who will gather in Salt Lake City, Utah, at the Hotel Utah for the 1985 Annual Meeting of the Society for Range Management. February 11 - 14,1985.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 537 Soil, Vegetation, and Hydrologic Responses to Grazing Management at Fort Stanton, New Mexico

N. DEDJIR GAMOUGOUN, ROGER P. SMITH, M. KARL WOOD, AND REX D. PIEPER

Abstract The purpose of this study was to evaluate vegetation, soils, in south-central New Mexico at latitude 33’30’ North and longi- infiltration rates, and sediment production as they relate to live- tude lOY’31’ West. The Range is in foothills transition between stock exclusion, continuous heavy grazing, continuous moderate shortgrass prairie and pinyon-juniper-oak types. Elevation of the grazing, and rotation grazing on a homogeneous plant-soil com- 10,522-ha ranch varies from 1,890 to 2,286 m and topography plex. The exclusion of livestock resulted in infIltration rates sign& consists of rolling hills and flat-topped mesas separated by deep cantly higher than when the pastures were grazed in any system. No canyons. differences were found between heavily and moderately stocked Mesa tops and bottomlands were dominated generally by her- pastures. This was attributed to organic matter additions from baceous vegetation, whereas slopes and rocky ridges were domi- forbs that replaced grasses when the area was heavily grazed. The nated by woody species (Pieper et al. 1971). The major forage rotation treatment had infiltration rates that were lower than the species was blue grama (Bouteloua grucilis (H.B.K.) Lag. ex exclosures or continuous grazing treatments. Sediment production Steud.), which was dominant or codominant on all sites. Other from interrill erosion was similar in all treatments except when the important grasses included wolftail (Lycurus phleoides H.B.K.), livestock were concentrated into a fourth of the rotation system’s sand dropseed (Sporobolus cryptandrus (Torr.) Gray), mat muhly area, which resulted in higher sediment levels. (Muhlenbergia richardsonis (Trin.) Rydb.), ring muhly (Muhlen- bergiu torreyi (Kunth) Hitchc.), red three awn (Aristidu longisera Better use and protection of limited soil, vegetation, and water Rydb.), galleta (Hiluriujumesii (Torr.) Benth.), Hall’s panic (Puni- resources are vital to the welfare, if not survival, of civilization as cum hallii Vasey), and sideoats grama (Bouteloua curtipendula global population continues to expand. Development of these (Michx.) Torr.). Common forbs were Carruth sagewort (Artemisiu resources on rangeland watersheds has been largely neglected, carruthii Wood), scarlet globemallow (Sphaerulcea coccinea (Pursh) although these watersheds occupy 40% of the earth’s land surface- Rydb.), Dakota vervain (Verbena bipinnatifida Nutt.) and Rocky 80% of which lies within arid and semiarid regions. In these drier Mountain zinnia (Zinnia grandifloru Nutt.). Important woody regions, overgrazing and short-term droughts drive a cycle of species included pinyon pine (Pinus edulis Engelm.), single-seeded decreasing soil surface micro-roughness and macroporosity, de- juniper (Juniperus monosperma (Englem.) Sarg.), walking stick creasing water infiltration, increasing surface runoff and evapora- cholla (Opuntiu imbricara (Haw.) DC.), and wavey-leaf oak tion, increasing sediment production, decreasing vegetation pro- (Quercus unduluta Torr.). Also present were bottlebrush squirrel- duction, and increasing land barrenness (Dixon and Simanton tail (Sitanion hystrix (Nutt.) J.G. Smith) and broom snakeweed 1977). This cycle leads to desertification and irreversible deteriora- (Xunrhocepholum surothrae (Pursh) Skinners). tion of soil, vegetation, and water resources. Infiltration rate is an Soils were of the Dioxice series which is in the fine-loamy, important process on rangeland watersheds because it influences mixed, mesic family of Aridic calciustolls. This soil was found in soil water content, sediment, and on-site vegetation production. It pastures of each treatment. This allowed a valid comparison to be should be controlled by resource managers through the manipula- made of treatments (Pieper et al. 1978). tion of vegetation. Grazing systems based on assumptions that they The average annual precipitation recorded at headquarters was improve production, cover, and composition of vegetation have 383.5 mm, 65% of which generally occurred from June through been recommended as “best management practices” for improve- September. Precipitation during summer was primarily from ment of rangeland watershed conditions (Wood 1980). Although convection-type storms and rainfall was often intense and local- this recommendation appears valid, only a small amount of inves- ized. Precipitation during 1979 fell earlier than usual (Fig. I). May tigation has been carried out to quantitatively describe effects of and June are normally dry (23 and 32 mm, respectively) but about grazing systems on rangeland watersheds through time (Knight et 46 and 70 mm was recorded during these 2 respective months in al. 1980, McCalla et al. 1984). No such investigation has been 1979. conducted in New Mexico. Temperatures were mild, with an annual mean of Il. 1 ‘C. The The objective of this study was to describe quantitatively the range was from a mean minimum of -6.7 ‘C in January to a mean influence of selected grazing schemes and grazing exclusion on maximum of 18.9 ‘C in July. The average frost-free period was 16 1 infiltration rates, sediment production, forage production, micro- days; the average date of last frost was May 2 and the average date relief, plant cover, and litter cover through time. of the first frost was October 10 (U.S. Weather Bureau). Study Area Methods This 3-year (1979, 1980, and 198 1) study was conducted on a mesa on the Fort Stanton Experimental Range in Lincoln County, Four grazed pastures and an area which had been excluded from livestock grazing since 1952 were selected as study sites. Grazed

At the time of this study the authors were graduate research assistants, associate pastures included 1 continuously grazed at a heavy stocking rate, I professor of watershed management, and professor of range ecology, Department of continuously grazed at a moderate stocking rate, and 2 pastures Animal and Range Sciences, New Mexico State University. Las Cruces 88003. This report is published with approval of the Director, New Mexico Agricultural from a four-pasture, one-herd rotation system stocked at a heavy Experiment Station. as JA 1007. The-authors wish toacknowledge the support of the rate. Stocking rate of the moderate continuous pasture (23.13 Range Improvement Task Force, and Richard Brown, Mark Wondzell, and Thomas Phillips for assistance in collecting data. ha/AU) was about 25% lower than the heavy continuous and Manuscript accepted March 12, 1984. rotation pastures. In the rotation system livestock were moved

538 JOURNAL OF RANGE MANAGEMENT 37(8), November 1984 Runoff collected from each plot was thoroughly agitated and a l-liter subsample removed. Sediment was filtered off the subsample, dried at 100°C for 24 hours, weighed, and converted to sediment yield in kilograms per hectare. Foliar cover estimates of grasses, forbs, rock, litter, and bare ground were made in each plot using the line intercept method. The tape measure was stretched across the metal runoff frame 6 times at 13-cm intervals. Readings were taken on all plots after the second simulated rainfall. Fohar cover was the cover that would intercept a raindrop. Rock was pebble sized particles or larger. Grasses and forbs, both standing live and dead, were clipped at 1S-cm stubble height, and litter was collected from each I-m* plot. The material was dried at 60” C for I week and weighed. A micro-relief meter was used to detemine surface roughness in each plot. The micro-relief meter consisted of 21 metal pins, spaced 4.35 cm apart on a vertical board which slid up and down according to the height of the ground. The top of the pins corresponded to

TIME a graduated scale which showed height difference in millimeters. The variation was obtained by averaging differences for 3 readings Fig. 1. Monrhlyprecipirationfor 1979, 1980, and 1981. taking per plot. about every 4 months, and data were collected immediately after Data normality was determined by tests for skewness and kurto- one pasture had been rested for 12 months and immediately after sis. Analysis of variance, based on a randomized design, was used another pasture had been grazed for the full grazing period of 4 to determine if any significant difference existed between means for months. This grazing scheme had been in effect since 1969 (Pieper both the dependent and independent variables. Within treatments et al. 1978). In 1980 and 198 1, the rotation system was changed to a variation (variation among subplots) was allocated to the residual four-pasture, three-herd system stocked at one-third the heavy for testing differences (p1.95) among treatments. If differences rate, except in summer (May through August) when the animals existed between the treatment or yearly means, Tuckey’s w- were concentrated in 1 pasture for breeding purposes. In this procedure was used to separate the means (Steel and Torrie 1980). four-pasture, three-herd system, livestock were moved about every Results and Discussion 3 months, and all the data were collected immediately after one pasture had been rested for 3 months and immediately after Foliar Cover another pasture had been grazed for the full grazing period of 9 Percentage of foliar cover across all treatments was highest in months (Pieper personal communication). 1979 (50%) and 1981 (54%) and lowest in 1980 (14%). In 1980 Within each pasture, 3 subsample areas were randomly located. precipitation did not fall until after the sampling period. Across all On each a mobile infiltrometer (Blackburn et al. 1974) was used to years foliar cover percentages were highest in the moderately simulate rainfall on 4, l-m2 plots. Simulated rainfall was applied to stocked pasture (50’?@, similar to the exclosure’s value of 43%. the plots at the existing soil moisture level until runoff occurred. Foliar cover in the exclosure was also similar to values in the Plots were then covered with clear polyethylene plastic to prevent heavily stocked pasture (40%) the rested rotation pasture (35%), evaporation, insuring fairly uniform soil moisture conditions for a and the grazed rotation pasture (36%). Table 1 shows foliar cover second simulated rainfall application which occurred 24 hours for each treatment within each year. later. Data collected following the first application of simulated Standing Biomass rainfall showed trends similar to data collected when the soil was at The average standing biomass in all pastures for 1979 (1,360 field capacity but they were more variable. Therefore, only results kg/ha) was significantly greater than for 1980 (750 kg/ha) and of the second readings are reported here. 198 1 (876 kg/ ha). This was the result of above-normal precipita- Simulated rainfall rate was 7.5 cm/ hr for a period of 45 minutes. tion which fell early in the 1979 growing season. When all years This rate represented a high intensity storm for the area. Prelimi- were combined, the most standing biomass was found in the exclo- nary tests showed 45 minutes was required for the infiltration rate sure (1,550 kg/ ha) followed by the heavily stocked pasture (1,163 to become fairly constant when simulated rainfall was applied to kg/ ha), the grazed rotation pasture (7 12 kg/ ha), the rested rotation plots at field capacity (Smith 1980). Run-off from plots was col- pasture (687 kg/ ha), and the moderately stocked pasture (625 lected and pumped into plastic bottles where it was weighed at kg/ ha). Yields of all of the grazed pastures were statistically sim- 5-minute intervals and converted into centimeters of runoff. Aver- ilar, and the heavily stocked pasture was similar to the exclosure. age infiltration rate was determined by subtracting measured pre- The large amounts of vegetation in the heavily stocked pasture can cipitation at the given time intervals. This procedure does not be explained by considering the species composition. Plant com- account for losses due to interception and evaporation, which were ponent of the moderately stocked pasture was chiefly grasses while probably not significant.

Table 1. Mean vegetation end soil characteristics within the exclosure and each grazing treatment.

1979 1980 1981

Venetation-we or soil variable EX’ M H RR RG EX M H RR RG EX M H RR RG Foliar cover (9%) 580b2 61’ 38’ 40k 47bc 18’ 16” 14’ 12’ II” 54” 73” 63” 46’ 42’ Standing biomass (kg/ ha) 2103’ 1127’ 1647b 987’ 950’ 997’ 256’ 672’ 962” 760’ 1548’ 527be 1203*b696 ” 351’ Bare ground (%) 36” 32’ 52” 48’b 44”b 79’b 79’b 75b 85’b 86’ 41’b 27b 30b 53’ 54’ Litter cover (%) 3* 2’ 4’ 2’ 2’ 3nb 1’ 6b lb lb 5’ lb 6” lb 4ab Micro-relief variabilitv (cm) .7’ .6’ .7’ .6’ .6’ .8’ .4’ .7’ .5a .6” 2.0” I.Ob 1. tb .9b l.lb

IEX,exclosurc; M, moderately stocked and continuously grazed; H. heavily stocked and continuously razed; RR, rested rotatio?; RG, grazed rotation. *Means followed by the same letter within years and vegetation or soil variable are not significantly dlf.% erent (P = <.05) as dctermmed by Tukey’s w-procedure.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 539 the heavily stocked pasture contained a high percentage of broadleaf forbs and broom snakeweed. The standing biomass was INFILTRATION RATE ICWHRI similar. In the wet year of 1979 the exclosure had a much higher production than the other pastures and an abundance of forbs emerged in the heavily stocked pasture (Table 1). Exclusion of grazing did not favor standing biomass over rotation and heavily stocked pastures 6 in the dry year of 1980. Standing biomass was lower in 1980 than 1979 in the exclosure and in the moderate and heavily stocked pastures, but production recovered in 198 1. Meanwhile, the rotation pasture continued to show decreased production from 1980 through 198 1. Litter Cover Litter cover was greatest in the heavily stocked, continuously grazed pasture each year (Table 1), although the differences were not significant. This was attributed to trampling, which detached the standing dead and live plant material. More litter cover was in the exclosure. Overall, the litter cover was small with the largest value being only 6%. Bare Ground Bare ground represented the reciprocal of all combined plant YEAR cover components. Bare ground percentages were highest during the driest year, 1980. Large areas of bare ground were found in the Rested rotation exclosure and were attributed to mature bunch grasses which may •nl be competing for soil moisture beyond the foliar area. The rotation Moderatelystocked, continuously grazed Grazed rotation pasture consistently had as much or more ground than other pastures. Heavily stocked, continuously grazed

Micro-relief Fig. 2. Mean wer final infiltration rates (cmlhr) in each pasture within Differences in micro-relief were not significant among pastures years (means indicated by the same Ierrer within each year are not across all years. Within years the micro-relief was greater in the significanily different at the 95% level as determined by Tukey’s exclosure, in the heavily stocked, and grazed rotation pastures w-procedure). than in the others (Table 1). Increased roughness in heavily stocked Similarity between the moderately and heavily stocked, contin- and grazed rotation pasture was the result of livestock hoof action uously grazed pastures can be explained by comparing percentages during wet periods. The rougher surface in the exclosure was the of foliar cover and standing biomass (Table 1). Although foliar result of blue grama plants attaining a bunchgrass form while the cover was highest in the moderately stocked pasture, standing same species was more of a flat sodgrass outside the exclosure. biomass was highest in the heavily stocked pasture due to the Infiltration Rate abundance of tall forbs with high leaf-area indices. Therefore, Averaged across all treatments, terminal infiltration rates (after standing biomass may offset influences of foliar cover and hydro- 45 minutes) were significantly lower in 1979 (6.0 cm/hr) than the logic condition, and may not necessarily correlate with range con- rates in 1980(6.9cm/hr)and 1981 (6.7cm/hr). Valuesduring these ditions that are suitable for livestock. In this situation, the heavily last 2 years were similar. These year-to-year fluctuations are stocked pasture exhibited a lower seral stage than did the moder- attributed to the different amounts of precipitation following the ately stocked pasture, but infiltration rates were similar. This spring thaw. Soils were fluffy or loose at the end of winter because situation persisted for all 3 years and was supported by correlation of the shrink-swelling effects of freezing and thawing. Late spring analysis (unpublished data on file with M.K. Wood). The rotation and summer rains changed this to a crusted and more dense epi- system was started in 1969 on an area that had not been grazed pedon. Most precipitation during 1980 fell in August and Sep- since 1952. The stocking rate was heavy, but these pastures have tember which rendered the loose soil conditions throughout June never regressed to the seral stage of the heavily stocked, continu- and July. Lack of precipitation during May and June of 1981 (40 ously grazed pasture. Some confounding effects existed because mm and 32 mm, respectively) was not as severe as in 1980, but not the heavily stocked, continuously grazed pasture received slightly nearly as great as in 1979. During the rainy summer of 1979, the soil more precipitation during some years because it is located closer to became crusted and acted as an impeding layer with higher bulk a mountain range. Also livestock movements within pastures were density and reduced porosity (unpublished data on file with M.K. variable with some higher concentrations of animals in the Wood). These conditions contributed to increased runoff. sampled area of the rotation pastures over the areas of the same soil Terminal infiltration rates were significantly higher (6.9 cm/ hr) type in the heavily and moderately stocked, continuously grazed in the exclosure for each year and when averaged across years (Fig. situations. This conclusion is supported by the similarity in infiltra- 2). The heavily and moderately stocked, continuously grazed pas- tion rates, and from vegetation cover, standing biomass, and soil tures had similar values (5.5 and 5.4 cm/ hr, respectively) which micro-relief data. were significantly higher than the rested and grazed pastures in the Sediment Production rotation system (4.7 and 43. cm/hr, respectively). Sediment production historically has been one of the compo- Meaningful comparisons were made by examining rates between nents of the ecosystem with the most variation (Wood and Black- treatments within the same year (Fig. 2). During the moist growing burn 1981). Our data concur (Fig. 3). This shows variations of season of 1979, infiltration rates in the exclosure were not affected sediment production within each year for each pasture. The impor- as much by changes in soil surface characteristics as were pastures tance of protective vegetation can be seen in year-to-year fluctua- subjected to grazing treatments. The exclosure had enough poten- tions, especially in the grazed rotation pasture. The rested rotation tial to continue to absorb most precipitation, as it did in 1980 and pasture had an opposite trend from wet to dry years. Of the 3 1981. The continuously grazed pastures in 1979 had infiltration pastures being rested in the rotation system, the 1 sampled had rates which were higher than those in rested and grazed rotation pastures. 540 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 When all years are combined, the lowest sediment levels were SEDIMENT PRODUCTION (KG/HA) produced in the exclosure, with a mean of 34 kg/ ha. Pastures with 50 sediment production that was higher than, but different from, the Or a exclosure and to each other include the moderately stocked, con- ...... : ...... tinuously grazed ( 155 kg/ ha), the heavily stocked, continuously ......

...... grazed (198 kg/ha), and the rested rotation (226 kg/ha). The ...... 40 ...... sediment produced as a result of concentrating the livestock into a ..: ::...... fourth of the rotation system was 367 kg/ ha, significantly higher ...... than the exclosure and moderately stocked, continuously grazed n areas. 30 Overall, the rotation treatment did not respond any better hydrologically than the continuously grazed pasture at the same heavy stocking rate. Exclusion of grazing did not give hydrologic responses which were superior or significantly different from mod-

20 erate stocking. Possibly, the exclosure is in a seral stage of rested disclimax and an optimum hydrologic condition could exist with some continuous level or systematic livestock grazing.

Literature Cited 101 Blnckburn, W.H., R.O. Meeuwig, and C.M. Sksu. 1974. A mobile infiltrometer for use on rangeland. J. Range Manage. 27:322-323. Dixon, R.M., and J.R. Simnnton. 1977. A land imprinter for revegetation - of barren land areas through infiltration control. Hydrology and Water Resources in Arizona and the Southwest. 7:79-88. 1979 1980 1981 Knight, R.W., W.H. Blackburn, and L.B. Merrill. 1980. Impacts of selected YEAR grazing systems on hydrologic characteristics, Edwards Plateau, Texas. Abs. 33rd Annual Meeting Sot. Range Manage. Rested rotation II4 McCalla, G.R., W.H. Blackburn, and L.B. Merrill. 1984. Effects of livestock grazing on infiltration rates, Edwards Plateau, Texas. J. Range Moderately stocked, continuously grazed Grazed rotation •a Manage (In Press). Pieper, R.D., G.B. Donart, E.E. Parker,and J.D. Wallace. 1978. Livestock Heavily stocked, continuously grazed and vegetational response to continuous and 4-pasture, l-herd grazing systems in New Mexico. Proc. 1st Int. Rangeland Congr. Denver, Colo. Fig. 3. Mean wet sedimentproduction (kglhr) in eachposrure within years Pieper, R.D., J.R. Montoya, and V.L. Grace. 1971. Site characteristics on (means indicated by the same letter within each year are not significantly pinyon-juniper and blue grama ranges in southcentral New Mexico. New different at the 5% levelas determined by Tukey ‘sw-procedure).pasrures. Mex. Agr. Exp. Sta. Bull. 573. Smith, R.P. 1980. The influence of different grazing practices on infiltra- been rested longest. It is assumed that sediment production from tion rates and sediment production at Fort Stanton, New Mexico. Mas- the other two pastures was proportionately between the longest ter’s thesis. New Mexico State Univ., Las Cruces, rested and the grazed rotation pastures. Therefore, the extremely Steel, R., G.D. and J.H. Torrie. 1980. Principles and procedures of statis- high levels of sediment production may be compensated for by the tics. McGraw-Hill Book Co., Inc., New York. resting phase of the system. The trends for the continuously grazed U.S. Weather Bureau. 1856-1976. Climatological data: New Mexico Sec- pastures are difficult to explain. Responses are at least partially tion. Summaries and annual reports. masked by the amount of variation. Infiltration rates were similar Wood, M.K. 1980. Impacts of grazing systems on watershed values p. 163-170. In: Proc. Grazing Manage Syst. for Southwest Rangelands between the continuously grazed treatments for the 3 years of the Symp., Range Improvement Task Proc. New Mexico State Univ. Las study. Cruces. Wood, M.K., and W.H. Blackburn. 1981. Sediment production as influenced by livestock grazing in the Texas Rolling Plains. J. Range Manage. 34:228-23 I.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 541 Forage Preferences of Livestock in the Arid Lands of Northern Kenya

W.J. LUSIGI, E.R. NKURUNZIZA, AND S. MASHETI

Abstract The desirability of forage plants by livestock or wildlife is an all major forage plants in the area so that range condition and trend important consideration in evaluating suitability of the range for can be assessed. grazing. This desirability may also in some cases be used in determining range condition. In many range types periodic determina- Study Area tion of plant species composition provides the best indication of The IPAL study area has been described in the many technical long-term trends. Evaluation of the effects of grazing on range reports published by the project (Lusigi 1981, 1983). It covers an flora usually requires that the vegetation be assigned to significant area of 23,000 km* in the Marsabit District of Northern Kenya, groups. This work represents the first attempt to make this kind of situated between 5 major mountain ranges: Hurri Hills, Marsabit classification for the arid zone of northern Kenya in the study area Mountain, Ndoto mountain range, Mount Kulal, and four impor- of the UNESCO Integrated Project in Arid Lands (IPAL). It was tant water catchment systems in this arid environment (Herlocker required for the preparation of grazing plans for the largely 1979). nomadic pastoralists there. Preferences for 250 plant species have The vegetation cover of the area is varied. There are small been assessed for camels, sheep, goats, and cattle. They are based montane forests at the peak of the mountains, mist forests on hills, on the best information presently available, and forms our basis for and perennial grasslands on lower slopes. In the lowlands are the classification of range condition for 147 range types in the study groundwater forests, semideciduous woodlands, deciduous wood- area. lands, semiarid to arid bushlands and shrublands, arid dwarf shrublands, annual grasslands, and barren lands. With exception The arid lands of Kenya, which cover over 75% of that country, of the mountains which receive about 700 mm average annual are being seriously degraded as a result of increased pressures from rainfall, the rest of the area receives about 200 mm. The rainfall is human livestock populations, periodic droughts, and an imbalance highly erratic. between modern and traditional pastoral practices. The producti- The area is inhabited by nomadic and seminomadic tribes, which vity of the range has been so reduced that pastoralists are under include the Rendille, Samburu, Gabbra, Borana, and Turkana. constant threat of famine. The Integrated Project on Arid Lands Apart from minimal agricultural activities on Marsabit mountain, (IPAL), of which this work is part, was established by UNESCO to the entire nomadic population derives its livelihood from livestock. reverse this degradation. The major livestock are camels, sheep, goats, and cattle. Research on the plant species preferred by livestock in Kenya, The main species of wild grazing herbivores in the study area are particularly in the arid north, has been limited. Since the estab- Beisa oryx (Oryx beisa beisa) and Grevy’s zebra (Equus grevyi). lishment of the IPAL program in 1976, the published vegetational Other wild herbivores include dik dik (Rhynchotragusguentheri), work has consisted of an annotated vegetation map (Herlocker genoruk (Litocranius walleri), Grant’s gazelle (Gazella granti), 1979a), problems concerning the status of montane and other reticulated giraffe (Giraffa camelopardalis reliculata), rhinoceros forests (Synott 1979a,b; Herlocker 1979b), an annotated check-list (Diceros bicornis), and elephant (Loxodonta africana). Other of the plants of Mount Kulal, Kenya (Hepper et al. 198 I), feeding animals include lion (Panthera lea), leopard (Panthera Pardus), observations of camels, cattle, sheep, and goats (Field 1978, 1979; and cheetah (Acinonyxjubatus). Kayongo-Male and Field 1981; Lewis 1977) and herb-layer pro- Numerically, sheep and goats are most abundant, followed by ductivity in relation to rainfall (Herlocker and Dolan 1980a, cattle and camels. The latest aerial survey of the study area esti- 1980b, Walther and Herlocker 1980). Work on woodland availa- mated 287,040 sheep and goats, 56,8 10 cattle, 4 1,400 camels, 1,150 bility, productivity and exploitation has also been reported (Lusigi donkeys, 1,840 oryxes, 1,380 zebras, 1,610 giraffes, 3,600 Grant’s 1981). gazelles, 1,150 gerenuks, and 2,070 ostriches, a total of just under The importance of forage plants for both livestock and wildlife is 400,000 animals (Lusigi 1981). The livestock in the study area determined by their availability, palatability, and nutritive value. (excluding donkeys) comprises about 44% of the total livestock A combination of various levels of these 3 factors should be consi- estimated for the entire Marsabit District. dered in the condition rating of a particular range site. In more Materials and Methods productive rangelands and pastures in Kenya, work has proceeded from determining the availability of forage species to analysis of Preferences of wild ungul;tes for various plants have been dis- the nutritive value of preferred forage species (Dougall) and Bog- cussed by Talbot (1962), Stewart (1966), and Field (1968). They dan 195la,b; Dougall et al. 1964; Field 1975; Gwynne 1969; Tae- consist mainly of direct feeding observations of plant species actu- rum 1970). In the less productive rangelands in the arid zone, ally ingested by the particular animal species, examination of diet information available on the nutritive value of indigenous forage samples collected through oesophogeal and rumen fistulas, ana- species and their preference for livestock is still very scanty except lyses of stomach contents by cropping animals, and microhistolog- for a generalized review of browse (Lamprey et al. 1980). The ical analyses of fecal material. objective of this study was to classify the preference of livestock for The method most commonly used by IPAL livestock scientists has been the direct observation of plants actually eaten. This Authors are project co-ordinator and range ecologist, plant ecologist and plant method is improved by determining the period of time an animal taxonomist with UNESCO Integrated Project on Arid Lands, UNESCO Regional Office for Science and Technology for Africa. P.O. Box 30592, Nairobi, Kenya. spends feeding on a particular plant species (Field 1978). Data have The authors wish to acknowledge contributions made to this study by all IPAL been collected in various parts of the study area both in the dry and Scientists and Field Support Staff. Manuscript accepted December 15, 1983. rainy season. The desirability of a plant species is best determined

542 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 during the rainy season, when plants are at optimal growth stages. may be toxic at particular growth stages. Caution must therefore The desirability of a species in the dry season largely reflects its be exercised in interpreting these data, since no attempt is made to availability (Field 1968). Data obtained from the above observa- give any information on toxicity or nutritive value in terms of tions are supplemented by interviews with herdsmen who have chemical composition. fairly accurate knowledge of plants eaten by their livestock. Ana- A total of 256 plant species consisting of 80 trees and large lyses done on fecal material of cattle (Kaycngo-Male and and Field shrubs, 34 dwarf shrubs, 22 perennial grasses, 68 herbs, 38 annual 1981) and goats (Said 1980) were incorporated into this study. grasses and sedges and 14 creepers and climbers were examined. This paper is a synthesis of available data in published and unpublished articles, and interviews with local people in the IPAL Results and Discussion study area. Seasonal Use of Plants Food preferences have been grouped into 4 categories-Very Since rainfall in the IPAL study area is low and temperatures Desirable (VD), Desirable (D), Intermediate (I), and Undesirable high (Edwards et al. 1979), a large number of annuals and decidu- (U). Very Desirable plants were those observed and reported to be ous plant species are present. A fair number of perennial species eaten every time an animal encountered them. Desirable plants have also been observed on the study area which are utilized in were those constituting 80% of the animal’s diet (Kayongo-Male both the dry and rainy seasons. During the drier parts of the year, and Field 1983) and, together with the very desirables, accounted annuals may still form an important part of the diet of livestock, in for 80% of the animal’s feeding time. Intermediate species were standing dead form or litter. The desirability of a plant species, or those comprising the remaining 20% of the animal’s diet and its undesirability, may reflect its availability rather than palatabil- feeding time, but this category also includes those plants known to ity. For example, the desirability of most Cndeba species seems to be eaten only when there was little or no alternative forage. Unde- improve in the dry season for camels; and though Balanites aegyp- sirables were those with no observed utilization. The last category time is recorded as undesirable in the wet season for the same includes toxic and noxious plants. However, a plant species pre- livestock species, it is intermediate in the dry season (Table I), ferred by an animal at a given time may, in some circumstances, because it remains green long after the desirable species have lost prove to be toxic, especially when eaten in large quantities. Others their leaves.

Table 1. Food preferenceof various livestock species in the Integrated Project on Arid Lands (I.P.A.L.) Study Area in the wet and dry seasons. (VD - Very desirable; D - Desirable; I- Intermediate; U - Undesirable)

Camels Sheep Goats Cattle Plant Species Wet Dry Wet Dry Wet Dry Wet Dry Trees and Large Shrubs Acacia brevispica D D I D D U U Acacia eletior D D u” U U I U U Acacia horrida U U U U U U U U Acacia mellifera 1 U U D D D U U Acacia nubica u I U U I U U Acacia paolii U u” U 1 U U U Acacia reficiens 1 U U U 1 I U U Acacia Senegal U U U U D U U U Acacia seyal var seyal U 1 U U U I U U Acacia tortilis D VD U D 1 VD U U Balanites aegypriaca U I U U U U U U Boscia angustifolia U U U U U U U I Boscia coriacea U I U U U U U U Boscia minimtyolia U U U U U U U U Boswellia hildebrandtii VD D I VD D U U Cadaba farinosa U D : U U D U U Cadabaglandulosa I D U D U D U U Cadaba mirabilis u D U U U I U U Cadaba ruspolii U I U U U U U U Cadaba sp.* U I U U D I U U Cambretun aculeatum D D 1 D D U U Commiphora africana D U z U D I U U Commiphora boiviniana U U U I U U U Commiphora candidula D U U U D U U U Commiphorajlavifora D U U U D U U U Commiphora paolii 1 U U U I U U U Commiphora P/Y+ (Walther et al. 1980) I U U I D I U U Commiphora rostrata U U U I U U U Commiphora sp. (green bark)* D 1 U U D I U U Commiphora sp. (red & white bark)* I U U U I I U U Cordia crenata U 1 U U U I U U Cordia sinensis D U U U D U U U Croton dichogamus U U U U U U U U Dobera glabra U U U U U U U U Euphorbia cuneata I U U U D U U U Grewia bicolor 1 U I I I U I Grewia trembensis D D U U D D I U Grewia tenax 1 U U U D D U U Grewia trichocarpa 1 D U U U D U U

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 543 Table 1. Continued.

Camels Sheep Goats Cattle PlantSpecies Wet Drv Wet Drv Wet DN Wet Drv

Grewia villosa D U U U D U U U Jatropha dichtar U U U U U U U U Lawsonia inermis D D U 1 D D U U Lycium europeaum I U U U U U U U Maerua crassifolia U D U U U D U U Maerua oblongtfolia D D U U U I U U Mawenus senegalensis U U U U D D U U Maringa sp.* U U U U U U U U Opilia campensrris I U U U D I U U Ormocarpum trichocarpum U U U U U U U U Premna resinosa D D U I VD D U U Rennea triphylla U U U U U U U U Salvadora persica I D U U U I U U Securinega virosa U U U U U U U U Sesamothamnus rivae U U U U D U U U Solanum arundo U U U U U U U U Sterculia stenocarpa U U U U U U U U Terminalis mollis D U U U VD U U U Terminalis orbicularis I U U U D U U U Trema orientalis I U U U I U U U Vangueria linearisepala U U U U U U U U Wrigtia demartianiana U U U U U U U U Ziziphus mautiana D D U U U I U U Shrubs Asparagus aethiopicus I U U U 1 U U U Asparagus buchananii U I U U U I U U Asparagus setaceus I U U U I U U U Aspiiia mossambicensis D U D U D U I U Blepharispermum fruticosum U 1 U U I I U U Clerodendrum myricoides U U U U U I U U Cyclocheilon erianthemum D U I U VD U U U Euclea schimperi U U U U U U U U Euphorbia scheffleri U U U U U U U U Hibiscus aponeurus D D U 1 D D U 1 Hibiscus sp.* D 1 1 1 I 1 U I Jatropha parvifolia U U U U U U U U Jatropha viilosa U U U U U U U U Lippia sp. I D U U I I U D Osyris abyssinica U U U U U U U U Tinnea aethiopica D I U U D I U U Tinnea sp.* VD D U U D D U U Veronia brachycalx D U 1 U I U U U Dwarfshrubs Acalypha fruticosa D D D 1 D I U D Aloe sp.* U U U U U U U U Barleria acanthoides D I U U D I U I Barleria eranthemoides D D U I D D U I Barleria proxima D D D D D D U U Barleris sp.* D U D U D U U U Crotalaria fasicularis U U U U 1 U U U Crotalaris laburnifolia I D I D 1 D I D Daysphaera prostrata D D U U U 1 U U Duosperma cremophilum 1 I U U I I I U Euphorbia heterochroma U U U U U U U U Helinus integrffolia D D U U D D U U Heliotropium albohispidum D 1 D I D I U U lndigofera cliffordiana D U D U D U U U Indofera spinosa VD VD D 1 VD I I U Ipomoea kituensis U U U U U U U U Justicia caerulea VD D I I VD D U 1 Justicia fischeri D I I U D 1 U U Justicia odora D I I U U I U U Justicia pinguor D D D D D D U I Kleinia kleinioidea U U U U U U U U Lantana rhodesiensis U I U 1 U I U I Leucas tomentosa U U U U U U U U Ocimum americanurn U I U I U 1 U 1 Ocimum sp* U I U I U I U 1 Ocimum suave U 1 U I U I U 1 Plectranthus igniarus D D U I D D D I

544 JOURNALOF RANGE MANAGEMENT37(6),November1984 Table 1. Continued.

Camels Sheep Goats Cattle Plant Species Wet Dry Wet Dry Wet Dry Wet Dry Salsola dendroides U Soturejo abyssinica I Seddera hirsuta U Sericocomopsis hilderbrandtii D Sericocomopsis pallida 1 Solanum sp. (IPAL)* 1 Suoeda monoica U

Perennial grasses Botheiochloa inseculpta I U VD D D U VD D Bruchiaria sp. (Olturot)* U 1 D D I I D D Brachiaria sp.* U U D D I U D D Cenchruspennisettformis 1 U D U U U D U Chrysopogon plumulosus U U D D U U D D Cynodon dactylon U U D D D I D D Dactyloctenium aegyptium U U D U 1 U D U Digitaria abyssinica U U D D D D D D Digitaria scalarum U U D D D D D D Hyparrhenia hirta U U D I U U D I Leptochloa abtustfolio U U D D D I D D Sehima nervosum U U D D D D D D Setaria sp.+ U U D D D I D D Sporolobus agrostoides U I D D 1 I D D Sporobolusfimbriatus U U D D I D D D Sporobolus hervolus U I D I U I D D Sporobolus nervosus U U D D U U D D Sporobolus sp. (Maikona)* U U D D U U D D Sporobolus spicatus U U U D U 1 U D Tetropogon tenellus U U 1 S U U I D Themeda triandra U U D D D D D D Tricholoena tenertffae I I D D U 1 D D Broadleaf herbaceous plants Abutilon hirtum U U U U U U U U Abutilon sp.* U 1 U I I D 1 D Acalypha indica U U U U D U U U Acalypha racemosa U U D U D U D U Aerva persica U U U U U U U U Barleria argenleo 1 I I 1 I I U U Becium oboratum U U U U U U U U Bidens incumbens U U U U I U U U Blepharis linaartfolia VD D U U VD U I U Cleome allamanii I U I U I U I U Cleome brachycarpa D U I U D U U U Cleome tenello U U U U U U U U Cleome usambarica D U U U D U 1 U Commelina africana D D VD VD D D VD VD Cammelina benghalensis D D VD VD D D VD VD Conyzo phrrhopappa U U U U U I U 1 Conyza stricta U U U U U I U I Crossandra sp.* U U U U U I U U Crotalaria deserticolor U U U U U U U U Crolalaria massaiensis U U D I D I D I Cynoglossum coeruleum U U U U U U II U Digera muricata U U I U 1 U U U Euphorbia lnaequilotera U U D U D U U U Geigeria acaulis U U U U U U U U Gesikia pharnacoides U U D U D U 1 U Gloriosa superba U U U U U U U U Gynandropsis cleome U U U U U U U U Heliotropium sp.* 1 U U U 1 U U U Heliltropium steudneri U U U U U U U U Hermonia borannensis D U D U D U D U Hermanis kirkii U U U U 1 U U U Indigofera brevicalyx D U D U D U D U Indigofera coerulea VD VD D D VD VD U I Indigofera hochsteteri U U U U 1 U U U Ipomoea cordofana D U I U U U U U Ipomoea erythrocephala I I I 1 I 1 1 I Ipomoea kotschyana U U U U I U U U Ipomoea sp.* D U D U D U I U Jasminum sp.* 1 U I U 1 U U U Justicia diclipteroides D I I U D I U U

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 545 Jusficia erigua D U I U D U U U Kouhatia caespitosa U U D U D U U U Leucas sp.* u U I U I U I U Leucas urticifolia U U U U U U U U Limeum prastermissum u U U U I U U U Limeum sp. U U U U I U U U Mollugo cerviana U u U U U U U U Monsonia senegalensis U U U U U U U U Ocimum basilicum U U U U U U U U Ocimum sp.* U U U U U U U U Oldenlandia fastigiata U U U U I U U U Oxygonum sinuarum U U I u I u U U Phyllanthus fischeri U U U U U U U U Phyllanthus rotundifolius U U U U U U U U Phyllanthus sp.* U U U U U U U U Plectranthus sylvestris 1 I I I I I I I Portulaca oleracea U U U I U I U U Portulaca quadrljida U U U I U I U U Pupalia lappacea U U U I U I U U Senecio lyratripatitus U U U U U U U U Sesamum alatum U U U U I U U U Sesamum sp.* U U U U U U U U Solanum dubium U U U U U U U U Stricta sp.* U U U U U U U U Tephrosia emeroides U I U U U I U U Tribulus terrestris U U U U I U U U Vernonia sp.* D U I U I U U I Zygophyllum simplex U U U U I U U U Annual grasses and sedges Aristida adscensionis U I I D I D D D Aristida mutabilis U D D D I D D I Brachiaria deflexa U I D D I U D D Brachiaria leersiodes U U D I U U D I Brachiaria ovalis U I D D I U D D Cenchrus ciliaris U U I D U I I D Chloris roxburghiana U U I U U U D U Chloris virgata U U VD D I U VD D Cyperus bulbosus U U D I U U D I Cyperus laevigatus U U D D I D D D Cyperus sp.* U U D D I D D D Digitaria sp.* U U VD VD I I VD VD Digitaria velutera U U VD VD I I VD VD Drake-Brockmania somalensis U U I I U U 1 1 Enneapogon cenchroides U I D D I I D D Enneapogon desvauxii U U D D U U D I Eragrostis cilianensis U U D D U I D D Eragrostis sp.* U I VD VD D D VD VD Heteropogon contortus IJ u D D u I D D Kyllinga alba U U D D D D D D Leptothrium senegalense I D D D I I I I Lintonia nutans U U D D U U D D Lintonia sp.* U U D D U U D D Microchloa kunthii U U D D D D D D Oropetium minimum U U I D U I U D Pennisetum mezianum U U D D U I D D Setaria pallidefusca U U D D U U D D Setaria plicatilis U U D D U U I D Setaria verticillata U U D D U U D I Sorghum sp. (I)* U I D D U U D D Sorghum sp.( I I)* U U D I I U D I Sorghum versicolor U I D D U U D D Sporobolus sp. (I)* U U D D D D D D Sporobolus sp. (1 I small)* U U D D U U D U Stipagrostis uniplumis U I I I I I I I Tetrapogon cenchriformis I U D D I I D D Tetrapogon spatheceus U I D D I I D D Tragus berteronianus U U D U I U U U Other important species (Creepers, Climbers etc.) Cissus rotundifolia U U U U U U U U Euphorbia samburuensis U U U U U U U U Euphorbia tirucalli U U U U U U U U Vigna membranacea D D D D D D D E *Specific name not derermined.

546 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 The annual plants require special consideration. They contrib- IPAL study area where they are frequently dominant (Table 1). ute to the diets of livestock, especially cattle and sheep, either in Dwarf shrubs are important camel and goat browse in all seasons their green form in the wet season or in the “standing dead” and and are also occasionally browsed by cattle and sheep during litter form in the dry season. Because their presence is unreliable periods of food scarcity. The most important species are Aculypha however, they are not normally used for assessing condition. Ble- fruticosu, Burleriu erunthemoides, Barleriu proximu, Indoferu spi- pharis lineurifoliu is highly desirable by camels in its dry form, nosu, Justicia ceuruleu, Justiciupinguor and Plectrunthus igniurus. however. It is common for species of low palatability in the wet season to become desirable in the dry period as illustrated by Crotulariu Trees and Shrubs luburnifoliu. and Sutureju ubyssinicu. Duospermu eremophilum, During wet seasons when a flush of green leaves appears, more which is one of the most common dominants of the dwarf shrubs, is than 50% of the trees were found to be utilized by camels and goats only occasionally browsed by camels, goats, and cattle. Other (Table 1). There are no records of sheep or cattle consuming trees important dwarf shrub species, both in terms of availability and or large shrubs in the wet season. Since most trees are deciduous, desirability, are Heliotropium ulbohispidum and Sericocomopsiu and lose their leaves in the dry season, they are not consumed by hildebrundtii. livestock. Approximately 50% of the tree species, however, are known to be utilized by camels, 50% by goats, and 16% by sheep, in Perennial Grasses the dry season. Even cattle have been observed to feed on Acacia Perennial grasses in the IPAL study area are mainly found on tortilis flowers, and Grewiu bicolor and Bosciu ungustifoliu which the mountains, though they also occur in limited areas of the arid remain green longer than other plant species. Kayongo-Male et al. plains. They form the most important plant group for sheep and (1981) recorded Cudubu species as constituting 8% of the cattle cattle (Table 1). Certain species have been included in this group diets in late October 198 1 in the Ngurunit/ Lependera area. Acacia because they tend to be either annual or perennial. Lewis (1977) tortilis pods are known to be eaten by goats and sheep. reported similar physiology for Erugrostis ciliunesis. More than Smaller shrubs (about 2 metres high) form an important part of 90% of the observed perennial grasses were desirable to both sheep camel and goat diets (Table 1). They are especially browsed in the and cattle in all seasons. The most desirable plants in this group for wet season when about 55% of the total number of shrub species sheep and cattle was Dichunthium insculptum (Table 1). In the observed are eaten by camels and 62% by goats. Sheep and cattle case of goats, as much as 60% of the perennial grasses present were eat approximately 22% and 6% of this plant category, respectively, found to be desirable (Table 2). Even for camels, which are gener- in the wet season. (Table 2). ally considered to be nongrazers, our observations indicated that Dwarf shrubs form a major component of the vegetation in the about 14% and 19% of the total number of perennial grasses were

Table 2. Desirability rating of tbe various plant species as percentage ($) of the total species in each class for different livestock species.

Wet Season Dry Season VD D I U VD D 1 U Camels Tree and large shrubs 2 26 29 43 2 24 I4 60 Shrubs 6 33 I7 44 - I7 22 61 Dwarf Shrubs 6 41 I5 38 3 41 29 27 Perennial grasses - - I4 86 - - I8 82 Herbs 3 I6 IO 71 2 4 9 85 Annual grasses - - 5 95 5 - 24 71 Others 7 43 29 21 I4 - - 86 Sheep Trees and large shrubs - - 100 - 5 II 84 Shrubs - T I7 78 - - II 89 Dwarf shrubs - 23 I5 62 - I2 35 53 Perennial grasses 5 86 5 5- 82 9 9 Herbs 3 I2 20 63 3 2 10 85 Annual grasses IO 74 I6 - 8 74 I3 5 Others 21 29 I4 36 - I4 - 86 Goats Trees and large shrubs 5 34 13 48 2 21 27 50 Shrubs 6 22 33 39 - II 33 56 Dwarf shrubs 7 33 18 42 28 42 29 Perennial grasses - 36 23 41 - 23 36 41 Herbs 4 24 32 40 2 I6 78 Annual grasses - II 42 47 - 2’: 32 47 Others I4 50 I4 22 I5 22 - 63 Cattle Trees and large shrubs - - - loo - 3 97 Shrubs - - 6 94 - 5 I2 83 Dwarf shrubs 3 9 75 I2 29 59 Perennial grasses 4 86 5 5 - 86 5 9 Herbs 3 5 13 79 3 2 IO 85 Annual grasses IO 75 13 2 9 61 24 6 Others - 36 64 - 7 I4 79 YD = Very desirable, D = Desirable, I = Intermediate, U = Undesirable.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 547 eaten by camels in the wet and dry seasons respectively. Edwards, K.A., C.R. Field, and I.M.M. Hogg. 1979. A preliminary analysis of climatological data from the Marsabit District of Northern Kenya. Broadleaf Herbaceous Plants IPAL Tech. Rep. No. B-l. UNEP-MAB Integrated Project in Arid The greatest proportion of the 68 broadleaf herbaceous plants Lands, Nairobi. encountered in our studies are annuals. Thus when they are eaten Field, A.C. 1978. ODM - IPAL sheepand goats project preliminary report in the dry seasons, they are usually dried. The species most pre- on the impact of sheep and goats on the vegetation in the arid zone of ferred are Cammelina africana and Cameliana benghalensis, which Northern Kenya. IPAL Tech. Rep. No. E-2. UNEP-MAP Integrated are preferred by cattle, sheep, camels, and goats in all seasons Project in Arid Lands, Nairobi. (Table 1). Zndigofera caerulea on the other hand, was found to be Field, C.R. 1968. Methods of studying the food habits of some wild very desirable for camels and goats and desirable for sheep in all ungulates in Uganda. Proc. Nutrition Sot. 27:172-177. Field, C.R. 1975. Climate and the food habits of ungulates on Galana seasons, although it was intermediate for cattle in the dry season Ranch. E. Afr. Wildl. J. 13:203-220. and undesirable for cattle in the wet season. Despite its spines, Field, C.R. 1978. The food habits of camels in Northern Kenya. IPAL Blepharis linaarzfolia is eaten by camels in the wet and dry seasons, Tech. Rep. No. E-lb. UNBP-MAB Integrated Project in Arid Lands, as well as cattle, sheep, and goats in the wet season. Because many Nairobi. of the species are desirable only in the wet season, they are not Field, C.R. 1979. Preliminary report on ecology and management of cam- available in the dry periods (Table 1). els, sheep and goats in Northern Kenya. IPAL Tech. Rep. No. E-la. UNEP-MAB Integrated Project in Arid Lands, Nairobi. Annual Grasses and Sedges Gwynne, M.D. 1979. The nutritive value of Acacia pods in relation to seed distribution by ungulates. E. Afr. Wildl. J., 7:176-178. The ground layer of the IPAL study area is dominated by annual Acacia Hepper, P.M.L., J.B. Gillet, and M.G. Gilbert. 1981. Annotated checklist grasses, notably Aristida spp. Annual grasses compose the bulk of of the plants of Mount Kulal, Kenya. IPAL Tech. Rep. No. D-3. UNEP- cattle and sheep diets, a substantial proportion of the diet of goats, MAB Integrated Project in Arid Lands, Nairobi. and a small but significant contribution to the diet of camels (Table Herlocker, D. 1979% Vegetation of southwestern Marsabit District, I). Digitaria sp., D. velutina, and Eragrostis sp. are very desirable Kenya. IPAL Tech. Rep. No. D-l. UNEP-MAB Integrated Project in for cattle and sheep in all seasons. Another highly preferred plant is Arid Lands, Nairobi. Chloris virgata. In fact, only a minor percentage of the plants in Herlocker, D. 1979b. Implementing forestry programmes for local com- this category was recorded as undesirable for cattle and sheep, munity development, southwestern Marsabit District, Kenya. IPAL mainly in the dry season. Undesirables include Oroperium min- Tech. Rep. No. D-2c. UNEP-MAB Integrated Project in Arid Lands, Nairobi. imum and Tragus berteronianus, which are both very small plants Herlocker, D., and R.A. Dolan. 1980a. Primary productivity of the herb that are not readily available or have sharp, prickly seeds. In the layer and its relation to rainfall. P. 22-27. In: IPAL Tech. Rep. No. A-3. dry season, annual grasses and sedges are eaten standing dead or as Proc. of a Scientific seminar, Nairobi 24-27 November 1980. UNEP- litter by livestock. In certain arid conditions such as are commonly MAB Integrated Project in Arid Lands, Nairobi. found in the IPAL study area, this may be the only food available Kayongo-Male, H., and C.R. Field. 1981. Feed quality and utilization by to the animals. All the sedges encountered, Cyperus laevigatus, C. cattle grazing natural pasture in the range of Northern Kenya. (Unpub- bulbosus, C. sp., Kyllinga alba, were desirable for cattle, sheep, lished). and goats, but not for camels. Lamprey, H.F., D.J. Herlocker, and C.R. Field. 1980. The state of knowledge on browse in East Africa. In: Proc. Intern. Conference on Browse Production: Addis Ababa. Other Important Species Lewis, J.G. 1977. Report of a short-term consultancy on the grazing The last category to be considered includes creeping and climb- ecosystem in the Mt. Kulal region, Northern Kenya. IPAL Tech. Rep. ing plants. Since the majority of these plants are drought- No, E-3. UNEP-MAB Integrated Project in Arid lands, Nairobi. deciduous, their use by livestock is largely confirmed to the wet Lusigi, W.J. 1981. Combatting desertification and rehabilitating degraded seasons (Table 1). However, a few of them are utilized by livestock production systems in Northern Kenya. IPAL Tech. Rep. A-4. in the dry season since their stems remain fleshy. The most pre- Said, A.N. 1980. Goat nutrition. P. 74-77. In: IPAL Tech. Rep. No. A-3. ferred species in this group is Kandrostis gijef which is highly Proc. of a scientific seminar, Nairobi, 24-27 November 1980. UNEP- desirable for camels, sheep, goats, and cattle even in its leafless MA3 Integrated Project in Arid Lands. Nairobi. form. Other preferred species are Rhynchosia minima, Rhyncho- Sands, E.B., D.B. Thomas, J. Knight, and D.J. Pratt. 1970. Preliminary selection of pasture plants for the semi-arid areas of Kenya. E. Afr. Agr. sia sp., Rhynchosia sublobata, Mormodica sessifolia, and Vigna For. J. 36:49-52. membranacea. The majority of the plants in this catagory are Stewart, D.R.M. 1966. A technique for studying the food preferences of desirable in the wet season, especially for camels, sheep, and goats. grazing herbivores. Ph.D. Thesis, Univ. East Africa. Synott, T.J. 1979a. A report on the status, importance and protection of the Summary montane forests. IPAL Tech. Rep. No. D-2a. UNEP-MAB Integrated Project in Arid Lands, Nairobi. A look at the classification indicates that most of the plants in Synott, T.J. 197913. A report on prospects, problems and proposals for tree the study area are used by livestock at different times of the year. planting. IPAL Tech. Rep. No. D-2B. UNEP-MAP Integrated Project There is a clear overlap in the diets of camels and goats and also in Arid Lands, Nairobi. cattle and sheep. Since this area is used traditionally by all these Taerum, R. 1970. A note on chemical content of some East African grasses. kinds of livestock, diet overlap should be considered while recom- E. Afr. Agr. For. J. 36:171-176. mending stocking rates. Manipulation of the range for better Talbot, L.M. 1962. Food peferences of some East African Wild ungulates. primary production should also take into account the needs of the E. Afr. Agr. For. J. 27: 131-138. different livestock species on a common use range. Waltber, D., and D.J. Herlocker. 1980. A preliminary study of the relationship between vegetation, soils and land use within southwestern Mar- sabit District. P. 41-54. In; IPAL Tech. Rep. No. A-3. Proc. of a Literature Cited scientific seminar, Nairobi, 24-27 November 1980. UNEP-MAB Inte- grated Project in Arid Lands, Nairobi. Dougall, H.W., and A.V. Bogdan. 1958a. Browse plants of Kenya with special reference to those occurring in South Baringo. E. Afr. Agr. For. J. 23:236-245. Dougall, H.W., and A.V. Bogdan. 1958b. The chemical composition of the grasses of Kenya-l. E. Afr. Agr. For. J. 24: 17-23. Dougall, H.W., V.M. Drysdale, and P.E. Glover. 1964. The chemical composition of Kenya browse and pasture herbage. E. Afr. Wildl. J. 2:86-92.

548 JOURNAL OF RANGE MANAGEMENT 37(6), November 1994 Cattle Distribution on Mountain Rangeland in Northeastern Oregon

R.L. GILLEN, W.C. KRUEGER, AND R.F. MILLER

Abstract

Cattle grazing distribution patterns were studied directly through 0 to 100%. observation and indirectly through plant utilization during 3 Annual average precipitation ranges from 500 mm in the valley summer grazing seasons under continuous and deferred-rotation bottom to 1,020 mm at the highest elevations. Most of the precipi- grazing systems. Small riparian meadows were the most preferred tation falls as snow from November through April. Little precipita- plant communities. Meadows covered 3-S% of the total observation is expected in July, August, and early September. During the tion area but 24-47% of all cattle were observed in those plant summer grazing season, daytime maximum temperatures reach communities. Logged forest communities ranked second in animal 20-30” C but frost may occur in any month. preferencewhen availabR.Relatively open-m Plant communities in the study area have been described by Hall menziesii plant communities were the most preferred forested hab- (1973). For this study, the plant communities were combined into 6 itats. Deferred grazing equalized cattle use between logged areas general groups. Meadows were narrow communities (less than 30 and P. ponderosa-P. menziesii forests and increased cattle use of m wide) found mostly along streams. Major dominants included riparian meadows. Heavily forested sites were least preferred by Kentucky bluegrass (Poapratensis), redtop (Agrostis stolonifera), cattle. Slope gradient was the only physical factor consistently various sedges (Carex spp.), and Baltic rush (Juncus balticus). associated with cattle grazing distribution. Water distribution was Grassland communities occupied upland sites with shallow soils not correlated with grazing patterns in uplant plant communities. and limited water storage capacity. Bluebunch wheatgrass (Agro- Multiple regression models could not predict grazing distribution pyron spicatum), Sandberg bluegrass (Poa sandbergii), and one- patterns with useful precision. spike danthonia (Danthonia unispicata) were major species on these sites. Proper distribution of livestock grazing is an integral part of The ponderosa pine (Pinus ponderosa)Douglas-fir (Pseudo- effective range management. The goal of livestock distribution tsuga menziesii) community group was located on the driest com- management is to gain maximum safe use over as wide an area as mercial timber sites in the area. It consisted of a variable mixture of possible without causing serious damage to any portion within it. ponderosa pine and Douglas-fir in the overstory with an under- Mountain rangelands often exhibit complex combinations of story dominated by elk sedge (Carex geyeri). Pinegrass (Calama- topography, plant communities and successional stages, water grostis rubescens) was nearly co-dominant with elk sedge on the distribution, and other habitat factors which create especially dif- better sites. The mixed conifer community group was similar to the ficult grazing distribution problems. For instance, utilization may ponderosa pine-Douglas-fir community group but occurred on reach 75 to 80% on gently sloping drainages while steep slopes 150 more productive sites. Grand fir (Abies grandis) was a major m away receive 5% use or less (Phillips 1965). On a relatively small overstory component along with ponderosa pine and Douglas-fir. range unit of 690 ha in northern Utah, as much as 62% of the area Pinegrass and elk sedge dominated the understory with pinegrass received no use by cattle (Gonzalez 1964). A better understanding generally contributing the larger amount. of the interactions between livestock behavior, natural habitat The grand fir community group occurred on the wettest forest factors, and management factors should aid in developing more sites at the higher elevations. Grand fir was the major tree species. effective methods of livestock distribution. Understory dominants included huckleberries (Vaccinium spp.), This paper reports on a 3-year study of cattle distribution behav- pinegrass, and various shade tolerant forbs. The logged forest ior. The objective of the study was to evaluate the effects of several community group consisted of ponderosa pine-Douglas-fir or physical, biological, and managerial factors on the patterns of mixed conifer sites that had received 90- 100% reductions in tree cattle use of mountain rangelands. Direct observation and forage canopy within the last IS years. Major herbaceous species included utilization sampling were used to quantify grazing patterns of pinegrass, elk sedge, mountain brome (Bromus marginatus), cattle under both continuous and deferred-rotation grazing schemes. orchardgrass (Dactylis glomerata), and intermediate wheatgrass (Agropyron intermedium). Study Area Data were collected in 2 of the 6 pastures that comprise the Upper Middle Fork Grazing Allotment. The Butte Pasture The study was conducted on the Upper Middle Fork Grazing included 4,675 ha with a general north-northeast aspect. This Allotment of the Malheur National Forest in north-central Oregon pasture was grazed continuously with 85 cow-calf pairs from early from 1979 to 1981. Elevations range from 1,160 m along the June to mid-October except in 1980, when grazing began in early Middle Fork of the John Day River, which flows northwest July. The Caribou Pasture included 3,610 ha with a south- through the Allotment, to over 2,300 m on the watershed boundar- southwest aspect. This pasture was grazed from early June to early ies. Topography is mountainous with slope gradients ranging from August in 1979 and 198 1 and from early August to mid-October in Authors are graduate research assistant, professor, and associate professor, respec- 198 1 as part of a two pasture deferred-rotation grazing system. The tively, Department of Rangeland Resources, Oregon State University, Corvallis Caribou Pasture was stocked with 200 cow-calf pairs. 97331. Gillen is currently assistant professor, Department of Agronomy, Oklahoma State University, Stillwater 74078 and Miller is at the Eastern Oregon Agricultural Methods Research Center, Burns, Oregon. This report was submitted as Technical Paper No. 6985. Oregon Agricultural Direct Cattle Observations Experiment Station. The work was supported by the USDA-Forest Service Range Evaluation Project through the Pacific Northwest Forest and Range Experiment Cattle distribution was determined by actual observation from Station. 1979 to 1981 using observation routes along existing roads. Start- Manuscript accepted March 19, 1984.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 549 Table 1. Proportion of cattle observed in different plant communities during three grazing periods. (PI q Preference Index).

Continuous grazing Early grazing Late grazing Plant community % Cattle % Area PI % Cattle % Area PI % Cattle % Area PI meadow 47a2 5 9.48’ 24a 3 8.0* 34a 3 1I.38 grassland 18b 6 3.0* 5b 9 0.6 5b 9 0.6 ponderosa pine- Douglas-fir 16b 17 0.9 23a 42 0.5* 24~ 42 0.6* mixed conifer 16b 35 o.s* IIC 24 0.5* 12d 24 0.5* grand fir 4c 37 0.1* - - - logged forest - - 37d 22 1.7* 25~ 22 I.1

‘based on 1001, 1906, and 583 observations for continuous, early, and late grazing, respectively. 2percentages wthin grazing period and column followed by different letters are significantly different, K.10, for the set of all simultaneous pairwise comparisons. ‘*significantly different from I .O, I’<.05 ing times and directions of travel for the observation routes were measured per transect at the peak of standing crop. The dominant allocated in a stratified random manner. Observations were made species were clipped and weighed for each plot. Secondary species every third day and during all time periods from dawn until dusk. were estimated by the weight-estimate-per-plot method (Pechanec The number of cattle, cattle activity, plant community occupied, and Pickford 1937b). Herbaceous production plots were not pro- and map Iocation were recorded for each cattle sighting. Addi- tected from grazing; standing crop data were adjusted by forage tional information calculated from topographic maps included utilization estimates to arrive at an estimate of ungrazed herbace- slope gradient (per cent) and trail distance to salt and water ous production. Samples of all species were collected and dried at (meters). 50°C to a constant weight to express production estimates on an Vegetation along the observation routes was mapped by com- oven-dry basis. Production transects were then pooled within plant munity type following the classification of Hall (1973). Vegetation communities to estimate annual herbage production and species mapping extended out to the estimated limit of cattle visibility. A composition within communities. random sample of 100 points was placed within the observation Physical att;butes estimated for all utilization sites included boundaries of each observation route and the habitat characteris- slope gradient; slope length to road, trail, or drainage bottom; tics of each sample point were determined to estimate the availabil- distance to water and salt along the probable trailing route; and ity of habitat factors used to describe cattle location. vertical distance to water and salt. Tree canopy cover was also Cattle observations were totaled by plant community and com- measured at each site with a spherical densiometer (LemmOn pared to available community distributions within pastures by 1957). chi-square analysis. When the hypothesis of equality among distri- For statistical analyses, utilization was expressed as a weighted butions was rejected, the normal approximation to the binomial percentage estimate of grass utilization. This estimate was calcu- distribution (Snedecor and Cochran 1967) was used to test pairs of lated by summing the products of percent utilization and percent proportions within categories. The simultaneous probability level composition by weight over all grass species. Upland sedges used for pairwise comparisons was 10%. A Bonferroni approach to (mainly elk sedge) were included within the grass component. simultaneous testing resulted in a pairwise probability level of Grass utilization within plant communities was analyzed by com- l-3% (Marcum and Loftsgaarden 1980). pletely randomized analysis of variance with unequal replications. Data were also expressed as preference indices since each class of The dependent variable of observed utilization and the quantita- the independent variables did not contain an equal proportion of tive independent site factors were subjected to correlation and the observation area. The preference index was defined as the ratio multiple regression analyses. The data were stratified by pasture, of the percentage of total animals observed within a particular class year, and forage types for the analyses. Forage types were based on to the percentage of the total observation area within that class. the major forage species and the general characteristics of the site This is the same concept as the relative preference index often used and consisted of 3 groups: grassland, forest, and logged forest. in animal diet selection studies (Krueger 1972). An index value Data from each year were analyzed separately with similar greater than one indicated that more cattle were observed within a results. The analyses were then combined over years in the Butte class than would have been expected by random use and was Pasture and over early and late grazing in the Caribou Pasture. considered to imply positive animal preference for that class. Cattle observation data from the Butte Pasture in 1979 were Values less than 1 implied a particular class was not preferred. deleted from the final analysis because of inadequate sample size. Upland Forage Utilization Results Fifty permanent sampling sites were located in upland plant communities in the Butte and Caribou Pastures by a stratified Direct Cattle Observations random procedure to represent the range of plant communities and Plant Communities topographic situations. In 1979,25 sites were sampled in the Butte The most obvious feature of cattle use of plant communities Pasture and 40 sites were sampled in the Caribou Pasture. Fifty under continuous grazing in the Butte Pasture was the high cattle sites were sampled in both pastures in 1980 and 1981. preference for meadow communities (Table 1). This vegetation Forage utilization was estimated on each site at the end of the group had a preference index (PI) of greater than 9.0 with almost respective grazing period in each pasture. Utilization was esti- half of the animals observed on 5% of the total area. The grassland mated by the ocular-estimate-by-plot technique (Pechanec and communities were also preferred although to a lesser extent than Pickford 1937a) for all species occurring within 20 OS-m2 circular the meadow communities. The ponderosa pine-Douglas-fir com- plots spaced along a 60-m pace transect. munities were used in proportion to availability. Cattle avoided the Herbaceous production was sampled on one-half of the perman- mixed conifer and grand fir forests. Logged forest communities ent study sites. Production measurements were collected on tran- were not present in the continuously grazed pasture. sects that ran parallel to and were within 10 m of the utilization Under deferred-rotation grazing in the Caribou Pasture, mea- sampling transects. Ten evenly spaced OS-m* circular plots were dow communities were also highly preferred with PI values rang-

550 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 ing from 7.7 to 11.3 (Table 1). Late season grazing increased cattle were little used. Under late grazing, the upland communities main- use of the meadows. The logged forest communities were preferred tained the same relative rankings for grass utilization. However, in the early grazing period but were used as available in the late utilization levels were not discrete but formed a continuum from period. The forested communities were not preferred; however, the the logged forest communities to the grasslands. ponderosa pine-Douglas-fir forests should still be considered an important cattle use area since almost one-quarter of all cattle Individual Site Factors observations were made in those communities. Use of grassland Although several site factors were correlated with grass utiliza- communities was not significantly different from availability. tion on forested sites, no single factor had a high association with Grand fir forest communities were not sampled in the deferred- grass utilization (Table 5). Many of the correlations in the continu- rotation pasture. ously grazed pasture can be explained by plant community preferences noted earlier. Within the forests, the relatively more pre- Physical Factors ferred ponderosa pine-Douglas-fir communities tended to have Slope gradient was inversely associated with cattle preference higher production, less pinegrass and more elksedge. The least for all grazing periods (Table 2). The high preference index for the preferred grand fir communities tended to be opposite in all of these categories. Correlations for the logged forest communities Table 2. Preference indices for slope gradient ciasses during three grazing were generally larger than for the forest sites in the early period but periods as determined by direct cattle observations. fewer logged sites were sampled. Only one site factor, slope gradient on forested sites, was significantly correlated with grass utilization in the late grazing period. The negative association Slope gradient Continuous Early Late between grass utilization and percent elk sedge on logged sites (%) grazing grazing grazing indicates higher utilization on sites with smaller proportions of elk 5 or less 3.0 3.7 4.0 sedge and larger proportions of introduced pasture grasses. This 6-10 5.3 2.9 4.8 relationship was reversed in the late period but the correlation is II-15 0.8 I.6 I.6 not significant. No site factors were significantly correlated with 16-20 0.8 I.1 0.6 21-30 0.4 0.5 0.5 grass utilization on grassland sites within any of the grazing 31-45 0.3 0.4 0.4 periods. greater than 45 0.1 0.1 0.1 Several transformations of site factors were included in the correlation analyses. These included the natural logarithms of all distance, slope, and tree cover variables, the square root of all 6- 10% slope class in the continuously grazed pasture was the result distance and slope variables, and the product of slope gradient and of the most heavily used meadow having a 6% slope gradient. In the slope length. These transformations did not improve the correla- deferred-rotation pasture, late grazing tended to shift cattle prefer- tion of any site factor with grass utilization. ence to the lower slope gradients. Cattle appeared to avoid slope The proportion of variation in grass utilization which could be gradients in excess of 2% during all grazing periods. accounted for by multiple regression models ranged from 0 to 55% Cattle preferred areas within 200 m of water and avoided areas for individual years and averaged 26% for the forest communities greater than 600 m from water in the continuously grazed pasture (K.05). The number of site factors included in the individual (Table 3). A similar relationship was present in the deferred- regression models ranged from 2 to 4. On the logged forest sites, the rotation pasture except for an unexplained avoidance of the 201- regression models accounted for 0 to 97% of the variability in grass 400 m distance class. Cattle preference shifted to the lower distance utilization. High coefficients of determination were the result of classes during late grazing. The shifts in both slope gradient and large single factor correlations in the early grazing period (Table 5). distance to water relationships can be attributed to the greater However, these relationships should be considered tentative preference for the meadow communities during late grazing. because of the limited sample size. Overall, it was felt that the Distance from salt had little effect on cattle preference in the regression models constructed for the forest and logged forest continuously grazed pasture (Table 3). In the deferred-rotation communities would not be useful in predicting grass utilization pasture, areas within 600 m of salt were slightly preferred or used as patterns in management situations. available while areas greater than 600 m from salt were not preferred. This relationship was similar between early and late Discussion grazing. The extreme preference cattle exhibited for meadow communi- Upland Forage Utilization ties in this study agrees with research reports from other mountain Plant Communities rangelands (Bryant 1982, Long and Irwin 1982, Roath and Considering only upland plant communities, grass utilization Krueger 1982) and makes these areas the major factor influencing was most intense in the ponderosa pine-Douglas-fir forest under grazing distribution. While these communities are certainly pre- continuous grazing (Table 4). The grasslands and more heavily ferred by cattle, it should be pointed out that the preference indices forested communities received half as much utilization as the pon- were calculated on the basis of plant community area, not herbage derosa pine-Douglas-fir forests. Under early grazing in the deferred- production. The meadow communities produced 12-16 times as rotation pasture, logged forest communities received the heaviest much herbage as the grasslands and forest communities and about grass utilization. The forested communities sustained about two- thirds as much utilization as the logged sites while the grasslands

Table 3. Preference indices for distance ciasses to water and salt during three grazing periods as determined by direct cattle observations.

Water Salt Distance Continuous Early Late Continuous Early Late (m) grazing grazing grazing grazing grazing grazing 200 or less 1.8 1.9 2.1 I.2 I.2 1.2 201-400 1.2 0.5 0.1 0.8 I.4 I.1 401-600 1.2 1.0 0.8 0.6 I.2 1.7 601-800 0.4 0.4 1.5 0.6 0.4 greater than 800 0.3 0.5 1.4 0.3 0.3

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 551 Table 4. Grass utilization (%), number of sample sites (N), and berbage production (kg/ha) for different upland plant communities during three grazing periods.

Continuous grazing Early grazing Late grazing Herbage Plant community Utilization N Utilization N Utilization N production grassland 8a’ 5 la 7 4a 7 240 ponderosa pine-Douglas-fir l4b I8 IOb 23 I Ibc 23 230 mixed conifer 8a 20 8b 13 Jab I3 209 grand fir 6a 7 - 184 logged forest - l5c 7 l4c 7 465

‘percentages within columns followed by different letters are significantly different, K.05.

Table 5. Simple correlations between upland grass utilization and individual site factors on forest and logged forest sites during continuous, early, and late grazing periods.

Forest Forest Logged forest Site factor Continuous grazing Early grazing Late grazing Early grazing Late grazing herbage production (kg/ha) .26*” .2l .I2 .29 -.45 Y0 pinegrass -.24* .I4 -.03 m.28 -.I4 $Yoelksedge .30** .02 .I5 -.68* .45 % tree canopy cover .Ol -.29’ -.I6 -.62 m.04 !Z slope -.24* m.so** -.52** m.58 .I5 slope length (m) -.08 -.04 .08 -.86** -.48 trail distance to water (m) -.02 .I6 m.18 -.24 -.36 vertical distance to water(m) -.08 -.25* -.22 m.33 .I2 trail distance to salt (m) -.I5 -.20 -.I8 -.93** m.25 vertical distance to salt (m) -.37** -.05 -.I4 m.46 m.02 number of sites sampled 45 36 36 7 I

“K.05. **p<.OI; correlations pooled over years.

6 times as much herbage as the logged forest and communities. The especially apparent for pinegrass, mountain brome and interme- meadows could be expected to support higher relative cattle use diate wheatgrass. Considerable amounts of useable herbage were because of their higher relative herbage production. However, lost because utilization on these species dropped nearly to zero. since overall forage utilization averaged about 75% on the mea- Utilization of the evergreen elk sedge also increased considerably dows versus 10% for the upland communities (Gillen 1982), differ- in the late grazing period. ences in relative forage production did not explain all of the The ability to predict upland forage utilization patterns from site disproportionate use. characteristics was limited under the conditions of this study. Significant differences in cattle use were found between upland Forage utilization levels were light on upland plant communities plant community groups, indicating overall vegetation structure with an average of 10% over all sites and a maximum of 36% on a and composition were major factors influencing upland cattle single site. Cattle sign, including fecal droppings and hoof prints, distribution. Cattle made heaviest use of the logged forest and was noted in virtually every part of the pastures, but large amounts more open forest communities, which agrees with other reports seldom occurred on any single upland site. from this region (Hedrick et al. 1968, Miller and Krueger 1976). It As utilization levels decrease, the relative variation of utilization was felt this behavior was related to lower tree canopy coverage estimates often increases. This additional sampling variation and increased forage production. Forage utilization was lower in makes the task of outlining utilization patterns more difficult. grassland communities compared to open forests, which does not Even though cattle would be expected to exhibit grazing preferen- agree with previous work (Harris 1954). Vegetation in these shal- ces most clearly under light stocking levels, practical sampling low soil communities matured in late June after only 1 month of limitations may prevent a clear delineation of these grazing prefer- grazing, while herbage in the forest was still green. This differential ences. This is one reason for the general lack of fit in the mathemat- forage maturity was probably most responsible for reduced graz- ical models describing upland grazing distribution. ing use of the grasslands, especially on south-facing sites in the Several investigators have developed regression models relating deferred-rotation pasture. Late summer showers which stimulate forage utilization and various site factors. Cook (1966) obtained R* green regrowth have been shown to increase cattle use on these values of 0.56 for a single year and 0.38 for 3 years of pooled data grassland sites (Bryant 1982), a phenomenon also observed during on Utah foothill range. McDaniel and Tiedeman (198 I) reported this study. R* values of 0.3%.63 in a study of sheep distribution on mountain Seasonal grazing appeared to affect plant community use in the range. Forage utilization ranged from 3% to 65%. With forage deferred-rotation pasture. Late grazing reduced the logged forest utilization of O-84%, Clary et al. (1978) reported an R* value of sites from preferred communities to communities used in propor- 0.79 for a cattle distribution model in northern Arizona. It would tion to availability. An increase in cattle use of riparian meadows appear that as the range in utilization increased, the proportion of resulted. Removal of the tree canopy accelerated plant growth on explained variation increased, possibly because distribution pat- logged areas. By the time cattle entered the pasture in the late terns were sampled with greater precision. period, many of the grasses were mature and unpalatable. This was The ability to predict cattle distribution also increases as the

552 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 number of major habitat factors influencing distribution decreases Harris, R.W. 1954. Fluctuations in forage utilization on ponderosa pine and interactions become lesscomplex. Mueggler (1965)controlled ranges in eastern Oregon. J. Range Manage. 7:250-255. all but 2 site factors and achieved an R? value of 0.81 for cattle use Hedrick, D.W., J.A. Young, J.A.B. McArthur,and R.F. Keniston. 1968. on mountain grasslands. In a relatively small mountain pasture of Effects of forest and grazing practices on mixed coniferous forests of northeastern Oregon. Oregon Agr. Exp. Sta. Tech. Bull. 103. 144 ha, Miller and Krueger (1976) developed a model accounting Julander, O., and D.E. Jeffery. 1964. Deer, elk, and cattle range relations for 99% of the observed variation in cattle distribution patterns. on summer range in Utah. Trans. N. Amer. Wildl. and Nat. Res. Conf. Correlation analyses pointed out slope gradient as the physical 29:404-414. factor most consistently associated with cattle use of the landscape. Krueger, W.C. 1972. Evaluating animal forage preference. J. Range Man- This negative effect of slope gradient on cattle distribution is well age. 251471475. known on mountain rangelands (Mueggler 1965, Cook 1966, Pat- Lemmon, P.E. 1957. A new instrument for measuring forest overstory ton 1971, Van Vuren 1980). Water distribution was moderately density. J. Forest 55:667-669. associated with cattle preference as determined by direct observa- Long, A.J., and L.L. Irwin. 1982. Elk-cattle interactions in the Bighorn tion. This was at least partly due to water being present in the Mountains, Wyoming. p. 553-563. In: J.M. Peek and P.D. Dalke, ed. Wildlife-Livestock RelationsSymposium: Proc. 10. Univ. Idaho, Forest, highly preferred riparian meadows, but water was not totally Wildl., and Range Exp. Sta., MOSCOW. responsible for this high preference. Water distribution had practi- Marcum, C.L., and D.O. Loftsparden. 1980. A nonmapping technique for cally no association with cattle grazing patterns when only upland studying habitat preferences. J. Wildl. Manage. 44:963-968. distribution ,was considered. Each pasture contained 3 perennial McDaniel, K.C., and J.A. Tiedeman. 1981. Sheep use on mountain winter streams and several upland water developments. Other workers range in New Mexico. J. Range Manage. 34: 102-104. have reported similar results on relatively well-watered ranges Miller, R.F., and W.C. Krueger. 1976. Cattle use on summer foothill (Julander and Jeffery 1964, Cook 1966, Clary et al. 1978). rangelands in northeastern Oregon. J. Range Manage. 29:367-371. Salt placement was inconsistently associated with cattle distri- Mueggler, W.F. 1965. Cattle distribution on steep slopes. J. Range Man- bution. Some research reports have indicated little influence of salt age, 18:255-257. An analysis of cattle grazing on steep slopes. Masters placement on grazing behavior (Cook 1966, Wagnon 1968, Bryant Patton, W.W. 1971. thesis, Brigham Young University, Proio, Utah. _ 1982) while others have reported a much larger influence (Cook Pechanec. , J.F.., and G.D. Pickford. 1937a. A comoarison of methods used 1967, Bjugstad and Dalrymple 1968, Patton 1971, Roath and in determining percentage utilization of range grasses. J. Agr. Res. Krueger 1982). More controlled experimentation with salt distri- 541753-765. bution on mountain rangelands seems warranted. Pechanec, J.F., and C.D. Pickford. 1937b. A weight estimate method for Some of the factors associated with upland grazing distribution the determination of range or pasture production. J. Amer. Sot. Agron. in this study can be modified through management activities. Some 29:894-904. degree of forest canopy removal will create preferred grazing areas, Phillips, T.A. 1965. The influence of slope gradient, distance from water, especially when palatable pasture grasses are established. This type and other factors on livestock distribution on national forest cattle allotments ofthe Intermountain region. USDA Forest Serv. Intermount. of manipulation should decrease cattle use of riparian zones and Forest and Range Exp. Sta., Range Improvement Notes. 10:9-19. partially overcome the negative effects of slope gradient (Cook and Ronth, L.R., and W.C. Krueger. 1982. Cattle grazing and behavior on a Jeffries 1963). Trail construction can also decrease the effects of forested range. J. Range Manage. 35:332-338. slope gradient (Patton 1971). Snedecor, G.W.,and W.C. Cochmn. 1%7.Statistical methods. Iowa. State Grazing distribution patterns on mountain rangelands are dif!i- Univ. Press. cult to predict with useful precision because they are influenced by Van Vuren, D. 1980. Ecology and behavior of bison in the Henry Moun- such a complex of physical and biological factors, including animal tains, Utah. Master of Sci. thesis, Oregon State Univ., Corvallis. social behavior. However, general relationships between plant Wagnon, K.A. 1968. Use of different classes of range land by cattle. Calif. communities, site factors, and cattle distributions have been dis- Agr. Exp. Sta. Bull. 838. cussed. Each of these factors should be considered when developing grazing plans, with the realization that every situation will have a different combination of factors. The use of grazing or land treatments which take advantage of the effects of desirable environmental factors on animal distribution or minimize the effects of undesirable factors could allow substantial improvement in the utilization and conservation of these mountain rangelands. Literature Cited

Bjugstad, A.J., and A. Dalrymple. 1968. Behavior of beef heifers on Ozark ranges. Missouri Agr. Exp. Sta. Bull. B-870. Bryant, L.D. 1982. Response of livestock to riparian zone exclusion. J. Range Manage. 35:780-785. Clary, W.P., P.F. Ffolliott, and F.R. Larson. 1978. Factors affecting forage consumption by cattle in Arizona ponderosa pirie forests. J. Range Manage. 31:9-IO. 0 Please send information about these titles: Cook, C.W. 1966. Factors affecting utilization of mountain slopes by cattle. J. Range Manage. 19:200-204. NMIle Cook, C.W. 1967. Increased capacity through better distribution on mountain ranges. Utah Sci. 28:39-42. Company/Institution Cook, C.W., and N. Jeffries. 1963. Better distribution of cattle on mountain ranges. Utah Farm and Home Sci. 24:48-49. Address Gillen, R.L. 1982. Grazing behavior and distribution of cattle on mountain rangelands. Ph.D. Diss., Oregon State Univ., Corvallis. City

Gonzalez, M.H. 1964. Patterns of livestock behavior and forage utilization state ZP as influenced by environmental factors on a summer mountain range. Phone 1 1 Ph.D. Diss., Utah State Univ., Logan. Cd t&free 500-521-3044. Or mail inquiry to: Hall, F.C. 1973. Plant communities of the Blue Mountains in eastern University Microfilms International. 300 North Oregon and southeastern Washington. USDA Forest Serv. Pac. NW Zeeb Road. Ann Arbor. MI 48106. Region, R6 Area Guide 3-1.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 553 Dietary Selection and Nutrition of Spanish Goats as Influenced by Brush Management

EXPEDITO A. LOPES AND JERRY W. STUTH

Abstract Botanical composition of Spanish goat diets was only different 1982a). A combination of cattle and goats might be used to effec- when diets selected from tebuthiuron-treated pastures were com- tively manipulate woody regrowth to sustain the beneficial aspects pared to those from untreated and mechanically treated areas in of mechanical treatments and maximize long-term returns. the Texas Post Oak Savannah. However, all brush management Use of goats in many situations is perceived to be harmful to the treatments significantly affected the browse component in summer environment. The goat controversy probably exists because of diets. Diets selected from untreated and mechanically treated pas- preconceived ideas and lack of reliable observations about this tures were dominated by browse, while grasses and grasslike plants species. This unfortunately impedes a better understanding of its occurred most in diets selected from the tebuthiuron-treated plots. role in land use. The counterpart of the “maligned”image of goats Yet, during fall and winter, vines comprised the bulk of diets is their capability to reduce brush cover on rangeland. Control of collected on these areas. Forbs were a minor dietary component. brush regrowth using goats following other brush treatments has Goat diets from untreated and mechanically treated pastures con- shown good results in maintaining fuelbreaks in the southern sistently shifted from browse to grasses and grasslike plants as California chaparral (Green et al. 1978), controlling brush species seasons advanced. Selection of grasses and grasslike plants on in Tanzania’s Massailand (Martin and Huss 1981) and in the tebuthiuron-treated pastures declined sharply from summer through northern and central rangelands of Mexico (Fierro et al. 1980), and winter and increased through spring. Similar but inverse trends in reducing Gambel oak (Quercus gambelii) sprouts in Colorado occurred in respect to vines and browse. Mean levels of crude (Davis et al. 1975). In western Texas, Spanish goats were useful in protein (CP) in diets selected by esophageally fistulated goats conjunction with prescribed burning for suppressing woody spe- grazing chemically treated pastures were significantly greater than cies during the growing season (Ueckert 1980). in diets from the other pastures in winter and spring. In summer The primary advantage of goats over other small ruminants lies and fall, dietary forage material from all pastures contained equi- in their multipurpose-utility. Since animal grazing preference is the valent levels of crude protein. Dietary in vitro digestible organic major determinant factor in a grazing program (Davis et al. 1975), matter (IVDOM) was higher in summer and winter from tebuthi- diet selection should be the first step in assessing the goat’s poten- uron-treated pastures compared to mechanically treated and tial as a complementary grazer and brush management agent. The untreated areas. In fall, diets from tebuthiuron-treated pastures primary objective of this study was to determine effects of brush were higher in IVDOM content than those from untreated ones but management practices and season on botanical composition and were similar to diets from mechanically treated pastures. However, nutritional content of Spanish goat’s diets and browsing preference in spring all pastures receiving brush management yielded diets of goats for forage classes and plant species in Texas Post Oak with higher IVDOM content than brush-treated areas. In general, Savannah. methods of brush control had greater effects on IVDOM than on Materials and Methods CP contents of diets. The research area was located in the Post Oak Savannah, Reduction of woody plant cover may be the primary step toward approximately 3 km west of College Station, Texas. The climate is increasing forage production of depleted rangelands in Texas. The subtropical with an average growing season of 274 days. Annual 2 most widely used approaches to brush management are mechani- precipitation averages 94 cm with peaks in May and September. cal methods and herbicide application (Scifres 1980). However, Mean temperature ranges from IO’ C in January to 30” C in July rising costs of fuel, equipment, and herbicides and lack of satisfac- (U.S. Department of Commerce 1975). Soils of the experimental tory control in some instances have stimulated interest in alterna- pastures were sandy loams to fine sandy loams of the Lufkin- tive control methods (Green et al. 1978). Axtel-Tabor series (Udertic paleustalfs, Scifres et al. 1981). Over- Uplands and river bottoms in the Post Oak Savannah region of story vegetation was dominated by oaks (Quercussp.). The princi- east-central Texas were originally characterized as open Savannah pal woody species in the understory included yaupon (Ilex with a ground cover of mid and tall grasses (Smeins and Slack vomitoria), possumhaw yaupon (Ilexdecidua), winged elm (Ulmus 1978). Yet, continued heavy yearlong grazing coupled with long- alata), gum bumelia (Bumelia lanuginosa), willow baccharis (Bac- term fire suppression has led to range deterioration. As a result, charis salinca), coralberry (Symphoricarpus orbiculatus), honey- millions of hectares now support dense thickets of brush which locust (Gleditsia triacanthos), flameleaf sumac (Rhus copallina), severely limit herbaceous forage production (Scifres and Haas sparkleberry (Vaccinium arboreum), Mexican plum (Prunus mex- 1974). Evaluation of cattle grazing practices in this region indi- icana), Texas persimmon (Diospyros texana), eastern redcedar cated that cattle will not maintain mechanically treated rangeland, (Juniperus virginiana), and downy hawthorne (Crataegus mollis). or make efficient use of brush-infested areas (Kirby and Stuth Primary grass species were little bluestem (Schizachyrium scopa-

Authors are EMBRAPA (Brazilian Agriculture Research Institution) fellow and rium), brownseed paspalum (Paspalum plicatulum), and Texas associate professor, Department of Range Science, Texas A&M University, College wintergrass (Stipa leuchotricha). Western ragweed (Ambrosiapsi- Station 77843. Lopes is pesently coordinator of goat research, EMBRAPA/- bitter sneezeweed plantago CNP-Caprinos, Sobral, Ceara,, Brazil, 62.100 lostachya), (Helenium amarum), (Plan- This article IS published wth the approval of the Dwector, Texas Agricultural tago sp.), blackeyed Susan (Rudbeckia divergens), croton (Croton Experiment Station,, as TA-l8090/ .92. sp.), southern thistle (Cirsium texanum), oxalis (Oxalis sp.), and Specml appreciatmn is extended to Dr. Charles Scifres for his professional support and suggestions provided throughout this study. throughwort weed (Eypatorium sp.) comprise the majority of Manuscript accepted February 29, 1984. forbs. Vines were saw greenbriar (Smilax bona-nox), southern dewberry

554 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 (Rubus trivialis), and peppervine (Ampelopis arborea). basis was determined by using the first stage of the in vitro tech- Two replications of the 3 treatments were fenced into equal nique of Tilley and Terry (1963) followed by neutral detergent units. Replication one pastures were 1.5 ha, while those of replica- analysis (Van Soest and Wine 1967) to complete the apparent tion two were 2.5 ha in size. The brush management treatments digestibility estimation. were (1) untreated, (2) mechanical dozing and piling of underbrush Grazing pressure was maintained by 6 open, fistulated cows in and small trees in spring-summer 1977, and (3) aerial application rotation through the treatment pastures each being grazed for 7 to of tebuthiuron N-(5-( l,ldimethylethyl)-1,3-4 thiadiazol-2yl-N,N- 19 days and rested for 60 to 115 days. The 6 fistulated goats were dimethyl) urea at 2.2 kg/ha (active ingredient) as 20% pellets in rotationally grazed 2 treatment pastures behind the cows resulting May 1977. in 7 to 19 days of rest prior to making ingesta collections. Availability of herbaceous vegetation was determined at the Ingesta and nutritional data were sorted by season and analyzed beginning of each of 4 seasonal collection periods. Herbaceous to detect differences (a = 0.05) among treatments using the General forage was clipped by species to ground level within 10, randomly Linear Model (Barr and Goodnight 1979). Duncan’s multiple selected, 0.25-m2 plots in each treatment pasture. range test was used to detect differences among significant treat- Browse species standing crop was determined by the crown- ment means (Steel and Torrie 1980). volume-weight technique described by Scifres et al. (1974). The point-centered quarter method (Cottam and Curtis 1956) was Results adapted for the basic evaluation method. Three, 30-point transects Woody species provided the bulk of available forage on were established in each treatment pasture. The nearest browse untreated and mechanically treated pastures, whereas grasses and plant encountered in each quarter of the point was visually grasslike plants were the major forage component on the tebuthi- appraised and assigned a geometric shape which best described the uron-treated pastures, regardless of season (Fig. 1). However, configuration of the canopy up to 1.5 m from the soil surface. The 20,000 appropriate axil measurements for each geometric shape were r made to determine the available canopy volume. Density was determined by species as described by Cottam and Curtis (1956). @fJ BROWSE Ten, 27,000-cm-’ samples were then randomly selected each season from the outer canopy edge of each brush species to calculate 1 HERBACEOUS weight of current year’s growth per unit-volume (g/cm3) of available canopy. Utilizable canopy was delineated by the depth (30 cm) and height (1.5 m) to which the grazing animal could effectively penetrate the plant canopy under moderate stocking regimes. Browse availability (kg/ ha) was derived by multiplying the average available canopy volume, the weight-volume factor (g/cm3), and the density (plant/ ha) for each species. Spanish goats with esophageal fistulae (Van Dyne and Torrel 1964) and fitted with removable cannulae as described by Taylor and Bryant (1977), were used to collect ingesta samples from each treatment pasture during summer (June 24 to July 3), fall (October 26 to November 5), winter (January 3 to 15), and spring(March 14 !? to 24) in 1980-8 I. An average of 10 animal-days of collections were fj made per treatment and replication over a IO-day samplng period 5,000 - across seasons. The fistulated goats were allowed to freely graze a 2 a given treatment pasture for 0.5 to 1.O hour after overnight fasting. Extrusa was collected in screenbottomed bags during early morn- ing and evening at least 3 times within the lOday period. Dietary material from each goat was thoroughly mixed, split into 2 subsamples for botanical and nutritional analysis, frozen for L storage, and subsequently freeze-dried at -50°C for 3 days. Nutri- “” mc T& tion subsamples were ground in a Wiley mill to pass a l-mm screen. SUMMER WINTER Botanical composition of the diet samples by species was deter- SEASON mined by the microscopic or macrofragment technique described Fig. 1. Seasonalstanding crop (kg/half or b rowse and herbaceous (grasses by Kothmann et al. (1972) except that 15 randomly selected micro- and grasslike planrs + forbs + vines) plant species on unrreared (Un). plots per extrusa sample were used rather than 20 microplots. mechanically treated (Met) and tebuthiuron-treated (Teb) pastures in Extrusa-fragment percentages for each species in the diets were east central Texas. Bar sections withinforage class with a common letter in a season are not significantly different (a = 0.05). expressed on a weight basis (Rector and Huston 1982). Derived categories included grasses and grasslike plants, forbs, vine, absolute quantities of phytomass were several orders of magnitude browse and unidentifiable fragments. lower on tebuthiuron-treated pastures than on others. Forbs and A selectivity ratio (SR, Hodgson 1982) was calculated by season vines contributed least to total standing crop but were always more and treatment for each forb, vine, and browse species, and for the available on pastures treated with tebuthiuron than on the other category, total grasses and grasslike plants. The formula described pastures (Table 1). Oaks dominated the browse stands in ali treat- by Taylor et al. (198 I) was used: ment pastures, followed by yaupons and willow baccharis (Table % diet - % available in field X 1o 1). Selectivitv . ratio q 70 diet + available in field Summer Ratings of + 10 and -10 are maximum and minimum selectivity Percentages of browse in the diets differed significantly (o = 0.05) values, while -1 to 1 indicates selection in proportion to availabil- among experimental pastures during summer (Table 1). Browse ity. Crude protein content (TO N X 6.25) was determined by the was the dominant dietary component on untreated and mechani- micro-Kjeldahl method (A.O.A.C. 1970) and expressed on an cally treated areas, comprising 77 to 5 lye, respectively, of the goats’ organic-matter basis. Digestible organic matter on an ash-free diets. Browse and forbs were minor components in goat diets on

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 555 Table 1. Botanical composition (%) by weight of forage stands in the field and in diets of Spanish goats following brush management’ in east-central Texas rangeland.

Composition on pastures* Composition in diets* Summer Fall Winter Spring Summer Fall Winter Spring - Un Met Teb Un Met Teb Un Met Teb Un Met Teb Un Met Teb Un Met Teb Un Met Teb Un Met Teb Grass and grasslike plants 7b 9b 80a 2b 4b 84a 3b 5b 84a 4b Sb 71a 13a 18a 64b 32a 25a 30a la t 12b 26a 25a 57b Forbs Western ragweed I 1 - t’ - t 1 2a 4a t la 2a 2a - - la - la Bitter sneezeweed t t 13 -1 -t 1 3a 6a 3a - -1 Blackeyed Susan t t t t 2 - - 1 - - 1 t la la la la 2a - - la la la Croton t t t -1 ---t t -t t - Plantago -I --1 117 --t t -7 - - la la 4b Southern thistle -_1 -t 2 -___-_ t --t t t la 4b 2b Oxalix -_1--2 ___1 __l la la 2a t t t - la la Throughwort weed -]__- - 1 Unidentified t -1 t t 2 t t 1 lb 2b 9a la 2a la 2a 4a 3a 2a 2a 4a la la 4a Total lb lb 14a -8 --8 2b 3b 20a 6a 13a 6a 5a 8a 16b 2a 2a 4a 5a 8a 15b Vines Saw greenbriar t t - t t --t -t 2 3a 17b 13b 7a 6a 17b 8a 8a 12a 2a 3a la Southern dewberry t t 1 t t 1 t t 1 1 1 3 la 8a 9a 13ab30b 14a 14a 42b 13a 21a 16a Total ~ 1 -1 t t I1 15 4a 18b 21b 16a 19a 47b 22a 22a 54b 15a 24a 17a

Woody plants Oaks 49b 61a 2c 52b 66a 2c 34a 49a lb 45b 63a lc 39a 30a 2b 31a 26a 3b 10a 22b la 15a lla lb Willow baccharis 9a 5ab 2b 10a 5ab 3b 15a 9ab 4b 7a 4ab Ic t t 1 5a 7a 2a 30a 38b 25a 10a 7a 9a Coralberry 1 t t 1 2 t t 1 -2 1 - 9a 4a la la 2a - t - la la - Gum bumelia 2 1 t 2 1 t 2 1 t 3 2 t 8a 14a t la 3a t t t t t t Yaupons 22a 19a lb 23a 20a lb 35a 32a 2b 25a 20a lb 4a la - 5a 5a 1 30a 14b 4b 18a 20a lb Winged elm 8 I t 8 t t 10 t t 6 t lla t t 1 t -t t 6a la - Honey locust t I t 1 - 1 1 - la la - Flameleaf sumac t t t 1 l-l t 11 t - Sparkleberry 1 I - Mexican plum t t - Texas persimmon t 1 - Eastern redcedar t t ---___ t - ~ 4a lat 1 - - Downy Hawthorne t t - Unidentified - _ - 4a la 5a 3a 5a la la la t 3a 3a t Total 92a 90a 5b 98a 96a 7b 97a 95a 7b 93a 91a 4b 77a 51b 9c 47a 48a 7b 75a 76a 30b 54a 43a llb

‘Brush treatments included: untreated (Un), mechanical (MS) and tebuthiuron (Teb).

*Percentage values in rows within a season followed by the same letter are not significantly (a q 0.05) different according to Duncan’s Multiple Range Test ‘Indicates composition less than 0.5.

tebuthiuron-treated pastures. Grasses and grasslike plants pro- Fall vided approximately 64% of the goat diets selected from the Diets selected from the tebuthiuron-treated pastures during the tebuthiuron-treated pastures and were present in significantly fall differed significantly from those on other treatments for all greater amounts than when the goats grazed untreated and classes of forage except for grasses and grasslike plants. All forage mechanically treated pastures. Vines, mainly saw greenbriar, con- classes were similar in diets collected from untreated and mechani- stituted one-fifth of the diet on the mechanical and tebuthiuron- cally treated areas. Browse was the major dietary component on treated pastures where these plants had been released from compe- both treatments, accounting for approximately 50% of the forage tition in response to canopy removal. consumed. However, on tebuthiuron-treated areas, browse dietary Primary browse species were ranked as follows in goat diets on contribution was minor. Grasses and grasslike plants were consist- untreated and mechanically dozed pasture during summer: oaks > ently an important dietary constituent, regardless of the brush winged elm > yaupons > willow baccharis (Table 1). If browse treatment, comprising 25 to 32% of the goat diets. availability was considered, then the selective order could be des- The proportion of vines in the diets were similar on untreated cribed as winged elm > oaks > yaupon > willow baccharis (Table areas (16%) and mechanically treated pastures (20%). However, 2). Selectivity for saw greenbriar and southern dewberry was high vines in diets of goats grazing the tebuthiuron-treated pastures during this period, particularly on the mechanically dozed and (47%) were greater (a ~0.05). Forbs were highest in diets from the tebuthiuron-treated pastures. tebuthiuron-treated plots (16%) compared with untreated and Although goat diets from the untreated and mechanically mechanically treated plots (5% and 8%, respectively). treated pastures were comprised mainly of browse while diets from The browse component of diets on the untreated pastres the tebuthiuron-treated pastures were grass dominated, dietary declined by 40%, relative to summer dietary levels, in the fall. values of CP were similar across treatments averaging 10% CP. Grasses and grasslike plants declined more than 50% on the (Fig. 2). However, IVDOM indiets was the highest from tebuthiuron- tebuthiuron-treated plots from the summer to fall. Little variation treated pastures during the summer period when compared to in dietary compositon occurred between summer and fall for the untreated and mechanically treated pastures (Fig. 3). The IVDOM mechanically treated pastures. Grasses and grasslike plants and value of goat diets from each treatment appeared to decrease with vines replaced browse in the diets from summer to fall on the increasing amounts of dietary browse. High dietary lignin content untreated pastures, and to some degree on the mechanically treated associated with cell walls of browse has been closely associated areas. Lush regrowth of grasses, grasslike plants, and vines was with lowered IVDOM in diets high in browse content. available on all pastures during the fall as a result of previous

556 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Table 2. Selectivity ratios for some key woody species, groups of species and classes of forage in Spanish goat diets selected from brush managed pastures in east-central Texas.*

Summer Fall Winter Spring Species/Class Un Met Teb Un Met Teb Un Met Teb Un Met Teb Woody plants Oaks -1.1 -3.4 0.0 -2.5 -4.3 t2.0 -5.5 -3.8 0.0 -5.0 -7.0 0.0 Willow baccharis -9.4 -9.2 -3.3 -3.3 t1.7 -0.2 +3.3 t6.2 t7.2 +I.8 +2.7 t8.0 Coralberry t8.0 t8.9 +6.6 0.0 0.0 -10.02 0.0 -10.0 3 -3.3 0.0 - Gum bumelia +6.0 +8.7 +5.0 -3.3 t5.0 0.0 -6.1 -10.0 0.0 -9.0 -7.4 0.0 Yaupon -6.9 -8.7 -10.0 -6.4 -6.0 0.0 -0.8 -3.9 +6.0 -1.6 0.0 0.0 Winged elm +I.6 t3.3 0.0 -7.8 t2.0 -10.0 -8.2 -10.0 +3.0 0.0 t6.0 -10.0 Eastern redcedar - - - - t9.9 t10.04 +4.3 t10.0 - Total -0.9 -2.8 +2.9 -3.5 -3.3 0.0 -1.3 -1.1 +6.2 -2.7 -;6 t4.7 Vines Saw greenbriar 9.3 t9.9 t10.0 t10.0 t9.7 t9.9 t10.0 tio.0 t9.7 +io.o +8.5 -3.3 Southern dewberry t8.2 +8.8 +7.8 t9.8 +9.7 +9.4 t9.9 t9.6 t9.5 +8.6 t9.1 +6.8 Total t9.0 +9.8 t9.1 t9.9 t9.7 +9.5 t9.7 t9.7 t9.6 t8.8 t9.2 t5.5 Grasses and grasslike plants +3.0 t3.3 -1.1 t8.8 t7.2 t4.7 -0.5 -9.2 -7.5 -7.3 +6.6 -1.1 Forbs +7.1 +8.6 -4.0 t10.0 tio.0 t2.5 t10.0 t10.0 -3.3 +4.3 +4.5 -1.4

‘Selecti\(itY ratios range from a maximum of +I0 to a minimum of -10 with -I to I being selection in proportion to availability. Brush treatments included: untreated (Un), mechamcal (Me@ and tebuthiuron (Teb). *Species were evident in clipped plois, bit not in diets. ‘Species were not evident in clipped plots or diets. ‘Species were not in clipped plots, but were identified in diets.

grazing by cattle and frequent rain showers. For example, during tebuthiuron-treated pastures during winter. Browse also was of this collection period the study area received 13.6 cm precipitation importance in diets selected from these pastures, increasing from over 30 days in 13 events. Decreased amounts of grasses and less than 10% during fall to approximately 30% in winter. More grasslike plants in the diets from the tebuthiuron-treated areas than 78% of the vine and browse components in diets resulted from were replaced by vines, and to a lesser degree by forbs. ingestion of southern dewberry and willow baccharis on the chemi- Primary browse species could be ranked in the diets of goats cally treated pastures. Grasses and grasslike plants consumed on grazing during fall as: oaks > willow baccharis > yaupons > these pastures during winter were comprised apparently of green winged elm. The following selective order is based on availability: regrowth of brownseed paspaium (Paspalum plicatulum). willow baccharis > oaks > yaupon > winged elm (Table 2). Primary browse species were ranked in the diets of goats in Winged elm and willow baccharis would replace yaupon and oaks winter as: willow baccharis > yaupon > oaks > winged elm. If in selective order during the fall period if available as sprouts. availability was considered, selective order was: willow baccharis Mature winged elm plants dropped their leaves early in the fall > yaupon > oaks > winged elm (Table 2). Willow baccharis was while sprouts maintained green leaves into winter. Rapid senes- readily consumed during winter while yaupon was consumed in cence of older willow baccharis leaves during the period, coupled proportion to its availability. Substantial grazing on oaks and with their overlapping growth habit would tend to reduce selectiv- winged elm would require forced consumption, since oaks and ity of more mature or ungrazed plants. winged elm lose most of their leaves during the winter period. Goats selected forage from all pastures in the fall, which aver- CP and IVDOM content of goat diets selected from tebuthiuron- aged 12% CP. Although dietary CP was not different across treat- treated pastures was greater than in these diets from untreated and ments, the diets varied across treatments. The lack of browse in the mechanically treated pastures during the winter. The higher qual- tebuthiuron-treated pasture was compensated for by increased ity diets from the tebuthiuron-treated pastures was associated with vine consumption, primarily southern dewberry. The greater con- a greater consumption of cool-season annual grasses, regrowth of sumption of grass, grasslike, and vines on the tebuthiuron-treated perennial grasses and southern dewberry leaves with a respective pastures resulted in higher IVDOM content in the goats’ diets as reduction in browse consumption as compared to diets from the compared to the greater browse-producing untreated pastures. other pastures. A greater CP content in diets in winter as compared to fall differs Winter from the results reported by Malechek and Leinweber (i972a) for Amounts of browse in diets from all pastures increased during Angora goats, and from those of Bryant et al. (1980) for Angora the winter, compared to amounts in fall diets. A similar trend and Spanish goats in the Edwards Plateau of Texas. Increasing CP existed for vines, though not as pronounced as with browse. The in diets of this study probably resulted from the constant and reverse occurred with herbs. Browse was the most important die- cumulative contribution of vines in the diet as seasons progressed. tary constituent in the untreated pastures, contributing 75% to the Saw greenbriar and principally southern dewberry leaves appeared diets, and approached the levels observed in summer. Willow to regrow rapidly even in late fall and winter when showers and baccharis and yaupon were the most important species in the diets, mild temperatures occurred. Continuous regrowth of willow bac- comprising 80% of the browse selected from untreated and charis leaves throughout winter also helped to maintain dietary CP mechanically treated pastures. Oaks in goat diets were represented levels in this normally deficient period. by dead, fallen leaves. Grasses and grasslike plants declined to their lowest dietary levels across treatments in winter, approaching zero Spring on untreated and mechanically treated pastures, whereas vines Differences in proportions to forage classes in diets from the increased in the diets. untreated and mechanically treated pastures was minimal during Vines were the major contributors to goat diets from the the spring. However, goat diets from the tebuthiuron-treated pas-

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 557 tures were higher in grasses, grasslikes and forbs and lower in coralberry, and gum bumelia. Depending on growth form, oaks, browse content. Vines were similar in composition across treat- winged elm and yaupons were generally selected in proportion to ments. their availability. Browse occurred in the diets of goa:s during spring as: yaupon > Because of the lack of a precise method of evaluation, current oaks > willow baccharis > winged elm. However, when availabil- year’s growth of woody vegetation was probably over-estimated, ity of browse was considered, then selection ranking for browse in mainly on untreated and mechanically treated areas. Conversely, the goats’ diet was as follows: willow baccharis > winged elm > experimental procedures most likely underestimated the standing yaupon > oaks. Selectivity of winged elm sprouts would be greater crop of saw greenbriar since these climbing vines are associated than willow baccharis when present. with tree trunks and may not have occurred in the sample plot- CP content of diets in spring from untreated and mechanically frames. These factors possibly influenced the generally low prefer- treated pastures was similar and averaged 4 or more percentage ence values of browse species and the always highly positive prefer- units less than that of diets from the tebuthiuron-treated areas. ence for saw greenbriar. Still, goat preference for browse was less This difference in CP for the spring period was probably a result of than expected. higher amounts of forbs and succulent grasses and grasslike plants Goat diets from the different brush-treated pastures apparently in diets from tebuthiuron-treated plots as compared with those contained adequate levels of CP for maintenance and production from the other treatment pastures. (NRC 1981). However, during dry summers a CP for lactation Dietary IVDOM was not different between the tebuthiuron- deficiency could occur in all pastures, especially in tebuthiuron- treated and mechanically treated pastures during spring. Goat treated pastures where the herbaceous material is in advanced diets from the untreated pastures were lower in IVDOM during stage of maturity and brush kill is near complete. this same period when compared to those of the tebuthiuron- Energy is potentially a limiting nutritive factor for Spanish goats treated and mechanically treated pastures. Careful review of the grazing Post Oak Savannah rangelands. This conclusion is based botanical composition of the diets from the mechanically treated upon nutritional requirement calculations for a 30-kg meat-type pastures did not reveal any explanation for these differences. In nanny in early pregnancy under medium activity conditions (NRC fact, the species composition in the diets from the mechanically 1981). The inadequacy of energy in goat diets is apparently an treated pastures would suggest IVDOM values more similar to intrinsic characteristic of this range situation, especially in wood- those of the untreated pastures. land situations during dry growing conditions. Malechek and Leinweber (1972a) on fair and good, and Bryant et al. (1980) on Discussion excellent condition rangelands, reported comparable energy shor- Grass may become an important component of goat diets when tages for goats in the Edwards Plateau of Texas. browse availability is decreased (Rector and Huston 1982) or when the availability of high quality grasses and forbs is increased Management Implications (Sidahmed et al. 1981). Comparison of these results with those In the Post Oak Savannah Region of Texas, yaupons, willow reported by Malechek and Leinweber (1972b) and Bryant et al. baccharis, and to some extent, saw greenbriar are serious range (1979) in the live oak Savannah of west Texas is somewhat difficult. management problems which resist control by broadcast applica- Fundamental differences existed among physiognomic structures tions of herbicides (Scifres et al. 198 1). Such troublesome species of the plant community of those trials and the present study. were the principal component of goat diets in this study. Overall, Moreover, goat grazing pressure was much reduced on the pas- they usually occurred most consistently in animal diets across tures during this study as compared with studies of Malechek and treatments and especially during winter when they formed the Leinweber ( 1972b) and Bryant et al. (1979). Despite these differen- primary food of the goats. Vines, yaupon, and willow baccharis ces, some basic points of these trials were similar. For instance, resprouted continuously after being grazed, even during the win- during winter browse was high in diets, regardless of treatment. ter. Goats persistently used these species, regardless of treatment or However, previous studies in live oak communities reported season. Since these species keep resprouting, and subsequent defo- browse to be the most important class of dietary forage in summer liation by goats occurs over time, plant vigor would likely decrease. and fall (Frapsand Cory 1940, McMahan 1964). Seasonal shifts in The final result would be a stand of woody plants with reduced selection of grasses and browse in diets were also evident among competitive capability. Therefore, controlled goat use applied in other studies. A basic deviation occurred because animals could this area especially during the winter and maintained through complement their diets with vines in this study. Forbs, especially spring could potentially reduce these species, with no relevant during the spring, occurred less frequently in diets in this study as harm to desirable range plants. This approach, if used as a fol- compared to those in the previously mentioned studies. This possi- lowup practice to conventional methods of brush management, bly could be due to the reduced diversity of forbs found in the Post would increase efficiency in controlling these troublesome plants, Oak Savannah as compared to the Edwards Plateau of Texas. thus sustaining the beneficial aspects of brush control methods Based on absolute plant values in dietary composition, goats with limited competition with cattle. An energy supplementation have been described traditionally as browsers (Fraps and Cory program would seem necessary for goats mainly during late gesta- 1940, Wilson 1957, McMahan 1964, Daviset al. 1975, Wilson et al. tion, if an overall high animal performance is desired in this region. 1975, Fierro et al. 1977, and Sidahmed et al. 1981), grazers (Knight 1965 and Somlo et al. 1981), and animals that readily consume Literature Cited both grass and browse (Malechek and Leinweber 1972b, Nge’the and Box 1976, and Bryant et al. 1979). However, goat dietary AOAC. 1970. Methods of analysis, I 17th ed. Ass. Official Agr. Chemists. selection assessed by the selectivity ratios tends to minimize Washington, D.C. extremes in their acceptability of a forage class or plant. In this Barr, A.J., and J.H. Goodnight. 1979. A user’s guide to the statistical study, vines were the forage class most highly preferred by goats, analysis system. Sparks Press, Raleigh, N.C. Bryant, F.C., M.M. Kothmann, and L.B. Merrill. 1979. Diets of sheep, regardless of treatment or season (Table 2). Preference for forbs, Angora goats, Spanish goats, and whitetailed deer under excellent range grasses and grasslike plants, and browse was inversely associated conditions. J. Range Manage. 32:412-417. with brush availability. Forbs, grasses and grasslike plants, in Bryant, F.C., M.M. Kothmann, and L.B. Merrill. 1980. Nutritive content general, had high selection ratios on untreated and mechanically of sheep, goat, and white-tailed deer diets on excellent condition range- treated pastures, whereas browse was highly selected on the land in Texas. J. Range Manage. 33:410-414. tebuthiuron-treated areas. Brush species with positive selection Cottam, G., and J.T. Curtis. 1956. The use of distance measures in phyto- ratios on all pastures were willow baccharis, eastern redcedar, sociological sampling. Ecology 37:45 I-460.

558 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 20 - .I” -.- TEBUTHIURON 70 c---4 MECHANICAL ./ ...... UNTREATED c.- TEBUTHIURON .-----a MECHANICAL / 65 *..****- UNTREATED a ,.$ a .“- --.,.>-/ ./ab ---___!y a,.‘,,5t ...... b

t Maintenance

I-

Maintenance. aI-

I I I I I I I 1 SPRING SUMMER FALL WINTER SPRING SUMMER FALL WINTER SEASON SEASON

Fig. 2. Seasonalfluctuation in % crudeprotein of Spanishgoat diets, from Fig. 3. Seasonalfluctuarion in % in vitro digestible organic matter of summer 1980 through spring 1981 on untreated, mechanically treated Spanish goat diets, from summer 1980 through spring 1981 on untreated, and tebuthiuron-treatedpastures in east central Texas. For each season, mechanically treated and tebuthiuron-treated pastures in east-central values with the same letter in a vertical direction are not significantly Texas. For each season, values with the same letter in a verticaldirection different (a = 0.05). Maintenance requirement is based on a 30-kgfemale are not significantly different (o = 0.05). Maintenance requirement is ond medium activity (NRC 1981). based on a 30 kg female and medium activity (NRC 1981).

Davis, G.G., L.E. Bartel, and C.W. Cook. 1975. Control of Gambel oak McMahan, C.A. 1964. Comparative food habits of deer and three sprouts by goats. J. Range Manage. 28:216218. classes of livestock. J. Wildlife Manage. 28:798-808. Fierro, L.C., F. Gomez, and M.H. Gonzales. 1977. Utilization de arbusti- vas indeseables por medio del pastoreo corn cabras. Bol. Pastizales, Nge’the, J.C. and Thadis W. Box 1976. Botanical composition of RELC-INIP-SARH, Mexico. 8:1-10. eland and goat on an acacia-grassland community in Kenya. J. Fierro, L.C., F. Gomez, and M.H. Gonzalez. 1980. Biological control of Range Manage. 29:290-293. undesirable brush species with the use of goats in Northern Mexico. NRC. 1981. Nutrient requirements of domestic animals; No. 15, (Unpub. data). Nutrient requirements of goats. Nat. Acad.f Sci., Washington, Fraps, G.S., and V.L. Cory. 1940. Composition and utilization of range D.C. vegetation of Sutton and Edwards Counties. Tex. Agr. Exp. Sta. Bull. Rector, B.S., and J.E. Huston. 1982. Composition of diets selected 586. by livestock grazing in combination. p. 320. In: Abstracts of Green, L.R., C.L. Hughes, and W.L. Graves. 1978. Goat control of brush Papers, 1982, Joint Meet. Amer. Sot. Anim. Sci. and Canad. regrowth on southern California chaparral fuelbreaks. p. 451-455. In: Proc. First Internat. Rangeland Cong., Denver, Cola. Sot. Anim. Sci., Guelph, Ont., Canada. Hodgson, J. 1982. Influence of sward characteristics on diet selection and Scifres, C.J. 1980. Brush Management. Texas A&M University herbage intake by the grazing animal. p. 153-166. In:Nutritional limits to Press. College Station. animal production from pastures. Commonwealth Agr. Bureau Slough Scifres, C.J., and R.H. Haas. 1974. Vegetation changes in a Post SL2 3BN, UK. Oak Savannah following plant control. Texas Agr. Exp. Sta. Kirby, D.R., and J.W. Stuth. 1982a. Botanical composition of cattle diets MP 1136. during brush managed pastures in east-central Texas. J. Range Manage. Scifres, C.J., M.M. Kothmann,and G.W. Mathis. 1974. Range site 35:434-436. and grazing system influence regrowth after spraying honey Kirby, D.R., and J.W. Stuth. 1982b. Brush management influences the mesquite. J. Range Manage. 29:97-100. nutritive content of cattle diets in east-central Texas. J. Range Manage. 35:43 l-433. Scifres, C.J., J.W. Stuth, and R.W. Bovey. 1981. Control of oaks Knight, J. 1965. Some observations on the feeding habits of goats in the (Quercus spp.) and associated woody species on rangeland with South Baring0 District of Kenya. East African Agr. and Forestry J. tebuthiuron. Weed Sci. 291270-275. 30:182-188. Sidahmed, A.E., J.G. Morris, and S.R. Radosevich. 1981. Sum- Kothmann, M.M., C.L. Leinweber, and L.B. Merrill. 1972. Nutrient con- mer diet of Spanish goats grazing chaparral. J. Range Manage. tent of the diet of range sheep. Proc: West. Sec. Amer. Sot. of Anim. Sci. 3433-35. 23:2ll-215. Smeins, F.E., and R.D. Slack. 1978. Fundamentals of ecology. Malechek, J.C., and C.L. Leinweber. 1972a. Chemical composition and in Kendall/ Hunt Publ., Dubuque, Iowa. vitro digestibility of forage consumed by goats on lightly and heavily Somlo, R., G. Campbell, and A. Pelliza-Shriller. 1981. Study of stocked ranges. J. Animal Sci. 35:1014-1019. the dietary habits of Angora goats in rangelands in Patagonia. p. Maleehek, J.C.,and C.L. Leinweber. 1972b. Forage selectivity by goats on lightly and heavily grazed range. J. Range Manage. 25:105-107. 525-544. In: Proc. Internat. Symposium on Nutrition and Sys- Martin, J.A., and D.L. Huss. 1981. Goats much maligned but necessary. tems of Goat Feeding (Vol. 2). Rangelands. 3: 199-201.

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 559 Steel, Robert G.D., and J.H. Torrie. 1980. Principles and proce- U.S. Department of Commerce. 1975. National Oceanic and dures of statistics. McGraw-Hill, New York. Atmospheric Administration Climatological Data for Texas. Tilley, J.A., and R.A. Terry. 1963. A two-stage technique for the in Van Dyne, G.N., and D.T. Torrel. 1964. Development and use of vitro digestion of forage crop. J. Brit. Grass. Sot. 18: 104-I 1I. the esophageal fistula: A review. J. Range Manage. 17:7-19. Taylor, C.A., M.M. Kothmann, L.B. Merrill, and D. Elledge. Van Soest, P.J., and R.H. Wine. 1967. Use of detergents in the 1980. Diet selection by cattle under high-intensity, low-frequen- analysis of fibrous feeds. IV. Determinations of plant and plant cy, short duration and Merrill grazing systems. J. Range Man- cell-wall constituents. J. Ass. Off. Agr. Chemists. 50:50-55. age. 33:428-434. Wilson, A.D., J.H. Leight, N.L. Hindley, and W.E. Mulham. Taylor, C.A., and F.C. Bryant. 1977. A durable esophageal can- l!XFZomparisonofdietsofg~~onaCasuanm~~ nula for sheep and goats. J. Range Manage. 30:397-398. oleifolium woodland community in Western New South Wales. Ueckert, D.N. 1980. Manipulating range vegetation with pres- Aust. J. Exp. Agr. and Anim. Hush. 15:45-53. cribed burning. p. 27-44. Proc. Symposium on Prescribed Range Wilson, P.M. 1957. Studies of the browsing and reproductive Burning in the Edwards Plateau of Texas. behavior of the East African dwarf goat. East African Agr. J. 23:138-148. Estimating Seasonal Diet Quality of Prong- horn Antelope from Fecal Analysis

B.H. KOERTH, L.J. KRYSL, B.F. SOWELL, AND F.C. BRYANT

Abstract Botanical composition of pronghorn antelope diets from fecal using ocular estimates of forage removed by free-ranging prongh- analysis and nutrient quality of samples of plants known to be used orns and subsequently hand-picked samples and composited by by pronghorn were evaluated from June 1979 to May 1980 in weight those plants consumed for later chemical analyses. Schwartz Oldham and Hartley counties of the Texas Panhandle. Pronghom et al. (1977) hand-picked plant samples to duplicate observations in this area consume forbs primarily throughout the year, followed of tame pronghorn and attempted to estimate dietary nutritional by browse and grasses. Pronghom exhibited an affinity for either content from mean bite-weight. The objective of this study was to Artemisia ludoviciana or Sphaeralcea coccinea, or both, in all evaluate nutrient content of pronghorn diets from known botani- seasons. Grass use was negligible. Seasonal crude protein estimates cal diet composition using fecal analyses, and nutrient consump- ranged from a low of 9.8% in winter to a high of 11.4% in spring. tion of herbages determined from hand-harvested samples. Estimates of phosphorus were loweffin winter (O.l5%)and highest in spring (0.18%) corresponding f&rapid plant growth. Digestible Study Area energy levels were lowest in the fall; approaching 2,227 kcal/kg, The study area included 1,821 ha on the Masten and Spring and highest in spring and summer, 2,656 and 2,631 kcal/kg, respec- Creek ranches in Oldham and Hartley counties of the Texas Pan- tively. Average in vitro digestible organic matter coefftcients for handle. Topography was level to rolling, broken only by the Cana- spring, summer, fall, and winter were 69%, 67%, 53%, and 6195, dian River and its tributaries. Elevations varied from 976 to I,28 1 respectively. The combination of fecal analysis for botanical com- m. Soils in the area were deep sands, sandy loams, and loams with position and nutrient content from samples of plants known to be small exposures of Permian Red Beds along the Canadian River. ingested provides at least an estimate of nutrient content of the Average annual precipitation was 49.5 cm (Soil Conservation Ser- diet. vice 1980). Evaluating the quality of a habitat for ungulates requires esti- Four vegetation types, juniper breaks, mesquite/shortgrass, mates of nutrient supply from the vegetation complex. For free- sand sagebrush, and catclaw acacia/yucca, cover most of the area ranging wild ungulates like pronghorn antelope (~nrrrocu~ru and have been described by Koerth (198 1). The study was grazed americana), gross estimates must sometimes be substituted where continuously, yearlong by cattle at moderate stocking levels. All use of refined techniques are limited. Chemical analyses of plants vegetation types also were used by mule deer (Odocoileus hem- collected from uneaten forage usually accompany some estimate of ionus). the diet. Smith and Malechek (1974) estimated pronghorn diets Nomenclature for grasses follows Gould ( 1975) whereas, nomenclature for forbs and browse follows Correll and Johnston (1970).

At the time of this research, the authors were research assistants and Bryant was assistant professor, Department of Range and Wildlife Management, Texas Tech Methods University, Lubbock. Koerth is currently research associate, Texas Agricultural Experiment Station, Rt. 2, Box 589, Corpus Christi, Tex. 78410. Krysl is now with Botanical Compostion of Diet Range and Animal Science Department, New Mexico State University, Las Cruces, 88001; Sowell is with Dept. Range Sci., University of Wyoming, Laramie, 82071. From observed defecations of pronghorn, fecal pellets were Partial funding of this project was provided by the USDA Forest Service, Rocky collected monthly from June 1979 through May 1980. Samples Mountain Forest and Range Experiment Station, Great Plains Wildlife Research (approximately 20 grams each from 43-57 single defecations/sea- Laboratory, Lubbock, Texas, the Ceasar Kleburg Foundation for Wildlife Conserva- !ion, an! the Noxious Brush and Weed Control Program, Texas Tech University. This son) were preserved in 95.0% ethanol prior to examination. 1sTechnIcal Article T-9-275, College of Agricultural Sciences, TexasTech University. Microscopic slides of reference and fecal material were prepared as Manuscript accepted April 27, 1984.

560 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 described by Free et al. (1970). Microhistological examination of ficient, corrected from IVOMD of a standard forage of known in samples followed procedures outlined by Sparks and Malechek vivo digestibility, by 4,000 kcal/ kg. (1968). Five microscopic slides were made from each sample and 20 Problems were encountered in trying to estimate nutrient con- fields/ slide were examined at 100X magnification. Relative density tent of each species found in the diet. Within a season, we some- of plant species in the diet was calculated for each month and times were unable to collect samples every month of each plant averaged across the following seasons: winter (December-February), species. Thus, a single month’s nutrient value for a plant sometimes spring (March-May), summer (June-August), and fall (September was used to represent the chemical content of the species for the -November). entire season. When more than one monthly sample was collected within a season for a particular plant, we used the highest monthly Nutrient Composition value of nutrient content found for that season. Last, for some Composite samples of individual plant species used by prong- species found in the diet, we were unable to collect a sample for horn were obtained each month by hand-picking from 20 or more chemical analysis because search time proved to be prohibitive. randomly selected plants. Plant parts (leaves and new growth Since nutrient intake could not be measured directly, the average twigs) were selected to simulate pronghorn grazing behavior. All percent a plant species contributed to seasonal diets was multiplied samples were air-dried in a forced air oven at 60°C for 48 hr, by its chemical content to provide an estimate of the weighted ground in a Wiley mill to pass a 40-mesh screen, and stored in air nutritional value of that species, similar to procedures of Urness tight jars. and McCulloch (1973). To estimate the nutrient content of sea- Percent nitrogen was determined for each composite plant sam- sonal pronghorn diets, weighted values for each nutrient were ple using the micro-Kjeldahl method of Ocherman (197 1). Percent summed across species and divided by the percent of the total diet phosphorus was calculated using standard A.O.A.C. (1970) proce- accounted for from plants analyzed for nutrient content. dures. In vitro organic matter digestibility (IVOMD) was determined by procedures outlined by Van Soest (1970), who modified Results and Discussion the Tilley-Terry 2-stage technique (Tilley and Terry 1963). The technique was a 48hr in vitro digestion with inocula from steer fed Botanical Composition alfalfa (Medicago sativa) hay, followed by neutral detergent to Yearlong, pronghorn diets were comprised of 57% forbs, 38% extract all available organic matter. Percent organic matter con- browse, 3% grass and sedges, and 2% unknown (Table 1). That tent was determined by ashing duplicate samples at 550” C for 4 hr. forbs dominated the diets agrees with data from Texas (Buechner Digestible energy was estimated by multiplying the IVOMD coef- 1950), Alberta (Mitchell and Smoliak 1971), and eastern New

Table 1. Vegetation (mean %) making up 2% or more of a seasonal diet for prongbom in the Texas Panhandle.

Season of Year spring Summer Fall Winter Annual Forage (43)” (43) (57) (46) (189) Grasses: Blue grama (Bouteloua gracilis) 2.5 I.1 I.1 T 1.3 Others 2.0 I.5 T I.1 1.2 Subtotal 4.5 2.6 I.4 I.3 2.5 Forbs: Scarlet globemallow (Sphaeralcea coccinea) 12.1 14.6 12.3 3.0 10.5 Bladder pods (Lesquereha sp.) 10.2 - 2.9 13.2 6.6 Whitesage (Artemisia ludoviciana) 8.6 7.3 9.7 5.1 7.7 Mentzelia (Mentzelia nuda) 4.9 3.1 2.9 10.6 5.4 Wooly plantain (Plantago patagonica) 2.9 3.6 1.7 I.9 2.5 Plains Zinnia (Zinnia grandiflora) 2.7 I.0 Tb T T Ragweed (Ambrosia psilostachya) 1.5 6.5 2.9 I.7 3.2 Texas croton (Croton texensis) T 6.5 10.8 3.4 5.3 Plains blackfoot (Melampodium leucanthum) 1.9 3.1 3.4 3.0 2.9 Buckwheat (Eriogonum sp.) 1.5 2.7 I.9 1.9 2.0 Ratany (Krameria kmceolata) T 2.3 1.4 I.1 T Gaura (Gaura sp.) T 2.0 T T T White milkwort (Polygala alba) I.5 - - 2.3 T Others 3.4 12.5 8.9 12.2 9.3 Subtotal 50.7 58.2 59.4 60.9 57.3 Browse: Half-shrub sundrop (Calylophus serrulatus) 17.9 19.9 14.7 9.6 15.5 Sand sagebrush (Artemisiafihfolia) IS.1 2.6 13.5 20.3 12.9 Skunkbush (Rhus aromatica) 5.0 6.3 3.9 I.5 4.2 Feather dalea (Dalea formosa) T 3.7 T 1.3 Honey mesquite (Prosopis glandulosa) - 3.0 2.4 - 1.4 Perennial broomweed (Xanthocephalum sarothrae) 1.2 T - 4.9 1.6 Others I.0 T T - T Subtotal 40.3 36.7 37.4 36.6 31.7 Unknown: 4.5 2.5 1.8 I.2 2.5 TOTAL loo.0 100.0 100.0 loo.0 loo.0

‘Number of samples bT = Traces (

JOURNAL OF RANGE MANAGEMENT 37(6). November 1984 561 be relative in nature depending upon their nutrient content. Furth- ermore, while Mitchell and Smoliak (1971) suggested managers of pronghorn habitat must consider the importance of Artemisia, data from the Trans-Pecos of Texas (Buechner 1950) and eastern New Mexico (Beasom et al. 1982) suggest pronghorn exist in habitats entirely void of Artemisia; thus its consideration as a mandatory component of the habitat is questionable. Oh et al. (1968) found volatile oils of A. douglasiana and A. tridentata had antibacterial properties; Nagy and Tengerdy (1968) believed that not over 50% of A. tridendata or A. nova would be tolerable in deer diets. Wallmo et al. (1977) used 20% dietary sagebrush in their mule deer model. Future research should focus on the optimal percent of Artemisia stands required by pronghorn, based on the nutrients, forage, and/ or thermal cover it provides. - ROADS Scarlet globemallow (Sphaeralcea coccinea) was the dominant _..-..-.._ STREAMS forb selected by pronghorn in every season but winter (Table 1) and SPECIFIC averaged 11% of the annual diet. The genus Sphaeralcea was an important food item in other pronghorn habitats. Globemallow ~ziY was the number 1 ranking forb in annual diets of pronghorn in eastern Colorado (Schwartz and Nagy 1976) and eastern New Fig. 1. Study area location in the Texas Panhandle. Mexico (Beasom et al. 1982), the number 1 ranking summerforb in Mexico (Beasom et al. 1982). From pronghorn antelope habitats in Utah (Beale and Smith 1970, Smith and Malechek 1974) the third Saskatchewan (Dirschl 1963), Oregon (Mason 1952), and Utah ranking forb annually in Utah (Beale and Smith 1970), the fifth (Beale and Smith 1970) browse played the dominant role in the ranking annual forb in Kansas (Hlavachick 1968), and was present annual diet of pronghorn. The negligible use of grass by pronghom in Alberta pronghorn diets (Mitchell and Smoliak 1971). in our study agrees closely with results from Texas (Buechner Other important forbs were bladderpods (Lesquerella sp.) in 1950), New Mexico (Beasom et al. 1982), Oregon (Mason 1952), winter and spring, Texas croton (Croton texensis) in fall, and and Utah (Beale and Smith 1970), but differs from data of bractless mentzelia (Mentzelia nuda) in winter (Table 1). As an Schwartz and Nagy (1976) in Colorado and Hlavachick (1968) in example of the seasonal importance, Texas croton contributed Kansas, who found pronghorn used considerably more grass. In more crude protein and more phosphorus to autumn pronghorn Colorado, grass-dominated annual pronghorn diets were attrib- diets than any other species. uted to a grass-dominated available forage (Schwartz and Nagy Nutrient Composition 1976). Crude Protein Seasonal consistency in pronghorn use of forbs (5060%) and Seasonal crude protein estimates ranged from a low of 9.8% in browse (3641%) (Table 1) agreed with findings from a similar winter pronghorn diets to a high of almost 11.4% in spring; habitat in eastern New Mexico (Beasom et al. 1982). However, summer and fall estimates were intermediate (Table 2). Smith and pronghorns from more northern and western habitats in the U.S. Malechek (1974) reported considerably higher spring and summer and Canada had seasonal peaks in forb use during spring and/ or crude protein levels in pronghorn diets in Utah. Compared with summer, and browse or grass replaced forbs as the primary food crude protein estimates for pronghorn diets in Colorado (Schwartz item during the rest of the year (Mason 1952, Dirschll963, Hlava- et al. 1977), our spring, summer, fall, and winter estimates were chick 1968, Severson et al. 1968, Beale and Smith 1970, Mitchell similar, respectively, to their March, August, October, and Janu- and Smoliak 1971, Schwartz and Nagy 1976). Where only fall ary estimates that varied from 7.3 to 10.4%. Spring crude protein (Couey 1946), winter (Bayless 1969), orsummer(Smith and Male- was lower in diets estimated in our study than for Colorado chek 1974) diets were reported, forbs also played secondary role, pronghoms during April and May. relative to browse, in pronghorn diets. Our data, supported by data Maintenance requirements for protein have been compared to of Buechner (1950) and Beasom et al. (1982) suggest pronghorns the 6-8% required by deer (Odocoileus sp.) (Schwartz et al. 1977). primarily are forb eaters in the southeastern part of their distri- Crude protein in the forage selected by pronghorn in our study was bution. adequate for maintenance, similar to the findings of Schwartz et al. Half-shrub sundrop (Calylophus serrulatus) was the primary (1977). browse in pronghorn diets in every season but winter and averaged over 15% of the annual diet (Table 1). More than any other Phosphorus individual species, half-shrub sundrop contributed the most crude Estimates of phosphorus in the diets of pronghorn were highest protein and phosphorus to spring and summer diets and the most in spring (0.18%), lowest in winter (0. IS~c), and intermediate in digestible energy to diets in every season of the year. summer and fall (0.16 and 0.17%, respectively) (Table 2). The high Throughout pronghorn range Artemisia has been listed as an value in spring was because half-shrub sundrop and sand sage- important browse (Couey 1946, Mason 1952, Dirschll963, Hlava- brush were relatively high in phosphorus content during that sea- chick 1968, Bayless 1969, Beale and Smith 1970, Mitchell and son and they comprised 33% of the diet. We found high levels of Smoliak 1971, Barrett 1974, Smith and Malechek 1974) or forb phosphorus in pronghorn diets during rapid plant growth in (Hlavachick 1968, Bayless 1969, Mitchell and Smoliak 1971, spring, as predicted by Smith and Malechek (1974), but Schwartz Schwartz and Nagy 1976) in pronghorn diets. The genus Artemisia et al. (1977) did not.. also contributed significantly to the diet of pronghorn in this study. Estimates of phosphorus for pronghorn diets in this study were Sand sagebrush (A. filifolia) was second-most-important browse, lower than those reported for Utah pronghorns (Smith and Male- comprising 13% of the annual diet and 20% of the winter diet of chek 1974), except in late summer. Our results were, however, (A. ludoviciana) was the second or third pronghorn. Whitesage compambletoresultsfromColoradounderuheavy”ca&grazing(Schwattz ranking forb in every season and averaged 8% in the annual diet et al. 1977) except during the third period of April-July. If prongh- (Table 1). Even though these Artemisia species were of dietary orn, like sheep, require 0.20 to 0.28% phosphorus as indicated by importance, their contribution to nutrient supply was moderate Schwartz et al. (1977), then pronghorn range on our study area was except during winter and spring. Thus, importance of a plant may

562 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Table 2. Estimated nutrient content of pronghom diets in the Texas Panhandle.

Season of Year Spring Summer Fall Wmter Annual Constituent (91.9)’ (87.8) (81.1) (70.5) (82.8) - Crude Protein (%) 11.4 10.5 10.1 9.8 10.4 Phosphorus (%) 0.18 0.16 0.17 0.15 0.17 IVOMDZ 6.90 67.0 53.0 61.0 63.0 Digestible energy (Kcal/ Kg) 2,656.0 2,63 I .O 2,227.0 2.424.0 2,482.0 rPercent of diet tested *In vitro organic matter digestibility. deficient in phosphorus, as are most western ranges. he considers will allow annual replacement for maintaining the Digestible Energy panhandle population under the current harvest levels. For the Digestible energy in pronghorn diets was lowest in the fall, period 1979-80, the 3-year Oldham County average fawn:doe ratio approaching 2,227 kcal/ kg (Table 2). Highest estimates of digesti- was ,0.21 (Kothmann, personal communication) and was consi- ble energy were found in spring and summer and were very similar dered poor production. The Texas panhandle regional average for to each other (2,656 and 2,63 I kcal/ kg, respectively), while winter 1979-81 was similar at a 0.18 ratio of fawns per doe (Kothmann estimates were intermediate (2,424 kcal/ kg) (Table 2). During 1982). rapid plant growth, higher digestibilities are expected, and usually Critical periods of nutritional stress for pronghorn would be late more energy would be available to pronghorns (Smith and Male- gestation, early lactation, and prior to ovulation for the doe; and at chek 1974, Schwartz et al. 1977). If pronghorns are comparable to lactation and weaning for the fawn. Nutrient content of pronghorn deer, low energy levels in autumn could affect ovulation in young diets in our study was similar to diets from Utah (Smith and females (Abler et al. 1976). Malechek 1974) and Colorado (Schwartz et al. 1977) except for our Average IVOMD digestion coefficients for spring, summer, fall, lower spring and fall values for protein and digestibility. Since and winter were 69%, 67%, 5370, and 61%, respectively. Digestion pronghorn have higher requirements for reproduction during coefficients reported for 2 Utah study areas (Smith and Malechek spring and fall, inadequate nutrition could be at least one of the 1974) averaged across spring and summer were 68% and 70%, very factors that operated to depress pronghorn reproduction in the similar to our results. Our spring and fall values for digestion Texas Panhandle during 1979-8 1. coefficients of pronghorn diets were the same as reported from Colorado (Schwartz et al. 1977) but our summer and winter values Literature Cited were much higher (Table 2). Abler, W.A., D.E. Buekland, R.L. Kirkpatrick, and P.F. Scanlon. 1976. Plasma progestins and puberty in fawns as influenced by energy and Conclusion protein. J. Wildl. Manage. 301442-446. Association of Official Agricultural Chemists. 1970. Official methods of Pronghorns primarily eat forbs in the panhandle of Texas just as analysis (I Ith ed.). Ass. Off. Agr. Chem., Washington, D.C. they do in eastern New Mexico (Beasom et al. 1982) and the Barrett,M.W. 1974.Importance, utilization, and quality of Arremisiucana Trans-Pecos of Texas (Buechner 1950). Browse is of secondary on pronghorn winter ranges in Alberta. Antelope States Workshop importance. Pronghorn seem to have a dietary affinity for either Proc. 6:26-56. Sphuerulcea or Artemisia, or both regardless of the habitat in Bayless, S.R. 1969. Winter food habits, range use, and home range of which the animals reside. But whether these are mandatory habitat antelone in Montana. J. Wildl. Manage. 33:538-55 I. components is questionable. Beale, D:Fvi.,,and A.D.. Smith. 1970. Forage use, water consumption, and producttvtty of pronghorn antelope in western Utah. J. Wildl. Manage. Potential bias in the technique used includes (1) the masking of 34570-582. the important relationship between pronghorn selection, plant Beasom, S.L., L. LaPIant, and V.W. Howard. Jr. 1982. Similarity of phenology and nutrient content, (2) the inability to account, nutri- pronghorn, cattle, and sheep diets in southeastern New Mexico. p. tionally, for an average of 17% of a pronghorn’s diet if plants 565-572. In: Wildlife-Livestock Relationships Symp. Proc. 10. Univ. of known to be eaten were difficult to find, and (3) the inability to Idaho, For., Wildl., and Range Exp. Sta. Moscow, 1D. adjust estimates of nutrient content based on weight of forage Bueehner, H.K. 1950. Life history, ecology and range use of the pronghorn consumed. Bias (1) could be improved by more intensive and antelope in Trans-Pecos Texas. Am. Midl. Nat. 43:257-354. frequent sampling. Although bias (2) might be improved through Correll, D.S., and M.C. Johnston. 1970. The manual of the vascular plants more intensive sampling, Schwartz et al. (1977) lacked an average of Texas. Texas Res. Foundation, Renner, Texas. Antelope food in southeastern Montana. J. Wildl. of 18% of the diet even though they used tame pronghorn for Couey, F.M. 1946. Manage. 101367. estimating botanical composition and hand-plucked plant samples Dirschl, H.J. 1963. Food habits of the pronghorn in Saskatchewan. J. for subsequent nutrient analyses. Bias (3) may be the most difficult Wildl. Manage. 27:81-93. to overcome. Free, C.J., R.M. Hansen, and P.L. Sims. 1970. Estimating dry weights of Yet, combination of fecal analysis for botanical composition food plants in feces of herbivores. J. Range Manage. 23:300-302. and nutrient content from samples of plants known to be ingested Gould, F.W. 1975. The grasses of Texas. Texas A&M University Press. provided a fair estimate of nutrient content of the diet. The esti- Hlavachick, B.D. 1968. Foods of Kansas antelopes related, to choice of mates indicated pronghorns received at least an adequate supply of stocking sites. J. Wildl. Manage. 32:399-401. nutrients from the range throughout the year to support mainte- Koerth, B.H. 1981. Habitat use, herd ecology, and seasonal movements of mule deer in the Texas Panhandle. M.S. Thesis. Texas Tech Univ. nance, if their requirements are similar to sheep. However, it is Kothmann, H.C. 1982. Pronghorn antelope harvest regulations. Perfor- difficult to evaluate whether nutrition is adequate to support opti- mance Rept. Job No. 6. Fed. Aid Project No. W-109-R-5. mum reproduction. H.G. Kothmann (personal communication) Mason, E. 1952. Food habits and measurements of Hart Mountain ante- believes that the fawn:doe ratio of 0.7-0.8 is necessary for popula- lope. J. Wild. Manage. 16:387-389. tion growth with ample hunter harvest, while 0.4-0.5 was the ratio

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 563 Mjtckejj, G.J., and S. Smoliak. 1971. Pronghorn antelope range character- Schwartz, CC., J.G. Nagy, and R.W. Rice. 1977. Pronghorn dietary istics and food habits in Alberta. J. Wild]. Manage. 35:238-250. oualitv relative to forage availability and other ruminants in Colorado. J. Nagy, J.G. and R.P. Tengerdy. 1968. Antibacterial action of essential oils Wildl:Manage. 41:161-168. of Artemisia as an ecological factor. II. Antibacterial action of the Smith, A.D. and J.C. Malechek. 1974. Nutritional quality of summer diets volative oils of Artemisia tridentuto (Big Sagebrush) on bacteria from the of prongbom antelopes in Utah. J. Wild]. Manage. 38:792-798. rumcn of mule deer. Appl. Microbial. 16441-444. Soil Conservation Service. 1980. Soil Survey of Oldham County, Texas. Ocherman, H.W. 1971. Quality control of post mortem tissue. Ohio State U.S. Dep. Agr. Soil Conservation Service. Univ. and Agr. Res. Develop. Center. Sparks, D.R. and J.C. Malechek. 1968. Estimating percentage dry weights in diets using a microscopic technique. J. Range Manage. 21:264-265. Oh, H.R.9 MB. Jones, and W.M. Longhurst. 1968. Comparison of rumen Tilley, J.M.A., and R.A. Terry. 1963. A two-stage technique for the in vitro microbial inhibition resulting from various essential oils isolated from digestion of forage crops. J. Brit. Grassland Sot. 18:104-l Il. relatively unpalatable plant species. Appl. Microbial. 16:39-44.Aid Rep. UmesqPJ.,and C.Y. MeCulloch. 1973. Deer nutrition in Arizonachapar- *v-m, K-9M. May, and W. Hapworth. 1%8. Food preferences, carrying ral and desert habitats. Ark Game and Fish Special Rep. No. 3. capacities, and forage competition between antelope and domestic sheep Van Soest, P.J. 1970. Chemical basis for the nutritional evaluation of in WYoming’sRed Desert. Univ. of Wyo. Agr. Exp. Sta. Sci. Monograph IO. forages. Proc. Nat. Conf. on Forage Qual. Eval. and Util., LincolnNeb., 1969. ScBwar% CC., and J.G. Nagy. 1976. Pronghorn diets relative to forage Wallmo, O.C., L.H. Carpenter, W.L. Regelin, R.B. Gill, and D.L. Baker. availability in northeastern Colorado. J. Wild]. Manage. 4k469-478. 1977. Evaluation of deer habitat on a nutritional basis. J. Range Manage. 30:122-127.

HERBAGE ABSTRACTS (grasses, pastures, rangelands, and fodder crops)

FIELD CROP ABSTRACTS (annual field crops)

for coverage of the world literature on agricultural research

For specimen copies of these computer-produced moo thly Commonwealth Bureau of jourr,als arjd for lists of Pastures and Field Crops aru>ota ted bibliographies and Hurley, Maidenhead, other publicatiorls write to. Berks SL6 5LR, UK I 1

564 JOURNAL OF RANGE MANAGEMENT 37(e), November 1984 Technical Notes: Technique to Separate Grazing Cattle into Groups for Feeding

J.F. KARN AND R.J. LORENZ

through the entryway to get the feed. After eating, the animals were moved back through the entryway into the holding pen. This procedure was repeated several times until cattle negotiated the entryway easily inspite of the wiggling spring. Cattle that were designated not to enter a specific pen were put into that pen, and Equipment and facilities utilized for individual supplementation the screen door spring in the entryway was electrified. The cattle of cattle in range and pasture experiments are expensive and the were then slowly driven out through the entryway so that each one procedures generally are labor intensive (Harris et al. 1967, Karn received a shock. This procedure was repeated a second time to and Clanton 1974). Thus, range and pasture supplementation make sure that all animals received a good shock. experiments are often conducted by maintaining animals on separate pastures to facilitate feeding. This approach generally results in differences in animal handling procedures and in forage composition and quality differences between pastures, especially where native range is involved. An inexpensive alternative procedure which allows cattle to be maintained together, yet separated into groups at feeding time, is described.

Materials and Methods Separation of animals into supplementation groups was effected by corralling animals in a holding pen and training them either to go through or avoid a restricted access entryway leading to a feeding pen. Thus, if a group of animals were to be separated into 2 groups for feeding, one group would be trained to go through the entryway into a feeding pen and the other group would be trained to avoid the entryway. The group trained to avoid the entryway could then be fed in the holding pen. The facility we used for separatmn of cattle into groups is shown in Figure 1. The holding pen is in the background, the supplementation pen is in the fore- ground, and the restricted access entry is to the right side of the supplementation pen. Although thisarrangementonlyfacilitated 2 separations, otherseparationscould bemadebyconstructingaddi- tional supplementation pens with restricted access entries leading out of the holding pen. Portable corral panels were used for the Results and Discussion pens and for the entryway when permanent pens were not availa- This technique has been used in studies at the Northern Great ble. The restricted access entryway was approximately 76 cm wide Plains Research laboratory, Mandan, N.Dak., where a group of and projected outward from the pen in a manner that normally 30 steers was separated into 3 smaller groups, 2 supplemented and would make it difficult for animals to find and use. Mid-way I unsupplemented and where a group of 53 first-calf heifers was through the entryway, a shock mechanism was fastened which separated into 2 groups, both supplemented. The heifers were fed consisted of a heavy duty screen door spring mounted approxi- by this procedure for 18 months, both before and after calving. mately 76 cm above the ground to one of the side corral panels (Fig. There were only 5 occurrences of unsupplemented steers going 2). The spring was held in place by putting it through a tight fitting into a feeding pen during 65 feeding days; however, there were piece of plastic pipe (l4mm I.D. by 100 mm long) which was in turn several occurrences of supplemented steers going into the wrong driven through a hole placed in a block of wood. The block of pen, especially during the first 2 weeks. The steers were fed 6 days wood was then wired to a side corral panel. The outside end of the per week. Throughout the first 40 feeding days with the heifers, spring was attached to an insulated wire which led to a fence which were fed only 3 days per week, there were 1I occurrences of charger, while the other end of the spring projected into the shocked animals going through the entryway into the wrong pen. entryway. Generally 2 shocks during training were adequate to discourage Animals to be supplemented in a particular pen were trained to animals from entering the wrong pen, but ifan animal persisted in go through the entryway into that pen by distributing a small entering an incorrect pen it received additional shocks. It was amount of feed in the feed bunks, and forcing the animals to go observed that it was easier to train animals not to enter a specific pen, byputtingthemintothepenasagroup,andshockingthemas theycameout,thanitwastoselectivelyshock themastheyentered a pen intermingled with animals designated to be fed in that pen.

JOURNAL OF RANGE MANAGEMENT Z(6). November ,984 565 The latter approach often resulted ineither the wronganimal being even through they sometimes made mistakes, it was comparatively shocked or else the reactions of a shocked animal would frighten easy to move them out of one pen and into another. No problems others causing them to balk at going through the entryway. were experienced with animals going into the wrong pen merely When animals were being trained, they seemed to go into the because they followed another animal in from the pasture. How- correctpen more readilywhenfeed wasdistributed astheycame in ever, since the passageway was designed to discourage entry it was from the pasture. Ifthey hesitated and did not godirectly into their sometimes more difficult to get animals to go through the entryway pen, they seemed to become confused and not sure where they were into a pen then to keep them out. supposed to go. However, during the experiment we wanted to make sure that animals were all fed correctly, thus, feed was not Literature Cited distributed untiltheywere intheircorrectpens. lfananimaldidget Harris, LE., G.P. Lofgrern, C.J. Kerehrr, R.J. Raleigh, and V.R. Boh- into the wrong pen, it was negatively reinforced by receiving no man. 1967. Techniques of research in range livestock nutrition. Utah feed and by being shocked as it was removed from the pen. Animals A@ Exp. sta. Bull. No. 471. seemed to learn where they were supposed to go very quickly and Kam, J.F., and DC Clanton. 1974. Electronically controlled individual cattle feeding. J. Anim. Sci. (Abstract) 39:1X1

JOURNAL OF RANGE MANAGEMENT 37(6), November ,984 Book Reviews:

The Genesis and Classification of Cold Soils. By Samuel dealing with a review of growth-promoting hormones that does not Rieger. 1983. Academic Press, Inc., 111 Fifth Avenue, fit the focus of the volume. The remaining chapters discuss animal production in regard to contemporary problems: energy resources, New York, New York 10003. 230 p. $32.00. world food situation, food animal substitutes, and economic The genesis and classification of cold soils is an area of increased development. Speculation on the future for and the use of grazing interest in recent years and this publication is very timely and animals during the next few decades concludes the volume. complete in content. The book is organized into thirteen chapters, The first chapters dealing with prehistoric domestication and with nine chapters devoted to the genesis and classification of cold dispersal are interesting reading and of technical value to scientists soils that are representative of six soil orders. Chapter 1 describes involved directly with grazing animals. Of particular interest was the temperature relationships in cold soils and the factors affecting the observation that upon domestication most animals showed a soil temperatures such as air temperature, vegetative cover, slope decline in body size as compared to their wild contemporaries. /aspect, snow cover, soil moisture and soil depth. This chapter is Lack of proper nutrition was stated as the causative agent. One quite basic and has several visuals to aid in the description and notable exception was the horse, which showed no body size explanation of specific phenomena. Chapter 2 discusses the effects decline with domestication. of freezing on: moisture migration, redistribution of fine soil parti- People undertaking foreign assignments in Africa or Asia would cles, coatings of coarse fragments, soil turbation, and the stability find the chapter on indigenous breeds of those regions necessary of sloping soils. Chapter 3 briefly reviews the United States soil reading. However, most scientists in the U.S. would consider it of taxonomy system in general terms and discusses terminology, the limited value. diagnostic horizons, and the soil unit or pedon used in the classifi- The chapter on conservation of genetic resources poses the cation of soils. Chapter 4 discusses the classification and genesis of interesting question of maintaining viable gene pools of animals Entisols; included in this soil order are the great groups, Cryoflu- that may not be obviously important at this time. The discussion of vents, Cryorthents, Cryopsamments and Cryaguents. Chapters 5, evolutionary adaptations is another chapter of excellent informa- 6, and 7 discuss in great depth the genesis and classification of tion for the scientist dealing with grazing animals. Spodosols, Alfisols and Mollisols, respectively. The genesis and Material covered in many of the remaining chapters is good classification of the Cryandepts, Cryochrepts, Cryumbrepts and reading from a philosophical standpoint. The chapter on energy Cryaquepts, great groups within the soil order Inceptisols, are resources related to animal production begins with the cliche con- covered in Chapters 8,9, 10, and I I respectively. The genesis and demnation of the developed world’s over consumption of unhealthy classification of Histosols is discussed in Chapter 12. The last food such as feedlotted beef and the great expense given to animals chapter, 13, discusses and compares the taxonomy systems of used soley for companionship and recreation, while the Third Canada, the U.S.S.R. and the FAOand theadvantages and disad- World starves. Despite the introduction the chapter contains excel- vantages of each, and compares them with the U.S. classification lent information on the world energy situation and the important system. role of grazing animals in the future. As energy becomes more In general the book is well written and each chapter contains a scarce and expensive in the world, grazing animals will play an lengthy list of references. Figures and illustrations are used increasing role in cheaply converting cellulose to useable human throughout the book to stress and clarify specific points or rela- food and fiber. One vital point is made: on a world-wide basis the tionships. Considerable detail is included in the genesis sections of two most important uses of domestic animals are draught power each soil order discussed, which provides the reader with a good and manure for fuel. background on soil formation processes and some general infor- Chapter 12 presents animal production as related to the world mation on the chemical and physical characteristics of those soils. food situation. Again an important point is made: improved The book will serve as a good reference for students and profes- hygiene, sanitation, vaccination and insect vector control has sionals interested in the genesis and classification of cold soils. The allowed a decreased death rate of humans and allowed the world book also supplies good reference information for research scient- population to increase to the next limiting factor-food resources. ists working in areas where cold soils are encountered.-Gerald E. We are faced with the factor and the responsibility of dealing with Schuman, Cheyenne, Wyoming. it. Domestication, Conservation and Use of Animal Resour- Each chapter in the book is authored or coauthored by different scientists. A biographical sketch of each is included in the book. It ces. Edited by L. Peel and D.E. Tribe. 1983. Elsevier is of particular note that no American scientists were included in Science Publishing Co. Inc., P.O. Box 1663, Grand Cen- the text. tral Station, New York, NY 10163. The book can be recommended reading for those active in This volume is part of a large series titled World AnimalScience. International Agriculture both from a technical and philosophical Subseries are Basic Information, Disciplinary Approach and view. Scientists in the United States should find it interesting Production-System Approach. The review volume is the first of the reading and a good reference. It is also an excellent reference for subseries Basic Information. teachers. The book is well referenced. A list of citations appears at The first chapter deals with the prehistoric domestication of the end of each chapter. The one chapter that appeared out of place various animals and their dispersion across Europe and Asia. The in the book was the one dealing with modifying animal growth and next two chapters deal with the same topic in the Americas, Austra- anabolic agents.-Murtin Vuvra, Union, Oregon. lia and Oceania. Chapter 4 discusses the indigenous animals of Africa and Asia and their uses. The following two chapters blend in well as they include the conservation of animal genetic resources and evolutionary adaptations in animal production. The direction of the book then becomes muddled. Chapter 7 gives a historical review of animal science research beginning with the seventeenth century. The ethics of animal use is covered in chapter 8 and the rational use of wild animals in chapter 9. The editors then throw in a technical chapter

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 567 Index, Volume 37

Abemethy, Rollin H., 509 Burning (prescribed), 44, II I, 116, 237, 277, Controlling individual junipers and oaks with Acceptable block beef from steers grazing 387,392,398,491 pelleted picloram, 380 range and crop forages, 122 Conversion of solar energy into digestible energy: Achouri, Moujahed, 451 C a basis for evaluating range forages?, 25 Adams, John C., 142,269 Calf crop, 127 Cook, C. Wayne, 122 Adaptability, 21 I Calf weaning, I27 Cooley, Keith R., 529 Agbakoba, A.M., 415 Callie bermudagrass yield and nutrient uptake Copeland, T.D., 406 Age prediction methods, 362 with liquid and soil N-P-K fertilizers, 496 Copper, 5 17 Agropyron desertorum, 353 Campbell, M.H., 87 Copper and molybdenum uptake by forages Agropyron spp., 25,206,365 Campa III, Henry, 468 grown on coal mine soils, 5 17 Albedo, 271 Campbell, M.H., 87 Costs and returns of Angora goat enterprises Alpaca food habits, 330 Campbell, W.F., 365 with and without coyote predation. 166 Anderson, D.L., 321 Campbell-Kissock, Linda, 442 Cover (plant) loss, 130 Anderson, Jay E., 79 Capra hircus, 172,340 Cox, J.R., 7,204,377,507 Andropogon gerardi, 147 Carbaryl, 200 Coyotes, 423 Ansley, R. James, 509 Carbohydrates, 28 Cramer, David A., 122 Antelope, 560 Carbohydrates, (total nonstructural), 465 Crested wheatgrass, 353 Aristida longiseta, 387 Cardenas, M., 22 Cultivars, 229 Armijo-T., J. Roberto, 195 Carexfilfolia, 25 Currie, Pat O., 503 Armour, Charles D., 44 Carex nebraskensis, 465 Cynodan dactylon, 496 Artemisia cana cana, 503 Carman, John Cl., 362 Cynomys ludovicianus, 325, 358 Artemisia nova, 370 Carrigan, Mike, 420 Artemisia spp., 262, 298, 353 Cattle behavior on a south Florida range, 248 Asay, K.H., 365 Cattle diets on seeded clearcut areas in central Aspen, 156,521 interior British Columbia, 349 Atrazine, 412 Cattle diets under continuous and four- Dahl, B.E., 152 Atriplex canescens, 22 pasture, one-herd grazing systems in south Dakessian, Suren, 3 I2 A triplex gardneri, 509 central New Mexico, 50 Dalen, Raymond S., 380 B Cattle distribution on mountain rangeland in Dalrymple, R.L., 285 northeastern Oregon, 549 Deer, 64,67 Bailey, A.W., 156 Cave, George H., 277,491 Deer food habits, 67 Balliette, John F., 218 Cenraurea spp., 50 1 Deer mice, 438 Baltensperger, Arden A., 77 Center, D.M., 476 Defoliation, 156,406 Beaver, I42 Ceratoides. spp., 2 18,222 Dehydration, 462 Bedunah, Donald J., 483 Cercocarpus montanus, 321 Desert ecosystems, 49 I Beetle, Alan A., 165, 269 Cervus elaphus nelsoni, 59 Desmanthus velutinus, I85 Berg, William A., 180 Chamrad, A.D., 340 Dettori, Michael L., 218 Bermudagrass, 496 Change in bacterial populations downstream Devaurs, Micheline, 523 Bethlenfalvay, Gabor J., 312 in a Wyoming mountain drainage basin, 269 Dicamba and 2,4-D, I59 Big sacaton, 377 Characteristics of oak mottes, Edwards Pla- Dietary selection and nutrition of Spanish Bighorn sheep food habits, 67 teau, Texas, 534 goats as influenced by brush management, Birds, 239 Chemically thinning blue grama range for 554 Bison and cattle competition, 260 increased forage and seed production, 412 Diets of black-tailed jack rabbits in relation to Black grass bug, 365 Child, R.D. 283 population density and vegetation, 79 Black-tailed prairie dog food habitsand forage Chlopyralid, 40 Diets of ungulates using winter ranges in relationships in western South Dakota, 325 Chloris gayana, 83 northcentral Montana, 67 Blackburn. W.H.. 265.291.303.534 Chow, Paul N.P., 159 Digestion trials, 468 Blankenship, Lytle H.; 442’ ’ Chrysorhamnus spp., 373 Diospyros texana. I89 Blue grama, 28,40,412,447, 514 Circumstances associated with predation rates Discussion of “Biomass and forage production Bock, Carl E., 239 on sheep and goats, 423 from reclaimed stripmined land and adjoin- Bock, Jane H., 239 Clark, William R., 438 ing native range in central Wyoming”, by Booth, D. Terrance, 222, 286 Claypool, P.L., 147 Lang JRM 35:755. A viewpoint, 280 Botanical composition of diets of cattle graz- Climax theories and a recommendation for Distribution of cattle, 549 ing south Florida ranaeland. 334 vegetation classification-A viewpoint, 427 Doerr, T.B., I35 Eorhriochloa spp., 180 - Coleman, GA., 295 Does summer range quality influence sex Bouteloua gracilis, 25, 28,4 12,447, 5 14 Coleman, SW., 243 ratios among mule deer fawns in Utah?, 64 Boureloua spp., 40 Collins, Alan R., 358 Donart, Gary B., 50 Bowman, R.A., 225 Communal farms, 195 Dormaar, J.F., 31 Bromus inermis, I59 Comparative diets of Rambouillet, Barbado, Douglasfir, 104 Bromus japonicus, 387 and Karakul sheep and Spanish and Angora Dry season forage selection by alpaca (Lama Brotherson, J.D., 321,362 goats, 172 paces) in southern Peru, 330 Brush management,83,87, 319,321,554 Comparison ofgrazed and protected mountain Dunn, G.L., 147 Brushlands, 298,3 I9 steppe rangeland in Ulukisla, Turkey, 133 Dystric Eutrocrepts, 402 Bryant, F.C., 330,420, 560 Competition (bison and cattle), 260 Buchloe dacryloides, 40,483 Competition (horses and cattle), 252 E Buckhouse, John C, 298 Conner, J. Richard, I I I, 166 Early weaning and part year confinement of Buffalograss, 40 Control of aspen regrowth by grazing with cattle on arid rangelands of the southwest, Bultsma. P.M.. 398 cattle, I56 127 Bunting,’ Stephen C., 44 Control of leafy spurge in pastures using Eck, H.V., 21 I Burning of northern mixed prairie during dicamba and 2,4-D, 159 drought, 398

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Economic analysis of black-tailed prairie dog Fire intensity effects on the understory in pon- Griffith, Larry W., 286 [Cynomys ludoviciunus] control, 358 derosa pine forests, 44 Economic analysis of two systems and three Fire potential, I04 H levels of grazing on ponderosa pine-bunch- Fire temperatures and physical characteristics Habitat relations of Cercocarpus montanus grass range, 309 of a controlled burn in the upper Sonoran (true mountain mahogany) in central Utah, Economic evaluation of chemical mesquite Desert, 277 321 control using 2,4,5-T, I52 Fistulated animals, 476 Habitat use, 79 Economic evaluation of fire-based improve- Fitzgerald, R.D., 156 Hafercamp, M.R., 185,406 ment systems for Macartney rose, I I I Flatwoods range, 36,334 Hagedorn, Charles, 229 Economic use, 200,358 Food habits (leparids), 256 Hake, D.R., 147 Economics of controlling serrated tussock in Forage collection, 476 Hakonson, T.E., I2 the southeastern Australian rangelands, 87 Forage preferences of livestock in the arid Hansen, H.D., 365 Ecophysiological studies of Eleusine indica lands of northern Kenya, 542 Hansen, R.M., 256,283 (L.) Gaertn. and Sporoboluspyramidalis P. Forage production estimates, 94 Hanson, Clayton L., 3 Beauv. at Ibadan. Niaeria. 275 Forage production model, 3,25 Harper, K.T., 64 Edaphic and microclimate ‘factors affecting Forage quality, 83, I22,21 I, 233,343 Harrison, A.T., 55 tobosagrass regrowth after fire, I16 Forage response of a mesquite-buffalograss Haufler, Jonathan B., 468 Effect of phytophagous nematode grazing on community following range rehabilitation, Haws, B.A., 365 blue grama die-off, 447 483 Hawthorne, Vernon M., 239 Effect of shade and planting depth on the Forage selection, 50, 172, 260, 330, 334, 340, Heat stress, 243 emergence of fourwing saltbush, 22 343,349,542,554,560 Hennessy, J.T., 22 Effects of fire on Texas wintergrass communi- Forage use, 140 Herbaceousvegetation-lotebush[Ziziphusob- ties, 387 Forage quality responses of selected grasses to tusifolia T. & G.) Gray var. obrusifolia] Effects of herbicides on germination and seed- tebuthiuron, 83 interactions in north Texas, 3 I7 ling development of three native grasses, 40 Forage yield and quality of dryland grasses Herbage yields and water-use efficiency on a Effects of livestock grazing on infiltration and legumes, 233 loamy site as affected by tillage, mulch, and rates, Edwards Plateau of Texas, 265 Forage yield of Japanese honeysuckle after seeding treatments, 180 Effects of livestock grazing on sediment pro- repeated burning or mowing, 237 Herbel, C.H., 127 duction, Edwards Plateau of Texas, 291 Forage yields, 3, 233,237,471 Herbicide, 159,483,488 Effects of planting depth and soil texture on Forest range, 156,229 Herbicide translocation, I59 the emergence of four lovegrass, 204 Foster, M.A., 3 I7 Higgins, Kenneth F., 100 Effects of soil disturbance on plant succession Frasier, G. W., 7 Hilaria murica. I I6 and levels of mycorrhizal fungi in a sage- Frazer, B.D., 501 Hingtgen, Terrence M., 438 brush-grassland community, I35 Fulbright, Timothy E., 462 Hobbs, N. Thompson, 402 Effects of temporary dehydration on growth G Honeysuckle, 237 of green needlegrass (Stipa viridula Trin.) Horn, F.P., 243 seedlings, 462 Gall fly, 50 I Horse and cattle plant competition, 72, I30 Eklaka rangeland, 529 Gambel oak, 380 Horses, 72, 130, 252 Eleusine indica, 275 Gamougoun, N. Dedjir, 538 Horsesand cattle grazing in the Wyoming Red Elk, 59,67 Garoian, L., I I I Desert, 1. Food habits and dietary overlap, Elk food habits, 67 Germination of knapweed, 501 72 Elymus spp., 206,225 Germination of lovegrasses, 507 Horses and cattle grazing in the Wyoming Red Emergenceand seedling survival of two warm- Germination of seeds of ‘Paloma’and ‘Nezpar’ Desert, II. Dietary quality, 252 season grasses as influenced by the timing of Indian ricegrass, I9 Horses and cattle grazing on the Wyoming precipitation: a greenhouse study, 7 Germination of Texas persimmon seed, I89 Red Desert, III., 130 Energy (radiant), 25 Germination profiles of introduced lovegrasses Hubbert, M.E., 72, 130,252 Engle, D.M., 140,398 at six constant temperatures, 507 Huffman, Anthony H., 40 Eragrostis spp., 204 Gibbens, R.P., 22 Hydrology, 265,269,291,295,523,529 Erosion, 291, 295, 298 Gifford, Gerald F., 45 I, 523 Hydrology and yield model, 529 Establishment of diffuse and spotted knap- Gillen, R.L., 549 weed from seed on disturbed ground in Brit- Giurgevich, Bob, 280 1 ish Columbia, Canada, 501 Glandular trichomes: a helpful taxonomic Impact of presowing seed treatment, tempera- Estimating-grazingland yield from commonly character for Arremisia nova (Black sage- available data, 47 I ture and seed coats on germination of velvet brush), 370 bundleflower, I85 Ethridge, D.E., 152 Glyphosate, 4 I2 Euphorbia escula, I59 Impact of small mammals on the vegetation of Goat food habits, I72 reclaimed land in the northern Great Plains, Euroria lanara. 286 Goats, 166,554 Evaluating soil water models on western range 438 Grass speciesadaptability in the southern high Improved esophageal fistula bag for sheep, lands, 529 plains-a 36-year assessment, 21 I Evaluation of air threshing for small lots of 476 Grasshoppers, 200 Infiltration rate, 45 I winterfat fruits, 286 Grasslands, 100, 398 Evans, Raymond A., l9,2 18,373 Infiltration (soil), 116, 133, 265, 534 Grazing distribution, 140 lnfiltration variability, 523 Everitt, J.H., I89 Grazing effects, 55, 104, 133, 142, 156, 239, Exclosures, 239 Influence of range seeding on rodent popula- 262,265,269,29 I, 303.3 I2,430,442,538 tions in the interior of British Columbia, 163 F Grazing effects on mycorrhizal colonization Insect control (chemical), 200 and floristic composition of the vegetation Insect resistance, 365 Farfan, R.D., 330 on a semi-arid range in northern Nevada, Insects (grazing), 365 Fecal analysis, 256, 340, 560 312 Introduced species, 21 I Fertilization, 406, 4 I5 Grazing habits, 248, 252, 549 Irby, Lynn R., 67 Fertilizer uptake, 496 Grazing management impact on quail during Finkner, M.D., 127 drought in the Northern Rio Grande Plain, J Fire effects, 44, 104, 277, 402 Texas, 442 Fire effects of nitrogen mineralization and fix- Grazing systems, 50, 309,442 Jackrabbit food habits, 79 ation mountain shrub and grassland com- Great Plains grasslands, 398 Jacoby Jr., Pete W., 40, 3 I7 munities, 402 Jameson, Donald A., I95 Jewett, T.K., 252

JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 569 Johnsen, Jr., Thomas N., 380 Marlow, Clayton B., 25 Oxytropis sericea, 59 Johnson, M.K., 334 Marsh (freshwater), 248 Ozone-treated mesquite for supplementing Johnson, Patricia S., 262 Martin, F.G., 36,248,334 steers in west Texas, 420 Johnston, A., 3 1 Martin, Martha H., 204,507 Johnston, Robert S., 521 Masheti, S., 542 P Jolley, Von D., 496 Masters, Robert A., 83 Pac. Helga B. Ihsle, 67 Jones, M.B., 476 McCalla II, G.R., 265,291 Palatability, 334, 349 Jones, V.E., 488 McGinnies, W.J., 225,412 Panicum hemitomon. 36, 334 Juniper, 380 McLean, A., 501 Panicum virgatum. 406 Juniperus spp., 380 McPherson, J.K., 147 Paspalum notatum, 83 Medicago spp., 77 Patten, Duncan T., 277,491 K Medina-T.. J. Galo. 195 Kalmbacher, R.S., 36,334 Pearson, Henry A., 229 Meeker, Jr., Donald O., 427 Pederson, Jordan C., 64 Kansas Flint Hills, 392 Menzel, R.G., 295 Karn, J.F., 565 Persistence and colonizing ability of rabbit- Kasworm, Wayne F., 67 Merkel, Daniel, 427 brush collections in a common garden, 373 Kauffman, J. Boone, 430 Merrill, L.B., 265,291,534 Pettit, R.D., 488 Kay, B.L., 373 Mesquite,ZZ, I 16, 152,420,483 Pfister, James A., 50 Kelsey, Rick G., 370 Mesquite dunelands, 22 Phenological development and water relations Kenney, William R., 239 Method for estimating injury levels for control in plains silver sagebrush, 503 Kenya, 542 of rangeland grasshoppers with malathion Phosphate mining, 52 I Knapweed, 50 I and carbaryl, 200 Physocarpus malvaceus, I04 Knight, R.W., 534 Meyerhoeffer, D.C., 243 Pickup grass seed stripper, 285 Knight, William E., 229 Microhistological analysis, 340 Picloram, 40 Koeleria pyramidata, 25 Miller, R.F., 549 Pieper, Rex D., 50,538 Koerth. B.H.. 560 \ Mills, Thomas, 420 P&us ponderosa, 44 Kossock, D.C.; -185 Mineral content, 36 Pitts. John S.. 420 Krueger, W.C., 430,549 Mining reclamation, 517,521 Plant communities, 542 Krysl, L.J., 72, 130, 252 560 Molydenosis, 517 Plant control, I I I, 237 Mooso, Galen D., 496 Plant control (biological), 156, 501 L Morphogenesis, 503 Labops hesperius, 365 Plant control (brush), 483,488,491 Morrison, Dennis, 447 Plant control (chemical), 152, 159,380 Lama paces, 330 Mountain range, 549 Lance, William R., 59 Plant growth, 206,377 Mountainmahogany, 32 I Plant hardiness, 373 Lane, L.J., I2 Mueller, D.M., 225 Laycock, William A., 447 Plant morphology, 275,365 Mulching, I80 Plant phenology, 503 Leaf area nonstructural carbohydrates, root Munshower, Frank F., 517 growth characteristics of blue grama seed- Plant production, 471 Murray, Robert B., 343 Plant succession, 44, 116, 135,21 I, 262,456 lings, 5 I4 Mycorrhizal fungi, I35,3 I2 Leafy spurge, I59 Plant survival, 373 Legumes, 77,229,415,517 N Plant water requirements, 7, I2 Leptochloa dubia, 83 Planting depth, 204 Lepus calijomicus, 79 Nass, Roger D., 423 Plumb, G.E., 72, 130,252 Lightning fires in North Dakota grasslands Nassella trichomata, 87 Poisonous plants, 59 and in pine-savanna lands of South Dakota Natural establishment of aspen from seed on a Ponderosa pine, 44 and Montana, 100 phosphate mine dump, 521 Ponderosa pine/bunchgrass range, 309 Litter, 55 Nava C., Roberto, 195 Populus tremuloides, 156, 52 I Livestock grazing influences on community Nebraska sedge, 465 Potvin, M.A., 55 structure, fire intensity, and fire frequency Needlegrass, 462 Powell, J., 147 within the Douglas-fir ninebark habitat Nelson, Sheldon D., 496 Prairie dog food habits, 325 Prairie dogs. 325. 358 type, 104 Nematodes, 447 Livestock impacts on riparian ecosystems and Neuenschwander, Leon F., 44, 104, I I6 Prairie (ta&ass), 392 streamside management implications. . . A Neuman, Dennis R., 5 I7 Predation, 423 review, 430 Neutron probe, 503 Predation (coyote), 166 Predication of sediment yield from Southern Locoweed poisoning in a northern New Mex- Nigeria, 4 I5 ico elk herd, 59 Ninebark, 104 Plains Grassland with the modified univer- Long, K.R., 36,334 Nitrogen fixation estimates for some native sal soil loss equation, 295 Long-term effects of annual burning at differ- and introduced legumes, forbs, and shrubs, President’s address, 99 ent dates in ungrazed Kansas tallgrass prairie, 77 Prosopis glandulosa, I 16, 152,420,483 392 Nitrogen mineralization, 402 Provenza, Frederick D., 262 Lonicera japonica, 237 Nkurunziza, E.R., 542 Pseudotsuga menziesii, 104 Nonstructural carbohydrates and root develop- Lopes, Expedito A., 554 Q Lorenz, R.J., 565 ment in blue grama seedlings, 28 Lotebush, 3 I9 Nutrient analysis, 560 Quail, 442 Lovearasses, 204.507 Nutrient content, 554 Quercus havardii, 488 Low rates of tebuthiuron for control of sand Nutrition of beef cattle, 122 Quercus spp., 380 shinnery oak, 488 0 Quercus virginiana, 534 Lusigi, W.J., 542 Quigley, Thomas M., 309 Lynch, Greg, 423 Oak motte, 534 Quinton, Dee A., 349 Ogle, Philip R., 280 M R ‘Macartney rose, I I I Omaliko, C.P.E., 415 MacCracken, James G., 256 Onobrychis viciaefolia, 438 Rabbitbrush, 373 Madany, Michael H., 456 Onsager, Jerome A., 200 Ranch economics, 358 Malathion, 200 Oryzopsis hymenoides, I9 Range economics, 127, 152, 166, 195, 309 Mammah, O.A., 415 Ovis aries, I72 Range improvements, 180 Marginal benefits of grazing and agricultural Owens, M. Keith, 262 Range renovation, 2 I I practices on a Mexican ejido, 195 Owensby, Clenton, 392 Ratliff, Raymond R., 465

570 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Reclamation, 135, 280 Seed scarification, 185, 509 Sporobolus wrightii. 377 Redente, E.F., 135,462 Seed threshing, 222,286 Stand establishment: the role of seedling size Reeves, F.B., 135 Seeded stands, 163 and winter injury in early growth of three Regression analyses, 47 I Seeding, 204,206 perennial grass species, 206 Reliability of captive deer and cow in vitro Seedling emergence, 189,509 Stanton, Nancy L., 447 digestion values in predicting wild deer Seedling establishment, 28, 514 Steele, Judith M., 465 digestion levels, 468 Semen characteristics and behavior of grazing Stem-diameter age relationship of Tamarix Relict areas, 456 bulls as influenced by shade, 243 ramesissima in central Utah, 362 Repellants (cattle), 140 Semen quality, 243 Stipa comata, 25 Repellent effects on distribution of steers on Serrated tussock, 87 Stipa leucotricha. 387 native range, 140 Sex ratio, 64 Stipa viridula, 462 Reproduction (bulls), 243 Shade, 243 Stocking rate, 309 Responses of birds, rodents, and vegetation to Shade tolerance, 229 Stranksy, John J., 237 livestock exclosure in a semidesert grass- Shade tolerance of grass and legume germ- Stream contamination, 142, 269 land site, 239 plasm for use in the southern forest range, Stream water quality as influenced by beaver Revegetation, 438,462, 509, 514, 521 229 within grazing systems in Wyoming, 142 Review of black grass bug resistance in forage Sharma, B.M., 275 Stuth, Jerry W., 554 grasses, 365 Sharrow, Steven H., 94 Sullivan, Druscilla S., 163 Ricegrass, 19 Sheer, food habits. 172.343.476 Sullivan. Thomas P.. 163 Riparian ecosystems, 269,430 Shelion, J.M., 172,340 ’ Summer.diets of bison and cattle in southern Ritenour, Gary L., 465 Shinnery oak, 488 Utah, 260 Rittenhouse, Larry, 122 Shoot growth and development of Alamo Summer range, 64 Robertson, David C., 529 switchgrass as influenced by mowing and Supplemental feeding, 420, 565 Rodent control, 358 fertilization, 406 Swanson, Sherman R., 298 Rodents, 163,239, 256, 353 Shoot production and biomass transfer of big Switchgrass, 406 Roehrkasse, Glenn P., 269 sacaton [Sporobolus wrighti& 377 Szyska, L.A., 321, 362 Romney, E.M., 12 Short-term vegetation responses to fire in the Root growth, 28,462,514 Upper Sonoran Desert, 491 T Roots (adventitious), 514 Shrub growth, 402 2,4-D, 412,483 Rosa bracteata, I I1 Shrublands, 321 2.4.5-T. 40, I52 Raze, L.D., 501 Sideoats grama, 40 Tallgrass prairie, 147 Rumen analysis, 283 Simple disc for measurement of pasture height Tamarix ramosissima, 362 Ruminal digestion consistency of Zebu cattle, and forage bulk, 94 Tanner, G.W., 248 283 Sims, Phillip L., 180, 21 I Taxonomy, 165,370 Ruminant nutrition, 420 Skinner, Quentin D., 142,269 Tebuthiuron, 83,488, 554 Runoff, 298,45 I Skovlin, Jon M., 309 Temperature profiles for germination of two Small mammal abundance on native and S species of winterfat, 2 I8 improved foothill ranges, Utah, 353 Texas persimmon, 189 Sagebrush, 298,353,370,503 Smith, Courtney B., 353 Theade, John, 423 Sagebrush-grass, 252 Smith, Mark A., 77, 130,252 Threshing damage to radical apex affects geo- Sagebrush-grass range, 343 Smith, Michael, 142 tropic response of winterfat, 222 Sainfoin, 438 Smith, Roger P., 538 Tilling, 180 Salt cedar, 362 Smith, S.J., 295 Tobosagrass, II6 Saltbush, 22,509 Smoliak, S., 3 I Towne. Gene, 392 Sampling methods, 94,283 Soil analysis, 3 I, 204 Trampling (by cattle and horses), 130 Sandoval, L.D., 248 Soil analysis (chemical), 225 Triclopyr, 40 Savannahs, 100,415 Soil and nitrogen loss from Oregon lands Tukel, Tuncay, 133 Schimel, David S., 402 occupied by three subspecies of big sage- Schimmel, J.G., 140 brush, 298 U Soil moisture, 7, 12 Schisachyrium scoparium, 83, 147 Ueckert, D.N., 172,340,387 Schizachyrium stoloniferum. 36,334 Soil nutrients, 3 I, 225 Soil-plant-air-water model, 529 Uresk, Daniel W., 325, 358 Scifres, C.J., 83, 1I I, 317, 387 Urness, Philip J., 353 Scifres, Charles, 83 Soil-plant factors in early browning of Rus- Using weather records with a forage produc- Scrivner, Jerry H., 166 sian wildrye on the natrustoll soils, 225 tion model to forecast range forage produc- Seasonal changes in carbon content, and de- Soil-plant relationships, 3 1,456 Soil (saline), 225 tion, 3 hydrogenase, phosphatase, and urease acti- Utilization cages, 565 vities in mixed prairie and fescue grassland Soil stabilization, 534 Ah horizons, 3 I Soil surface treatments, 180 V Seasonal foods of blacktail jackrabbits and Soil temperature, 22, II6 nuttall cottontails in southeastern Idaho, Soil texture, 204 VanVuren, Dirk, 260 256 Soil water, 147, 529 Variability of infiltration within large runoff Seasonal mineral concentration in diets of Soil water repellancy, 277 plots on rangelands, 523 esophageally fistulated steers on three range Soil, vegetation, and hydrologic responses to Vegetation and litter changes of a Nebraska areas, 36 grazing management at Fort Stanton, New sandhills prairie protected from grazing, 55 Seasonal variation, 36,45 1,465 Mexico, 538 Vegetation and soil responses to cattle grazing Seasonal variation in total nonstructural car- Some aspects of rangeland improvement in a systems in the Texas Rolling Plains. 303 bohydrate levels in Nebraska sedge, 465 derived savanna ecosystem, 4 15 Vegetation change after 13 years of livestock Sediment yields, 29 I Sosebee, R.E., 152,483 grazing exclusion on sagebrush semidesert Seed germination rates, 19, 2 18, 222 Sowell, B.F., 72, 252, 560 in west central Utah, 262 Seed harvest, 285,286 Spanish goat diets on mixed-brush rangeland Vegetation classification, 427 Seed polymorphism, 19 in the South Texas Plains, 340 Vegetation of two relict mesas in Zion National Seed pretreatments and their effects on field Spatial and seasonal variability and field mea- Park, 456 establishment of spring-seeded Gardner salt- sured infiltration rates on a rangeland site in Velvet bundleflower, 185 bush, 509 Utah, 45 I Speck, Jr., John E., 142 Sporoboluspyramidalis, 275

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 571 Vere, D.T., 87 Wheatgrasses, 365 Workman, John P., 309,358 Viewpoint: Trinomials are in poor taste, 165 Whisenant, S.G., 387 Wright, Henry A., I16 White. Larrv D.. 442 W White; La& M.1 233 X Waggoner, J.W., 130, 252 White, Richard S., 206, 503 Xeric sites, 362 Wallace, Joe D., 50, 127 Whitmer, Duane, 3 Wanyama, J.M., 283 Whittington, D.L., 283 Y Wiggers, Ernie P., 420 Wapiti, 59 Yarbrough, CC., 127 Warren, L.E., 172,340 Wight, J. Ross, 3, 233 Williams. Brvan D.. 521 Yields, nutrient quality, and palatability to Water balance calculations and net produc- sheep of fourteen grass accessions for poten- tion of perennial vegetation in the Northern Williams., J . R ., 295. Wilson, A.M., 28,462, 514 tial use on sagebrush-grass range in south- Mojave Desert, 12 eastern Idaho, 343 Water deficiency, 225 Winter grazing, 412 Winter injury, 206 Young, James A., 19, 116, 122,218,373 Water quality, 269 Youtie, Berta A., 23 Water stress, 12, 147 Winter range, 67 Water stress of tallgrass prairie plants in cen- Winterfat. 2 18. 222. 286 Z tral Oklahoma, 147 Wisiol, K&in, 47 I Watson, Vance H., 229 Wolfe, Gary J., 59 Zebu cattle, 283 Webster, R.D., 185 Wood, M. Karl, 303,538 Zimmerman, G., Thomas, 104 Webb, Bruce L., 496 Woodyard, David K., 468 Ziziphus obtusifolia, 3 19 West, Neil E., 262,456 Woolhiser, D.A., 7

572 JOURNAL OF RANGE MANAGEMENT 37(6), November 1964 Official bimonthly publication of Society for Range Management

Volume 37,1984 Table of Contents

Volume 37,1984

Number 1, January Number 2, March Using Weather Records with a Forage Production Model to Forecast President’s Address ...... 99 Range Forage Production-J. Ross Wighl, Clayton L. Hanson, Lightning Fires in North Dakota Grasslands and in Pine-Savanna and Duane Whirmer ...... 3 Lands of South Dakota and Montana-Kenneth F. Hig- Emergence and Seedling Survival of Two Warm-season Grasses as gms ...... loo Influenced by the Timing of Precipitation: A Greenhouse Study- Livestock Grazing Influences on Community Structure, Fire Inten- G. W. Frasier, D.A. Woolhiser, and J. R. Cox ...... 1 sity, and Fire Frequency within the Douglas-fir/Ninebark Habitat Water Balance Calculationsand Net Production of Perennial Vegeta- Type-G. Thomas Zimmerman and L. F. Neuenschwander . . . . 104 tion in the Northern Mojave Desert-L.J. Lane, EM. Romney, Economic Evaluation of Fire-based Improvement Systems for Macart- and T.E. Hakonson ...... *... 12 ney Rose-L. Garoian. J. R. Conner, and C.J. Scl$res ...... 111 Germination of Seeds of ‘Paloma’ and ‘Nezpar’ Indian Ricegrass- Edaphic and Microclimate Factors Affecting Tobosagrass Regrowth James A. Young and Raymond A. Evans ..,...... 19 after Fire-Leon F. Neuenschwander and Henry A. Wright . . . 116 The Effect of Shade and Planting Depth on the Emergence of Four- Acceptable Block Beef from Steers Grazing Range and Crop For- wing Saltbush-J.T. Hennessy, R.P. Gibbens, and M. Car- ages-C. Wayne Cook, David A. Cramer, and Larry Rirten- denas ...... 22 house ...... 122 Conversion of Solar Energy into Digestible Energy: a Basis for Eva- Early Weaning and Part-year Confinement of Cattle on Arid Range- luating Range Forages?-Clayton B. Marlow ...... 25 lands of the Southwest-C.H. Herbel, J.D. Wallace, M.D. Nonstructural Carbohydratesand Root Development in Blue Grama Finkner, and C. C. Yarbrough ...... 127 Seedlings-AM Wilson ...... 28 Horses and Cattle Grazing on the Wyoming Red Desert, III-G.E. Seasonal Changes in Carbon Content, and Dehydrogenase, Phospha- Plumb, L.J. Krysl, M.E. Hubberr, M.A. Smith, and J. W. Wag- tase, and Urease Activities in Mixed Prairie and Fescue Grassland goner ...... 130 Ah Horizons-J.F. Dormaar. A. Johnston, and S. Smo- Comparison of Grazed and Protected Mountain Steppe Rangeland in liak ...... 31 UlukiJa, Turkey-Tuncay Tukel ...... 133 Seasonal Mineral Concentration in Diets of Esophageally Fistulated Effects of Soil Disturbance on Plant Succession and Levels of Mycor- Steers on Three Range Areas-R.S. Kalmbacher, K. R. Long, and rhizal Fungi in a Sagebrush-Grassland Community-T. B. Doerr, F. G. Martin ...... 36 E. F. Redente. and F. B. Reeves ...... 135 Effects of Herbicides on Germination and Seedling Development of Repellent Effects on Distribution of Steers on Native Range-D.M. Three Native Grasses-Anthony H. Huffman and Pete W. Jac- Engle and J.G. Schimmel ...... 140 oby, Jr...... I...... 40 Stream Water Quality as Influenced by Beaver within Grazing Sys- Fire Intensity Effects on the Understory in Ponderosa Pine Forests- tems in Wyoming-Quentin D. Skinner, John E. Speck,Jr., Charles D. Armour, Stephen C. Bunting, and Leon F. Neuen- Michael Smith, and John C. Adams ...... 142 schwander ...... 44 Water Stress of Tallgrass Prairie Plants in Central Oklahoma-D. R. Cattle Diets Under Continuous and Four-pasture, One-herd Grazing Hake, J. Powell, J.K. McPherson. P.L. Claypool, and G.L. Systems in Southcentral New Mexico-James A. plister, Gary B. Dunn ...... 147 Donart, Rex D. Pieper, Joe D. Wallace, and Eugene E. Economic Evaluation of Chemical Mesquite Control Using 2,4,5-T- Parker ...... 50 D. E. Erhridge, B. E. Dahl, and R. E. Sosebee ...... * 152 Vegetation and Litter Changes of a Nebraska Sandhills Prairie Pro- Control of Aspen Regrowth by Grazing with Cattle-R. D. Fitzgerald tected from Grazing-M.A. Porvin and A. T. Harrison ...... 55 and A. W. Bailey ...... 156 Locoweed Poisoning in a Northern New Mexico Elk Herd-Gary J. Control of Leafy Spurge in Pastures Using Dicamba and 2,4-D-Paul Wolfe and William R. Lunce ...... * 59 h’.P. Chow ...... a...... 159 Does Summer Range Quality Influence Sex Ratiosamong Mule Deer Influence of Range Seeding on Rodent Populations in the Interior of Fawns in Utah?-Jordan C. Pederson and K. T. Harper . . . . . 64 British Columbia-mornas P. Sullivan and Druscilla S. Sulli- Diets of Ungulates Using Winter Ranges in Northcentral Montana- van ...... 163 Wayne F. Kasworm, Lynn R. Irby, and Helga B. Ihsle Pat . . . 67 Viewpoint: Trinomials Are in Poor Taste-Alan A. Beetle ...... , 165 Horses and Cattle Grazing in the Wyoming Red Desert, 1. Food Costs and Returns of Angora Goat Enterprises with and without Habits and Dietary Overlap-L.J. Krysl, M.E. Hubberr, F.B. Coyote Predation-Jerry H. Scrivner and J. Richard Con- Sowell, G.E. Plumb, T.K. Jewert, M.A. Smith, and J. W. Wag- ner ...... 166 goner ...... 72 Comparative Diets of Rambouillet, Barbado, and Karakul Sheep and Nitrogen Fixation Estimates for Some Native and Introduced Spanish and Angora Goats-L.E. Warren, D.N. Ueckerr. and Legumes, Forbs, and Shrubs-Arden A. Bahenspergerand Mark J. M. Shelton ...... * 172 A. Smith ...... 77 Herbage Yields and Water-use Efficiency on a Loamy Site as Affected Diets of Black-tailed Jack Rabbits in Relation to Population Density by Tillage, Mulch, and Seeding Treatments- William A. Bergand and Vegetation-Rondo1 D. Johnson and Jay E. Anderson . . . 79 Phillip L. Sims ...... 180 Forage Quality Responses of Selected Grasses to Tebuthiuron- Impact of Presowing Seed Treatments, Temperature and Seed Coats Robert A. Masters and Charles J. Scifes ...... 83 on Germination of Velvet Bundleflower-A4.R. Haferkamp, D.C. Economics of Controlling Serrated Tussock in the Southeastern Aus- Kissock. and R. D. Webster ...... *...*...... 185 tralian Rangelands-D. T. Yere and M. H. Campbell ...... 87 Germination of Texas Persimmon Seed-J. H. Everirr A Simple Disc Meter for Measurement of Pasture Height and Forage ...... a...... *...... 189 Bulk-&even H. Sharrow ...... I...... 94 Book Reviews ...... t...... 192 Book Reviews ...... *...... 96

574 JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 Number 3, May Number 4, July Marginal Benefits of Grazing and Agricultural Practices on a Mexi- Effects of Livestock Grazing on Sediment Production, Edwards can Ejido-Donald A. Jameson, J. Roberto Armijo-T, J. Go10 Plateau of Texas-G. R. McCallo II, W. H. Blockburn, and L. B. Medina-T., and Roberto Nova-C. ,...... 195 Merrill ...... 291 A Method for Estimating Economic Injury Levels for Control of Prediction of Sediment Yield from Southern Plains Grasslands with Rangeland Grasshoppers with Malathion and Carbaryl-Jerome the Modified Universal Soil Loss Equation-S.J. Smith. J.R. A. Gnsager ...... a...... 200 Williams, R. G. Menzel, and G. A. Coleman ...... 295 Effects of Planting Depth and Soil Texture on the Emergence of Four Soil and Nitrogen Loss from Oregon Lands Occupied by Three Sub- Lovegrasses-Jerry R. Cox and Martha H. Martin ...... 204 species of Big Sagebrush-Sherman R. Swanson and John C. Stand Establishment: The Role of Seedling Size and Winter Injury in Buckhouse ...... 298 Early Growth of Three Perennial Grass Species-Richard S. Vegetation and Soil Responses to Cattle Grazing Systems in the White ...... 206 Texas Rolling Plains-M. Karl Wood and Wilberr H. Bloek- Grass Species Adaptability in the Southern High Plains-a 36-Year burn ...... 303 Assessment-H. V. Eck and P. L. Sims ...... 211 An Economic Analysis of Two Systems and Three Levels of Grazing Temperature Profiles for Germination of Two Species of Winterfat- on Ponderosa Pine-Bunchgrass Range-77romos M. Quigley. Jon Michael L. Dettori, John F. Balliette, James A. Young, and M. Skovlin, and John P. Workman ...... 309 Raymond A. Evans ...... 218 Grazing Effects on Mycorrhizal Colonizationand Floristic Composi- Threshing Damage to Radicle Apex Affects Geotropic Response of tion of the Vegetation on a Semi-arid Range in Northern winterfat-D. Terrance Booth ...... 222 Nevada-Gobor J. Bethlenfalvay ond Suren Dakessian ...... 312 Soil-Plant Factors in Early Browning of Russian Wildrye on Natrus- Herbaceous Vegetation-Lotebush [Ziziphus obtusifolio (T. & G) toll Soils-R.A. Bowman, D.M. Mueller, and W.J. McGin- Gray var. obrustfolio] Interactions in North Texas-M.A. Foster, rues ...... 225 C.J. Scifes, and P. W. Jocoby, Jr...... 317 Shade Tolerance of Grass and Legume Germplasm for Use in South- Habitat Relations of Cercocarpus montanus (true mountain maho- ern Forest Range-Vance H. Worsen, Charles Hagedom. Wil- gany) in Central Utah-J.D. Brotherson, D.L. Anderson, and liam E. Knight, and Henry A. Pearson ...... *.. 229 L. A. Szysko ...... , ...... 321 Forage Yield and Quality of Dryland Grasses and Legumes--Larry Black-tailed Prairie Dog Food Habits and Forage Relationships in M. White and J. Ross Wight ...... 233 Western South Dakota-Daniel W. Uresk ...... 325 Forage Yield of Japanese Honeysuckle after Repeated Burning or Dry Season Forage Selection by Alpaca (Luma pocos) in Southern Mowing-John J. Srronsky ...... 237 Peru-R.D. Far-an and F.C. Bryant ...... 330 Responses of Birds, Rodents, and Vegetation to Livestock Exclosure Botanical Composition of Diets of Cattle Grazing South Florida in a Semidesert Grassland Site-Curl E. Bock, June H. Bock, Rangeland-R.S. Kalmbacher. K. R, Long, M, K. Johnson, and William R. Kenney. and Vernon M. Hawthorne ...... 239 F. G. Martin ...... 334 Semen Characteristics and Behavior of Grazing Bulls as Influenced by Spamsh tioat Diets on Mixed-brush Rangeland in the South Texas Shade-S. W. Coleman, D. C. Meyerhoeffer. ond F. P. Horn Fcrnsr;f E. Warren, D. N. Ueckert, J. M. Shelton, and A. D...... 243 ...... , ...... 340 Cattle Behavior on a South Florida Range-G. W. Tanner, L.D. Yields, Nutrient Quality,and Palatability to Sheep of Fourteen Grass Sandoval, and F. G. Martin ...... * 248 Accessions for Potential Use on Sagebrush-Grass Range in South- Horses and Cattle Grazing in the Wyoming Red Desert, 11. Dietary eastern Idaho-Robert B. Murray ...... 343 Quality-L.J. Krysl, B.F. Sowell, M.E. Hubbert, C.E. Plumb, Cattle Diets on Seeded Clearcut Areas in Central Interior British T K. Jewett, M.A. Smith, and J. W. Waggoner ...... 252 Columbia-Dee A. Quinton ...... 349 Seasonal Foods of Blacktail Jackrabbits and Nuttall Cottontails in Small Mammal Abundance on Native and Improved Foothill Southeastern Idaho-James G. MacCracken and Richard M. Hon- Ranges, Utah-Courtney B. Smith and Philip J. Urness ...... 353 sen ...... 256 An Economic Analysis of Black-tailed Prairie Dog [Cynomys ludovi- Summer Diets of Bison and Cattle in Southern Utah-Dirk cionus] Control-R.Collins. John P. Workman, and Daniel W. Van Vuren ...... 260 Uresk ...... 358 Vegetation Change after 13 Years of Livestock Grazing on Sagebrush Stem-Diameter Age Relationships of Tomarix ramosissimo in Cen- Semidesert in West Central Utah-Neil E. West, Frederick D. tral Utah-Jock D. Brotherson, John G. Cormon, and Lee A. Provenm, Patricia S. Johnson, and M. Keith Owens ...... 262 Szyska ...... 362 Effects of Livestock Grazing on Infiltration Rates, Edwards A Review of Black Grass Bug Resistance in Forage Grasses-W. F. Plateau of Texas-G. R. McCalla IL W. H. Blackburn. and L. B. Campbell. B.A. Haws, K.H. Asay, and H.D. Hansen ...... 365 Merrill ...... 265 Glandular Trichomes: A Helpful Taxonomic Character for Artemisio Change in Bacterial Populations Downstream in a Wyoming Moun- nova (Black Sagebrush)-Rick G. Kelsey ...... 370 tain Drainage Basin-Quentin D. Skinner, John C. Adams, Alan Persistence and Colonizing Ability of Rabbitbrush Collections in a A. Beetle. and Glenn P. Roehrkasse ...... *...... *. 269 Common Garden-James A. Young, Raymond A. Evans, and Ecophysiological Studies of Eleusine indico (L.) Gaertn. and Sporo- B.L Kay *...... *...... 373 bolus pyramidalis P. Beauv. at Ibadan, Nigeria-B.M. Shoot Production and Biomass Transfer of Big Sacaton [Sporobolus Sharma ...... 275 wright+Jerry R. Cox ...... 377 Fire Temperatures and Physical Characteristics of a Controlled Burn Controlling Individual Junipers and Oaks with Pelleted Picloram- in the Upper Sonoran Desert-Duncan T. Potten and George H. Thomas N. Johnsen, Jr., and Raymond S. Dalen ...... 380 Cave ...... 217 Book Reviews ...... 384 Discussion of “Biomass and Forage Production from Reclaimed Stripmined Land and Adjoining Native Range in Central Wyom- ing” by Lang JRM 35:755. A Viewpoint-Philip R. Ogle ond Bob Giurgevich ...... 280 Ruminal Digestion Consistency of Zebu Cattle-R. M. Hansen, D.L. Whittington, R.D. Child, and J.M. Wanyamo ...... 283 Pickup Grass Seed Stripper-R. L Dolrymple ...... *.*...... 285 Evaluation of Air Threshing for Small Lots of Winterfat Fruits-D. Terronce Booth and Larry W. Griffh ...... 286 BookReviews ...... 288

JOURNAL OF RANGE MANAGEMENT 37(6), November 1984 575 Number 5, September Number 6, November Effects of Fire on Texas Wintergrass Communities-S. G. Whiseruznt, Forage Response of a Mesquite-Buffalograss Community Following D. N. Ueckert, and C. J. Scijres ...... 387 Range Rehabilitation-Donald J. Bedunah and Ronald E. Sose- Long-term Effects of Annual Burning at Different Dates in Ungrazed bee ...... 483 Kansas Tallgrass Prairie-Gene Towne and Clenton Low Rates of Tebuthiuron for Control of Sand Shinnery Oak-V.E. Owensby ...... 392 JonesandR.D.Pettit ...... 488 Burning of Northern Mixed Prairie during Drought-D.M. EtrgC Short-term Vegetation Responses to Fire in the Upper Sonoran and P.M. Bultsma ...... 398 Desert-George H. Cave and Duncan T. Putten ...... 491 Fire Effects on Nitrogen Mineralization and Fixation in Mountain Callie Bermudagrass Yield and Nutrient Uptake with Liquid and Shrub and Grassland Communities-N. Thompson Hobbs and Solid N-P-K Fertilizers-Gulen D. Mooso, Von D. Jolley, Shel- DuvidSSchimel ...... 402 don D. Nelson, and Bruce L. Webb ...... 496 Shoot Growth and Development of Alamo Switchgrass as Influenced Establishment of Diffuse and Spotted Knapweed from Seed on Dis- by Mowing and Fertilization-M. R. Huferkump and T. D. Cope- turbed Ground in British Columbia, Canada-L.D. Roze, B.D. lnnd ...... 406 Fruzer, and A. McLean ...... , ...... 501 Chemically Thinning Blue Grarna Range for Increased Forage and Phenological Development and Water Relations in Plains Silver Seed Production- William J. McGinnies ...... 412 Sagebrush-Richard S. White ond Put 0. Currie ...... 503 Some Aspects of Rangeland Improvement in a Derived Savanna Germination Profiles of Introduced Lovegrasses at Six Constant Ecosystem-C.P.E. Omuliko. O.A. Mummoh, and A.M. Agbu- Temperatures-Martha H. Martin and Jerry R. Cox ...... 507 kobu ...... 415 Seed Pretreatments and Their Effects on Field Establishment of Ozone-treated Mesquite for Supplementing Steers in West Texas- Spring-Seeded Gardner Saltbush-R. James Ansley and Rollin Fred C. Bryant, Thomas Mills, John S. Pitts, Mike Currigun. and H. Abernethy ...... , ...... 509 Ernie P. Wiggers ...... 420 Leaf Area, Nonstructural Carbohydrates, and Root Growth Charac- Circumstances Associated with Predation Rates on Sheep and teristics of Blue Grama Seedlings-A.M. Wilson ...... 514 Goats-Roger D. Nuss, Greg Lynch, und John Theude ...... 423 Copper and Molybdenum Uptake by Forages Grown on Coal Mine Climax Theories and a Recommendation for Vegetation Classifi- Soils-Dennis R. Neumnn and Frunk F. Munshower ...... 5 17 cation-A Viewpoint-Donald 0. Meeker. Jr., and Daniel L. Natural Establishment of Aspen from Seed on a Phosphate Mine Merkel ...... 427 Dump-Bryun D. Williams ond Robert S. Johnston ...... 521 Livestock Impacts on Riparian Ecosystems and Streamside Man- Variability of Infiltration within Large Runoff Plots on Rangelands- agement Implications. . . A Review-J. Boone Kuuffmon und Micheline Devours and Gerald F. Gifford ...... 523 W. C. Krueger ...... 430 Evaluating Soil Water Models on Western Rangelands-Keith R. Impact of Small Mammals on the Vegetation of Reclaimed Land in Cooley and David C. Robertson ...... 529 Characteristics of Oak Mottes, Edwards Plateau, Texas-R. W. the Northern Great Plains-Terrence M. Hingtgen and William Knight, W.H. Blackburn, and L. B. Merrill ...... 534 R.Clurk ...... 438 Soil, Vegetation, and Hydrologic Responses to Grazing Management Grazing Management Impacts on Quail during Drought in the at Fort Stanton, New Mexico-N. Dedjir Gamougoun. Roger P. Northern Rio Grande Plain, Texas-Linda Campbell-Kissock, Smith, M. Karl Wood, and Rex D. Pieper ...... 538 Lytle II. Blankenship, und Larry D. White ...... 442 Forage Preferences of Livestock in the Arid Lands of Northern The Effect of Phytophagous Nematode Grazing on Blue Grama Die- Kenya- W.J. Lusigi, E. R. Nkurunziza. and S. Musheti ...... 542 off-Nancy L. Stonton, Dennis Morrison, and William A. Lay- Cattle Distribution on Mountain Rangeland in Northeastern Ore- cock ...... 447 gon-R. L. Gillen, W. C. Krueger, and R. F. Miller 549 Spatial and Seasonal Variability and Field Measured Infiltration . . . . . , . . . . . Dietary Selection and Nutrition of Spanish Goats as Influenced by Rates on a Rangeland Site in Utah-Moujuhed Achouri und Brush Management-Expedito A. Lopes and Jerry W. Stuth GeruldF.Gifford ...... 451 ...... 554 Vegetation of Two Relict Mesas in Zion National Park--M&hoe1 H. Muduny and Neil E. West ...... 456 Estimating Seasonal Diet Quality of Pronghorn Antelope from Fecal Effects of Temporary Dehydration on Growth of Green Needlegrass Analysis-B.H. Koerth. L.J. Krysl, B.F. Sowell, and F.C...... *...... 560 (Stipa viridulu Trin.) Seedlings-Timothy E. Fulbright, A.M. Bryant for Feeding-J. F. Wilson, and E. F. Redente ...... 462 Technique to Separate Grazing Cattle into Groups Seasonal Variation in Total Nonstructural Carbohydrate Levels in Kurn and R.J. Lorenz ...... 565 BookReviews ...... 267 Nebraska Sedge-Judith M. Steele, Raymond R. Rutliff; and 568 Gory L. Ritenour ...... 465 Index ...... I...... Reliability of Captive Deer and Cow In Vitro Digestion Values in Table of Contents .,...... *...... 574 Predicting Wild Deer Digestion Levels-Henry Campu, III, David K. Woodyurd, and Jonathan B. Huufkr ...... 468 Estimating Grazingland Yield from Commonly Available Data- Karin Wisiol ...... 471 An Improved Esophageal Fistula Bag for Sheep-D. M. Center and M.B.Jones ...... 476 BookReviews ...... 478

576 JOURNAL OF RANGE MANAGEMENT 37(6). November 1984