Published bimonthly—January, March, May, July, 537 Tiller recruitment patterns and biennial tiller production in prairie sandreed by September, November J.R. Hendrickson, L.E. Moser, and P.E. Reece Copyright 2000 by the Society for Range 544 Seed biology of rush skeletonweed in sagebrush steppe by Julia D. Liao, Stephen Management B. Monsen, Val Jo Anderson, and Nancy L. Shaw INDIVIDUALSUBSCRIPTION is by membership in the Society for Range Management. 550 Seed production in sideoats grama populations with different grazing histories by Steven E. Smith, Rebecca Mosher, and Debra Fendenheim LIBRARY or other INSTITUTIONAL SUBSCRIP- TIONS on a calendar year basis are $105.00 for the United States postpaid and $123.00 for other 556 Hoary cress reproduction in a sagebrush ecosystem by Larry Larson, Gary countries, postpaid. Payment from outside the Kiemnec, and Teresa Smergut United States should be remitted in US dollars by international money order or draft on a New York bank. Book Review 560 Old Fences, New Neighbors, by Peter R. Decker BUSINESS CORRESPONDENCE, concerning subscriptions, advertising, reprints, back issues, and related matters, should be addressed to the Managing Editor, 445 Union Blvd., Suite 230, Lakewood, Colorado 80228.

E D I TO R I A L CORRESPONDENCE, concerning manuscripts or other editorial matters, should be addressed to the Editor, Gary Frasier, 7820 Stag Hollow Road, Loveland, Colorado 80538. Page proofs should be returned to the Production Editor, 445 Union Blvd., Lakewood, Colorado 80228.

INSTRUCTIONS FOR AUTHORS appear on the inside back cover of most issues. THE JOURNAL OF RANGE MANAGEMENT (ISSN 0022-409X) is published bimonthly for $56.00 per year by the Society for Range Management, 445 Union Blvd., Ste 230, Lakewood, Colorado 80228. SECOND CLASS POSTAGE paid at Denver, Colorado and additional offi c e s . POSTMASTER: Return entire journal with address change —RE T U R N PO S TA G EG U A R A N - TEED to Society for Range Management, 445 Union Blvd., Suite 230, Lakewood, Colorado 80228.

Managing Editor Associate Editors ROBERT GILLEN FAISAL K. TAHA J. CRAIG WHITTEKIEND VIVIEN G. ALLEN USDA-ARS KISR-Aridland Agriculture 445 Union Blvd., Ste 230 Texas Tech University Southern Plains Range Res. Sta. Box 24885 Lakewood, Colorado 80228 Dept. of Plant & Soil Science 2000 18th Street 13109 Safat (303) 986-3309 Box 42122 Woodward, Oklahoma 73801 KUWAIT Fax: (303) 986-3892 Lubbock, Texas 79409-2122 e-mail address: ELAINE E. GRINGS ALLEN TORELL [email protected] YUGUANG BAI USDA-ARS New Mexico State University Editor/Copy Editor Kamloops Range Res. Sta. Fort Keogh-LARRL Agricultural Economics GARY FRASIER/JO FRASIER Agr. and Agri-Food Canada Route 1, Box 2021 Box 3169 7820 Stag Hollow Road 3015 Ord Road Miles City, Montana 59301 Las Cruces, New Mexico 88003 Loveland, Colorado 80538 Kamloops, British Columbia e-mail address: V2B 8A9 MARK JOHNSON MIMI WILLIAMS [email protected] CANADA Louisiana State University USDA-ARS Forestry and WIldlife Management Production Editor 22271 Chinsegut Hill Rd PATTYRICH DAVID BELESKY Baton Rouge, Louisiana 70803 Brooksville, Florida 34601-4672 Society for Range Management USDA-ARS 445 Union Blvd., Ste 230 1224 Airport Road PAUL OHLENBUSCH Lakewood, Colorado 80228 Beaver, West Virginia 25813 Kansas State University e-mail address: Department of Agronomy [email protected] ROBERT R. BLANK Throckmorton Hall Manhattan, Kansas 66506 Book Review Editor USDA-ARS 920 Valley Road DAVID L. SCARNECCHIA ROBERT PEARCE Dept of Natural Res. Sci. Reno, Nevada 89512 Resources Concept Washington State University 340 N. Minnesota St. Pullman, Washington 99164-6410 JOE E. BRUMMER Carson City, Nevada 89703 e-mail address: Mt. Meadows Res. Ctr. [email protected] P.O. Box 598 MICHAEL H. RALPHS Gunnison, Colorado 81230 Electronic JRM Editor USDA-ARS M.KEITH OWENS Poisonous Plant Lab Texas A&M University DAVIDGANSKOPP USDA-ARS 1150 E 1400 N Research Center Logan, Utah 84341-2881 1619 Garner Field Road HC-71 4.51 HWY 205 Ulvade, Texas Burns, Oregon 97720 e-mail address: [email protected] THE SOCIETY FOR RANGE MANAGEMENT, founded in 1948 as the American Society of Range Management, is a nonprofit association incorporated under the laws of the State of Wyoming. It is recognized exempt from Federal income tax, as a scientific and educa- tional organization, under the provisions of Section 501(c)(3) of the Internal Revenue Code, and also is classed as a public foundation as described in Section 509(a)(2) of the Code. The name of the Society was changed in 1971 by amendment of the Articles of Incorporation. The objectives for which the corporation is established are: —to properly take care of the basic rangeland resources of soil, plants, and water; President JOHN L. McLAIN —to develop an understanding of range ecosystems and of the principles applicable to the 340 N. Minnesota St. management of range resources; Carson City, Nevada 89703-4152 —to assist all who work with range resources to keep abreast of new findings and 1st Vice-President techniques in the science and art of range management; JAMES T. O’ROURKE 61 Country Club Road —to improve the effectiveness of range management to obtain from range resources the Chadron, Nebraska 69337 products and values necessary for man’s welfare; 2nd Vice-President Rodney K. Heitschmidt —to create a public appreciation of the economic and social benefits to be obtained from USDA-ARS the range environment; Ft. Keogh LARRL Rt 1, Box 2021 —to promote professional development of its members. Miles City, Montana 59301-9801 Membership in the Society for Range Management is open to anyone engaged in or interested in Executive Vice-President any aspect of the study, management, or use of rangelands. Please contact the Executive Vice- J. CRAIG WHITTEKIEND President for details. 445 Union Blvd., Ste. 230 Lakewood, Colorado 80228 (303) 986-3309 Fax: (303) 986-3892 e-mail address: [email protected] The Journal of Range Management is a publication of the Society for Range Directors Management. It serves as a forum for the presentation and discussion of facts, ideas, and philosophies 1998-2000 pertaining to the study, management, and use of rangelands and their several resources. Accordingly, all PATRICK L. SHAVER material published herein is signed and reflects the individual views of the authors and is not necessari- 2510 Meadow Lane ly an official position of the Society. Manuscripts from anyone—nonmembers as well as members—are Woodburn, Oregon 97071-3727 welcome and will be given every consideration by the editors. Editorial comments by an individual are CAROLYNHULL SIEG also welcome and, subject to acceptance by the editor, will be published as a “Viewpoint.” Forest &Range Experiment Station 501 E. Saint Joseph Street In Cooperation With: Some of the articles appearing in The Journal of Range Management (JRM) School of Mines Campus are presented in cooperation with The American Forage and Grassland Council (AFGC). This coopera- Rapid City, South Dakota 57701-3901 tion consists of JRM acceptance of professional papers in forage grazing management and related sub- 1999-2001 ject areas from AFGC members and the appointment of two AFGC affiliated associate editors JAMES LINEBAUGH to JRM’s Editorial Staff. The American Forage and Grassland Council Offices: P.O. Box 94, 3 Yhvona Dr. Georgetown, Texas 78627; Larry Jeffries, President; Dana Tucker, Executive Secretary. Carson City, NV89706-7717 GLENSECRIST Idaho Dept. of Agriculture 3818 S. Varian Ave. Boise, Idaho 83709-4703

2000-2002 RICHARDH. Hart Contribution Policy: The Society for Range Management may accept donations of real USDA-ARS and/or personal property subject to limitations set forth by State and Federal law. All donations shall be High Plains Grasslands Station 8408 Hildreth Rd. subject to management by the Executive Vice President as directed by the Board of Directors and their Cheyenne, Wyoming 82009-8809 discretion in establishing and maintaining trusts, memorials, scholarships, or other types of funds. Individual endowments for designated purposes can be established according to Society policies. Gifts, DONKIRBY bequests, legacies, devises, or donations not intended for establishing designated endowments will be North Dakota State University deposited into the SRM Endowment Fund. Donations or requests for further information on Society poli- Animal &Range Science cies can be directed to the Society for Range Management, Executive Vice-President, 445 Union Blvd., Fargo, North Dakota 58105 Lakewood, Colorado 80228. We recommend that donors consult Tax Advisors in regard to any tax con- The term of office of all elected officers and sideration that may result from any donation. directors begins in February of each year during the Society’s annual meeting. J. Range Manage. 53: 466–470 September 2000 Grazing impacts on selected soil parameters under short- term forage sequences

E. MAPFUMO, D.S. CHANASYK, V.S. BARON, AND M.A. NAETH

Post-doctoral fellow and professor with the Department of Renewable Resources, University of Alberta, Alberta, Canada T6G 2H1, and research scientist with Lacombe Research Centre, Agriculture and Agri-Food Canada, 6000 C&E Trail, Lacombe, Alberta, Canada T4L 1W1, and professor, Dept. Renewable Resources University of Alberta, Alberta, Canada T6G 2H1. Corresponding author, Chanasyk, (Email address: [email protected]; Fax no (780) 492-4323.

Abstract Resumen

Long-term cultivation is known to change soil physical and Se sabe que el cultivar los terrenos por un largo tiempo cambia chemical properties, but little is known about whether short- la propiedades físicas y químicas del suelo, sin embargo pero term agricultural practices, such as rotational grazing, can initi- poco se sabe acerca de que si la practicas agrícolas de corto ate such changes. This study investigated the impacts of 3 graz- plazo, tales como el apacentamiento rotacional, puede iniciar ing intensities (heavy, medium, and light) and 4 forages on select- dichos cambios. En este estudio se investigo el impacto de 3 ed soil physical and chemical parameters of a Typic Haplustoll at intensidades de apacentamiento (alta, media y ligera) y 4 especies Lacombe, Alberta. Measurements were conducted on soil sam- forrajeras en los parámetros físico-químicos de un suelo "Typic ples collected at the beginning (1993) and the end (1996) of the Haplustoll" en Lancombe, Alberta. La mediciones se realizaron study. Two perennial forages, smooth bromegrass (Bromus iner - en muestras de suelo colectadas al inicio (1993) y al final (1996) mis cv. ‘Carlton’) and meadow bromegrass (Bromus riparius cv. del estudio. En este experimento se utilizaron dos especies ‘Paddock’), and 2 annuals, a mixture of triticale (X Triticosecale perennes, "Smooth bromegrass" (Bromus inermis cv. ‘Carlton’) Wittmack cv. ‘Pika’) and barley (Hordeum vulgare L. cv. ‘AC y "Meadow bromegrass" (Bromus riparius cv. ‘Paddock’) y dos Lacombe’) and triticale alone were used for the study. Grazing especies anuales, una mezcla de triticale (X T r i t i c o s e c a l e intensity or forage species did not affect carbon-to-nitrogen ratio. Wittmack cv. ‘Pika’) y cebada (Hordeoum vulgare L. cv. ‘AC Grazing intensity influenced changes in available water holding Lamcombe’) y triticale solo. Ni la intensidad de apacentamiento capacity for the 0–5 cm interval, soil nitrogen for the 30–45 cm ni las especies vegetales afectaron la relación carbón-nitrógeno. interval, soil pH for the 5–15 cm interval and electrical conduc- La intensidad de apacentamiento indujo cambios en la capacidad tivity for all depth intervals except for the 0–5 cm interval (P de retención de agua disponible del estrato de 0-5 cm, en el 0.05). Forage species affected changes in soil carbon in the 0-5 cm nitrógeno del suelo del estrato 30–45 cm , en el pH a la profundi- interval, soil pH between 0 and 15 cm, and electrical conductivity dad de 0–15 cm y en la conductividad eléctrica de todos los between 5 and 45 cm (P 0.05). Soil electrical conductivities for estratos, excepto el de 0-5 cm (P#0.05). Las especies vegetales all grazing levels and forage treatments were within the range afectaron el carbón del suelo a la profundidad de 0–5 cm, el pH (i.e. 0–2 dS m- 1) considered to have negligible effects on plant del estrato 0–15 cm y la conductividad eléctrica del estrato de 5– growth. The minimal effects of grazing and plant species on soil 45 cm (P#0.05). Las medidas de conductividad eléctrica del suelo parameters in this study may have been due to the resilient para todas las intensidades del apacentamiento y especies vege- intrinsic properties of the soil and/or the short study length. tales estuvieron dentro del rango (esto es 0–2 dS m-1) que se con- sidera no detrimental para el crecimiento de las plantas. El efec- to mínimo de la intensidad de apacentamiento y especies vege- Key Words: Carbon, electrical conductivity, nitrogen, potas- tales en los parámetros del suelo medidos en este estudio se pudo sium, soil pH deber a las propiedades intrínsecas del suelo para soportar situa- ciones de disturbio o a que el período de tiempo que comprendió el estudio fue muy corto. Cultivation accelerates mineralization of organic matter (Broersma 1991) and may consequently alter water retention properties, and carbon (C) and nitrogen (N) content of the soil. The combined effect of dung and urine depositions with cattle Yearly cultivation in annual cropping systems may pulverize soil grazing may alter soil acidity (Johnston et al. 1971) and salinity particles and thus weaken soil structure and render the soil more (Chaneton and Lavado 1996), and may also increase the loss of vulnerable to erosion. Although long-term cultivation is known to nitrogen via volatilization (Holland and Detling 1990). change many soil physical and chemical properties (Johnston et Furthermore, heavy grazing is known to increase surface soil al. 1971) as little as 5 years of cropping can initiate such changes bulk density (Krenzer et al. 1989, Mapfumo et al. 1999) which (Dormaar et al. 1989, Dormaar and Willms 1990). may in turn hinder root growth. Perennial forages reduce soil erosion due to the maintenance of Funding from the Canada-Alberta Environmentally Sustainable Agriculture Agreement protective cover, improved soil structure and aggregate stability (CAESA) is gratefully acknowledged. Thanks are extended to David Young, Kelly and increased litter on the soil surface. Indeed, most studies of the Ostermann, Pola Genoway, Kirsten Gregorwich and Mae Elsinger for field assistance. grazing impacts on soil physical and chemical properties have Manuscript accepted 8 Dec. 1999. been conducted on perennial grasses. In many parts of the aspen

466 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 parkland of Alberta, however, pasturing a depth of 10 cm, and seeded to triticale at auger to a depth of 60 cm. Sampling at the annual crops by swath grazing is being 135 kg ha-1 in 2 passes, or triticale at 90 kg end of the study was conducted in October used to extend the grazing season ha-1 and barley 50 kg ha-1 in separate oper- 1996 to the same depth using a tractor- (Lagroix-McLean and Naeth 1997), with ations. Plant residue remaining after the mounted hydraulic auger. For both sam- annuals such as winter triticale and barley last grazing of the previous season was left plings, three samples per plot were taken becoming common pasture species (Baron in place until spring (end of April) seed- and separated into increments of 0–5, et al. 1993). However, little is known ing. The herbicide, MCPA (2-methyl-4- 5–15, 15–30, 30–45, and 45–60 cm. All about the direct impacts of grazing of chlorophenoxyacetic acid), was applied to samples were air-dried and ground to pass annual forages on soil parameters. annual plots after the crop emerged at a 2-mm sieve. Before determination of Our working hypothesis considered that rates of 600 g a.i. ha-1 in 1994 to 1996. No physical and chemical properties, the 3 grazing differentially impacts particular soil weed control was used in perennial plots. samples from the same depth interval and parameters and the extent of impact is relat- within each plot were mixed to form a composite sample. ed to the intensity of grazing and the type of Grazing Systems forage being grazed. The objective of this Beginning in 1994, up to 6 animals were study was to investigate the effects of heavy, placed on a treatment at 1 time, depending Soil Physical and Chemical Properties medium and light intensity grazing of annual on the intensity of grazing desired. Each Water retention properties were deter- and perennial forages on soil parameters grazing event was less than 24 hours. mined on these disturbed soils using pres- over a relatively short period of time. Forage height was used to define grazing sure plates (Topp et al. 1993). Field capac- intensity. Forage heights were determined ity was the gravimetric moisture content using a disk meter (Bransby et al. 1977) retained at a pressure of 0.033 MPa and Materials and Methods and the average of 10 disk heights was wilting point was moisture content calculated. Averaged over 3 grazing years, retained at 1.5 MPa. Available water hold- Experimental Design the number of grazing events per season ing capacity (AWHC) was the difference The study was conducted at the for perennials were 7, 5, and 3 for heavy, between field capacity moisture and wilt- Lacombe Research Station, Alberta, medium and light grazing, respectively, ing point moisture. Canada on a Typic Haplustoll derived from whereas comparable figures for annuals Electrical conductivity (EC) was deter- glacio-lacustrine parent material. On aver- were 4, 4, and 2 for heavy, medium and mined using the saturation extract method age the soil (0–15-cm depth) contained light grazing, respectively (Table 1). (Janzen 1993). Soil pH was determined 15% clay, 34% silt, and 51% sand. The Averaged over 3 years, mean cow-days using distilled water in a soil: water ratio experimental design was a randomized per season for perennials was 37, 20, and of 1:2. Total C in soil was determined complete block with 3 grazing treatments, 16 and for annuals 19, 11, and 8 for heavy, using the Leco carbon analyzer. Initial soil 4 plant species and 4 replicated blocks, for medium and light grazing, respectively. tests did not indicate the presence of free a total of 48 plots. Plots, 33 ´ 9 m, were Mean residue levels for perennials aver- carbonates. Soil samples for determina- aged over 3 years were 0.89, 1.71, and established in 1992 by tilling a 15-year old tions of total N were digested using the 2.57 Mg ha- 1 and for annuals 0.91, 1.22, growth pasture which had been under and 2.23 Mg ha- 1 for heavy, medium and standard Kjeldhal method with concentrat- extensively managed grazing. The old- light grazing, respectively. ed sulfuric acid (McGill and Figuerido growth pasture was composed of smooth 1993). Total N concentration was then bromegrass (Bromus inermis L.), Kentucky determined using the autoanalyzer bluegrass (Poa pratensis L.), quackgrass Soil Sampling (Technicon Autoanalyzer II). Extractable (Agropyron repens L), bluegrass (P o a Soil sampling was conducted prior to K using extractant solution of 1N spp.) with common occurrences of dande- grazing in September 1993 using a hand N H4OAC adjusted to pH 7 with NH4O H lion (Taraxacum officinale L.).

Table 1. Number of grazings, cow-days for each grazing season and the total number of cow-days Plant Species for each grazing intensity and plant species for the entire study period. The 4 forage species used to seed plots included ‘Carlton’ smooth bromegrass Grazing Number of Grazings Cow-days Total (Bromis inermis), ‘Paddock’ meadow Level 1994 1995 1996 1994 1995 1996 Cow-days bromegrass (Bromus riparius R e h m . ) , Smooth bromegrass ‘Pika’ triticale (X Triticosecale Wittmack), Heavy 7 7 6 34.8 45.0 32.5 112.5 or a mixture of ‘Pika’ triticale/‘AC Medium 6 5 4 21.7 21.7 16.3 59.7 Lacombe’ barley (Hordeum vulgare L . ) . Light 3 3 3 15.5 18.5 15.5 49.5 Perennials (i.e. smooth bromegrass and Meadow bromegrass meadow bromegrass) were seeded on 31 Heavy 8 6 5 43.3 36.8 27.0 107.1 May 1993. Smooth bromegrass was seed- -1 Medium 6 5 4 23.8 19.3 15.5 58.6 ed at 11.2 kg ha and meadow bromegrass Light 3 3 3 15.0 15.3 13.5 43.8 at 16.8 kg ha-1. Before seeding the experi- mental area received a broadcast applica- Barley/Triticale tion of 8, 14, 26, and 5 kg ha-1 of N, P, K, Heavy 5 3 3 23.3 16.5 15.8 55.6 Medium 4 4 3 11.3 11.3 9.8 32.4 and S, respectively. In subsequent years, Light 2 2 2 6.5 9.8 5.8 22.1 perennial and annual plots were fertilized at the same time. Each spring 100, 22, and Triticale Heavy 5 4 3 25.5 18.8 12.5 56.8 41 kg ha- 1 of N, P, and K, respectively Medium 4 4 3 13.5 10.0 8.5 32.0 were broadcast over the experimental area. Light 2 2 2 9.8 9.3 6.3 25.4 Annual plots were rototilled each spring to

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 467 or acetic acid (Hendershot et al. 1993) was Table 2. Change in available water holding -1 - 1 Table 3. Change in nitrogen content (g 100 g ) determined using atomic absorption spec- capacity (g 100 g ) after 3 years relative to after 3 years relative to corresponding pre- trophotometry. corresponding pre-grazing values under dif- grazing values under different grazing inten- ferent grazing intensities at different soil sities at different soil depths. Values pooled depths. Values pooled over plant species. over plant species. Statistical Analysis To test for significant (P £ 0.05) Depth Interval Grazing intensity Depth Interval Grazing intensity changes in soil parameters, pre- versus Heavy Medium Light Heavy Medium Light -1 post-treatment, a paired t-test was used for (cm) ------(g 100 g ) ------(cm) ------(g 100 g-1) ------each combination of grazing intensity, - plant species and soil depth. For individual 0–5 –1.25 *b 1.26 a 1.86 *a 0–5 0.043 *a 0.035 a 0.043 *a 5–15 0.12 a –1.64 *a –0.15 a 5–15 0.046 *a 0.068 *a 0.051 *a depths, we tested for significant effects of 15–30 -0.018 b 0.049 *a 0.003 ab grazing intensity and plant species on soil 15–30 2.37 *a 0.28 a 0.25 a 30–45 0.49 a 0.03 a 0.61 a 30–45 -0.011 a 0.001 a –0.026 a parameters with a 2-way ANOVA using 45–60 –0.57 a –0.13 a –0.33 a 45–60 -0.004 a 0.001 a 0.003 a the SAS institute generalized linear mod- Average 0.23 –0.04 0.45 Average 0.011 0.031 0.015 els procedure (SAS Institute 1989). Means followed by an asterisk are significantly (P £ 0.05) Means followed by an asterisk are significantly (P £ 0.05) Change (post-treatment–pre-treatment) in different than their corresponding pre-grazing values different than their corresponding pre-grazing values (t- available water holding capacity, organic (t–test). Row means followed by different small letters test). Row means followed by different small letters indi- indicate significant differences among grazing intensities cate significant differences among grazing intensities carbon, total soil nitrogen, extractable (LSD test). (LSD test). potassium, soil pH, and electrical conduc- tivity were dependent variables. Where the compaction observed in a concurrent study Extractable K content ranged between F-test indicated a significant (P £ 0.05) (Mapfumo et al. 1999) may have reduced 73 and 650 ppm for the depths between 0 effect, means were separated using the macro-porosity and thus AWHC at low and 30 cm, whereas for intervals between least significance difference (LSD) test moisture tension. 30 and 60 cm, K content ranged between (Gomez and Gomez 1984). 68 and 97 ppm. Change in extractable K content was significantly affected by the Chemical and Nutrient Properties interaction between grazing and species Results Total soil nitrogen content in all depths only for the 30–45 cm interval (Table 4). and treatments ranged between 0.06 and - 1 For heavy grazing, meadow bromegrass 0.50 g 100 g of soil. Species did not had the largest change in K content com- Water Retention affect soil N content. Generally grazing For all species and grazing intensities, pared to other forages. For the medium altered soil N content between 0 and 30 grazing treatment, change in extractable K water contents at field capacity (FC) for cm (P £ 0.05) (Table 3). Averaged over all content was similar among all species. soil depths between 0 and 30 cm ranged depths the change in N content was posi- - 1 However, the light grazing barley/triticale between 21 and 31 g 100 g of soil, tive in all 3 grazing treatments. whereas water contents at wilting point -1 (WP) ranged between 11 and 18 g 100 g Table 4. Change in potassium content (mg kg-1 ) after 3 years relative to corresponding pre-grazing of soil. For depth intervals between 30 and values for different grazing intensities, plant species and soil depths. 60 cm, FC ranged between 15 and 22 g 100 g - 1 of soil, whereas WP ranged Depth Interval Grazing intensity between 7 and 10 g 100 g-1 of soil. Change Depth Heavy Medium Light in available water holding capacity (cm) ------(g 100 g-1) ------(AWHC) after 3 years of grazing was not Smooth bromegrass 0–5 63 aA 18 aA 139 aA significantly different among species. For Meadow bromegrass 64 *aA 15 aA –22 aA the 0–5 cm depth interval the heavy graz- Barley/triticale 121 aA – 6 aA 100 aA ing intensity had a significantly lower Triticale 40 aA 25 aA 12 aA AWHC after 3 years of grazing (Table 2). Smooth bromegrass 5–15 cm 30 aA 119 aA 32 aA Change in AWHC for medium and light Meadow bromegrass 57 aA 10 aA 41 aA Barley/triticale 35 aA 28 aA 85 *aA grazing in the 0–5 cm interval was posi- Triticale 7 aA 71 *aA 11 aA tive and significantly greater than that for Smooth bromegrass 15–30 cm –17 aA 4 aA 25 aA the heavy grazing. Water retention in light Meadow bromegrass 30 aA –12 aA –11 aA and medium grazing may have increased Barley/triticale –3 aA –2 aA 20 aA relative to heavy grazing, possibly because Triticale 14 aA 4 aA –4 aA the former treatments had greater organic Smooth bromegrass 30–45 cm –7 aB 6 aA 3 aB matter input and incorporation into the soil Meadow bromegrass 22 aA 0 bA 0 bB compared to the latter. In fact results from Barley/triticale 0 bB 1 bA 25 aA a concurrent study indicate that the aver- Triticale 3 aB 0 aA 2 aB age litter (all dead plant material) mass Smooth bromegrass 45–60 cm 1 aA 8 aA 4 *aA measured in the fall over 3 years of graz- Meadow bromegrass 0 aA –2 aA 15 aA ing were 2.8, 3.4, and 5.4 Mg ha- 1 f o r Barley/triticale 5 aA 5 aA 6 aA heavy, medium, and light grazing, respec- Triticale 9 aA 2 aA 4 aA tively (Baron et al. 1999). Litter masses Means followed by an asterisk are significantly (P £ 0.05) different than their corresponding pre–grazing values (t–test). Row means followed by different lower case letters indicate significant difference among grazing intensities (LSmeans, measured in spring for the respective graz- P £ 0.05) within species. Column means followed by upper case letters indicate significant difference among species ing treatments were 1.6, 2.5, and 4.2 Mg (LSmeans, P £ 0.05) within grazing intensity. h a- 1. It is also possible that surface soil

468 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Table 5. Change in carbon content (g 100 g-1 ) after 3 years relative to corresponding pre-grazing val- Discussion ues under different plant species at different soil depths. Values pooled over grazing intensities. The small differences in water retention Depth Smooth Meadow Barley/ Triticale interval bromegrass bromegrass Triticale among grazing treatments generally reflect -1 the minimal impacts of grazing, regardless (cm) ------(g 100 g ) ------of grazing intensity, on soil parameters 0–5 0.12 a 0.08 a –0.29 *b –0.39 *b 5–15 0.21 a –0.12 a 0.12 a –0.04 a evaluated in this study. Results from a 15–30 0.22 a –0.36 a –0.09 a –0.31 a concurrent study also indicate minimal 30–45 –0.10 a –0.48 *a –0.42 *a –0.29 a grazing effects on soil bulk density, rela- 45–60 –0.05 a –0.17 a 0.00 a –0.09 a tive compaction (Mapfumo et al. 1999) Average 0.08 –0.21 –0.14 –0.22 and water-stable aggregate size distribu- Means followed by an asterisk are significantly (P £ 0.05) different than their corresponding pre–grazing values tion (Mapfumo 1997). (t–test). Row means followed by different small letters indicate significant differences among grazing intensities (LSD The carbon:nitrogen ratio reflects the test). potential for soils to mineralize and immo- bilize N. Generally a C:N ratio less than treatment had the largest change in K con- increase in EC for heavy grazing was five 20:1 indicates a potential for net mineral- tent compared to other treatments. times greater than that for medium and ization whereas C:N ratio greater than Soil carbon content generally ranged light grazing. -1 30:1 indicates potential immobilization between 3 and 6 g 100 g of soil for depth Soil pH ranged between 5.1 and 6.4 for (Tisdale et al. 1993). In our study, for all intervals between 0 and 30 cm, and from depth intervals between 0 and 30 cm, and - 1 grazing levels, plant species and years, <1 to 2 g 100 g of soil for intervals ranged between 6.6 and 7.0 for depth C:N ratios were less than 20:1, suggesting between 30 and 60 cm. Change in soil car- intervals between 30 and 60 cm. The that N release through mineralization like- bon (C) content after 3 years of grazing change in pH after 3 years of grazing was ly occurred. This process results in was not significantly affected by grazing significantly different among forage increased mineral nitrogen (NH -N and intensity, but was affected by forage species for the 0–5 and 5–15 cm intervals 4 N O3-N) which may contribute to species only for the 0–5 cm depth interval (Table 7). Under both smooth bromegrass increased solute concentration and hence (Table 5). Soil carbon content under peren- and meadow bromegrass changes in pH electrical conductivity (EC) of the soil nials slightly increased after 3years of were greater than that for barley/triticale, solution. grazing whereas under annuals soil carbon but were not different from soil pH change Extractable K includes soil solution K as content decreased. These differences may under triticale. Grazing intensity affected well as K that is adsorbed onto exchange be attributed to a combination of opposing soil pH change for the 5–15 cm depth sites on the clay surfaces (exchangeable factors; yearly cultivation of annuals which interval. Under heavy grazing soil pH K). It is difficult to ascertain why grazing likely promoted organic matter decomposi- change was significantly greater (and neg- and forage species had an interactive tion, and accumulation of litter under ative) than that for medium and light graz- effect only for the 30–45 cm interval. perennials which likely increased soil car- ing. Overall soil pH change for all depths, However, processes such as leaching, fixa- bon. In perennials, annual cultivation was grazing intensities and forage species, did tion by micaceous clays, weathering of not conducted such that rate of litter incor- not exceed 0.2 units; such a change feldspars and crop uptake occur concur- poration into the soil carbon pool was most unlikely influenced plant growth and soil rently and are known to affect exchange- likely greater than the rate of organic mat- chemical and microbial processes. able K (Tisdale et al. 1993). Changes in ter breakdown. In spite of the differences in change in carbon content among Table 6. Change in electrical conductivity (dS m-1) after 3 years relative to corresponding pre-graz- species, grazing intensity or forage species ing values under different plant species (pooled over grazing intensities) and under different did not affect the change in carbon-to- grazing intensities (pooled over different plant species) at different soil depths. nitrogen (C:N) ratio. For all grazing inten- sities, forage species and soil depths, C:N Depth Smooth Meadow Barley/ Triticale ratios ranged between 9:1 to 18:1. interval bromegrass bromegrass Triticale Electrical conductivities (EC) ranged (cm) ------(dS m-1)------between 0.16 and 0.75 dS m- 1 for depth 0–5 –0.09 a –0.27 *a –0.25 *b –0.08 *a intervals between 0 and 30 cm and 5–15 0.22 *ab 0.11 *b 0.27 *a 0.15 *b between 0.12 and 0.30 dS m- 1 for depths 15–30 0.20 *ab 0.11 *b 0.27 *a 0.18 *ab 30–45 0.19 *ab 0.11 *b 0.23 *a 0.15 *b between 30 and 60 cm. Change in EC was 45–60 0.10 *a 0.06 *a 0.14 *a 0.09 *a significantly influenced by forage species Average 0.12 0.02 0.13 010 for the 5–15, 15–30, and 30–45 cm inter- vals. Barley/triticale had a greater increase Depth Grazing intensity in electrical conductivity compared to interval Heavy Medium Light meadow bromegrass (Table 6). However (cm) ------(dS m-1)------smooth bromegrass had a similar change 0–5 –0.03 *a –0.25 *a –0.23 *a in EC to meadow bromegrass, barley/triti- 5–15 0.36 *a 0.12 *b 0.08 *b cale and triticale. Grazing intensity also 15–30 0.29 *a 0.17 *b 0.11 *b 30–45 0.24 *a 0.12 *b 0.14 *b influenced change in EC for all depth 45–60 0.16 *a 0.06 *b 0.07 *b intervals between 15 and 60 cm. The Average 0.20 0.04 0.03 heavy grazing treatment had a significant- Means followed by an asterisk are significantly (P £ 0.05) different than their corresponding pre–grazing values ly greater increase in EC than medium or (t–test). Row means followed by different small letters indicate significant differences among plant species and among light grazing. Averaged over all depths the grazing intensities (LSD test).

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 469 Table 7. Change in soil pH after 3 years relative to corresponding pre-grazing values under differ- Broersma, K. 1991. The effects of different crop- ent plant species (pooled over grazing intensities) and under different grazing intensities (pooled ping systems on a luvisolic soil in the Peace over different plant species) at different soil depths. River region. Ph.D. Diss., Univ. Alberta, Edmonton, Canada. Depth Smooth Meadow Barley/ Triticale Chaneton, E.J. and R.S. Lavado. 1996. S o i l interval bromegrass bromegrass Triticale nutrients and salinity after long-term grazing exclusion in a Flooding Pampa grassland. J. (cm) Range Manage. 49:182–187. 0–5 0.16 *a 0.12 *a –0.06 *b 0.04 *ab Dormaar, J.F. and W.D. Willms. 1990. Effect of 5–15 0.04 *a 0.03 *a –0.18 *b 0.04 *a grazing and cultivation on some chemical prop- 15–30 0.18 a 0.20 a 0.13 a 0.18 a erties of soils in the mixed prairie. J. Range 30–45 0.04 *a 0.11 *a –0.01 *a 0.04 *a Manage. 43:456–460. 45–60 –0.06 *a 0.00 *a –0.23 *b –0.12 *ab Dormaar, J.F., S. Smoliak, and W.D. Willms. Average 0.07 0.09 –0.07 0.04 1 9 8 9 . Vegetation and soil responses to short- duration grazing on fescue grasslands. J. Range Depth Grazing intensity Manage. 42:252–256. interval Heavy Medium Light Gomez, K.A. and A.A. Gomez. 1984. Statistical (cm) procedures for agricultural research. John Wiley 0–5 0.08 *a 0.02 *a 0.11 *a and Sons, New York, N.Y. 5–15 –0.17 *b 0.05 *a 0.07 *a Hendershot, W.H., H. Lalande, and M. 15–30 0.13 a 0.18 a 0.21 a Duquette. 1993. Ion exchange and exchange- 30–45 –0.04 *a 0.07 *a 0.10 *a able cations, p. 167–176. In: M.R. Carter (ed.), 45–60 –0.15 *a –0.04 *a –0.12 *a Soil sampling and methods of analysis. Lewis Average –0.03 0.06 0.07 Publishers, Boca Raton, Fla. Holland, E.A. and J.K. Detling. 1990. P l a n t Means followed by an asterisk are significantly (P £ 0.05) different than their corresponding pre-grazing values (t-test). response to herbivory and belowground nutrient Row means followed by different small letters indicate significant differences among plant species and among grazing intensities (LSD test). cycling. Ecol. 71:1040–1049. Janzen, H.H. 1993. Soluble salts, p 161–166. In: M.R. Carter (ed.), Soil sampling and methods of soil macroporosity and rooting patterns Electrical conductivity in all treatments analysis. Lewis Publishers, Boca Raton, Fla. -1 might influence preferential flow paths was between 0 and 2 dS m , which is the Johnston, A., J.F. Dormaar, and S. Smoliak. and fertilizer distribution. Annuals might range within which salinity has negligible 1 9 7 1 . Long-term grazing effects on fescue uptake more K before it leaches very far. effects on plant growth (Bernstein 1975). grassland soils. J. Range Manage. 24:185–188. Changes in soil pH among grazing treat- The greater increase in EC under heavy Krenzer Jr., E.G., C.F. Chee, and J.F. Stone ments within species were small (£ 0.20 compared to medium and light grazing 1 9 8 9 . Effects of animal traffic on soil com- paction in wheat pastures. J. Prod. Agr. units), and soil depth seemed to have more treatments for depth intervals between 15 2:246–249. influence on pH than either plant species and 60 cm, may have been due to a combi- Lagroix-McLean, R.L. and M.A. Naeth. 1997. or grazing level. However, the decrease in nation of yearly application of fertilizer, Forages in the aspen parkland, A literature and soil pH after 3 years of heavy grazing increased urine and dung loading rates as data review. Final Report Prepared for Ducks obtained for the 5–15 cm interval may be well as organic-N mineralization to pro- Unlimited Canada, North Amer. Waterfowl due to increased NH -N from dung and duced mineral N. These processes are Manage. Plan and Conservation and Dev. 4 Branch. Dep. of Renewable Resources, Univ. urine depositions, which through nitrifica- known to increase total solutes in soil Alberta, Edmonton, Canada. tion produced NO3-N plus hydrogen ions. solution which is directly related to elec- Mapfumo, E. 1997. Soil and plant response to In fact results from a concurrent study trical conductivity. compaction. Ph.D. Diss., Univ. Alberta, indicate that both mineral N and NO3- N In general, grazing intensity nor the type Edmonton, Canada. levels were greater under heavy compared of forage grazed has insignificant effects Mapfumo, E., D.S. Chanasyk, M.A. Naeth, and to medium and light grazing (Baron et al. on the soil parameters measured after three V.S. Baron. 1999. Soil compaction under graz- ing of annual and perennial forages. Can. J. Soil unpublished data). Estimates of urine years of grazing. These early results may Sci. 79:191–199. loading rates from that study were 18,615, be refined after comparisons are made fol- McGill, W.B. and C.T. Figuerido. 1993. T o t a l 10,075 and 7,919 L h- 1 y r - 1 for heavy lowing 10 to 20 years of treatments. nitrogen, p. 201–212. In: M.R. Carter (ed.), Soil medium and light grazing, respectively sampling and methods of analysis. Lewis with dung loading estimates for the Publishers, Boca Raton, Fla. respective grazing intensities of 1489, 806 References SAS Institute 1989. SAS/STAT user’s guide. -1 -1 SAS Inst., Cary, N.C. and 634 kg DM ha yr . The lack of soil Tisdale, S.L., W.L. Nelson, J.D. Beaton, and pH differences among grazing intensities Baron, V.S., E. Mapfumo, M.A. Naeth, and J.L. Havlin. 1993. Soil fertility and fertilizers. for the surface (0–5 cm) interval can be D.S. Chanasyk. 1999. Sustainable grazing sys- McMillan Publ. Co., New York. tems for perennial and annual forages on sloped partly attributed to NH3-N loss from fresh Topp, G.C., Y.T. Galganov, B.C. Ball, and lands. Canada-Alberta Environmentally M.R. Carter. 1993. Soil water desorption cattle wastes through volatilization which Sustainable Agr. Agreement Final Rep. 161 pp. some researchers have reported to be more curves, p. 569–580. In: M.R. Carter (ed.), Soil Baron, V.S., H.G. Nadja, D.F. Salmon, J.R. sampling and methods of analysis, Lewis Publ., than 50% only three days after deposition Pearen, and A.C Dick. 1993. Cropping sys- Boca Raton, Fla. (Westerman et al. 1985). Despite the tems for spring and winter cereals under simu- Westerman, P.W., L.M. Safley Jr., J.C. Barker, observed changes after 3 years of grazing, lated pasture: sward structure. Can. J. Plant Sci. and G.M. Chescheir. 1985. Available nutrients the soil pH remained within the non- 73:947–959. in livestock waste. Agr. Waste Utiliz. and Bernstein, L. 1975. Effects of salinity and sodici- restrictive range for growth of annual or Manage. Proc. Fifth Int. Symp. Agr. Wastes, ty on plant growth. Ann. Rev. Phytopathol. Dec. 16–17, Chicago, Ill. perennial grasses. Overall, if we assume a 13:295–312. constant rate of change of pH over time it Bransby, D.I., A.G. Matches, and G.F. Krause. would take at least 15 years to effect a unit 1 9 7 7 . Disk meter for rapid estimation of change in soil pH. herbage yield in grazing trials. Agron. J. 69:393–396.

470 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 J. Range Manage. 53: 471–478 September 2000 Vegetation response to stocking rate in southern mixed- grass prairie

ROBERT L. GILLEN, JOHN A. ECKROAT, AND F. TED MCCOLLUM III

Authors are rangeland scientist, USDA-ARS, Woodward, Okla. 73801; graduate student, Department of Plant and Soil Sciences, Oklahoma State University, Stillwater, Okla. 74078-6028; and beef cattle extension specialist, Texas A&M University, Amarillo, Tex. 79106. At the time the research was conducted Gillen was professor, Department of Plant and Soil Sciences, and McCollum was professor, Department of Animal Science, Oklahoma State University, Stillwater, Okla.

Abstract Resumen

Stocking rate directly influences the frequency and intensity of La carga animal influye directamente en la frecuencia e inten- defoliation of individual plants which, in turn, impacts energy sidad de defoliación de las plantas individuales, lo cual a su vez, flow and plant succession in grazed ecosystems. The objective of impacta en el flujo de energía y la sucesión vegetal de los ecosis- this study was to determine the effect of stocking rate on stand- temas bajo apacentamiento. El objetivo de este estudio fue deter- ing crop dynamics and plant species composition of a southern minar, durante un período de 7 años (1990 a 1996), el efecto de mixed-grass prairie over a 7-year period (1990 through 1996). la carga animal en la dinámica de la producción de forraje en pie Long-term (30-year) mean precipitation has been 766 mm per y la composición botánica de un pradera de zacates mixtos del year. Growing conditions were generally favorable for the study sudeste. La media de precipitación a largo plazo (30 años) ha period. Yearling cattle (initial weight 216 kg, SD = 12 kg) grazed sido de 766 mm por año. Durante el período de estudio, las at 6 stocking rates, ranging from 23 to 51 AUD ha- 1, from 14 condiciones de crecimiento para las plantas generalmente fueron April to 24 September (162 days). The currently suggested year- favorables. Ganado de año (peso inicial 216 kg, SD = 12 kg) long stocking rate is 25 AUD ha- 1. Herbage standing crop was apacentó bajo seis cargas animal que variaron de 23 a 51 UAD measured in July and September every year while species com- h a- 1, el período se apacentamiento fue del 14 de Abril al 24 de position was determined in July in even years. Total and dead Septiembre (162 días). La carga animal que actualmente se standing crop declined as stocking rate increased but live stand- recomienda es de 25 UAD ha-1. La producción de forraje en pie ing crop was not related to stocking rate. Slopes of regression se midió en Julio y Septiembre de cada año, mientras que la lines relating standing crop and stocking rate were constant over composición botánica se determinó cada dos años en Julio. La years, indicating no response for plant productivity. The major cosecha total de forraje en pie y la cantidad de forraje muerto vegetation components, sideoats grama [Bouteloua curtipendula disminuyó conforme la carga animal aumento, pero el forraje (Mich.) Torr.], shortgrasses, and forbs were not affected by vivo en pie no se relaciono con la carga animal. Las pendientes stocking rate over years. Tallgrasses responded by increasing at de las líneas de regresión que relacionan la cosecha en pie y la the lower stocking rates over the study period. However, these carga animal fueron constantes a través de los años, indicando grasses contributed less than 5% of the total standing crop. Red una falta de respuesta en la productividad de las plantas. Los and purple threeawn (Aristida longiseta Steud. and A. purpurea componentes principales de la vegetación, "Sideoat grama" Nutt.) increased at all stocking rates from 1990 to 1996 but the [Bouteloua curtipendula (Mich.) Torr.], zacates cortos y hierbas increase was greater at the lower stocking rates. This mixed- no fueron afectados por la carga animal a través de los años. grass vegetation showed little response to stocking rate over the Durante el período de estudio, los zacates altos respondieron 7-year study period. The vegetation may have been in equilibri- incrementandose en las cargas animal bajas. Sin embargo, estos um with previous heavy stocking rates so that little change would zacates contribuyeron con menos del 5% del total de forraje en be expected at those rates. Increases in grazing sensitive species pie. Los zacates "Red threeawn"y "Purple threeawn" (A r i s t i d a at lighter stocking rates may occur over longer time intervals. longiseta Steud. y A. Purpurea Nutt.) se incrementaron en todas las cargas animal evaluadas, pero el aumento fue mayor en las cargas animal bajas. La vegetación de zacates- mixtos mostró Key Words: standing crop, plant succession, grazing impacts, poca respuesta a la carga animal durante el período de estudio Bouteloua, Aristida de 7 años. La vegetación pudo haber estado en equilibrio con cargas animal altas previas al periodo de estudio de tal manera que se esperarían solo cambios menores con las cargas animal Vegetation dynamics on Great Plains grasslands are a function altas. El aumento de especies sensitivas a la carga animal que se of grazing, fire, climate, and soils. Of these factors, managers can registró con cargas animal bajas puede ocurrir en períodos de tiempo mas largos. The authors thank Charles Worthington, Brock Karges, and John Weir for livestock management; Kenneth Tate, Justin Derner, Lance Vermeire, Mike Lohmann, and Randy Donges for assistance with vegetation sampling; and D. M. Engle, S. D. Fuhlendorf, and J. D. Derner for constructive reviews of the manuscript. exert a major influence on grazing and fire. Over the great major- This research is a contribution from the Oklahoma Agricultural Experiment Station ity of these grasslands, grazing is currently the primary manage- and the USDA-ARS. All programs and services of the U.S. Department of Agriculture are offered on a nondiscriminatory basis without regard to race, color, national origin, ment factor. The fundamental principle of grazing management is religion, sex, age, marital status, or handicap. to control the frequency and intensity of defoliation of individual Manuscript accepted 5 Dec. 1999.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 471 plants (Heitschmidt and Walker 1996). Red Shale ecological sites dominate the among treatments. To do this, steers were The primary method for controlling the study pastures. This site supports mixed- first classified into weight groups based on frequency and intensity of defoliation has grass prairie as the potential natural vege- 23 kg increments. We then calculated the been to control the stocking rate or the tation with mean forage production of number of steers required from each number of animals present per unit land 1,050 kg ha-1. Major grass species include weight group for each treatment based on area for a specified time. sideoats grama [Bouteloua curtipendula proportional representation of all weight Mixed-grass vegetation in west central (Michx.)Torr.], blue grama [B. g r a c i l i s groups within each treatment. Finally, spe- Kansas changed considerably as stocking (Willd. ex Kunth) Lag. ex Griffiths], hairy cific steers from each weight group were rate increased (Launchbaugh 1967) but lit- grama [B. hirsuta Lag.], buffalograss randomly allocated to treatments. The tle change was observed in western [Buchlöe dactyloides (Nutt.) Englem.], sil- steers had access to a free-choice mineral Nebraska (Burzlaff and Harris 1969). ver bluestem [Bothriochloa saccharoides supplement during the grazing season. All Differences between studies can be attrib- (Sw.) Rydb.], red threeawn [A r i s t i d a procedures for animal care and manage- uted to weather, the levels of stocking rate longiseta Steud.], purple threeawn [A. pur - ment were in accordance with accepted included, the length of the study, and the p u r e a Nutt.], and little bluestem guidelines (Consortium 1988). particular mix of plant species at a loca- [Schizachyrium scoparium ( M i c h x . ) Herbage standing crop was measured tion (Hart and Norton 1988). If the range Nash]. Major forbs include western rag- each year around 20 July and 24 of stocking rates is wide enough in mixed weed [Ambrosia psilostachya DC.] and September. For each sample date, 50 prairie, midgrasses are generally more curlycup gumweed [Grindelia squarrosa quadrats were clipped to ground level to abundant at lower stocking rates while (Pursh) Dun.]. Scattered populations of the determine total standing crop in each pas- shortgrasses are more abundant at higher half-shrub broom snakeweed [G u t i e r r e z i a ture. Quadrat area was 0.1 m2. Quadrats stocking rates (Launchbaugh 1967, Sims sarothrae (Pursh) Britt. & Rusby] are were systematically spread across each et al. 1976, Thurow et al. 1988, found on shallow sites. pasture along pace transects. The same Fuhlendorf and Smeins 1997). The specif- Approximately 30% of the Klemme approximate transects were sampled on ic objective of this study was to measure Station was previously cultivated. Old each date. In each pasture, 2 to 3 samples the effect of cattle stocking rate on stand- fields (3 to 27 ha) are scattered across all of pure live and dead herbage were col- ing crop dynamics and species composi- study pastures. About two-thirds of these lected to determine ratios of live and dead tion of a southern mixed-grass prairie. old fields were reseeded to mixtures of tissue in the total herbage (Gillen and Tate native grasses 25 to 35 years ago. The 1993). exact seed mixtures used are not known Species composition was determined Materials and Methods but likely included sand bluestem around mid-July every other year from [Andropogon hallii Hack.], little bluestem, 1990 to 1996. Species composition was Study Area indiangrass [Sorghastrum nutans ( L . ) measured using the dry-weight rank The study area is located 15 km south- Nash], sideoats grama, and blue grama. method (Gillen and Smith 1985) with 100 west of Clinton, Okla. on the Marvin The reseeded fields are now dominated by quadrats per pasture (50 of which were sideoats grama. The remaining fields also used for standing crop sampling). The Klemme Range Research Station (35° 50' 2 N, 99° 8' W), a unit of the Oklahoma revegetated naturally and supported 0.1 m quadrats were systematically dis- Agricultural Experiment Station. The mixed-grass vegetation. tributed across the pastures along pace study area is characterized by rolling Grazing management on the study pas- transects. Species groups were: sideoats uplands (2–15% slopes) dissected by mod- tures is not documented prior to 1988. grama, shortgrasses (buffalograss, blue erately deep, steep-walled drainages. The Livestock have grazed the pastures since grama, and hairy grama), silver bluestem, mean elevation is 490 m. The predominant at least 1900. Stocking rates since 1965 threeawns (red and purple threeawn), tall- soil of the experimental pastures is the are estimated to have been moderately grasses (little bluestem and sand Cordell silty clay loam (Loamy, mixed, heavy to heavy. bluestem), other perennial grasses, annual thermic Lithic Ustochrepts). Cordell soils grasses, forbs, and broom snakeweed. are shallow with a depth of 25 to 36 cm Methods For the statistical analysis, we used an (Moffatt and Conradi 1979). Rocky out- Stocking rates of 23, 26, 34, 41, 48, or analysis-of-covariance model (Littell et al. crops of hard red siltstone make up 0–25% 51 AUD ha -1 were randomly allocated to 6 1991). The model contained terms for of Cordell mapping units. pastures. The currently recommended sus- stocking rate, year, and the stocking rate x The nearest reporting station for long- tainable stocking rate is 25 AUD ha-1 (SCS year interaction. Year was a classification term climatological data (1961–1990) is 1960). Average pasture area was 65 ha. variable while stocking rate was a quanti- located at Clinton, Okla. (35° 22' N, 99° The pastures were continuously stocked tative covariate. Since stocking rate varied 04' W), 14 km northeast of the research with mixed breed yearling beef steers (Bos slightly in a given pasture over years station. The climate is continental with taurus x Bos indicus; maximum 1/8 B . because of differences in the length of the cool winters and hot summers. The mean indicus) typical of commercial stocker cat- grazing season or initial steer weight, we annual precipitation is 766 mm, ranging tle originating in the southeastern U.S. used the mean stocking rate over the 7- from 510 to 817 mm. The mean precipita- Herd size varied from 10 to 25 steers year study period. Dependent variables tion from April through September is 518 depending on stocking rate. During the 7- were standing crop or relative species mm or 68% of the annual precipitation. year study period (1990 through 1996), the composition. The mean frost-free growing period is 206 mean start and end of the grazing season The analysis-of-covariance model fit a days from April to October. The mean was 15 April and 24 September. linear regression between the dependent temperature is 16.1°C with January as the Initial body weight of the steers aver- variable and stocking rate for each year. coldest month (mean 8.9°C) and July as aged 216 kg (SD = 12 kg). We attempted For instance, if the stocking rate effect the hottest month (mean 28.4°C). to equalize average body weight of steers was significant for July standing crop,

472 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 there was a linear relationship between Table 1. Precipitation during the study period Live Standing Crop. Averaged over July standing crop and stocking rate. If the and 30-year average precipitation (1961–1990) stocking rates and years, live standing year effect was significant, the intercept at Clinton, Okla. herbage averaged 930 kg ha-1 in July and for the regression lines was different 720 kg ha- 1 in September. Live standing between some years. If the stocking rate x Precipitation crop in July was not related to stocking year interaction was significant, the slope Year Annual Apr. through Sep. rate and there was no stocking rate by year of the line relating the July standing crop (mm) (mm) interaction (Table 2). Live standing crop and stocking rate was different for some of 1990 670 514 varied among years, ranging from 630 kg the years in the analysis. If the stocking 1991 733 535 ha-1 in 1995 to 1,410 kg ha-1 in 1992. 1992 774 484 rate x year interaction was not significant, 1993 709 496 There was a stocking rate by year inter- the slope of the regression relating July 1994 575 361 action for live standing crop in September. standing crop and stocking rate was con- 1995 1195 935 From 1990 to 1995, live standing crop in stant over years. 1996 817 781 September was not affected by stocking We initially fit a full covariance model Study average 813 591 rate (Fig. 2). In 1996, live standing crop to our data for standing crop and relative 30-year average 766 518 was negatively related to stocking rate. species composition. If the stocking rate x Precipitation for the July to September year interaction was significant in the ini- period in 1996 totaled 294 mm, 40% 1995 and 1996 although much of the grow- tial model, we examined all regression above the long-term mean, and was well ing season precipitation in 1995 fell in large coefficients for the full model. We then distributed. The favorable precipitation intense storms with considerable runoff. dropped all coefficients that were not sig- regime may have delayed transfer of Overall growing conditions were average to nificant (P > 0.10) and fit a reduced herbage from the live to the dead compo- favorable during the study period. regression model. If the stocking rate x nent and allowed an accumulation of live year interaction was not significant (P > herbage as stocking rates decreased. 0.10), we fit a reduced model containing Standing Crop Percentage live standing crop was signif- only stocking rate and year. If year was Total Standing Crop. Total standing icantly related to stocking rate in both July significant in the reduced model, we made crop, averaged over stocking rates and and September (Table 2). Averaged over pair-wise comparisons of least squares years, was similar in July (1,920 kg ha- 1) sample dates and years, percentage live means between years (P = 0.05). and September (1,870 kg ha- 1, P = 0.96). standing crop ranged from 42% at 23 AUD With this procedure, we essentially fit a Total standing crop was only different h a- 1 to 52% at 51 AUD ha- 1. Percentage linear regression between a dependent among months (P < 0.05) in 1991, with live standing crop increased by 0.4% for variable and stocking rate for different 2,460 and 1,880 kg ha-1 present in July and each increase of 1 AUD ha-1 (P = 0.05), a years, for instance 1990 and 1996 as a September, respectively. Herbage growth relationship that was constant over sample simplified example. We then compared and consumption were in rough balance in dates and years. Percentage live standing the lines for the 2 years. If the slopes of the later portion of the growing season. crop was similar (P = 0.96) for July and the lines for 1990 and 1996 were different, Total standing crop and stocking rate September at 48% and 46% of total stand- stocking rate had a significant effect over had an inverse relationship because of ing crop, respectively, when averaged over years. If the slopes of the lines for 1990 greater total forage demand at higher stocking rates and years. Live standing and 1996 were not different, stocking rate stocking rates (Table 2, Fig. 1). Averaged crop ranged from a low of 28% in had no effect over years, regardless of the over months, standing crop decreased September 1990 to a high of 67% in July relationship in 1990. We assumed any sig- approximately 17.5 kg ha-1 for every AUD 1992. nificant relationship present between ha-1 increase in stocking rate. We found no Dead Standing Crop. Dead standing species composition and stocking rate in interaction between stocking rate and year crop was related to stocking rate (P < July 1990 to be a chance relationship. This (Table 2). This suggests stocking rate did 0.01) in both July and September (Table 2, was the first year of the study and we not affect the herbage production potential Fig. 3). As stocking rate increased, the would not expect stocking rate to have an of the pastures over the term of the study. standing crop of dead herbage decreased. effect on plant populations within 3 months. To address our research objective, Table 2. P-values from analyses of variance for standing crop components in July and September we emphasized interactions of stocking from a 7-year stocking rate study on mixed-grass prairie in western Oklahoma. rate with year. Significant effects of stock- ing rate were not considered important Stocking rate x unless they changed over years. Item Stocking rate Year year Total Standing Crop July <0.01 <0.01 0.31 Results and Discussion September <0.01 <0.01 0.55 Live Standing Crop July 0.96 <0.01 0.92 Precipitation September 0.43 <0.01 0.08 Means of annual and growing season precipitation during the study were 6% % Live Standing Crop July <0.01 <0.01 0.20 and 14% above the long-term means, September <0.01 <0.01 0.24 respectively (Table 1). The most stressful Dead Standing Crop conditions occurred in 1994 when growing July <0.01 <0.01 0.08 season precipitation was 69% of the mean. September <0.01 <0.01 0.48 This was offset by high precipitation in

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 473 Fig. 1. Total standing crop of herbage in July and September as Fig. 2. Standing crop of live herbage in September as affected by stock- affected by stocking rate. Points are averages of 7 years. For ing rate. For 1990-95, standing crop = 721 + 1.4 (stocking rate). For July, standing crop = 2483 – 15.2 (stocking rate). For 1996, standing crop = 2222 – 21.0 (stocking rate). For the combined September, standing crop = 2458 – 14.4 (stocking rate). For the model, R2 = 0.92, P < 0.01. Intercepts and slopes are different between combined regression model, R2 = 0.59, P = 0.06. Slope and inter- sets of years (intercepts P < 0.01, slopes P = 0.02). cept terms for July and September are not different (P > 0.90).

The only exception to this relationship Standing crop of dead herbage averaged stocking rate reduced both live and dead occurred in July 1990, the first year of the 1,660 kg ha- 1 in July and 1560 kg ha- 1 i n standing crop of the grass component study. On that date, there was no relation- September (P = 0.85). (Heitschmidt et al. 1989). ship between dead standing crop and Effects of stocking rate on standing crop With 2 minor exceptions, the relation- stocking rate. Grazing treatments had been components were manifest in the dead ship between standing crop and stocking imposed for about 3 months at that time herbage component. This suggests the rate did not change over the 7-year study and had not exerted a measurable impact amount of herbage transferred from the period. This suggests the impact of stock- on dead standing crop. By the end of the live to the dead component was greater at ing rate was simply increased herbage first year in September of 1990, a relation- lower stocking rates because more herbage consumption due to greater numbers of ship had been established between dead would reach physiological maturity and livestock as stocking rate increased. There standing crop and stocking rate and this move into the dead component. is no evidence of an effect on plant pro- relationship remained constant over the Conversely, at higher stocking rates a por- ductivity at the community level. If plant following years and sample dates. Dead tion of this herbage would be consumed productivity had declined over years at the standing crop declined 17.4 kg ha- 1 f o r before it could transfer to the dead com- higher stocking rates, the slope of standing each increase of 1 AUD ha- 1 (P < 0.01). partment. In north central Texas, a high crop versus stocking rate would have become more negative.

Table 3. Coefficients and statistics of regression models describing response of species composition (%) to stocking rate between 1990 and 1996. Year is coded 0 for 1990 and 1 for 1996. Stocking rate is AUD ha-1.

Period N Intercept Year Stocking Year R2 C.V. MSE rate x Stocking rate Sideoats grama 12 14.8** 0.1+ 0.27 13.0 6.7 (2.8)1 (0.8) Shortgrasses 12 37.4** –0.3* 0.39 18.0 21.1 (4.9) (0.1) Silver bluestem 12 1.0 0.1* 0.41 31.0 3.8 (2.0) (0.05) Threeawns 12 5.8** 11.7** –0.14+ 0.84 20.5 3.0 (0.7) (2.7) (0.07) Other grasses 12 9.8** 0.00 39.4 14.8 (1.1) Annual grasses 12 2.7** –2.7** 0.81 51.8 0.5 (0.3) (0.4) Forbs 12 22.1** 0.00 23.7 27.5 (1.5) Broom snakeweed 12 4.4** 0.00 58.0 6.7 (0.7) **, *, + significant at the 0.01, 0.05, and 0.10 levels respectively. 1Standard error of estimate.

474 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 of the vegetation. Buffalograss was the major species in this category. As with sideoats grama, shortgrasses did not respond to either stocking rate or year (Tables 3 and 4, Fig. 4). Other studies have reported increases in the relative composition of shortgrasses as stocking rate increased (Klipple and Costello 1960, Launchbaugh 1967, Thurow et al. 1988, Heitschmidt et al. 1989). Silver bluestem was most abundant at the higher stocking rates (Table 3, Fig. 4). This relationship was present at the begin- ning of the study and remained constant over years indicating no effect of stocking rate on silver bluestem. There was also no relationship with year (Fig. 4). Silver bluestem percentage composition aver- aged 6% and remained unchanged over the 7 years of the study. Silver bluestem is also a secondary species in Texas mixed- grass prairies and increased slightly as stocking rate was decreased (Fuhlendorf and Smeins 1997) or else was not affected by stocking rate (Heitschmidt et al. 1989). There was no relationship between threeawns and stocking rate in 1990. By 1996, threeawns were more abundant at the lower stocking rates (Table 3, Fig. 4). The response of threeawns to stocking rate has been variable in other studies. Studies Fig. 3. Standing crop of dead herbage as affected by stocking rate. For 1990, slopes are not in Texas found either little response of different from 0 (P > 0.96) and are not different between July and September (P = 0.83). perennial threeawns to stocking rate For 1991 to 1996, July standing crop = 1655 – 18.0 (stocking rate); September standing (Heitschmidt et al. 1989, Taylor et al. crop = 1640 – 16.8 (stocking rate). For the combined model, 1991 to 1996, R2 = 0.82, P < 1997) or higher abundance of threeawns at 0.01. Intercept and slope terms for July and September are not different (P > .67). intermediate stocking rates (Fuhlendorf

Species Composition Table 4. Relative composition (%) for species or species groups over years, averaged over stocking Sideoats grama was the most abundant rate from a 7-year stocking rate study on mixed-grass prairie in western Oklahoma. single species, comprising approximately 20% of the vegetation, and was not affect- Year ed by stocking rate. While the relative Species 1990 1992 1994 1996 composition of sideoats grama was greater ------(%) ------at the higher stocking rates, this positive Sideoats grama 20.8ab1 22.5a 17.6c 19.1bc relationship was already present in 1990 at (2.9)2 (3.0) (3.6) (2.8) the initiation of the study and did not Shortgrasses 27.1 22.4 26.7 23.9 change over 7 years (Table 3, Fig. 4). This (6.5) (2.9) (5.3) (4.6) was unexpected because sideoats grama is Silver bluestem 6.8 5.2 5.6 5.6 often considered a decreaser in mixed- (2.7) (1.5) (1.5) (2.0) grass prairie (SCS 1960, Taylor et al. Threeawns 5.2c 6.0c 9.0b 11.7a 1997). Heitschmidt et al. (1985, 1989) (2.1) (3.0) (2.7) (1.9) also reported the composition of sideoats Tallgrasses 2.0 3.1 3.2 2.5 grama was not influenced by stocking rate (0.8) (1.6) (2.4) (1.9) in mixed-grass prairie. In central Texas, Other perennial grasses 11.2 10.7 7.8 8.3 sideoats grama increased as stocking rate (3.4) (4.4) (3.6) (4.0) was decreased but changes were not Annual grasses 2.7a 2.4a 0.8b 0.0c immediate and were mediated by soil (1.0) (0.7) (0.2) (0.0) depth (Fuhlendorf and Smeins 1997, Forbs 20.2 19.0 21.8 24.0 1998). Sideoats grama declined slightly on (6.3) (2.4) (3.8) (3.5) all pastures in 1994 but returned to initial Broom snakeweed 3.9c 8.6a 7.5ab 5.0bc levels in 1996 (Table 4). (1.7) (2.9) (3.3) (3.3) The shortgrasses were prominent in all 1Means within a row followed by different letters are different using protected pair-wise t-tests, P < 0.05. pastures, contributing approximately 25% 2Standard deviation.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 475 Tallgrasses were negatively related to stocking rate but there was an interaction of stocking rate and year. In the first year of the study (1990), tallgrasses were not related to stocking rate (Fig. 5). By 1992, there was a negative relationship between the tallgrass component and stocking rate. This relationship remained constant through 1996. Although tallgrasses increased at the lower stocking rates, they still contributed no more than 5 to 6% of the total herbage. Descriptions of the potential natural community for the Red Shale ecological site suggest that little bluestem could be co-dominant on this site with sideoats grama (SCS 1960). More than 7 years of moderate stocking rates would be required for little bluestem to increase to such levels. It may be that the initial response of tallgrasses was due to the short-term process of increasing the size of existing plants while further increases in relative composition must wait for the long-term process of estab- lishing new plants. Other perennial grasses were about 10% of the vegetation. Hairy tridens [Erioneuron pilosum (Buckl.) Nash], tall dropseed [Sporobolus asper ( M i c h x . ) Kunth], and windmillgrass [Chloris verti - cillata Nutt.] were the major grasses in this category. This component was not affected by stocking rate or year and there was no interaction between stocking rate and year (Table 3, Fig. 4). Different responses among species could have pre- vented any discernible group response. The annual grass category was primarily Japanese brome [Bromus japonicus Thunb.]. This component was not promi- nent in the study pastures. The annual Fig. 4. Relative composition of vegetation components as affected by stocking rate in 1990 and 1996. Regression models for each component are found in Table 4. A single line indi- grasses were not affected by stocking rate cates no difference between years. (Table 3, Fig. 4) but declined significantly from 2.8% to less than 0.1% over the study period (Table 4). Japanese brome and Smeins 1997). In contrast, threeawns the last plants to be grazed as stocking fluctuates greatly over years in response to decreased at higher stocking rates in short- rates increase (Klipple and Costello 1960). favorable winter and spring precipitation grass prairie in northeastern Colorado As the more palatable plants are grazed, (Launchbaugh 1967). (Klipple and Costello 1960). Later work at threeawns become more visually promi- Forbs were one of the larger vegetation the same site found heavy grazing after nent. This may explain why they have categories, contributing 20 to 24% of the June was generally detrimental to red been considered increasers. herbage. Forbs were not affected by stock- threeawn (Hyder et al. 1975). Threeawns increased at all stocking ing rate or year (Table 3, Fig. 4). This sup- Threeawns are classified as increasers or rates from 1990 to 1996 (Table 4) but the ports the results of previous studies, which invaders on virtually all ecological site increase was greater at the lower stocking reported no difference in forb abundance descriptions used by the Natural rates (Fig. 4). The reason for this general due to stocking rates (Sims et al. 1976, Resources Conservation Service in west- increase is not known. Red threeawn Hart et al. 1988, Heitschmidt et al. 1989). ern Oklahoma. This means threeawns appeared to be favored in wet years but Launchbaugh (1967), in contrast, found should increase in abundance as stocking reduced in dry years in shortgrass prairie that forbs decreased as stocking rate rates increase. Results from the current (Hyder et al. 1975). The increase in three- increased. This is contrary to much popu- and previous studies suggest this classifi- awns in this study cannot be easily attrib- lar opinion that holds that forbs increase as cation should be reconsidered. These uted to weather since precipitation was stocking rates increase. Differences in the grasses appear to remain unchanged or well above average only in years 6 and 7 effect of stocking rate on forb composition decrease as stocking rate increases. of the study while threeawns increased among studies may be attributable to indi- Threeawns are unpalatable and are often between years 3 and 5 (Table 4).

476 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 precipitation during the last 2 years of the study was much greater than in the earlier years. During the first 5 years, annual pre- cipitation was 85% of average and grow- ing season precipitation was 81% of aver- age. These precipitation levels would not constitute a severe drought but were below average. We expected these precipitation levels in combination with high stocking rates to trigger vegetation changes but the mixed-grass vegetation was resistant under the observed stocking rates and weather conditions. A third hypothesis considers the influ- ence of initial conditions. Significant changes may have already occurred in the 80 years the pastures were grazed at high stocking rates before the study began and the vegetation may have already been in equilibrium with the higher stocking rates used in our study. In that case, little change would be observed at high stock- ing rates but change could potentially occur at lower stocking rates. This is sug- gested by the slight increase in tallgrasses at the lower stocking rates. Since this component was at low levels initially, increases might take many more years to be practically significant. A similar situa- tion has been reported from central Texas (Fuhlendorf and Smeins 1997). Fig. 5. Relative composition of tallgrasses as affected by stocking rate. Intercept coefficients are 0.9, 7.1, 9.7, and 6.3 for 1990, 1992, 1994, and 1996, respectively. Slope coefficients are Additionally, grazing sensitive species 0.03, –.11, –.17, and –.10 for 1990, 1992, 1994, and 1996, respectively. For the combined may not increase immediately because model, R 2 = 0.60, P = 0.02. Coefficients for 1990 are different from 1992–1996 (P < 0.05), lower stocking rates alone are not enough which are not different from each other (P > 0.24). of a disturbance to cause grazing resistant species to relinquish resources vidual species responses. Species within as stocking rate increased rather than to (Fuhlendorf and Smeins 1998). Increases this group should probably be studied long-term impacts on plant vigor. Stocking in grazing sensitive species may be individually for a better understanding of rate affected the relative species contribu- delayed until a climatic stress reduces the the influence of stocking rate on their rela- tion of only 2 vegetation components, grazing resistant species and opens space tive composition (Heitschmidt et al. 1985). threeawns and tallgrasses. There may be for colonization. We believe this third The half-shrub broom snakeweed was potential for further changes in tallgrasses hypothesis best fits the observed vegeta- not affected by stocking rate (Table 3, Fig. over longer time periods. Broom snake- tion responses from this study. 4). Broom snakeweed composition more weed fluctuated over time, but this effect than doubled from 1990 to 1992 before was independent of stocking rate. The veg- declining back to near initial levels in etation was, for all practical purposes, Literature Cited 1996 (Table 4). These changes over years unchanged at the end of this study. were independent of stocking rate. Klipple Several hypotheses could be advanced Burzlaff, D.F. and L. Harris. 1969. Yearling and Costello (1960) reported that broom to explain the moderate vegetation steer gains and vegetation changes of western snakeweed was not affected by stocking response. First, the stocking rates studied Nebraska rangeland under three rates of rate and was cyclic in nature. In addition, may have been too light to cause signifi- stocking. Nebraska Agr. Exp. Sta. SB 505. broom snakeweed populations fluctuated cant effects. We place lesser weight on Consortium. 1988. Guide for care and use of dramatically in response to weather in this hypothesis because the stocking rates agricultural animals in agricultural research New Mexico (McDaniel et al. 1993). used were substantially higher than the and teaching. Consortium for Developing a Guide for the Care and Use of Agr. Animals recommended rates. The impact of these in Agr.Res.and Teaching. Champaign, Ill. stocking rates was further compounded Fuhlendorf, S.D. and F.E. Smeins. 1997. Conclusions because all of the grazing occurred during Long-term vegetation dynamics mediated by the growing season. herbivores, weather and fire in a J u n i p e r u s - Stocking rate had few impacts on our Second, weather conditions during the Quercus savanna. J. Veg. Sci. 8:819–828. experimental pastures over the course of study period may have ameliorated the Fuhlendorf, S.D. and F.E. Smeins. 1998. The this study. Effects on herbage standing impact of stocking rate. Both annual and influence of soil depth on plant species crop can be attributed to the simple fact growing season precipitation were above response to grazing within a semi-arid savan- that livestock demand for forage increased average for the study period. However, na. Plant Ecol. 138:89–96.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 477 Gillen, R.L. and K.W. Tate. 1993. The con- Heitschmidt, R.K., S.L. Dowhower, W.E. McDaniel, K.C., L.A. Torell, and J.W. Bain. stituent differential method for determining Pinchak, and S.K. Canon. 1989. Effects of 1993. Overstory-understory relationships for live and dead herbage. J. Range Manage. stocking rate on quantity and quality of avail- broom snakeweed-blue grama grasslands. J. 46:142–147. able forage in a southern mixed-grass prairie. Range Manage. 46:506–511. Gillen, R.L. and E.L. Smith. 1985. Evaluation J. Range Manage. 42:468–473. Moffatt, H. H. and A.J. Conradi. 1979. S o i l of the dry-weight-rank method for determin- Hyder, D.N., R.E. Bement, E.E. Remmenga, survey of Washita County Oklahoma. USDA ing species composition in tallgrass prairie. J. and D.F. Hervey. 1975. Ecological respons- Soil Conserv. Serv., Stillwater, Okla. Range Manage. 39:283–285. Hart, R.H. and B.E. Norton. 1988. G r a z i n g es of native plants and guidelines for man- Sims, P.L., B.E. Dahl, and A.H. Denham. management and vegetation response, p. agement of shortgrass range. USDA Tech. 1976. Vegetation and livestock response at 493-525. I n: P.T. Tueller (ed.), Vegetation Bull. 1503. Washington, D.C. three grazing intensities on sandhill range- science applications for rangeland analysis Klipple, G.E. and D.F. Costello. 1960. land in eastern Colorado. Colorado Agr. Exp. and management. Kluwer Academic Vegetation and cattle responses to different Sta. Tech. Bull. 130. Ft. Collins, Colo. Publishers, London. intensities of grazing on shortgrass ranges on [SCS] Soil Conservation Service. 1960. Hart, R.H., M.J. Samuel, P.S. Test, and M.A. the Central Great Plains. USDA Tech. Bull. Technical range site descriptions. USDA- Smith. 1988. Cattle, vegetation, and eco- 1206. Washington, D.C. Soil Conserv. Serv., Stillwater, Okla. nomic responses to grazing systems and Launchbaugh, J.L. 1967. Vegetation relation- Taylor, C.A., Jr., M.H. Ralphs, and M.M. grazing pressure. J. Range Manage. ships associated with intensity of summer Kothmann. 1997. Vegetation response to 41:282–286. grazing on a clay upland range site in the increasing stocking rate under rotational Heitschmidt, R.K. and J.W. Walker. 1996. Grazing management: technology for sus- Kansas 20- to 24-inch precipitation zone. stocking. J. Range Manage. 50:439-442. taining rangeland ecosystems? Rangel. J. Kansas Agr. Exp. Sta. Bull. 154. Manhattan, Thurow, T.L., W.H. Blackburn, and C.A. 18:194–215. Kan. Taylor, Jr. 1988. Some vegetation responses Heitschmidt, R.K., S.L. Dowhower, R.A. Littell, R.C., R. J. Freund, and P.C. Spector. to selected livestock grazing strategies, Gordon, and D.L. Price. 1985. Response of 1991. SAS system for linear models. 3rd Ed. Edwards Plateau, Texas. J. Range Manage. vegetation to livestock grazing at the Texas SAS Institute, Inc. Cary, N.C. 41:108-114. Experimental Ranch. Texas Agr. Exp. Sta. Bull. 1515. College Station, Tex.

478 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 J. Range Manage. 53: 479–482 September 2000 Effects of roundups on behavior and reproduction of feral horses

KYLE V. HANSEN AND JEFFREY C. MOSLEY

Abstract Resumen

Roundups are used to maintain feral horse populations in bal- Las redadas se utilizan para mantener las poblaciones de ance with rangeland grazing capacity, but little is known about caballos silvestres en balance con la capacidad de apacentamien- short-term and long-term effects of roundups on horses. We eval- to, pero poco se sabe acerca de los efectos a corto y largo plazo uated the effects of roundups on behavior and reproduction of de las redadas en los caballos. Evaluamos el efecto de las redada feral horses. The study was conducted near Challis, Ida. during en el comportamiento y reproducción de los caballos silvestres. 1994–1995, and repeated near Lander, Wyo. during 1995–1996. El estudio se condujo cerca de Challis, Ida. durante 1994 y 1995 Horses were randomly assigned to 3 different treatment groups. y se repitió cerca de Lander Wyo. durante 1995 y 1996. Los One group (ADOPTED) was gathered by a Bureau of Land caballos se asignaron aleatoriamente a tres grupos de tratamien- Management roundup crew using a helicopter. These horses to. Un grupo (ADOPTADO) fue reunido y removido por una were removed and placed in the Adopt-A-Horse Program. The cuadrilla de redadas del Buro de Manejo de Tierras la cual uti- second group (SIMULATED) consisted of horses that were gath- lizó un helicóptero. Estos caballos fueron después colocados en el ered by helicopter, but these horses evaded capture and programa "Adopta un Caballo". El segundo grupo, (SIMULA- remained in the study area after the roundup. Horses in the third DO) consistio de caballos que fueron reunidos con helicóptero, group (CONTROL) were not herded by helicopter. Horse behav- pero estos caballos evadieron la captura y permanecieron en el ior was monitored in the SIMULATED and CONTROL groups área de estudio después de la redada. El tercer grupo (CON- before and after roundups. Behavioral variables analyzed were TROL), fueron caballos que no se pastorearon con helicóptero. the percentage of time spent resting, feeding, vigilant, traveling, El comportamiento de los caballos se monitoreo en los grupos and engaged in agonistic encounters. Neither foraging or social SIMULADO y CONTROL antes y después de las redadas. Las behavior of feral horses was affected by roundups in either study variables de comportamiento se analizadas fueron el porcentaje area (P > 0.10). Reproduction was monitored within the SIMU- de tiempo utilizado en descanso, alimentación, vigilancia, viajar, LATED, CONTROL, and ADOPTED groups during the year y encuentros agonísticos. Las redadas no afectaron ni el compor- following roundups. The percentages of mares with live foals did tamiento de alimentación ni el comportamiento social de los not differ (P > 0.10) among the 3 treatment groups in Idaho or caballos salvajes en el área de studio (P > 0.01). Al año siguiente Wyoming. Foaling success rates in Idaho were 29%, 31%, and de las redadas se monitoreo la reproducción dentro de los grupos 43% for CONTROL, ADOPTED, and SIMULATED mares, SIMULADO, CONTROL y ADOPTADOS. El porcentaje de respectively. In Wyoming, foaling success rates were 29%, 42%, yeguas con potros vivos no difirió entre los tres tratamientos and 48% for CONTROL, ADOPTED, and SIMULATED (P>0.10) y los resultados fueron similares en Idaho y Wyoming. groups, respectively. We found no evidence that roundups had Las tasas de potros vivos en Idaho fueron 29%, 31% y 43% para deleterious effects on behavior or reproduction of feral horses. los tratamientos CONTROL, ADOPTADO y SIMULADO respectivamente. En Wyoming los porcentajes de potros vivos fueron 29%, 42% y 48% para lostratamientos CONTROL, Key Words: wild horses, Equus caballus, A d o p t - A - H o r s e ADOPTADO y SIMULADO respectivamente. No encontramos Program, foraging behavior, social behavior evidencia de que las redadas tuvieran un efecto nocivo en el com- portamiento o la reproducción de los caballos salvajes. About 46,000 feral horses (Equus caballus) live in the western United States (USDI-BLM and USDA-FS 1993). The Wild Free- Roaming Horse and Burro Act, Public Law 92-195 (U.S. Code causing rangeland degradation (Smith 1986). Roundups remove 1988), obligates the Bureau of Land Management (BLM) and the excess animals, after which horses are placed in private owner- U.S. Forest Service to protect feral horses on public lands of the ship via the Adopt-A-Horse Program. Roundups obviously dis- western United States from unauthorized capture or killing. These turb the daily routine and behavior of feral horses, but little is agencies must also maintain these populations in balance with known about short-term and long-term effects of roundups on rangeland carrying capacity (USDI-BLM and USDA-FS 1993). horses. When perturbations occur, horses may shift their behavior Uncontrolled horse populations lead to excessive grazing, in turn to minimize the effects. For example, feral horses alter their habi- tat use to seek refuge from biting insects (Keiper and Berger 1982). Behavioral responses to perturbations also may include Authors wish to thank Kendall Johnson, Gordon Woods, Carl Hunt, and BLM increased time spent in agonistic encounters with other horses, personnel in Idaho and Wyoming, in particular Mike Courtney, Roy Packer, and Phyllis Stark, for their assistance. such as when mares enter new areas and encounter resident mares Manuscript accepted 5 Dec. 1999. (Tyler 1972). Perturbations may also induce stress that decreases

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 479 reproduction. For example, forced copula- vations and Douglas-fir/Idaho fescue at istics of each horse were recorded to aid tion by invading males may induce abor- the highest elevations (BLM-SCS 1979). their identification. In Idaho, the SIMU- tion in females unprotected by their resi- LATED, CONTROL, and ADOPTED dent stallion (Berger 1983), and abortions Wyoming Study Area groups totaled 14, 14, and 39 mares, may occur in response to male harassment The Wyoming study area was in respectively. In Wyoming, there were 33, or other stresses imposed by changes in a Muskrat Basin, 202-km2 of public land 14, and 34 mares, respectively, in the herd’s social structure (Berger 1983). The administered by the BLM, located SIMULATED, CONTROL, and ADOPT- purpose of this study was to document the between State Highways 789 and 136. ED groups. In each study area the roundup effects of roundups on the foraging behav- Vegetation was dominated by Wyoming and simulated roundup occurred the same ior, social behavior, and reproductive suc- big sagebrush, bluebunch wheatgrass, day, on 27 August 1994 in the Idaho study cess of feral horses. Our specific null western wheatgrass, Sandberg bluegrass, area and 13 August 1995 in Wyoming. hypotheses were: 1) behavior does not dif- Idaho fescue, and needle-and-thread grass fer between horses rounded up and horses (Stipa comata Trin. & Rupr.). Elevations Horse Behavior not rounded up; and 2) reproductive suc- ranged from 1,820–2,100 m. Annual pre- Horse behavior was monitored in the cess does not differ among horses rounded cipitation averages 25 cm with about 50% SIMULATED and CONTROL groups for up, horses not rounded up, and horses occurring during May and June, and 50% 3 weeks immediately before and 3 weeks placed in the Adopt-A-Horse Program. occurring during November through after the simulated roundups. Monitoring March in the form of snow (Perrich 1992). before the simulated roundup occurred Soils were a loamy-skeletal, mixed (cal- from 25 July through 26 August 1994 on Study Area careous), frigid, shallow Ustic the Idaho study area and 10 July through Torriorthent (Cragosen) in most of the 12 August 1995 for the Wyoming area. The study was conducted in 2 separate area, and a fine, montmorillonitic Borollic Data following the simulated roundup was areas, one in central Idaho about 3 km Paleargid (Milren) on the lower sites. collected from 28 August through 20 southwest of Challis, Ida., and one in cen- The dominant range site for this study September 1994 for the Idaho area and 14 tral Wyoming about 32 km north of area was Loamy (24–35 cm precipitation) August through 28 August 1995 for the Jeffrey City, Wyo. with associated range sites of Shallow Wyoming area. Behavioral observations Sandy (24–35 cm precipitation) and Sandy spanned more than 600 hours during each Idaho Study Area (25–35 cm precipitation) (Soil Conservation year of the study. Service 1988). We recorded the percentage of time that The Idaho study area was 150-km2 o f a band of horses spent in the following public land administered by the BLM, activities: feeding (primarily grazing with located north of Spar Canyon Road, east little or no travel other than to move to a of State Highway 75, and west of U.S. Methods new feeding station); resting (both recum- Highway 93. Elevations ranged from bent and standing); traveling (including 1,675–2,745 m. Annual precipitation aver- This experiment was conducted in Idaho incidental grazing while walking); aggres- ages 18 cm, with about 50% occurring as from July 1994 to July 1995 and in sion within the band (defined as threat snow during November through March Wyoming from July 1995 to July 1996. postures, striking with the feet, or biting); (Perrich 1992). Major plant species pre- Field procedures and statistical methods aggression between bands; and vigilance sent were bluebunch wheatgrass were identical for both study areas. (head and neck erect, ears forward, active- (Agropyron spicatum (Pursh.) Scribn. & ly looking around). Data were recorded Smith), western wheatgrass (A g r o p y r o n Roundups every 10 minutes during 4-hour observa- smithii Rydb.), Sandberg bluegrass (P o a Horses in each study area were random- tion periods using instantaneous scan sam- s a n d b e r g i i Vasey), Idaho fescue (F e s t u c a ly assigned to 3 different treatment groups. pling (Altmann 1974). Each band of hors- i d a h o e n s i s Elmer), and Wyoming big One group (ADOPTED) was gathered by es was observed 3 times before and 3 sagebrush (Artemesia tridentata wyomin - a BLM roundup crew using a helicopter. times after the simulated roundups, with g e n s i s Beetle & Young). Limber pine These horses were removed and placed in one of the 3 observation periods each (Pinus flexilus James), Douglas-fir the Adopt-A-Horse Program. The second occurring in early morning (0700–1100 (Pseudotsuga menziesii (Mirb.) Franco) group (SIMULATED) consisted of horses hours), mid-day (1100–1500 hours), and and curlleaf mountain mahogany that were gathered by helicopter, but these early evening (1500–1900 hours). (Cercocarpus ledifolius Nutt. ex. Torr. & horses evaded capture and remained in the Gray) occurred at higher elevations within study area after the roundup. Horses in the the study area. third group (CONTROL) were not herded Reproduction Soils ranged from a loamy-skeletal, by helicopter. In the Idaho study area, the Adopted Groups. A veterinarian drew 10 ml of blood from each mare in the mixed frigid Xerollic Haplargid ADOPTED, SIMULATED, and CON- ADOPTED groups. These 10-ml samples (Dawtonia) at the lower elevations to a TROL groups were composed of 12, 3, were immediately frozen, and later in the loamy skeletal, mixed Calcic Cryoborol and 4 bands (i.e., social subgroups) of laboratory blood serum was separated (Gany) at the highest elevations. The dom- horses, respectively. The number of bands from plasma. Blood serum samples were inant range site for the study area was in the ADOPTED, SIMULATED and then sent to the Veterinary School at the Gravelly Loamy (20–30 cm precipitation) CONTROL groups in Wyoming were 10, University of California-Davis for analy- with associated range sites of Gravelly 8, and 5, respectively. Bands contained 2 Loam (33–40 cm precipitation) and sis. Mares were considered pregnant if to 10 adult horses, and all bands were pho- estrogen blood serum levels were 7 ng/ml Flagstone (20–28 cm precipitation). The tographed to document their membership. habitat type was Wyoming big sage- or greater and progesterone levels were 6 Individual members of each band were ng/ml or greater. In July of the year fol- brush/bluebunch wheatgrass at lower ele- also photographed, and physical character-

480 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 lowing roundups, each adopter of a mare Table 1. Mean percentage (±SE) of time that bands of feral horses spent in each behavioral activity was sent a questionnaire asking for foaling following simulated roundups. information. Non-respondents were con- tacted in a follow-up survey by telephone. Idaho Site Wyoming Site A 100% response rate was achieved. Activity Simulated Control Simulated Control Simulated and Control Groups. H o r s e s ------(%) ------1 in the SIMULATED and CONTROL Vigilance 12.9 ± 0.8a 8.3 ± 0.4a 1.3 ± 0.7a 1.8 ± 0.9a groups were observed about 10 months Feeding 58.5 ± 17.3a 71.4 ± 9.4a 46.9 ± 5.9a 38.0 ± 5.3a Traveling 24.0 ± 17.0a 17.4 ± 10.la 5.5 ± l.7a 18.7 ± 3.8a after the roundups. The number of mares Resting 4.6 ± 3.0a 2.9 ± 2.0a 46.1 ± 6.4a 41.3 ± 0.1a and foals in each band was counted on 30 Aggression within bands 0.0 ± 0.0a 0.0 ± 0.0a 0.0 ± 0.0a 0.0 ± 0.0a June and 1 July 1995 on the Idaho site and Aggression between bands 0.0 ± 0.0a 0.0 ± 0.0a 0.1 ± 0.0a 0.2 ± 0.0a 1 July through 5 July 1996 on the 1Within each study area, values within rows followed by the same lowercase letter are not different (P > 0.10). Wyoming site.

Statistical Analysis Results monal changes induced by stressful condi- A completely random experimental tions (Traub-Dargatz et al. 1988, Baucus design was used. Data from the Idaho and Horse Behavior et al. l990a, 1990b, Colburn et al. 1991, Wyoming sites were analyzed separately. Individually, none of the behavioral Ferlazzo et al. 1991, Wong et al. 1992). All differences were considered signifi- variables differed between SIMULATED Similarly, in our study roundups did not cant at P £ 0.10. For the behavioral data, a and CONTROL treatments in either study decrease reproductive rates of feral horses. band of horses was the experimental unit. area (Table 1). Multivariate analyses also Foaling rates of the SIMULATED, CON- Individual horses within the bands were found no differences between the SIMU- TROL, and ADOPTED mares all com- subsamples. Behavioral data were trans- LATED and CONTROL groups in either pared favorably with a 10-year average of formed using arcsine transformation to the Idaho study area (P = 0.67) or the 35% documented within a feral horse herd meet conditions for normality. Each Wyoming study area (P = 0.58). in Montana (Garrott and Taylor 1990). behavioral variable (feeding, resting, trav- The ADOPTED mares experienced the eling, aggression within the band, aggres- most perturbation but had an overall foal- sion between bands, and vigilance) was Reproduction ing rate of 33%. analyzed using analysis of covariance The percentage of mares with live foals Feral horses in our study apparently (Johnson and Wichern 1992). did not differ (P > 0.10) among the 3 treat- adapted easily to any stress caused by Observations collected before the simulat- ments in either study area. By July 1995 in roundups. We found no evidence that ed roundups were the covariables (e.g., the Idaho study area, 29%, 31%, and 43% roundups had deleterious effects on behav- vigilance before the simulated roundups of the mares had live foals in the CON- ior or reproduction of feral horses. was the covariable when analyzing vigi- TROL, ADOPTED, and SIMULATED lance after the simulated roundups). Data groups, respectively. In the Wyoming were then analyzed collectively using mul- study area, by July 1996 the percentage of Literature Cited tivariate analyses. Data from the mares with live foals in the CONTROL, Wyoming site were analyzed using multi- ADOPTED, and SIMULATED groups variate analysis of covariance (MANCO- was 29%, 42%, and 48%, respectively. Altmann, J. 1974. Observational study of behav- ior: Sampling methods. Behav. 49:227–265. VA). Dependent variables in the MAN- At the time of roundup, 27 of 39 mares (69%) were pregnant in the Idaho Baucus, K.L., S.L. Ralston, C.F. Nockels, A.O. COVA were the behavioral variables vigi- McKinnon, and E.L. Squires. 1990a. E f f e c t s lance, feeding, traveling, resting, and ADOPTED group, and 18 of 34 mares of transportation on early embryonic death in aggression between bands following the (53%) were pregnant in the Wyoming mares. J. Anim. Sci. 68:345–351. simulated roundup. The independent vari- ADOPTED group. Successful pregnancies Baucus, K.L., E.L. Squires, S.L. Ralston, A.O. able used was group (i.e., SIMULATED numbered 12 of 27 (44%) and 14 of 18 McKinnon, and T.M. Nett. 1990b. Effects of or CONTROL). Covariables used were (77%) for Idaho and Wyoming ADOPT- transportation on the estrous cycle and concen- vigilance, feeding, traveling, resting and ED groups, respectively. trations of hormones in mares. J. Anim. Sci. aggression between bands before the sim- 68:419–426. ulated roundup. For the Idaho site there Berger, J. 1983. Induced abortion and social fac- tors in wild horses. Nature 303:59–61. were too few degrees of freedom to allow Discussion and Conclusions BLM-SCS (Bureau of Land Management-Soil the observations before the simulated Conservation Service). 1979. Range site roundup to be used as covariables, so Roundups did not increase agonistic descriptions. B12. March 1979. Internal, these data were analyzed using multivari- behavior among feral horses, nor did Boise,Ida. ate analysis of variance. Aggression with- roundups cause feral horses to reallocate Colburn, D.R., D.L. Thompson, Jr., T.L. Roth, in the band was not used as a variable or their time spent in various daily activities. J.S. Capehart, and K.L. White. 1991. covariable in the multivariate analyses Responses of cortisol and prolactin to sexual Feral horses in our study allocated their because no agonistic encounters within excitement and stress in stallions and geldings. time similarly to undisturbed free-roaming bands were recorded. J. Anim. Sci. 69:2556–2562. horses in southern France (Duncan 1980), The percentage of mares with live foals Crane, K.K., M.A. Smith, and D. Reynolds. western Alberta (Salter and Hudson 1979), in the SIMULATED, CONTROL, and 1997. Habitat selection patterns of feral horses and central Wyoming (Crane et al. 1997). in south central Wyoming. J. Range Manage. ADOPTED groups was compared with Feeding was the primary activity. 50:374–380. chi-square analysis (Johnson and Wichern Duncan, P. 1980. Time budgets of Camargue 1992). Domestic horses are usually able to maintain their pregnancies despite hor- horses. II. Time budgets of adult horses and weaned sub-adults. Behavior 72:26–49.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 481 Ferlazzo, A.M., R. Vinci, M. Panzera, A. Salter, R.E., and R.J. Hudson. 1979. F e e d i n g Tyler, S.J. 1972. The behaviour and social organi- Ferlazzo, and A. Calatroni. 1991. Glyco- ecology of feral horses in western Alberta. J. zation of the New Forest ponies. Anim. Behav. saminoglyan concentrations in horse plasma and Range Manage. 32:221–225. Mono. 5:85–196. serum. Differences with other animal species Soil Conservation Service. 1988. T e c h n i c a l US Code. 1988. Title 16 sec. 1331 Dec. 15, 1961 and identification of affecting factors. Biol. guide, section IIB major land resource area (34) Public law 92-195, 85 STAT. 649. U.S. Code. Comp. Biochem. l00b:4:745–751. Range site descriptions. USDA-SCS-WY. Washington, D.C. Garrott, R.A., and L. Taylor. 1990. Dynamics of Cheyenne, Wyo. USDI-BLM and USDA-FS. 1993. Ninth report to a feral horse population in Montana. J. Wildl. Manage. 54:603–612 Smith, M.A. 1986. Impacts of feral horses grazing Congress on the administration of the Wild Johnson, R.A., and D.A. Wichern. 1992. on rangelands: An overview. J. Equine Vet. Sci. Free-Roaming Horse and Burro Act. Bureau of Applied multivariate statistical analysis. 6:236–238. Land Management, Washington, D.C. Prentice Hall, Englewood Cliffs, N.J. Traub-Dargatz, J.L., A.O. McKinnon, W.J. Wong, C.W., S.E. Smith, Y.H. Thong, J.P. Keiper, R., and J. Berger. 1982. Refuge-seeking Bruyninckx, M.A. Thrall, R.L. Jones, and B. Opdebeeck, and J.R. Thornton. 1992. Effects and pest avoidance by feral horses in desert and Blancquaert. 1988. Effects of transportation of exercise stress on various immune functions island environments. Appl. Anim. Ethol. stress on bronchoalveolor lavage fluid analysis in horses. Amer. J. Vet. Res. 53:1414–1417 9:111–120 in female horses. Amer. J. Vet. Res. Perrich, J.R. 1992. The ESE national precipita- 49:1026–1029. tion databook. Cahner’s Publishing Co., New York.

482 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 J. Range Manage. 53: 483–488 September 2000 Rotational stocking and production of Italian ryegrass on Argentinean rangelands

E.J. JACOBO, A.M. RODRIGUEZ, J.L. ROSSI, L.P. SALGADO, AND V.A. DEREGIBUS

Authors are assistant research scientists and professor, Department of Animal Science. Facultad de AgronomÌa, Universidad de Buenos Aires. C1417DSE. Buenos Aires. Argentina.

Abstract Resumen

The decreased carrying capacity of Argentinian Flooding La reducción de la receptividad de los pastizales naturales de Pampa rangelands through the reduction of density of C3 grasses la Pampa Deprimida (Argentina) como consecuencia de la dis- may be partially attributed to continuous stocking. The objective minución de la densidad de gramÌneas C3 ha sido atribuida al of this study was to evaluate the effectiveness of rotational vs. pastoreo continuo. El objetivo de este estudio fue evaluar la continuous stocking to improve winter forage production by capacidad del pastoreo rotativo con respecto al continuo de incrementing the density of Italian ryegrass (Lolium multiflorum incrementar la producción invernal de forraje mediante la pro- Lam.). Under rotational stocking, ryegrass seedlings established moción de ryegrass anual (Lolium multiflorum Lam.). Bajo pas- almost 2 months earlier in the fall and tiller density was 3-fold toreo rotativo, las plántulas de ryegrass se estalecieron dos meses higher in winter than under continuous stocking. Aerial net pri- más temprano en el otoño y la densidad de macollos fue 3 veces mary productivity of C3 grasses was approximately 2-fold más alta en invierno que bajo pastoreo continuo. La productivi- greater under rotational compared with continuous stocking in dad primaria neta aérea de las gramíneas C3 fue aproximada- the first and second years. This substantial increase in winter mente 2 veces más alta bajo pastoreo controlado que bajo pas- productivity supported almost 2-fold increase in stocking rate toreo continuo. Este significativo incremento en la productividad (from 0.6 to1.0 AU ha-1). invernal permitió duplicar la carga animal (de 0.6 a 1 EV ha-1).

Key Words: winter forage production, temperate-humid range- This restriction may be overcome by incrementing the density land, grazing systems, Lolium multiflorum. of Italian ryegrass (Lolium multiflorum Lam.), an annual grass with high quality forage, native to the Mediterranean region. This The native temperate grasslands of the Argentinean Flooding species, together with wheat (Triticum aestivum L. em Tell.), has Pampa occupies 5,000,000 ha in the eastern portion of central become widespread in Flooding Pampa grasslands since the Argentina. This region is characterized by a mild, humid climate beginning of this century. with low fertility soils (Soriano et al. 1991). This vast forage Italian ryegrass seeds disperse late in spring and germinate the resource is utilized by cow-calf operations to produce yearlings following autumn. During autumn and spring, competitive inter- that are weaned in autumn and fattened on cultivated pastures in actions among C3 and C4 grasses occur. The intensity is con- the dryer Western Pampa. Frequent flooding prevents the replace- trolled by the availability of resources. Light quality, determined ment of native grasslands by cultivated pastures or crops. These by canopy density, plays an important role in these interactions grasslands support C and C grasses during the cool and warm (Deregibus et al. 1983, Casal et al. 1985, 1986). Soil moisture 3 4 and temperature allow the onset of the germination of Italian rye- seasons, respectively. Since the productivity of C4 grasses is 3- fold greater than C grasses (Sala et al. 1981), forage production grass seeds exposed on soil surface (Rodriguez et al. 1998), and if 3 it is irradiated by direct sunlight, the germination is increased follows a seasonal pattern. During fall and spring, C3 and C4 grasses vegetate coincidently. Early in the fall, C grasses start to (Deregibus et al. 1994). This occurs whenever the canopy is 3 intensively defoliated through high stocking rate at the end of the grow while C4 grasses end their growing period with a great quantity of accumulated biomass. summer. As a consequence it would be possible to increase win- Continuous stocking of these grasslands by domestic herbi- ter forage production through a grazing strategy designed to pro- vores since the last century has resulted in degradation mote Italian ryegrass germination and seedling establishment (as (Deregibus and Cahuepé 1983). Grazing progressively reduced shown in experiment I/II). The aim of this study was to evaluate the effectiveness of rotational vs. continuous stocking to improve C3 grasses such as Stipa neesiana Trin, Melica brasiliana Arduinus and Piptochaetium montevidense Par.(León and winter Italian ryegrass production. Oesterheld, unpublished), which magnifies the seasonality of for- age production. Methods and Materials

We are grateful to Rodolfo Golluscio for his review, which greatly improve an Study area earlier draft of this manuscript. This research was funded by the National Council of Research in Argentina (CONICET). The study sites were located on 2 commercial farms of the Manuscript accepted 17 Jan. 2000. Flooding Pampa region. Each farm consisted of approximately

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 483 Table 1. Monthly precipitacion registered in the study site during the experimental period. 15 frames(10 x 20 cm) were placed ran- domly beneath the canopy in 50 ha pad- First period Precipitation Second period Precipitation docks that were either continuously or rotationally stocked (3 paddocks per treat- (mm) (mm) ment). To determine Italian ryegrass seed Mar 89 105 Mar 90 65 bank, a soil sample (20 x 10 x 5 cm) was Apr 89 226 Apr 90 372 May 89 105 May 90 10 taken next to each frame. These 15 sam- Jun 89 126 Jun 90 46 ples were gathered to obtain a composite Jul 89 5 Jul 90 0 sample for each paddock and Italian rye- Aug 89 0 Aug 90 7 grass seeds were extracted by sieving. Sep 89 72 Sep 90 12 Immediately after this, a 4-day grazing Oct 89 109 Oct 90 36 period was implemented on rotationally Nov 89 191 Nov 90 109 stocked paddocks, during which cattle Dec 89 50 Dec 90 28 removed nearly all the available biomass. Jan 90 72 Jan 91 51 At the end of this period, canopy height Feb 90 34 Feb 91 51 inside each frame was measured. Italian TOTAL 1,095 TOTAL 787 ryegrass seedling and tiller densities were recorded after approximately 4, 45, 75, 1,200 ha of native grasslands for cow-calf Angus and Hereford cow/calf pairs for and 120 days from the end of the grazing operations. Annual precipitation (average more than 25 years. The mean weight of period on rotationally stocked paddocks. from 1958 to 1988) is 920 mm evenly dis- cows was 420 kg, the breeding season Measurements on continuously stocked tributed throughout the year. Monthly occurred from November to January and paddocks were performed on the same temperatures (average from 1958 to 1988) calves were weaned during April at 6 to 8 sampling dates. The experiment was per- range from 6.8°C in July-August to months of age with an average weight of formed in 1989 and repeated in 1990. 21.8°C in January. Because of the flat 170 kg. Because the physiological status Regression analysis was performed relief and the occurrence of a high water of herds changed throughout the year, the between the seed bank and seedling densi- table, soils belong to the halo-hydromor- stocking rate gradually decreased from 0.8 ty recorded during autumn in 1989 and phic complex and associations (INTA animal unit (AU) ha-1 in March when cows 1990. To analyze the response curves of 1977). Studies were conducted in a hydro- weaned calves in the last month of lacta- seedling density over time under different -1 morphic community, where dominant C4 tion to 0.5 AU ha in winter (from May to grazing treatments and years, a nonpara- grasses were Paspalidium paludivagum July) after weaning of calves. Since March metric split-plot analysis using the (Hitchc. et Chase) Parodi and L e e r s i a 1989, this grazing method was replaced by Kruskal-Wallis rank analysis to test the h e x a n d r a Sw.; dominant C3 grasses were rotational stocking on both farms except among-experimental-units source of varia- Chaetotropis elongata (H.B.K.) Björkm. for paddocks that were maintained under tion and the Friedman statistic for the and Lolium multiflorum Lam.; and domi- continuous stocking as a control treatment. within-experimental-unit variation was nant spring-summer dicots were Rotational stocking was performed by utilized (Potvin and Lechowicz 1990). Alternanthera philoxeroides ( M a r t . ) concentrating the cow-calf herd and mov- Data recorded in the first 3 sampling dates Griseb., Leontodon nudicaulis L. and ing it through a series of 10–12 paddocks were analyzed as repeated measures. In Mentha pulegium L. with an average size of 45 ha. The annual the fourth sampling date, when Italian rye- During the experimental period (March average stocking rate was 1 AU ha-1, vary- grass plants were tillering, tiller density 1989–February 1991) precipitation was ing between 1.3 AU ha- 1 in March to 0.8 was recorded and compared using Tukey’s recorded monthly (Table 1). AU ha-1 in winter (from May to July). This test (P < 0.05). represented an increase of 54% in the average stocking rate compared to contin- Research layouts C3 and C4 grass productivity Two experiments were conducted simul- uous stocking. Occupation period in each This experiment was performed using a taneously. Experiment I was located on a paddock varied between 3 to 15 days and single replicate per grazing treatment, farm (35E°30'S, 57°15'W) 160 km south rest period between 25 to 90 days accord- rotationally or continuously stocking, in of Buenos Aires. It was designed to study ing to the growth rate of forage species. 40 and 60 ha paddocks respectively. whether the increment of Italian ryegrass During the occupation period imposed in Total aboveground biomass was har- germination under rotational stocking led early autumn, nearly all the available bio- vested from March 1989 to May 1991 to an earlier and denser establishment of mass was removed to increase Italian rye- before and after each occupation period in ryegrass than did continuous stocking. grass germination followed by about 90 the rotationally stocked paddock. Thirty Experiment II was conducted on a farm days rest to allow seedling development. 0.25 m2 areas were randomly located and (36°25'S, 59°05'W) 240 km southwest of During the occupation periods imposed in clipped to ground level on each date. In Buenos Aires with the objective of deter- spring only around 50% of the available the continuously stocked paddock, thirty, mining whether rotational stocking was biomass was grassed. This allowed the 1 m2 cages were randomly placed to avoid able to improve winter Italian ryegrass cattle to make the selection and the rye- the access of large herbivores. Biomass productivity compared to continuous grass plants to flower and seed. samples were obtained inside each 0.25 m2 stocking. cage and outside by clipping to ground Germination and establishment of level whenever the rotationally stocked Grazing treatments Italian ryegrass. paddock was grazed. After each clipping, Grasslands of both farms had been con- In late summer (end of February), cages were moved to a new location. tinuously stocked at 0.65 cows ha- 1 b y approximately 2 months after seed shed, Those 30 samples were composited then

484 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 separated randomly in 3 new groups. These samples were hand separated into 6 components (green and standing dead material of C3 and C4 grasses, dicots and grasslikes), oven dried at 70°C and weighed. Within the C3 component, dry matter of Italian ryegrass was weighed separately. Above-ground net primary productivity (ANPP) of C3 and C4 grasses in rotation- ally stocked paddocks was estimated as the difference between the accumulated green biomass for each component at the end of each rest period and the remnant biomass left after the previous occupation period, divided by the number of days of the rest period. In the continuously stocked paddock, ANPP was similarly cal- culated as the difference between the green biomass accumulated within the Fig. 1. Pattern of ryegrass seedling appearance during autumn under continuous or rotation- cages and the existing biomass when the al stocking. cages were placed, divided by the number of days of the period. To reduce the mask- density is plotted in Figure 1. As the 73%) during early autumn shaded a great ing effect in the estimation of senescence repeated measures analysis showed that proportion of seeds. These seeds would in the productivity of each group of grass- the factor “Year” and the interactions germinate only when their period of light- es, any daily increment of standing dead “Year x Grazing” and “Year x Time” were sensitivity is over in late autumn biomass for both compartments was added not significant (Table 2), each point in (Rodriguez et al. 1998), therefore maxi- to the respective green productivity value mum germination was achieved 1 month (Sala et al. 1981). Annual production was later than in rotationally stocked paddocks estimated as the summation of the ANPP Table 2. Non-parametric split-plot analysis of (Fig.1). of each period multiplied by the duration ryegrass seedling density over time under The earlier establishment of Italian rye- of the corresponding period. different grazing treatments and years. grass seedlings under rotational stocking To analyze the overall pattern of ANPP may account for the large differences in of C and C grasses over time and the Sources Statistic P 3 4 ryegrass tiller density between grazing differences in their trends between grazing Among experimental units treatments found in early winter (end of treatments, the multivariate procedure for Year 0.0256 0.8728 June), 3,094 seedlings m- 2 in rotational repeated measures experiments proposed Grazing 8.3077 0.0039 stocking compared to 1,173 seedlings m-2 by Gurevitch and Chester (1986) and the Year x Grazing 0.0476 0.8273 in continuously stocking. Similar values of correction method for unequally spaced Within experimental units tiller density were previously reported measurements proposed by Robson (1959) Time 8.101 0.0411 (Jacobo et al. 1995). Those seedlings, were employed. This procedure was Year x Time 6.001 0.1116 Grazing x Time 0.606 0.8981 established earlier under rotational stock- applied for the first (26 April 1989-10 ing, developed under low competition December 1989) and the second (13 April without being disturbed by grazing during 1990-7 January 1991) sampling period of Figure 1 is the average of data recorded in the following rest period (90 days). This C grasses and for the first (12 July 1989- 3 the corresponding sampling dates of 1989 occured in the wet, temperate autumn sea- 18 March 1989) and second (28 August and 1990. The highly significant effect of son. Nevertheless, a long fall rest (135 1990–25 March 1991) sampling period of factors “Grazing” and “Time” (Table 2) days) without removal of the canopy is not C grasses. A Tukey’s test (P < 0.05) was 4 indicates that grazing treatments modified enough to increase Italian ryegrass density applied to compare annual production of the pattern of Italian ryegrass seedling (Hidalgo and Cahuepé 1991). C and C grasses and total forage produc- 3 4 appearance. In rotationally stocked pad- tion. docks seedlings established earlier and in C3 and C4 grass productivity. a greater abundance than in continuously Grazing treatment resulted in different stocked paddocks (Fig.1). This was par- Results and Discussion patterns of ANPP for C3 grass (Fig. 2). tially due to the promotion of germination Different linear trends were found for the obtained through the removal of the dense first (26 April 1989–10 December 1989) Germination and establishment of C4 grass canopy to a 3.5 cm stubble height and the second (13 April 1990–7 January Italian ryegrass. (CV = 12%) during the grazing period 1991) sampling period (Table 3). The The seed bank of Italian ryegrass performed in early autumn, because greater ANPP values obtained from April recorded in late summer ranged from canopies shorter than 5 cm height allow to September for rotational stocked pad- 13,500 to 31,800 seed m- 2 and showed a sunlight illumination of Italian ryegrass dock in 1989 and 1990 (Fig. 2) resulted in low association (–0.23>r>0.35) with seeds (Deregibus et al. 1994). Conversely almost 2-fold greater annual production seedling density recorded during autumn. in continuously stocked paddocks a (1,377 and 1,135 kg dry matter ha-1 in the The pattern of Italian ryegrass seedling canopy of 8.3 cm stubble height (CV = first and second growing period respec-

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 485 Table 3. ANOVA’s for linear to fifth degree contrasts generated from data of ANPP of C3 and C4 grasses under rotational or continuous stocking.

tively) compared to continuous stocking results, we conclude that earlier seedling This result shows that rotational stocking (610 and 731 kg DM ha- 1). Under both establishment and an increase in Italian may overcome the major negative effect grazing treatments, Italian ryegrass was ryegrass density obtained under rotational of continuous stocking in these grasslands. the major component of this group of stocking, are the major determinants of the Patterns of ANPP of C4 grasses were grasses (avg. 77%). In view of these substantial increase in winter production. different between grazing treatments (Fig. 3), as significantly different linear trends

486 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Although the scope of this experiment must be considered very carefully due to limited replication, the results are useful because it is the first comparison of ANPP under rotational and continuous grazing in this region. We conclude that the earlier establish- ment of Italian ryegrass and the conse- quent greater winter forage production explains the increase of stocking-rate sup- ported under rotational stocking compared to continuous stocking (1.0 vs 0.6 average AU ha- 1). Under rotational stocking the period of low forage availability may be overcome earlier, so cows may lose less weight during winter supporting improved body condition prior to breeding.

Fig. 2. Aboveground net primary productivity of C3 grasses under continuous or rotational stocking. Literature Cited were found for the first (12 July 1989–18 the first (3,504 and 3,860 kg DM ha-1) and March 1989) and the second (28 August in the second year (2,223 and 2,577 kg Biondini, M.E. and L. Manske. 1996. Grazing 1990–25 March 1991) sampling period DM ha- 1). Therefore, grazing treatment frequency and ecosystem processes in a (Table 3). The lower values of ANPP had no relevant impact on overall ANPP, a northern mixed prairie, USA. Ecol. Appl. observed under rotational stocking (Fig.3) 6(1):239–256. well documented fact in grazing system Casal, J.J., V.A. Deregibus, and R.A. resulted in a decrease in annual production literature (Van Poollen and Lacey 1979, of C grasses (2,127 and 1,088 kg DM ha-1 Sanchez, 1985. Variations in tiller dynamics 4 Hart et al. 1989, O'Reagain and Turner and morphology in Lolium multiflorum Lam. in the first and second year) respective to 1992, Biondini and Manske 1996). The vegetative and reproductive plants as affect- continuous stocking (3,250 and 1,846 Kg reduction of annual production under both -1 ed by differences in Red/Far-Red irradiation. DM ha ). This decrease may be a conse- grazing treatments in 1990 (Table 3) was Ann.of Bot., 56:553–559. quence of the increasing competition from due to the lower precipitation registered in Casal, J.J., R.A. Sanchez, and V.A. the greater population of Italian ryegrass that year (Table 1). Deregibus, 1986. The effect of plant density during spring, when ryegrass biomass The distribution of total forage productiv- on tillering: The involvement of R/FR ratio increases exponentially because of repro- and the proportion of radiation intercepted ity (C3 and C4 grasses) was affected by ductive growth. rotational stocking as the changes in C an d per plant. Environ. and Exp. Bot., Because the increase of C grass pro- 3 26(4):365–371. 3 C4 grass production reduced seasonal vari- duction was accompanied by a reduction ability in forage on offer. This contrasts Deregibus, V.A. and M.A. Cahuepé. 1983. Pastizales naturales de la Depresión del in C4 grass production under rotational with Heitschmidt et al. (1987) who reported Salado: utilización basada en conceptos stocking, total annual production of all no difference between rotational and con- grasses was similar among treatments in ecológicos. Rev. de Invest. Agrop. INTA, Bs. tinuous stocking in herbage distribution. As., Argentina, XVIII:47-78. Deregibus, V.A., R.A. Sanchez, and J.J. Casal, 1983. Effects of light quality on tiller production in L o l i u m spp. Plant Phys. 72:900–902. Deregibus, V.A., J. Casal, E. Jacobo, D. Gibson, M. Kauffman, and A. Rodriguez. 1994. Evidence that heavy grazing may pro- mote the germination of Lolium multiflorum seeds via phytochrome-mediated perception of high red/far-red ratios. Funct. Ecol. 8:536–542. Gurevitch, J. and S.T. Chester. 1986. Analysis of repeated measures experiments. Ecol. 67(1):251–255. Hart, R.H., M.J. Samuel, J.W. Waggoner, and M.A. Smith. 1989. Comparisons of grazing systems in Wyoming. J. Soil and Water Conserv. 44(4):344–347. Heischmidt, R.K., S.L. Dowhower, and J.W. Walker. 1987. Some effects of a rotational grazing treatment on quantity and quality of available forage and amount of ground litter. J. Range Manage. 40(4):318–321. Fig. 3. Aboveground net primary productivity of C4 grasses under continuous or rotational stocking.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 487 Hidalgo, L.G. and M.A. Cauhépé. 1991. O'Reagain, P.J. and J.R. Turner. 1992. A n Sala, O.E., V.A. Deregibus, T. Schlichter, Effects of seasonal rest in aboveground bio- evaluation of the empirical basis for grazing and H. Alippe. 1981. Productivity dynamics mass for a native grassland of the flooding management recommendations for rangeland of a native temperate grassland in Argentine. Pampa, Argentina. J. Range Manage. in southern Africa. J. Grassl. Soc. South. Afr. J. Range Manage. 34:48–51. 44(5):471–475. 9(1):38–49. Soriano, A. León, R.J.C., O.E. Sala, R.S. INTA. 1977. La Pampa Deprimida. Potvin, C. and M. Lechowicz. 1990. The sta- Lavado, V.A. Deregibus, M.A. Cahuepé, Condiciones de drenaje de sus suelos. tistical analysis of ecophysiological response O.A. Scaglia, C.A. Velazquez, and J.H. Publ.154. Departamento de Suelos, Instituto curves obtaines from experiments involving Lemcoff. 1991. Rio de la Plata grasslands. repeated measures. Ecol. 7(14):1389–1400. Nacional de TecnologÌa Agropecuaria, I n : RT Coupland (ed). Ecosystems of the Buenos Aires, 167 pp. Rodríguez, A.M., E.J. Jacobo, and V.A. Deregibus. 1998. Germination behaviour of world. Elsevier, Amsterdam, 367–407 pp. Jacobo, E., A. Rodríguez, and V.A. Van Poollen, H.W. and J.R. Lacey. 1979. Deregibus. 1995. Effect of the nature and Italian ryegrass in flooding pampa range- lands. Seed Sci. Res. 8(4):521–528. Herbage response to grazing systems and stock- intensity of autumn grazing events on estab- ing intensities. J. Range Ma n a g e . 3 2 ( 4 ) : 2 5 0 – 2 5 3 . lishment of Italian Ryegrass in humid tem- Robson, D.S. 1959. A simple method for con- perate rangelands. Proc. Fifth International struction of orthogonal polinomials when the independent variable is unequally spaced. Rangel. Congr., Salt Lake City, Utah. Biometrics. 15: 187–191.

488 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 J. Range Manage. 53: 489–498 September 2000 Flow processes in a rangeland catchment in California

ROHIT SALVE AND TETSU K. TOKUNAGA

Authors are principal research associate and staff scientist, Earth Sciences Division, Lawrence Berkeley National Laboratory, I Cvclotron Road, Berkeley, Calif. 94720. At the time of the research, the senior author was a Ph.D candidate, Department of Environmental, Science, Policy and Management, University of California, Berkeley, Calif.

Abstract Resumen

Emerging hydrology-related issues in California grasslands have Los problemas que están surgiendo en relacion a la hidrología directed attention towards the need to understand subsurface de los pastizales de California, se han enfocado hacia la necesi- water flow within a complex, dynamic system. Tensiometers and dad de entender el flujo subterráneo del agua dentro de un sis- neutron probes evaluated the subsurface hydrology of a rangeland tema complejo y dinámico. En una área de captación del pastizal catchment. Hydrological processes within the catchment varied se evaluó la hidrología subterránea mediante el uso de tensiómet- both in space and time. Spatial variability was evident along the ros y dispersores de neutrones. Los procesos hidrológicos dentro vertical profile and between the catchment slopes. Temporal vari- del área de captación variaron en espacio y tiempo. La variabili- ability in processes coincided with the seasons (i.e., wet winter, dry dad espacial fue evidente a lo largo del perfil vertical y entre las summer, and spring). From a water-balance equation developed pendientes del área de captación. La variabilidad temporal de los for the catchment, we determined that there was significant vari- procesos coincidió con las estaciones (esto es, invierno húmedo, ability both spatial and temporal in the amount of soil moisture verano y primavera secos). A partir de una ecuación del balance lost to evapotranspiration and deep seepage. During the 16 month hídrico desarrollada para esta área de captación, determinamos monitoring period there was a total of 50 cm of rainfall that fell in que hubo una variabilidad temporal y espacial significativa en la the catchment of which 9–55 cm was lost to evaporation and 37–79 cantidad de humedad del suelo perdida por evaporación y fil- cm to deep seepage. A simple deduction of the losses (evaporation tración. Durante los 16 meses del período de monitoreo la pre- and deep seepage) from the input (rainfall) shows that all moni- cipitación recibida en el área de captación fue de 50 cm de los tored locations had a substantial decrease in the amount of water cuales de 9–55 cm se perdieron por evapotranspiración y de that was stored in the soil profile. 37–39 cm se perdieron por filtración. Un simple substracción de las perdidas (evapotranspiración y filtración) de lo recibido (llu- via) muestra que todas la localidades monitoreadas tuvieron un Key Words: California rangelands, subsurface flow, water budget decremento un substancial enla cantidad de agua que fue alma- cenada en el perfil del suelo. A common omission in rangeland hydrology studies has been a rigorous treatment of the subsurface component of the hydrologic cycle. A need exists for the understanding of flow path dynamics In recent years both physical and chemical methods have been (surface and subsurface) and the spatial and temporal variations used to estimate recharge in arid and semiarid regions. Physical in the hydrology of rangelands. methods have included both direct and indirect measurements. The movement of soil water in semiarid and arid climates has While direct measurements of deep seepage have mainly involved been investigated since at least 1949 when Maxey and Eakin the use of lysimeters (Allen et al. 1991, Gee et al. 1993), indirect attempted to measure ground water recharge in the desert basins physical methods have included the use of soil water balances of Nevada. By assuming basin recharge to equal basin discharge (e.g. Rushton and Ward 1979), the zero flux plane method they concluded that in areas with less that 20 cm yr-1 of precipita- (Wellings 1984), and estimates of water fluxes from solutions to tion, recharge was essentially negligible. Nixon and Lawless either Darcy’s law or Richards’ equation (Sophocleous and Perry (1960) monitored the movement of soil moisture from rainfall up 1985, Stephens and Knowlton 1986). Johnston (1987) and Sharma to depths of 6.0 m in a region 250 km northwest of Los Angeles, and Huges (1985) demonstrated the considerable variation in the California. Using a neutron probe and a nest of sparsely distrib- rates of local recharge over a scale of a few meters in many soil uted tensiometers (up to a depth of 1.0 m), they concluded that types. Assessment of this spatial variability in recharge has been 31% of the rainfall migrated to the deep profile over a period of attempted with frequency-domain electromagnetic (EM) or tran- 240 days following the last rainfall event. Holmes and Colville sient electromagnetic methods (Allison et al. 1994). Table 1 sum- (1970), while investigating the water balance of a grassland in marizes some estimates of deep seepage for semiarid regions. southern Australia with lysimeters and neutron probes, deter- This paper presents results from the investigation of the subsur- mined that less than 10% of the 63 cm of precipitation recorded face hydrology of a rangeland watershed in California. The cen- in 5 years recharged a low lying water table. tral thesis presented is that seasonal changes in the hydrologic cycle cause important variations in the flow dynamics in range- land catchments in California. In this paper, observations of soil Thanks to Dan Hawkes, Nigel W. Quinn and Peter Zawislanski for their careful moisture content and associated energy levels are used to develop review of this manuscript. Manuscript accepted 19 Dec. 1999. a water balance for a catchment dominated by annual grasses.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 489 Table 1. Estimates of deep seepage in semi-arid regions. Source Stephans et al. 1986.

Author Method Location Estimate recharge rate Enfield et al. 1973 Soil water potential Washington state (south central) 1 cm yr-1 Klute et al. 1972 Neutron moisture logging High Plains, Colorado 0 cm yr-1 Dincer et al. 1974 Tritium tracer Saudi Arabia 25% annual precipitation Allison et al. 1985 Chloride concentration South Australia 1.4 cm yr-1 Meyboom 1966 Water level hydrographs Saskatchewan, Canada 7.5% annual precipitation Boyle and Salem 1979 Temperature profiles Illinois (north west) 7.8–31 cm yr-1 Maxey and Eakin 1949 Steady State flow Nevada 0 cm yr-1 Watson et al. 1976 Steady State flow Nevada 3.4% annual precipitation Sammis et al. 1982 Hydraulic gradients Phoenix, Arizona 18 cm yr-1 Sammis et al. 1982 Temperature profiles Phoenix, Arizona 9 cm yr-1 Sammis et al. 1982 Tritium tracer Phoenix, Arizona 40 cm yr-1 Stephen and Knowlton 1986 Hydraulic gradients Socorro, New Mexico 4 cm yr-1

Materials and Methods with elevations ranging from 230 m above ly in the winter months. While most cli- mean sea level (msl) in the valley bottom matic parameters remain relatively con- to nearly 360 m above msl on the ridge stant from year to year, precipitation in the Study Site top. The slopes range from 2 to 75%, and region shows variability both within and Soil water potential and content was the catchment contains at least 2 perennial between seasons (Fig. 2). measured within a watershed in north-cen- springs. Annual grasses dominate the veg- Thirteen sites were located along the tral California over a period of 16 months. etation cover in the catchment. Sporadic slopes for measurement of soil moisture The data was used to identify processes clusters of predominantly coyote bush content and potentials. These sites were influencing the hydrology of the catch- shrubs (Baccharis pilularis, D.C.) are selected to represent the 3 dominant vege- ment, and to determine quantitatively, the found scattered throughout the catchment. tation types found in the catchment (annu- moisture status along vertical soil profiles The other major vegetation type is the al grasses, baccharis shrubs, and oak trees) at specific locations. California oak, which is restricted to the and 3 broad elevations viz. low, medium, The watershed is located within the higher elevations close to the ridge and is and high. Russell Tree Farm, a 115-hectare (ha) dominated by coast live oak (Q u e r c u s Water content of soils was determined research station located, in Contra Costa a g r i f o l i a, Nee). The Mediterranean cli- by the neutron-probe method (Gardner County (Lat. 37° 54' N, Long 122° 03' W), mate of the region results in warm to hot 1986). A Campbell Pacific Nuclear Model California (Fig. 1). The amphitheater- summers and winters that are relatively 503 with a 50 millicurie americium-beryl- shaped catchment has an area of 20 ha, mild. Precipitation occurs as rainfall main- lium source neutron probe was used. The probe was calibrated by a gravimetric analysis that utilized in situ samples from the access tube holes. Thirteen neutron probe access tubes (Sites 1–13) were installed in February 1993. In most cases, the depth of augering was limited by bedrock. The deeper pro- files monitored ranged from 3.75 to 5.70 m in depth and included 5 sites (6–9, and 13) located on the southeast corner of the catchment. Of the 7 shallow profiles, Sites 1 and 2 extended to 2.00 m, while the remaining were restricted to 1.20-m depths. During installation of the access tubes for the neutron probe, a 0.05-m diameter hole was augered, into which a 0.04-m inner diameter, schedule 40 PVC pipe was inserted. The PVC pipe in each hole extended 0.15 m above the ground surface. Bentonite clay was poured around the PVC pipe, 0.10 m below the surface to prevent surface run-off water from flow- ing down the sides of the tube. Soil water potential along the catchment slopes was determined using tensiometers. Thirteen nests of tensiometers (Sites 1–13) were installed at a distance of about 1.0-m from the neutron probe access tubes (Fig. Fig 1. Location of the Russell Tree Farm with monitored sites.

490 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 powder and pellets) was packed to a height of 0.10 m and wetted to provide a vertical hydraulic seal. Above this seal native soil was packed to the original den- sity. Water was injected into the tensiome- ters through narrow diameter tubing which extended down to the ceramic cup. At each site, tensiometers were installed in nests such that the ceramic tips were located at approximately 0.25-m intervals down to a depth of 1.0 m from the surface. Between 1.0 and 2.0 m, tensiometers were installed at 0.5 m intervals, and at greater depths the interval was increased to approximately 1.0 m (Fig. 3b). Of the 13 locations, 6 had tensiometers located at depths greater than 3.0 m (Sites 6–9, 12, and 13). The shallowest nest, Site 3, extended to a depth of 1.0 m, and the deep- est, Site 6, extended to a depth of 6.0 m. Hydraulic head measurements were determined as the sum of gauge pressure at the ceramic tip and the elevation of the ceramic tip relative to the ground surface at the nest. Thus, the hydraulic head mea- surements in each nest are referenced to the local elevation of the nest. Gauge pres- sure readings were taken using a portable pressure transducer, commercially referred to as the Tensimeter (Soil Measurement Fig. 2. (a) Annual rainfall at Walnut Creek located 10 Km east of Russell Tree Farm. (b) Systems, Tucson, Ariz., 2 Marthaler et al. Variability in monthly rainfall 10 Km east of study site. 1983). Water potential readings were taken at weekly intervals during the winter 3a). A 0.10-m diameter soil auger was This soil was repacked to a density similar and monthly intervals late in the summer. used to create the hole into which ten- to that of the undisturbed native soil and siometers were inserted. The soil removed filled to approximately 0.10 m above the Water balance estimates during augering was collected in 0.15-m center of the ceramic tip (The original Parameters used to evaluate the water sections and stored in sealed plastic bags. packing density was achieved by reintro- balance of a location estimate how much The area around the ceramic tip of each ducing the augered soil from the 0.15 m usable energy and water are available (for tensiometer was back-filled with native sections back into the same zone). Above evaporation and transpiration), how much soil removed from that particular depth. this backfill, bentonite (as a mixture of evaporation demand is met by available water, and how much water is usable excess. In its simplest form, the equation describing annual conservation of water mass for a unit watershed can be written as: P - Q - E = DS (1) Where P is the average annual precipita- tion, Q is the average annual discharge, E is the annual evapotranspiration, and DS is the change in the amount of moisture stored. The signs indicate whether water is entering (positive) or leaving (negative) the system. If an average is taken over many years of record, then the DS term can be assumed to equal zero, and Equation [1] becomes (Freeze and Cherry 1979): P = Q + E (2) The terms on the right in Eq. [1] can be modified to various levels of detail, Fig. 3. Schematic of monitoring station at Site 6: (a) the location of nested tensiometers with depending on the important characteristics respect to the neutron probe access tube and (b) vertical extent of tensiometer and neutron of a region or on the available database. probe measurements.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 491 For example, Eagleson (1978) presented the water-balance equation for a unit watershed as:

PA = QSA + QGA + ETA + DSS + DSg (3) Where PA is the annual (seasonal) precipi- tation, QSA is the annual (season) surface runoff, QG A is the annual (seasonal) sub- surface flow through the watershed, ETA is the annual (seasonal) total evapotranspira- tion, DSS is the annual (seasonal) change in surface storage, and DSG is the annual (seasonal) change in soil moisture and groundwater storage. For rangeland catchments in central California, large annual fluctuations in precipitation (Fig. 2a) results in large vari- ability in the amount annual of recharge, which in turn results in large changes in the amount of moisture stored in the soil profile (DSG). In the absence of surface Fig. 4. Migration of the zero-flux boundary in the soil profile at Sites 9 and 13. Vertical runoff (QS A) and surface storage (SS) on arrows indicate the direction of the hydraulic gradient. catchment slopes, Eq. [3] can be re-written for these rangelands as: seepage was estimated by assuming that of the profile at all sites were directed P = Q + E + DS (4) positive flow gradients lying below the towards the surface. A small rainfall event A GA TA G zero flux plane directed flow into the deep (12 mm) during the third week of April, In this study the moisture status for profile. For example, in Site 9 the zero which was spread over a period of 48 locations within the catchment was deter- flux plain migrated to a depth of 5.0 m hours, did not change this trend. By the mined using a modified version of Eq.[1] over a period of 7 months while in Site 13 end of April, most of the sites were losing where the dynamic water content for a it was located 3.0 m below the ground sur- water to the surface from the top 0.5 m, vertical soil profile was evaluated using face (Fig. 4). while deeper in the soil, flow was directed the equation: into the profile. PP = DSE + DSQ (5) Observations During the winter of 1993–94, the first rainfall (12 mm) was recorded in mid- Here water infiltrating the surface (PP) The data collected includes measure- was partitioned as surface evapotranspira- ments of soil matric potential from 13 sites October. This event was not detected as tion or deep seepage based on the direc- and water content from 12 sites within the soil moisture changes by any of the ten- tion flow gradients as indicated by the slopes, and from 16 locations along the siometers located in the catchment. The position of a zero-flux boundary (Fig. 4). catchment valley. Site 12 was not included second event (20 mm) occurred approxi- (The zero-flux boundary is determined in the analysis because of installation mately 1 month later was insufficient to along the vertical profile as the location errors in the neutron access tube. bring potentials at the near surface within where hydraulic gradients are directed, in Monitoring began in April 1993, follow- functioning range of the tensiometers. At opposite directions, away from the point). ing a winter during which the recorded the end of November, 2 storms of 42 mm total precipitation briefly increased All changes in moisture content (DSP) rainfall in the region was the highest in at above the zero-flux boundary were least the past 7 years( 324 mm), and con- hydraulic potentials, but in the next 10 days these values again decreased below assigned to ET (DSE), while all losses tinued through September 1994, a year below the boundary were assigned to deep which was much drier (Fig. 2). the tensiometer range. Surface water recharge deep enough to penetrate to seepage (DSQ). Since soil moisture content was moni- depths below 0.20 m was observed follow- tored at monthly intervals and hydraulic Recharge events in the catchment ing 5 rainfall clusters which occurred gradients were measured at weekly inter- Effectively saturated conditions close to between December and April. The first vals, the monthly changes in moisture con- the surface were evident in the tensiometer recharge event was initiated after 74 mm tent were proportioned according to the measurements when monitoring began in of precipitation over a period of 6 days in length of time the gradient direction was April 1993, and recharge along the vertical mid-December. The second event observed. For example, if the change in profile of the monitored sites was indicat- occurred almost 45 days later following a moisture content at a given depth, over a ed by positive hydraulic gradients (Table 6-day period in late January when 66 mm single month, was 10.0 cm and hydraulic 2). Following the rains in early April there of rain occurred. A week later, 40 mm of gradients were positive for 3 of the 4 was a 2-week precipitation free period rain, which occurred over 3 days, provided weeks, it was assumed that 7.5 cm of the when tensiometers close to the surface sufficient moisture for the third recharge moisture was lost to deep seepage, while began to record decreases in hydraulic event. The largest rainfall of the season 2.5 cm was released as ET to the surface. potentials. At depths below 0.2 m from the (90 mm), recorded over a 5 day period in From data collected on moisture content surface, potentials also began to decrease late February resulted in the fourth and associated potentials in the soils the slower such that by the second week of recharge event. During this period, pro- amount of moisture contributing to deep April hydraulic gradients in the top 0.5 m files 1.5 m below the surface were close to

492 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Table 2. Hydraulic gradients measured along the vertical profile of 3 sites. Bold arrows indicate saturation, the deepest observed during the direction of flow. Clear arrow indicate hydrostatic conditions. entire wet season. The final recharge occurred 2 months later when at the end of Site 6 April a cluster of rainfall events provided 45 mm of precipitation. Depth (m) 4/1/93 7/21/93 2/25/94 7/18/94 Thus during the wet season which extended over 206 days, there were only 5 significant periods (totaling 21 days, and 71% of the season’s precipitation) when moisture deposited at the soil surface migrated below the top 0.50 m of the soil profile. Each of these periods had more than 40 mm of rainfall deposited. In 4 of these events, groundwater recharge (as detected in the magnitude and direction of hydraulic gradients) significantly altered the moisture profile of the top 1.0 m of the soil. During a single cluster of events in late February, 20% of the seasons’ rainfall was recorded, near saturated profiles were detected up to 1.5-m depths. However at greater depths the changes in potential val- Site 9 ues were small during the entire winter season.

Hydrologic Processes on Catchments Slopes. The vertical soil profiles of the 12 moni- tored sites showed a sinusoidal pattern of wetting and drying (Fig. 5). Early in May 1993 all the sites recorded high amounts of moisture at all depths (~0.15 to 0.35 c m3 c m- 3). Over the next month all the shallow profiles recorded losses in mois- ture content, and in the ensuing months moisture losses from the profiles contin- ued, but at decreasing rates. With the start of the winter rains, the profiles began to moisten with increases in wetness occur- ring during the first wet month (i.e. December 1993). In the next 2 months, Site 13 when the bulk of the season’s rain fell, small increases in soil moisture were detected in all profiles, but these were much smaller than those observed early in the winter (~0.20 to 0.25 cm3 c m - 3) . Shortly after the wet season ended in early March 1994, the shallow profiles began to record losses in moisture. The largest decreases were observed in April, and smaller reductions occurred in the follow- ing precipitation free months. This drying pattern was similar to that of the previous year with the exception that the drying process in 1994 began almost 60 Julian days earlier (Fig. 5). At each of the monitored sites, the total moisture lost from the near surface profile during the summer of 1993 was replen- ished during the following winter. Similar amounts of moisture were then lost from the profiles by the end of August 1994. In essence, these profiles reached a fixed

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 493 In the subsequent months this pattern was repeated such that the center of the drying zone was located at a depth of 3.3 m in Site 6 and 2.0 m in Site 8. This ‘step-wise’ drying pattern is significantly different from that observed in Site 9 and 13, where the drying was relatively uniform along the entire profile (Fig. 6b). During the following winter (1993–1994), the depths to which the deep profiles wetted were less than that observed at the end of the previous winter rains (Fig. 7). Except for Site 13, in which the wetting front migrated to 2.4 m, wet- ting in the remaining sites was restricted to depths less than 2.0 m. Immediately fol- lowing the last winter rains in March 1994, all profiles began to lose moisture along the entire wetted length. Unlike the previous summer, where there was ‘step- wise’ drying in some profiles, moisture losses were recorded at near equal rates at all the wetted lengths. In the deeper sec- tions, which were not wetted by the winter rains, small decreases in moisture content continued to occur during the summer. For the first 6 dry months the loss in moisture in all the sites was much greater than the increases recorded during the wet period. Comparing the moisture levels of differ- ent sampling dates with those observed in early May 1993, provides a measure of the ‘relative wetness’ within the catchment (Table 3). The relative wetness recorded in the deep profiles during May 1993 was never exceeded and only 1 profile (7N) Fig. 5. Changes in volumetric moisture content at various depths along the vertical profiles registered levels similar to those at the of (a) Site 6 and (b) Site 8. start of the recording period. The peak in upper and lower limit in storing soil mois- ture towards the end of each season irre- spective of the amount of rainfall received the previous wet season. At depths greater than 2.0 m, the amount and depth to which changes in volumetric moisture content occurred varied considerably with each profile showing a distinct response to sea- sonal changes in precipitation. Large loss- es in moisture, following a wet winter (1992–93), extended beyond the 2.0-m depth in 4 of the 5 deep sites. The single exception, Site 7, recorded losses in mois- ture, which were largely restricted to the top 1.9 m of the profile. There were 2 distinct patterns of drying observed in the 5 deep profiles. At Sites 6, 7, and 8 drying began at the near surface zones early in the summer and migrated downwards as the summer progressed (Fig. 6a). In this downward migration, the Fig. 6. Patterns of drying in 2 deep sites (6 and 9) during the summer of 1993. (a) At Site 6 near surface zones reached a critical mois- drying was stepwise while (b) at Site 9 drying was relatively uniform along the entire ver- ture level early in the summer after which tical soil profile. drying was observed lower in the profile.

494 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 Table 3. Percentage of moisture present relative to early March 1993 along the vertical profile of monitored sites.

Site 1 May 1993 1 Jul. 1993 1 Sept. 1 Nov. 1993 1 Jan. 1993 1 Mar. 1994 1 May 1994 1 Aug. 1994 ------(%) ------1* 100 82 79 77 95 102 91 82 2* 100 83 74 72 86 98 85 67 3* 100 70 67 64 101 108 80 57 4* 100 67 65 62 86 108 84 61 5* 100 78 74 71 90 100 85 70 10* 100 68 59 55 92 110 67 53 11* 100 75 73 69 86 109 92 65 6 100 85 73 71 77 84 80 69 7 100 88 88 88 97 100 93 83 8 100 92 89 83 88 89 82 77 9 100 91 85 79 81 88 83 75 13 100 88 86 84 86 92 84 75 *Shallow Profiles moisture levels was in early March 1994, difference in wetness between these 2 sites near-surface profiles began to dry as the when the soil wetness ranged between remained fairly constant. In the other 3 summer progressed, a zero-flux boundary 84–100% of the May 1993, values. The deep sites the relative difference in wet- migrated into the deeper zones (Fig. 5). moisture content at all sites early in ness, however, continued to change at dif- The zone above the boundary had a net August 1994 had reached levels lower ferent times of year. Among the shallow negative gradient, resulting in moisture than those recorded prior to the start of the profiles the pattern of changes in moisture moving towards the surface while the pro- wet season in November 1993 indicating content were similar to that observed for file below the boundary continued to that the catchment was already drier than the deeper profiles. release moisture to deep seepage. it had ever been in the previous year. Early in the summer of 1993, when Within and among the 3 zones dominat- Spatial variability in total soil water steady vertical flow in the profile (which ed by transpiration from grasses and sur- content was detected both among and originated as surface infiltration) ceased, face evaporation, transpiration from within all 12 profiles. Of the deep profiles the contribution to deep seepage was shrubs and trees, and deep seepage, monitored, Site 7 consistently contained through the draining of the soil profile. respectively, continuous fluctuations the most water per unit volume of soil This draining process began close to the occurred in the magnitude and directions while Site 8 was always the driest (Fig. surface and as the summer progressed, of hydraulic gradients, resulting in the 8). Throughout the monitoring period the migrated deeper into the profile. As the continuous redistribution of moisture in

Fig. 7. Extent of migration of the wetting front in 4 sites (6, 7, 9, and 13) following a wet winter in 1993 and a significantly drier winter in 1994.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 495 movement of soil moisture between the surface and shrub/tree root zone. Within the shallow profile sites (where the bedrock lies between 1.0 and 2.0 m from the surface) there was a different hydrolog- ic dynamic, with moisture migrating to the deeper profile in the winter. As the winter ended, the soil began to dry from gravita- tion drainage and from losses to ET. Once the grasses died, moisture losses were reduced, and the subsurface profile approaches a near-hydrostatic condition. Toward the end of the summer, most move- ment of soil water in the slopes was restrict- ed to the deep profiles. In the shallower sec- tions of the deep profile, the roots of tran- spiring trees and shrubs drew water from the soil. In the deeper profile moisture, con- Fig. 8. Monthly changes in soil moisture content along the vertical profile of Sites 6, 7, 8, 9, tinued to be lost to deep seepage. and 13. Results the soil (Table 2). Early in the summer, zones, where the soil was drier from tran- Parameters used in determining the site- large negative gradients close to the sur- spiration losses that occurred before the specific water balance for Site 9 are sum- face resulted in the movement of water grasses died. As this moisture was being marized in Figure 9. Similar parameters from below the grass root zone towards the redistributed, the continuous transpiration were used in the other sites for which the surface. However, as the annual grasses losses in the shrub/tree root zone allowed dynamic moisture content was evaluated. dried and mulch covered the surface, little moisture from the near-surface profile and In all 6 sites, for a brief period between of the soil moisture was lost to the atmos- the deeper profiles to migrate toward this January and February 1994, the amount of phere. A significant portion of soil mois- dry zone. This shifting of energy gradients precipitation recorded was greater than the ture was relocated within the grass root in the profile resulted in a continuous calculated ET and seepage losses during

Table 4. Changes in moisture content at 6 locations resulting from precipitation, ET and deep seepage.

May–July ‘93 Aug–Oct ’93 Nov ‘93–Jan ‘94 Feb–Apr ‘94 May–July ‘94 Cummulative Site 2N ------(cm) ------Rainfall 0 1 21 25 3 50 Change water content –13 –2 9 –3 –9 –17 ET loss –13 –2 0 –7 –9 –31 Deep seepage loss 0 –1 –12 –21 –3 –37 Site 6N Rainfall 0 1 21 25 3 50 Change water content –33 –12 12 1 –17 –50 ET loss –22 –10 0 –4 –20 –55 Deep seepage loss –12 –4 –9 –20 –1 –45 Site 7N Rainfall 0 1 21 25 3 50 Change water content –17 –1 0 –11 –14 –43 ET loss –10 –1 0 –6 –15 –32 Deep seepage loss –7 –1 –21 –30 –3 –62 Site 8N Rainfall 0 1 21 25 3 50 Change water content –6 –8 4 –5 –4 –20 ET loss –5 –3 0 –6 –4 –18 Deep seepage loss –2 –6 –16 –24 –3 –51 Site 9N Rainfall 0 1 21 25 3 50 Change water content –16 –15 5 1 –12 –37 ET loss –3 –1 0 –1 –3 –8 Deep seepage loss –13 –15 –15 –23 –13 –79 Site 13N Rainfall 0 1 21 25 3 50 Change water content –15 –3 4 –5 –10 –28 ET loss –8 –2 –1 –7 –10 –28 Deep seepage loss –8 –2 –15 –23 –3 –51

496 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 Fig. 9. Example of parameters used to calculate the (a) water balance Fig. 10. a. Cumulative moisture losses to ET at Sites 2, 6, 7, 8, 9, at and (b) cumulative water losses and gains (for Site 9). and 13. b. Cumulative moisture losses to seepage at Sites 2, 6, 7, 8, 9, and 13. the same period. For the remaining 1993, deep seepage continued at a steady tion of the losses (evaporation and deep 14–month period, the monthly net mois- rate throughout the monitoring period. By seepage) from the input (rainfall) shows ture losses were much greater than net the end July 1994, the largest amount of that all monitored locations had a substan- gains in all the sites. However, among moisture lost to deep seepage was at Site 9 tial decrease in the amount of water that sites where the monthly precipitation was (79 cm) and the least was at Site 2 (37 was stored in the soil profile. assumed to be the same (since they were cm). The pattern of ET losses from these 5 located in the same catchment), the sites was similar to that observed at Site 2. monthly losses to ET and seepage varied The largest ET losses were recorded in Literature Cited significantly (Table 4). Site 2 (65 cm) and the least were at Site 9 At Site 2, ET losses were large (13 cm) (<10 cm), with the remaining sites show- Allen, R. G., T.A. Howell, W.O. Pruitt, I.A. early in the summer of 1993, and then ing losses ranging between 30 and 50 cm. Walter and M.E. Jensen. 1991. Lysimeters gradually ceased over the next 6 months for evapotranspiration and environmental before increasing again in February 1994 measurements. In Pro. Int. symp. Lysimetery Conclusions Honolulu, Hawai. 32–25 July 1991. Amer. until the end of the monitoring period Solutions to a simple water-balance Soc. Civ. Eng. New York. (Fig. 10a). Deep seepage was detected equation, comprising of precipitation, Allison, G.B., W.J. Stone, and M.W. Hughes. with the start of the winter rains in 1994, evapotranspiration, runoff, and storage as 1985. Recharge in karst and dune elements of and observed to continue until the end of the key parameters, suggest that there was a semi-arid landscape as indicated by natural April 1994 (Fig. 10b), during which time significant variability both spatial and isotopes and chloride. J. of Hydrol. 76:1–25. 33 cm of water was lost from this vertical Allison, G.B., G.W. Gee, and S.W. Tyler. temporal in the amount of soil moisture 1994. Vadose-zone techniques for estimating profile. By late July1994, there was a lost to evapotranspiration and deep seep- groundwater recharge in arid and semiarid cumulative loss of 68 cm of water to ET age. During the 16-month monitoring peri- regions. Soil Sci. Soc. Amer. J. 58:6–14. and deep seepage at this location. od there was a total of 50 cm of rainfall Boyle, J.M. and Z.A. Saleem. 1979. In the remaining 5 sites, deep seepage that fell in the catchment. Measurements Determination of recharge rates using tem- losses were observed throughout the moni- of water potential and soil moisture con- perature-depth profiles in wells. Water Resour. Res. 15(6):1616–1622. toring period. In 4 of these sites (6, 7, 8, tent at various locations within the catch- and 13) the seepage losses were also Dincer, T., A. Al-Murgin, and U. Zimmerman. ment suggest that during this time losses 19 7 4 . Study of infiltration and recharge through largest (between 20–30 cm) immediately to evaporation ranged from 9 to 55 cm sand dunes in arid zones with special reference after the wet season in 1994. At Site 9, while those to deep seepage ranged to the stable isotopes and thermo-nuclear tri- except for a brief period in November between 37 and 79 cm. A simple deduc- tium. J. of Hydrol. 23: 79–109.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 497 Eagleson, P.S., 1978. Climate, soil, and vege- Klute, A., R.E. Danielson, D.R. Linden, and Sammis, T.W., D.D. Evans, and A.W. tation: 6. Dynamics of the annual water bal- P. Hamaker. 1972. Ground water recharge Warrick. 1982. Comparison of methods to ance, Water Resour. Res. 14(5):749–764. as affected by surface vegetation and man- estimate deep percolation rates. Water Enfield, C.G., L.J.C. Hsieh, and A.W. agement. Colorado State University, Resour. Bull. 18(3):465–470. Warrick. 1973. Evaluation of water flux Colorado Water Resour. Res. Inst. Compl. Sharma, M.L. and M.W. Hughes. 1985. above a deep water table using thermocouple Rep. No. 41. Groundwater recharge estimation using chlo- psychrometers. Soil Sci. Soc. Amer. Proc. Marthaler, H.P., W. Vogelsanger, F. ride, deuterium and oxygen-18 profiles in the 37:968–970. Richard, and P.J. Wierenga, 1983. A pres- deep coastal sands of Western Australia. J. Freeze, R.A., and J.A. Cherry, 1979. sure transducer for field tensiometers. Soil Hydrol. 81:93–109. Groundwater. Prentice Hall Inc., New Jersey. Sci. Soc. of Amer. J. 47:624–627. Sophocleous, M. and C. A. Perry. 1985. Gardner, W.H., 1986. Water Content. In Maxey, G.B. and T.E. Eakin. 1949. G r o u n d Experimental studies in natural groundwater Methods of soil analysis, Part 1. Ed. A. water in White River Valley, White Pine, recharge dynamics: The analysis of observed Klute, Agronomy Monograph No. 9, 2nd edi- Nye, and Lincoln Counties, Nevada. Nevada recharge events. J. Hydrol. 8:297–332. tion, Amer. Soc. Agron., Madison, Wisc. State Engineer, Water Resour. Bull. No. 8. Stephens, D. B. and R. Knowlton. 1986. Soil Gee, G.W., P.J. Wierenga, B.J. Andraski, Meyboom, P. 1966. Esitmates of ground water water movement and recharge through sand M.H. Young, M.J. Fayer, and M.L. recharge on the Praisres. Symp. on Hydraulic at a semi-arid site in New Mexico. Water Rockhold. 1993. Variations in water balance Resour. of Canada. Water Resour. of Canada, Resour. Res. 22:881–889. and recharge potential at three western desert University of Toronto Press. Watson, P., P. Sinclair, and R. Waggoner. sites. Soil Sci. Soc. Amer. J. 58:63–72. Nixon, P.R. and G.P. Lawless. 1960. 1976. Quantitative evaluation of a method Holmes, J. W. and J. S. Colville. 1970. Translocation of moisture with time in unsat- for estimating recharge to the desert basins of Grassland hydrology in a karst region of urated soil profiles. J. of Geophys. Res. Nevada. J. of Hydrol. 31: 335–337. Southern Australia. J. of Hydrol. 10:38–58. 65(2):655–661. Wellings, S. R. 1984. Recharge of the Chalk Johnston, C. D. 1987. Preferred water flow Rushton, K.R. and C. Ward. 1979. The esti- aquifer at a site in Hampshire, England, 1. and localized recharge in a variable regolith. mation of groundwater recharge. J. Hydrol. Water balance and unsaturated flow. J. J. Hydrol. 94:129–142 41:345–361. Hydrol. 69: 259–273.

498 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 J. Range Manage. 53: 499–505 September 2000 Comparison of 3 techniques for monitoring use of western wheatgrass

LACEY E. HALSTEAD, LARRY D. HOWERY, AND GEORGE B. RUYLE

Authors are former graduate research assistant, assistant rangeland management specialist, and rangeland management specialist, School of Renewable Natural Resources, The University of Arizona, Tucson, Ariz., 85721. Current address for senior author is Nuisance Wildlife Coordinator, New Mexico Department of Game and Fish, 408 Galisteo, Santa Fe, N.M. 87504.

Abstract Resumen

Forage use data can help rangeland and wildlife managers make Los datos acerca de la utilización del forraje pueden ayudar a informed decisions. However, managers need to know if forage use los manejadores de pastizales y de fauna silvestre a tomar deci- techniques that are commonly used to estimate ungulate herbivory siones bien fundamentadas. Sin embargo, los manejadores nece- under field conditions produce comparable results. The objective sitan saber si las técnicas que comúnmente se utilizan en campo of this 2-year study was to directly compare forage use measure- para estimar el uso de forraje de los ungulados producen resulta- ments obtained via the paired-plot method and 2 height-weight dos comparables. El objetivo de este estudio de 2 años fue com- methods (using on-site height-weight curves and the pre-estab- parar directamente las medidas de uso de forraje obtenidas lished United States Forest Service height-weight gauge). In June, mediante el método de parcelas apareadas y 2 métodos de altura- July, and October of 1997 and 1998, we measured forage use of peso (usando curvas de altura-peso y el medidor altura peso western wheatgrass (Pascopyrum smithii Rydb.) by cattle (Bos tau - preestablecido por el Servicio Forestal de Estados Unidos ru s L.) and wild ungulates, mainly elk (Cervus elaphus L.). On-site (USFS). En Junio, Julio y Octubre de 1997 y 1998 medimos el height-weight curves and the USFS gauge consistently produced uso del forraje del "Western wheatgrass"(Pascopyrum smithii lower estimates (overall means = 8 and 7%, respectively) than the Rybd.) por el ganado (Bos taurus L.) y por ungulados silvestres, paired-plot method (overall mean = 31%). Height-weight estimates principalmente alce, (Cervus elaphus L.). Las curvas de altura- did not differ (P > 0.05) when calculated with either on-site curves peso obtenidas en el sitio y el medidor de altura-peso del USFS or the USFS gauge. Within sampling areas, paired-plot estimates produjeron estimaciones consistentemente mas bajas (media were relatively more precise (mean CV = 63%) than on-site curves general = 8 y 7% respectivamente) que el método de parcelas (mean CV = 238%) or the USFS gauge (mean CV = 271%). apareadas (media general = 31%). Las estimaciones de altura- Selective grazing likely contributed to higher CVs for height- peso no difirieron (P > 0.05) cuando se calcularon por las curvas weight techniques. Our findings are important for rangeland and de altura-peso en el sitio o por el medidor del USFS. Dentro de wildlife managers because the forage monitoring technique they las áreas de muestreo, las estimaciones del método de parcelas use may influence the results obtained and, consequently, grazing apareadas fueron relativamente mas precisas (CV de la media = management and wildlife harvest decisions. Managers should 63%) que las de las curvas en el sitio (CV de la media = 238%) o ensure that chosen monitoring techniques provide an appropriate que las del medidor del USFS (CV de la media = 271%). El evaluation of management goals and objectives. apacentamiento selectivo probablemente contribuyó para obten- er CV altos en los métodos de altura-peso. Nuestros hallazgos son importantes para los manejadores de pastizales y fauna sil- Key Words: Arizona, height-weight, herbivory, paired-plot, stub- vestre porque la técnica de monitoreo de utilización de forraje ble height que ellos usan puede influir en los resultados obtenidos y, conse- cuentemente, el manejo del apacentamiento y de las decisiones de cosecha de la fauna silvestre. Los manejadores deben asegurarse As demands upon rangeland resources increase, managers must de escoger la técnica de monitoreo que provee una evaluación monitor rangeland uses and resource responses to satisfy a grow- apropiada para las metas y objetivos de manejo. ing number and variety of stakeholders. Methods such as the paired-plot and height-weight techniques are commonly used for monitoring forage use on rangelands (Stoddart et al. 1975, Rasmussen 1998). Forage palatability, relative abundance of for- Bonham 1989). Forage use estimates provide neither complete age species, plant growth and morphology, and interspecific dif- nor infallible information on herbivore activities or resource ferences in foraging behavior may confound forage use estimates impacts (Cook 1962, Sharp et al. 1994, Laycock 1998, obtained by different monitoring techniques (Rechenthin 1956, Cook 1962, Hanley 1982, Zhang and Romo 1995). With so many Research was funded by a grant from The University of Arizona Agriculture variables influencing the efficacy of forage monitoring, managers Experiment Station. We acknowledge the Coconino National Forest and the need to know if techniques commonly used to estimate ungulate Arizona Game and Fish Department for their support of this project. Authors herbivory under field conditions produce comparable results acknowledge Matt Barnes, Alex Connley, Tom DeLiberto, Vicki Gempko, Arlo Halstead, Elizabeth Howery, Dan Koepke, and David Womack for assistance in (Smith and Ruyle 1997). the field and laboratory. We thank Dr. Mitchel McClaran and 2 anonymous The paired-plot technique uses protected plots (small grazing reviewers for reviewing earlier drafts of this paper. exclosures), each paired with a similar unprotected (grazed) plot Manuscript accepted 24 Jan. 2000.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 499 via ocular estimate. The ratios of unpro- tected and protected clipped dry forage weights are used to calculate percent for- age use (Bonham 1989, Interagency Technical Reference 1996). The height- weight technique involves measuring grazed and ungrazed plant heights along a transect and calculating forage use based on height-weight relationships (Inter- agency Technical Reference 1996). These relationships are established by either clip- ping ungrazed plants and developing height-weight regression equations for a particular range site and season or by using pre-established USFS height-weight gauges (Interagency Technical Reference 1996). The paired-plot method is consid- ered objective and quantitative but requires Fig. 1. Pasture layout for the study site on the Walker Basin Allotment, central Arizona, 1997/1998. Cattle grazed southern pastures (3 and 4) in 1997 and northern pastures (1 and considerable time and expense (Klingman 2) in 1998. Elk had access to study pastures year-round. et al. 1943). The height-weight technique requires less equipment, and if pre-estab- precipitation is mainly in the form of snow. mals,” measured at the end of the growing lished height-weight gauges are used, Our study area (about 6,750 ha) was a pin- season (Glossary Revision Special much less time than the paired-plot tech- yon-juniper (Pinus edulis E n g e l m . — Committee 1989). nique (Interagency Technical Reference Juniperus osteosperma Torr.) savanna with 1996). However, the height-weight tech- a herbaceous understory dominated by west- nique is considered more subjective and Year 1 (1997) ern wheatgrass, the key forage species. The We randomly located 3 sampling areas qualitative than the paired-plot method study area consisted of 4 pastures ranging (Lomasson and Jensen 1943, Mitchell et al. (about 50 x 600 m) in each pasture with from 850–2,100 ha in size (Fig. 1). Each the restriction that they were ³ 0.4 km 1993). Furthermore, some question year, 2 pastures (hereafter called grazed pas- whether the USFS gauge is appropriate to from well-traveled roads, fences and water tures) were grazed during summer and fall and ³ 0.3 km from each other. Six paired- use across seasons and sites (Mitchell et al. by cattle (Bos taurus L.) and 2 pastures 1993). plot units (protected and unprotected (hereafter called rested pastures) were rested 2 Although the pros and cons of the macroplots, 1.7 m ea) and 1 height- from cattle grazing. Grazed pastures were weight line transect were placed within paired-plot and height-weight techniques occupied by 400–450 mature, cross-bred have been debated, there have been no each sampling area (Fig. 2), for a total of Hereford cows with calves for about 14 days 72 paired-plot units in the entire study field experiments that have directly com- during the growing season. All pastures pared whether these techniques provide area. Seventy-two paired-plot units within were subject to year-round grazing by wild the 6,750-ha study area were considered to similar forage use estimates for free-rang- ungulates, mainly elk (Cervus elaphus L. ) . ing wild and domestic ungulates. The be the optimum trade-off between the objective of this 2-year study was to com- number necessary to provide a quantitative pare western wheatgrass (P a s c o p y r u m General Sampling Procedure seasonal comparison between the paired- s m i t h i i Rydb.) forage use estimates col- We used the paired-plot and 2 height- plot and height-weight techniques, and the lected with the paired-plot and 2 height- weight techniques (Interagency Technical number of plots that a resource manager weight methods (i.e., using on-site height- Reference 1996) to evaluate forage use in could realistically sample within a year. weight regression curves and the pre- all pastures 3 times each year: 1) in Height-weight transects were about 400 m established USFS height-weight gauge) early/mid-June, immediately-before-cattle long, located between protected and under field conditions. entered grazed pastures, 2) in mid- unprotected macroplots and ³ 10 m from June/early July, immediately-after-cattle protected macroplots. exited grazed pastures, and 3) in mid- To avoid attracting animals to sampling Materials and Methods October, at the end of the growing season, areas, each protected macroplot was ³ 100 about 3 months-after-cattle had left the m from the others. Protected and unpro- Study Area study area. The grazing schedule allowed tected macroplots were ³ 50 m apart and us to monitor various levels (13–61%) of unprotected plots within a paired-plot unit This study was conducted from March relative (immediately-before and immedi- were ³ 10 m apart. Two unprotected 1997 to October 1998 on the Walker Basin ately-after-cattle) and total forage use macroplots, rather than the traditional 1, Livestock Grazing Allotment, Coconino (months-after-cattle) by domestic and wild were ocularly matched with each protected National Forest in central Arizona. The ungulates. Relative use describes the macroplot to account for possible patch Walker Basin Allotment is comprised of amount of forage consumed or destroyed grazing by cattle and elk (Klingman et al. about 31,000 ha of private and USFS range- up to a certain time during the growing 1943, Grelen 1967). Each unprotected land. Annual precipitation averages 33 cm season but prior to peak standing crop macroplot was marked with 1 wooden and typically occurs in a bimodal pattern (e.g., June or July) (Frost et al. 1994). stake that protruded about 24 cm above from December to February and July to Total forage use, or simply forage use, is ground level. To minimize bias due to September (National Oceanographic and the “proportion of current-year’s forage enhanced growth within protected Atmospheric Administration 1997). Winter consumed or destroyed by grazing ani- macroplots (Owensby 1969), 72 new

500 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 mean of the 3 matched subsample use esti- mates. Means for sampling areas and pas- tures were calculated as described in 1997.

Forage use via the height-weight tech - ni q u e s Methodology used for the height-weight transect was the same as in 1997. However, 1997 paired-plot use estimates were consistently higher (P < 0.0001, n = 36) than height-weight use estimates (Halstead et al. 1998). This discrepancy may have been due to herbivores being attracted to the wooden stakes used to Fig. 2. Typical sampling area layout used to estimate forage use with the paired-plot and delineate unprotected macroplots. If ani- height-weight techniques, 1997/1998. Each sampling area contained 6 paired-plot units (1 mals were attracted to stakes, this could 2 protected and 2 unprotected 1.7-m macroplots (Klingman et al. 1943)). have led to higher use within unprotected macroplots than along height-weight tran- paired-plot units were established in During each sampling period, we mea- sects and help explain the difference March of both years. sured average heights of 60 western between paired-plot and height-weight Before initially establishing paired-plot wheatgrass plants along a 400-m transect forage use estimates. We tested this units, the precision of ocular pairings was (1 transect/sampling area = 3/pasture). hypothesis by applying the height-weight tested by clipping and weighing western About every 3 m, the grazed or ungrazed method (with on-site curves) to 60-point wheatgrass from 29 pairs of 0.25-m2 circu- western wheatgrass plant nearest the transects within unprotected macroplots. lar subplots that were later used to collect observer’s toe was measured. A plant was Each sampling period, the heights of 30 forage data (described below). A paired t- defined as the vegetation occupying a cir- western wheatgrass plants were measured test revealed no difference among subplot cle of turf at least 2 inches in diameter. within each of the 4 unprotected pairings (P ³ 0.2). The same observer Each sampling period, 10 ungrazed west- macroplots scheduled to be clipped within always selected paired-plot units. ern wheatgrass plants per sampling area a sampling area (two, 60-point macroplot were clipped to develop 36 on-site height- transects/sampling area). Heights of grazed or ungrazed plants were systemati- Forage use via the paired-plot tech - weight curves. We used these height- weight curves to estimate forage use for cally measured at 15-cm intervals within nique the 4 macroplots. In each sampling area, During each sampling period, we each transect. Forage use was also estimat- ed along each transect with the pre-estab- we compared mean use from the 2 clipped western wheatgrass from 2 ran- macroplot height-weight transects to the domly selected paired-plot units in each lished USFS culmless western wheatgrass height-weight gauge. Mean use for a pas- mean use calculated from the 400-m sampling area (i.e., 6 paired-plot units/pas- height-weight line transect. ture/sampling period). A 0.25-m2 c i r c u l a r ture was calculated from the 3 sampling plot frame was used to delineate the 4 sub- areas with both techniques. Finally, each plots to be clipped within each 1.7-m2 year’s immediately-before-cattle, immedi- Statistical Analysis macroplot. We averaged the 4 subplot dry ately-after-cattle, and months-after-cattle We used a 3 x 3 x 4 factorial analysis of weights to obtain 1 mean protected weight on-site height-weight curves from each variance (ANOVA) to examine differ- for each protected macroplot. For the 2 sampling area were combined and com- ences across forage use techniques corresponding unprotected macroplots, we pared to the pre-established USFS height- (paired-plot, on-site height-weight, and averaged the 8 subplot weights (4 subplots weight gauge curve (Fig. 4). USFS height-weight), sampling periods, x 2 macroplots) to obtain 1 mean unpro- and pastures. Variances were unequal for tected weight. Percent use for a paired-plot Year 2 (1998) the height-weight and paired-plot forage unit was calculated as the ratio of mean Forage use via the paired-plot technique use estimates. Therefore, the arc-sine unprotected and protected dry weights. In 1998, the paired-plot sampling proce- transformation was applied to forage use Negative utilization values from subplots dure was altered slightly to address high data prior to analysis as recommended by were zeroed (Werner and Urness 1998). standard errors (range = 0–35%) for 1997 Steel and Torrie (1980) for percentage Mean percent use for a sampling area was subplot use estimates within sampling data. A 2 x 4 ANOVA was used to detect calculated from the 2 randomly selected areas. Three subsamples were taken in differences among use estimates from line paired-plot units. Mean percent use for a each paired-plot unit by ocularly matching and macroplot height-weight transects pasture was calculated from the 3 sam- 0 . 2 5 - m2 subplots. We ocularly matched between locations (line vs macroplot tran- pling areas in that pasture. three, 0.25-m 2 subplots within a protected sects) and among pastures. When F-tests macroplot to three, 0.25-m2 subplots with- were significant (P < 0.05), LSD tests were used to detect differences among Forage use via the height-weight in each of the 2 unprotected macroplots. Percent use for a subsample was the ratio means (P < 0.05). Coefficients of variation techniques were calculated for each technique within Height-weight sampling occurred con- of dry weights clipped from the 2 unpro- tected 0.25-m2 subplots (averaged) and the sampling areas as an indicator of tech- currently with paired-plot sampling and 2 nique precision. followed guidelines as outlined in the corresponding 0.25-m protected subplot. Interagency Technical Reference (1996). Percent use for a paired-plot unit was the

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 501 Results

P a i red-Plot vs Height-We i g h t Techniques The paired-plot technique produced sig- nificantly higher (P < 0.05) use estimates than either of the 2 height-weight tech- niques across all pastures (Fig. 3). Compared across pastures and sampling periods, paired-plot estimates were about 23 percentage points (or about 4 times) higher than height-weight estimates using on-site curves and the USFS gauge (Fig. 3). Matching individual subplots within macroplots in 1998 did not reduce stan- dard errors of paired-plot estimates within sampling areas (SE range = 0-35% and 0- 36% in 1997 and 1998, respectively). Standard errors within sampling areas for height-weight techniques were relatively lower than paired-plot estimates both years (SE range = 0–3% and 0–6% in Fig. 3. Western wheatgrass use (%) in 4 pastures measured using on-site height-weight (HW) 1997 and 1998, respectively). However, curves, the USFS height-weight gauge, and the paired-plot technique, averaged over 1997 lower standard errors for the height-weight and 1998. Pasture means with the same upper case letters are not significant at P 0.05. Technique means across pastures with the same lowercase letters are not significant at P techniques were a function of their lower 0.05. Analysis of variance was performed on arc-sine transformed data, however, actual means relative to the paired-plot tech- percent forage use data are presented for ease of interpretation. nique. Mean coefficients of variation with- in sampling areas were higher for on-sight height-weight curves (CV = 238%) and study were about 9 percentage points Table 1). The USFS gauge curve generally the USFS gauge (CV = 271%) than for the higher than non-zeroed paired-plot data (P showed a more linear relationship than on- paired-plot technique (CV = 63%). < 0.0001, n = 72; Wilcoxon signed rank site curves (Fig. 4, Table 1). However, Selective grazing, a common occurrence test). However, non-zeroed paired-plot mean use calculated with the USFS gauge on light- to moderately-stocked ranges, estimates were still about 12 percentage did not differ (P > 0.05) from forage use evidently contributed to higher CVs for points higher than use estimates obtained calculated with on-site curves in any pas- the height-weight techniques. For exam- with either of the 2 height-weight tech- ture (Fig. 3) or during any sampling period. ple, a few individual plants measured niques (P < 0.0001). along height-weight transects were typi- Line vs Macroplot Height-We i g h t cally heavily-grazed (up to 82% use), On-Site Height-Weight Curves vs Transects however, the majority of plants were the USFS Gauge There was no significant difference either ungrazed or lightly-grazed. On-site height-weight curves fluctuated between line and macroplot height-weight The effect of sampling period on forage with plant phenological development: transect estimates (P = 0.23), indicating use estimates was significant (P < 0.001). months-after-cattle curves showed propor- herbivores were not attracted to the unpro- Immediately-after-cattle use estimates tionately more weight in the lower half of tected macroplots marked by wooden were generally higher (mean = 21± 3 plants than immediately-before-cattle or stakes. (SEM) percentage points) than immediate- immediately-after-cattle curves (Fig. 4, ly-before-cattle (mean = 9 ± 3 (SEM) per- centage points) and months-after-cattle (mean = 16 ± 3 (SEM) percentage points) Table 1. Third-order polynomial regression equations and r2-values for 6 on-sight height-weight use estimates with all techniques (P < curves, and the USFS height-weight gauge curve for culmless western wheatgrass. On-site curves were developed for 3 sampling periods (immediately-before-cattle or IBC, immediately-after-cat- 0.05). This difference was expected tle or IAC, and months-after-cattle or MAC) in 1997 and 1998 in central Arizona. Sampling because immediately-after-cattle use esti- periods were early/mid-Jun. (IBC), mid-Jun./early Jul. (IAC), and mid-Oct. (MAC). See Figure 4 mates were made 1–2 days after cattle had for height-weight graphs. exited grazed pastures and around the mid-point of the growing season (i.e., Curve, Year Regression equation r2 before any regrowth had occurred and IBC, 1997 y = 0.0997x3 – 0.9213x2 + 9.1949x + 0.875 0.9999 before total forage production had been 3 2 achieved). IAC, 1997 y = 0.0528x – 0.2861x + 7.4149x + 2.2889 0.9996 MAC, 1997 y = 0.1621x3 – 1.253x2 + 6.3614x – 0.0194 0.9999 Bork and Werner (1999) suggested that 3 2 zeroing “negative” paired-plot data IBC, 1998 y = 0.0996x – 0.7972x + 8.1091x – 0.3697 0.9997 3 2 between protected and unprotected plots IAC, 1998 y = 0.1209x – 1.1111x + 8.6832x + 4.2499 0.9994 3 2 as done in our study inflates forage use MAC, 1998 y = 0.1709x – 1.5511x + 8.496x – 0.6955 0.9997 estimates on spatially heterogeneous USFS y = -0.1329x3 + 3.037x2 – 8.1332x + 10.55 0.9987 ranges. Zeroed paired-plot data in our

502 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 with allometric utilization monitoring techniques, in accurately converting height remaining to weight removed (Bonham 1989, Mitchell et al. 1993, Laycock 1998). The height-weight technique does not account for the selective way that most ungulates graze individual plants (McKinney 1997). Ungulates typically remove only some of a plant’s leaves or tillers which results in a single plant of varying heights (McKinney 1997). However, the height-weight model assumes that herbivores “clip” all parts of a single plant to the same height (Reid and Pickford 1941). To illustrate, assume half of a plant has been grazed to the ground while the other half is left ungrazed. Following the height-weight methodology (Interagency Technical Reference 1996), the plant would be recorded as having 50% of its height removed, as if the entire plant had been grazed to half its height. Based on western wheatgrass height- weight curves, 50% height removed is equivalent to about 35% utilization (Fig. 4). However, because proportionally more weight is in the lower portion of the plant (and was removed on the grazed half), the hypothetical plant has actually had about 50% of its weight removed (50% utiliza- tion). This bias would ostensibly occur more on lightly to moderately-stocked ranges (as in our study; Fig. 3) than on heavily-stocked ranges because animals tend to be less selective (i.e., graze more uniformly) on heavily-stocked ranges (Holechek et al. 1998). Three-dimensional height-volume relationships more accu- rately predict forage production than 2- dimensional height-weight regression equations (Johnson et al. 1988) and may also more accurately assess forage removed. However, quantifying the vol- ume of the specific sections removed from individual plants (e.g., lighter upper sec- tions vs heavier lower sections) would be prohibitively cumbersome for manage- Fig. 4. On-site height-weight curves vs the USFS height-weight gauge curve for culmless ment purposes. western wheatgrass for 3 sampling periods (immediately-before-cattle or IBC, immediate- The paired-plot method may overesti- ly-after-cattle or IAC, and months-after-cattle or MAC) in 1997 (a) and 1998 (b) in central mate forage use if cages create microcli- Arizona. Sampling periods were early/mid-Jun. (IBC), mid-Jun./early Jul. (IAC), and mates that enhance forage growth mid-Oct. (MAC). See Table 1 for regression equations and r2-values. (Owensby 1969, Sharrow and Motazedian 1983). Microclimates may arise when Discussion all pastures and sampling periods. This perching birds “fertilize” exclosures, in finding is important information for range- years with above average precipitation, or Paired-Plot vs Height-Weight land and wildlife managers because the when exclosures are not moved each year, Techniques forage use monitoring technique they allowing more than 1-year’s production to choose may influence the results obtained Differences in forage use estimates accumulate (Grelen 1967, Owensby 1969, and, consequently, grazing management obtained with the paired-plot and the 2 Sharrow and Motazedian 1983). We and wildlife harvest decisions. height-weight methods were substantial. addressed the latter concern by moving We propose several hypotheses to help The paired-plot technique consistently cages at the beginning of each growing explain the disparity between paired-plot produced higher use estimates than either season so that they protected forage for £ and height-weight forage use estimates. of the 2 height-weight techniques across 7 months. However, this effort did not pre- Part of the problem lies, as it usually does

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 503 clude the possibility of enhanced growth quantify the selective way that free-rang- Literature Cited within protected plots, especially in 1998, ing ungulates graze under field conditions. when precipitation was higher as evi- The paired-plot technique consistently Bonham, C. D. 1989. Measurements for terres- denced by greater western wheatgrass produced higher forage use estimates than trial vegetation. John Wiley and Sons, New standing crop and stubble heights both height-weight techniques across all York, N.Y. (Halstead 1998). For example, during pastures and sampling periods during both Bork, E. W. and S. J. Werner. 1999. 1998, western wheatgrass standing crop years of the study. On the other hand, on- Viewpoint: implications of spatial variability averaged about 131 kg ha- 1 ± 18 (SEM) site and pre-established USFS western for estimating forage use. J. Range Manage. higher and stubble heights averaged about wheatgrass curves consistently produced 52:151–156. 3 cm ± 1 (SEM) higher than in 1997 similar forage use estimates. We believe Cook, C. W. 1962. An evaluation of some (Halstead 1998). the differences among the paired-plot and common factors affecting utilization of height-weight techniques, and the similari- desert range species. J. Range Manage. 15:333–338. On-Site Height-Weight Curves vs ties between the height-weight techniques Frost, W. E., E. L. Smith, and P. R. Ogden. the USFS Gauge are real for western wheatgrass in central 1994. Utilization guidelines. Rangelands 16: Forage use estimates made with on-site Arizona because of the consistency of 256–259. height-weight curves and the pre-estab- results. Glossary Revision Special Committee. 1989. lished USFS height-weight gauge were As this study showed, popular forage A glossary of terms used in range manage- use monitoring techniques can produce ment 3r d ed. Soc. Range Manage., Denver, remarkably similar across pastures (Fig. 3) Colo. and sampling periods. Mitchell et al. different results on the same range site. Managers should consider that different Grelen, H. E. 1967. Comparison of cage meth- (1993) concluded that height-weight ods for determining utilization on pine- curves changed statistically with pheno- forage use techniques might produce dif- ferent estimates, and if used alone, will bluestem range. J. Range Manage. 20:94–96. logical development and location. Hall, F. C. and R. Lindenmuth. 1998. However, their analysis was conducted on result in different grazing and wildlife Developing and achieving management height-weight regressions and not on use management decisions. The monitoring objectives on National Forest System lands, estimates derived from those equations. technique(s) chosen must address the p. 47–49. I n: Stubble height and utilization On-site height-weight curves in this study resource objectives that are formulated in measurements: uses and misuses. Oregon changed across time (Fig. 4), but not management plans. If it is important to State Univ. Agr. Expt. Sta., Corvallis, Ore. Halstead, L. E. 1998. Monitoring elk and cat- enough to produce use estimates different know the relative or total percentage of current-year’s growth removed by herbi- tle forage utilization under a specialized from those calculated with the pre-estab- grazing system in Arizona. M.S. Thesis, lished USFS gauge. Time required to vores on a management unit for a particu- lar season or year, forage use techniques Univ. Arizona, Tucson, Ariz. develop height-weight curves made forage Halstead, L. E., L. D. Howery, G. B. Ruyle, use calculation with on-site curves slower might be used with the caveat that the and P. R. Krausman. 1998. Comparison of than with the USFS gauge. However, data paired-plot technique may result in higher the paired-plot and height-weight methods from this study suggest developing west- use estimates than height-weight tech- for measuring forage use. Abstr. 51st Ann. ern wheatgrass height-weight curves for a niques. Meeting Soc. Range Manage. Guadalajara, particular site or season may be unneces- Effective rangeland management does Jalisco, Mex. 51:31. not always require, or does not only Hanley, T. A. 1982. The nutritional basis for sary. Other researchers have indicated that food selection by ungulates. J. Range height-weight relationships for other for- require, measurements of forage use. In some situations, residual vegetation data Manage. 35:146-151. age species on the same range site are fair- Heady, H. F. 1950. Studies on bluebunch ly consistent across years (Heady 1950, (e.g., stubble height) may provide more useful information than forage use data wheatgrass in Montana and height-weight McDougald and Platt 1976). Nevertheless, relationships of certain range grasses. Ecol. those monitoring other forage species or (Smith 1998, Scarnecchia 1999) because it Monog. 20:56–80. dissimilar sites should field check on-site is correlated with erosion protection, soil Holechek, J., R. D. Pieper, and C. H. Herbel. height-weight curves against the USFS moisture retention, forage regrowth poten- 1998. Range management, principles and gauge before using pre-established gauges tial, and small animal and insect habitat practices, 3rd ed. Prentice Hall, Upper to calculate forage use. (Papanastasis 1985, Hall and Lindenmuth Saddle River, N.J. 1998, Holechek et al. 1998). Stubble Interagency Technical Reference. 1996. height data can be collected simultaneous- Utilization studies and residual measure- ments. BLM National Appl. Resour. Sci. Conclusions and Management ly with height-weight data and has the advantage of directly and quantitatively Center. Coop. Ext. Ser., USFS, NRCS, BLM. Implications Johnson, P. S., C. L. Johnson, and N. E. assessing plant material left after grazing West. 1988. Estimation of phytomass for (Hall and Lindenmuth 1998). Use pattern The primary goal of this study was to ungrazed crested wheatgrass plants using maps (e.g., stubble height and/or height- allometric equations. J. Range Manage. directly compare 3 forage use techniques weight data) coupled with relevant trend 41:421–425. under field conditions commonly encoun- data (e.g., plant species composition, Klingman, D. L., S. R. Miles, and G. O. tered by land managers on arid and semi- herbaceous and shrub cover, vegetation Mott. 1943. The cage method for determin- arid rangelands (i.e., seasonal, irregular structure) can help managers to more com- ing consumption and yield of pasture grazing by more than one large ungulate prehensively evaluate management strate- herbage. J. Amer. Soc. Agron. 35:739–746. species in a heterogeneous, patchy envi- gies. More complex resource issues (e.g., Laycock, W. A. 1998. Variation in utilization ronment). Technique accuracy was not management of threatened or endangered estimates caused by differences among meth- ods, years, and observers, p. 17–24. I n: addressed because such studies typically species) may require more intensive moni- involve an artificial population and simu- Stubble height and utilization measurements: toring programs. uses and misuses. Oregon State Univ. Agr. lated herbivory (e.g., mowing or clipping) Expt. Sta., Corvallis, Ore. which would have negated our ability to

504 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 Lomasson, T. and C. Jensen. 1943. Rasmussen, G. A. 1998. Interpretation of uti- Smith, E. L. 1998. Seasonal effects on the Determining use of grasses from height- lization and long-term frequency measure- measurement and interpretation of utiliza- weight tables. J. Forest. 41:589–593. ments for rangeland management, p. 25–28. tion, p. 9-16. I n: Stubble height and utiliza- McDougald, N. K. and R. C. Platt. 1976. A I n: Stubble height and utilization measure- tion measurements: uses and misuses. method of determining utilization for wet ments: uses and misuses. Oregon State Univ. Oregon State Univ. Agr. Expt. Sta., mountain meadows on the Summit allotment, Agr. Expt. Sta., Corvallis, Ore. Corvallis, Ore. Sequoia National Forest, California. J. Range Rechenthin, C. A. 1956. Elementary morphol- Smith, E. L. and G. B. Ruyle. 1997. Mange. 29:497–501. ogy of grass growth and how it affects uti- Considerations when monitoring rangeland McKinney, E. 1997. It may be utilization, but lization. J. Range Manage. 9:167–170. vegetation, p. 1-6. I n: G. B. Ruyle (ed.), is it management? Rangelands 19:4–7. Reid, E. H. and G. D. Pickford. 1941. A com- Some methods for monitoring rangelands Mitchell, J. E., R. E. Elderkin, and J. K. parison of the ocular-estimate-by-plot and and other natural area vegetation. USDA Lewis. 1993. Seasonal height-weight dynam- the stubble-height methods of determining Coop. Ext. Rep. 9043, Univ. Arizona, ics of western wheatgrass. J. Range Manage. percentage utilization of range grasses. J. Tucson, Ariz. 46:147-151. Forest. 39:933–941. Steel, R. G. D. and J. H. Torrie. 1980. National Oceanic and Atmospheric Scarnecchia, D. L. 1999. Viewpoint: the range Principles and procedures of statistics, a bio- metrical approach, 2nd ed. McGraw Hill Administration. 1997. Climatological data utilization concept, allocation arrays, and Book Co., New York, N.Y. annual summary Arizona. Nat. Environ. range management science. J. Range Stoddart, L. A., A. D. Smith, and T. W. Box. Satellite, Data and Info. Serv., Nat Climatic Manage. 52:157–160. 1975. Range management, 3rd ed. McGraw Data Ctr., Vol. 101(13), Asheville, N.C. Sharp, L., K. Sanders, and N. Rimbey. 1994. Hill Book Co., New York, N.Y. Owensby, C. E. 1969. Effect of cages on Management decisions based on utilization— Werner, S. J. and P. J. Urness. 1998. Elk for- herbage yield in true prairie vegetation. J. is it really management? Rangelands age utilization within rested units of rest- Range Manage. 22:131–132. 16:38–40. rotation grazing systems. J. Range Manage. Papanastasis, V. P. 1985. Stubble height, basal Sharrow, S. H. and I. Motazedian. 1983. A 51:14–18. cover, and herbage production relationships comparison of three methods for estimating Zhang, J. and J. T. Romo. 1995. Impacts of in grasslands of northern Greece. J. Range forage disappearance. J. Range Manage. defoliation on tiller production and survival Manage. 38:247–250. 36:469–471. in northern wheatgrass. J. Range Manage. 48:115–120.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 505 J. Range Manage. 53: 506–510 September 2000 A method for estimating cattle fecal loading on rangeland watersheds

KENNETH W. TATE, EDWARD R. ATWILL, NEIL K. MCDOUGALD, MELVIN R. GEORGE, AND DAVID WITT

Authors are rangeland watershed specialist, Agronomy and Range Science, University of California, Davis, Calif. 95616-8515; environmental health spe - cialist, School of Veterinary Medicine, University of California, Veterinary Medicine Teaching and Research Center, Tulare, Calif. 93274; natural resources and livestock advisor, University of California Cooperative Extension, Madera, Calif. 93637; range and pasture specialist, Agronomy and Range Science, University of California, Davis, Calif. 95616-8515; and range conservationist, USDA-NRCS, Madera, Calif. 93637.

Abstract Resumen

Water quality contamination by pathogens and nutrients from La contaminación del agua por patógenos y nutrientes prove- cattle fecal deposits is a concern on rangeland watersheds. The nientes de heces fecales del ganado es motivo de preocupación en temporal and spatial deposition of fecal material relative to las cuencas hidrológicas de pastizal. La deposición espacial y storm events and water-bodies determines much of the risk a temporal del material fecal en relación a los eventos de tormen- grazing scheme presents to water quality. The objective of this tas y cuerpos de agua determina mucho del riesgo que un esque- study was to develop and evaluate a comparative technique to ma de apacentamiento presenta para la calidad del agua. El estimate cattle fecal loading across a watershed through time. objetivo de este estudio fue desarrollar y evaluar una técnica Once the method was developed, dry and wet season trials were comparativa para estimar la carga de material fecal a lo largo de conducted on a 138 ha experimental rangeland watershed at the una cuenca hidrológica a través del tiempo. Una vez que el méto- San Joaquin Experimental Range in 1996–97. Fifty-four perma- do se desarrollo, se condujeron ensayos en las épocas seca y de nent 40 m2 belt transects were established across the watershed. lluvia de 1996 y 1997, los ensayos se llevaron a cabo en una cuen- Observers ocularly assigned a rank of 1 (smallest diameter) to 5 ca hidrológica de pastizal de 138 ha situada en "San Joaquín (largest diameter) to each fecal deposit within a transect. A Experimental Range". A lo largo de la cuenca se establecieron 54 regression relationship was developed to predict fecal deposit dry transectos permanentes de 40 m2. Los observadores asignaron en weight by rank. Load per transect was calculated as the total forma ocular una clasificación del 1 (diámetro mas pequeño) al 5 weight of all fecal deposits in a transect. All fecal deposits in (diámetro mas grande) a cada uno de los depósitos fecales encon- transects were collected and actual fecal load determined. The trados dentro del transecto. Se desarrollo una relación de regre- comparative yield methodology was successfully adapted to esti- sión para predecir el peso seco del depósito fecal por categoría. mate rangeland fecal loading. Regression relationships predict- La carga por transecto se calculó como el peso total de los ing fecal deposit dry weight by ranks were highly significant for depósitos fecales dentro del transecto. Todos los depósitos fecales all observers (p < 0.001). The R2 values ranged from 0.97 to 0.99 del transecto se colectaron y se determinó la carga fecal actual. in the dry season and 0.89 to 0.94 in the wet season. There was no La metodología comparativa se adapto exitosamente para esti- significant difference between the weighed fecal load estimate mar la carga fecal del pastizal. Las relaciones de regresión para and the estimates of observers using the comparative method (p predecir el peso seco de los depósitos fecales por categoría fueron < 0.05). This method provides a rapid, simple method for esti- altamente significativos para todos los observadores (P < 0.001). mating spatial and temporal livestock fecal loading on rangeland Los valores de r2 variaron de 0.97 a 0.99 en la época seca y de watersheds. 0.89 a 0.94 en la época de lluvia. No hubo una diferencia signi- ficativa entre la estimación de la carga fecal obtenida mediante pesaje y la estimada por los observadores utilizando el método Key Words: water quality, pathogens, nutrients, grazing comparativo (P < 0.05) Este método provee un método rápido y simple para estimar la carga espacial y temporal de depósitos fecales de ganado dentro de las cuencas hidrológicas de pastizal. Water quality contamination by pathogens and nutrients in live- stock fecal deposits is a subject of concern on rangeland water- sheds (Atwill 1996, Nader et al. 1998). Recent events in ment alternatives. If managers do not take this initiative, regulato- California illustrate the need for land managers to evaluate the ry control of grazing on priority watersheds is likely. risk of pathogen and nutrient loading that their grazing animals Estimating the spatial and temporal loading of livestock fecal present to water quality on priority watersheds that supply munic- material, thus nutrients and pathogens, on a watershed is a critical ipal drinking water (Barry et al. 1998). Managers need methods step in determining the risk to water quality presented by grazing that provide site specific data to confirm or negate the risk of management schemes. The t i m i n g of fecal loading relative to contamination. In situations where the risk is unacceptable, man- runoff as well as its location relative to riparian areas and water- agers must be able to propose and defend viable grazing manage- bodies determines much of the potential for contaminants in live- stock fecal deposits to reach source water. Research was funded by UC Division of Agriculture and Natural Resources Hand collection of fecal deposits from plots across a landscape 1995-96 Competitive Grants Program, Grant #020. Manuscript accepted 12 Dec. 1999. is one method of estimating fecal loading. The process is time

506 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 and labor intensive and limits sample size. included. Transect starting points were fecal deposit dry weight was the depen- A rapid, simple method is needed to esti- randomly positioned within each stratifi- dent variable and fecal deposit rank the mate fecal loading. Observational methods cation via random selection from a grid. independent variable. Calibration was based upon ranking systems have been Orientation of the transects was deter- accomplished using the Type 3 method successfully used to facilitate herbage mined via random compass bearing. The described by Haydock and Shaw (1975). yield estimation (Mannetje and Haydock watershed was grazed continuously with a Approximately half way through sam- 1963, Haydock and Shaw 1975, Reese et 30 unit commercial, cow-calf herd from 1 pling, observers collectively selected a al. 1980, Gillen and Smith 1986, Frost et January 1996 through 31 May 1997. minimum of 15 fecal deposits that covered al. 1990, USDI-BLM 1996). These meth- the range of ranks from 1 to 5. These cali- ods have lower costs per plot which allows Comparative Fecal Load Method bration fecal deposits were ranked by each for increased sample size and improved The comparative fecal load method observer and subsequently collected for estimates of population parameters at the described in this section closely follows dry weight determination. landscape scale. Decreases in precision the comparative herbage yield procedures The fourth step was to develop a cali- occur at the plot scale. Many accept this as outlined by Haydock and Shaw (1975) and bration curve for each observer using a worth-while trade-off (Haydock and addresses the first objective of our study. regression procedures. Steps 3 and 4 were Shaw 1975, Reese et al. 1980). The comparative herbage yield method extremely critical because they quantified The first objective of this study was to uses the quadrat as both the basic unit for the individual biases of each observer. adapt the comparative herbage yield ranking and as a sample unit. Our compar- The final step was to use the calibration methodology to a method for estimating ative fecal load method used the individ- curve to estimate dry weight (g) of each cattle fecal loading on rangelands. The ual fecal deposit as the basic unit for rank- fecal deposit in a transect. Total fecal load second objective was to test the newly ing and the 40 m2 transect as the experi- (kg) for the transect was the sum of all developed comparative fecal load method mental unit. A transect could contain 0, 1, individual fecal deposit weights. Fecal by comparing watershed scale loading or numerous fecal deposits of various loading per unit area was calculated from estimates by several observers to estimates sizes. The transect fecal load (kg) was the the known area of the transect. Fecal load- from hand collection and weighing of sum of the weight of the fecal deposits in ing parameters (mean, standard deviation, fecal deposits during wet and dry seasons. the transect. The comparative yield range, etc.) for this area were then estimat- Our null hypothesis was that this method method bases rank (1, 2, 3, etc.) on esti- ed from the entire set of transects dis- would not accurately estimate watershed mated herbage dry weight (g). Our com- persed across the watershed. fecal loading across seasons or observers. parative fecal load method used fecal deposit diameter (cm) to establish rank (1, Testing the Method Methods 2, 3, etc.). A dry and wet season trial were con- The first step was to select 5 reference ducted to address our second objective of fecal deposits to set the ranking scale for evaluating the method. The dry season Study Site the sample period. Rank 1 (smallest) and 5 trial began on 31 May 1996 when all fecal The study was conducted on a 138 ha (largest) were established by surveying the deposits were cleared from each transect. experimental rangeland watershed at the general sampling area and selecting fecal The dry season trial ended and the wet San Joaquin Experimental Range in deposits encompassing the range of sizes season trial began when all transects were Madera County, Calif. Vegetation is oak within the area. Once reference fecal sampled and subsequently cleared on 1 savanna with annual grassland understory. deposits 1 and 5 were established, refer- November 1996. The wet season trial, and Climate is Mediterranean with mean annu- ence fecal deposit 3 was defined to have a the study, ended 31 May 1997 when all al precipitation of 485 mm. Precipitation diameter half-way between 1 and 5. transects were re-visited and sampled. All falls almost entirely as rainfall from Reference fecal deposits 2 and 4 were then transects were sampled once at the end of November through May (wet season) with selected to have diameters half-way each trial. Each observer worked indepen- almost no precipitation falling from June between 1 and 3 and 3 and 5, respectively. dently. Following ranking of all fecal through October (dry season). In Following establishment of the reference deposits in a transect, all fecal material November 1995, 54 permanent 40 m2 belt fecal deposit dimensions, each observer was collected, dried at 60° C for 96 hours, transects were established across the spent 15 to 30 minutes calibrating their and weighed to determine total dry weight watershed. An individual transect served mind and eye to the reference set and the fecal load. There were 4 observers in the as the experimental unit in this study. ranking scale it established for that specif- dry season and 3 observers in the wet sea- Transects were 30.5 m long by 1.3 m wide ic sample period and location. son trial. Observers A and B participated and staked at both ends. Following a strat- The second step was to record the rank in both trials. Observers C and D only par- ified-random design, the transects were of each fecal deposit in a transect based ticipated in the dry season trial and distributed proportionally across the upon the 5 references established in step 1. Observer E participated only in the wet watershed to encompass slope classes of Half scores were also utilized (1.5, 2.5, season trial. Calibration curves predicting 0–10, 11–20, 21–30, >30%, with addition- etc.). Data was recorded using the sheet fecal deposit dry weight from ranks were al transects placed in the riparian zone and developed by Frost et al. (1990) for com- developed for each observer and the rela- around livestock concentration areas. parative herbage yield. tionship used to calculate individual Transect numbers per stratification were The third step was to rank and collect a watershed fecal load estimates. 14, 15, 5, 5, 10, and 5, respectively. The set of fecal deposits for development of Calibration curves for each observer in stratified sampling design was used to the calibration curve to calculate fecal each season were developed using regres- insure that important topographic and deposit dry weight (g) from individual sion, with adjusted multiple coefficients of managerial areas representing the range of fecal deposit ranks. The calibration curve determination (R2) and analysis of vari- fecal loading on the watershed were was derived by regression analysis where ance for the regression coefficients used to

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 507 characterize the significance of each observer’s ability to predict fecal deposit weight as a function of rank. Both linear and quadratic regression models were test- ed. Regression models were forced through the origin because a zero rank must equal a zero fecal deposit dry weight. To determine the need to develop separate calibration curves for each observer within a season, the equality of curves developed for Observers A, B, C, and D for the dry season and Observers A, B, and E for the wet season was tested. To determine the need to develop separate calibration curves for the dry and wet season, the equality of dry and wet season calibration curves for Observers A and B was tested. Given that the intercept of all calibration curves was forced through 0, only the equality of slopes could be tested. The equality of the regression lines was tested by conducting pairwise F- tests following Fig. 1. Theoretical calibration curve for fecal deposit dry weight predicted by rank when ref- Dixon (1992) with the F-test calculated as, erence fecal deposit ranks are established from fecal deposit diameter. F = (Regression SS across groups)/( pi + g - p - 1) 2 (Residual SS within groups)/(N - g - p ) that the relationship between quadrat ter (2r) and fecal deposit dry weight (hpr i herbage dry weight and rank was x fecal deposit dry weight specific gravity) where Residual Sums of Squares ( S S ) explained by simple linear regression. is not linear. Figure 1 illustrates the theo- within groups = å Residual SS of group i, This is logical given that they selected retical relationship between fecal deposit Total SS = Residual SS of the data prior to their reference quadrats to insure a linear rank and fecal deposit dry weight at a con- grouping, Regression SS across groups = relationship between rank and quadrat dry stant fecal deposit depth of 2 cm, a specif- (Total SS - Residual SS within groups), p weight. In developing their reference set, ic gravity of 0.2, and with diameters for = number of independent variables in the Haydock and Shaw (1975) specifically reference fecal deposit ranks 1, 3, and 5 regression equation for all groups taken state that reference quadrat 5 be selected equal to 2, 6, and 10 cm, respectively. to have 5 times the dry weight as the During the dry season, we found that 3 together, pi = number of independent vari- ables in the equation for the ith group, g = quadrat for reference rank 1, and that ref- of the 6 pair-wise comparisons of calibra- number of groups, and N = number of erence quadrat 3 be selected to have dry tion curves among the 4 observers were cases in all the groups combined. weight half-way between ranks 1 and 5. significantly different (Table 1, Fig. 2). Analysis of variance was used to test for This leads to the linear relationship for For example, the calibration curve for differences between mean watershed fecal calibration curves that they were seeking Observer A compared to Observer B was load estimates from hand collected sam- for their method. significantly different based upon the pair- ples and comparative load based estimates In our adaptation of the comparative wise F-test for the equality of regression from ranks for each observer. Differences yield method, the first step is to select ref- lines (P = 0.013). The predicted dry between mean watershed fecal load esti- erence fecal deposits 1–5 based upon weight for a fecal deposit of rank 2 was 96 mates among observers was tested in the diameter. Considering that the shape of a g compared to 99 g, or 394 g compared to same model. Each trial (season) was sub- fecal deposit can be approximated as a 330 g for a fecal deposit of rank 4 for jected to an individual ANOVA. short cylinder, it becomes clear that the Observer A and B, respectfully. This relationship between fecal deposit diame- result indicated that it was important to

Results and Discussion Table 1. Comparative fecal load calibration curves developed for each observer for the dry and wet season trials that were used to predict fecal deposit dry weight from observer rank, n = 15 for Calibration Curves each observer. Samples were taken on a 138 ha watershed at the San Joaquin Experimental Calibration curves for each observer in Range in 1996–97. each season are presented in Table 1. The 2 R2 values for linear calibration curves Season Observer* Regression Equation R across observers ranged from 0.84 to 0.94 Dry Aa,1 –2.95(rank) + 25.36(rank2) 0.97 and from 0.75 to 0.90 for the dry and wet Bb,c,1 16.59(rank) + 16.50(rank2) 0.99 season trials, respectfully. Based upon a 5 Cb –3.62(rank) + 19.07(rank2) 0.98 a,c 2 percentage point or greater improvement D 20.35(rank) + 16.16(rank ) 0.97 a,2 2 in R 2 values, the quadratic equation with Wet A 71.45(rank) + 16.78(rank ) 0.94 Ba, 2 19.01(rank) + 24.52(rank2) 0.93 observer rank and the square of the 2 observer rank best fit the data (Table 1). Ea 35.10(rank) +23.40(rank ) 0.89 While there is no rule in comparative *Within each season, calibration curves for observers with the same letter were not significantly different (P < 0.05). For yield, Haydock and Shaw (1975) found Observers A and B across seasons, calibration curves with the same number were not significantly different (P < 0.05).

508 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Given the differences in calibration curves for observers in the dry season and for Observers A and B across seasons, we recommend that individual calibration curves be developed for each observer in each season. We recommend that fecal deposit diameter be used as the principle variable in determining fecal deposit size and that the quadratic form be investigated when developing the best fit regression- based calibration curve. Reference fecal deposits should be reexamined each day of consecutive sampling. A new reference set should be developed when sampling is spread discontinuously over several days or when a change in fecal deposit appear- ance occurs. Changes in sample site, sea- son, cattle class or age, feed type or mois- ture, or a rainfall event can all alter the relationship between appearance and dry weight of fecal deposits and new calibra- tion curves should be developed.

Fig. 2. Calibration curves used to predict fecal deposit dry weight from ranks for Observers Table 2. Watershed fecal loads estimated by A, B, C, and D for the dry season trial conducted on a 138 ha watershed at the San each observer using the comparative fecal Joaquin Experimental Range on 1 November 1996. load method and by hand collection for both dry and wet season trials, n = 54 for each estimate. Samples were taken on a 138 ha develop separate calibration curves for Calibration curves for the dry and wet watershed at the San Joaquin Experimental each observer during the dry season. season were significantly different for Range in 1996–97. There were no significant differences both Observer A (P = 0.0004) and B (P = among calibration curves for Observers A, 0.001), which indicated it was important Season Estimate Mean* SE B, and E during the wet season (Table 1, to develop separate calibration curves for (kg ha-1) Fig. 3), which indicated it was not as each observer for each season. Fecal Dry Hand Collected 37.2 8.2 important to develop separate calibration deposit dry weight for a given rank was Observer A 53.9 10.9 curves for each observer during this sea- consistently heavier during the wet season Observer B 50.1 10.6 son. In this season, a composite calibra- trial for both observers. This could be due Observer C 43.4 10.0 Observer D 48.8 10.1 tion curve could have been developed and to higher fecal deposit density under green used for all observers. versus dry feed conditions. Wet Hand Collected 27.6 7.3 Observer A 34.8 6.8 Observer B 23.3 5.0 Observer E 31.7 7.1 *There were no significant differences (P<0.05) between methods nor among observers within each season of sam- pling.

Watershed Fecal Load Estimation There were no significant differences in watershed fecal load estimated by each observer using the comparative fecal load method and by hand collection for either the dry (P = 0.70) or the wet (P = 0.49) season trial (Table 2). Our null hypothesis that the comparative fecal load method developed in this study would not accu- rately estimate average fecal loading on the watershed across seasons or observers was rejected. Regardless of method, there was high variability associated with watershed fecal load estimates. The observed variability verifies the need for a rapid, simple sam- pling method that allows increased sample Fig. 3. Calibration curves used to predict fecal deposit dry weight from ranks for Observers size. Fifty-four transects represented the A, B, and E for the wet season trial conducted on a 138 ha watershed at the San Joaquin largest number that a 2 person team could Experimental Range on 31 May 1997.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 509 physically harvest in a single day from the observers are used, load can be averaged Barry, S.J., E.R. Atwill, K.W. Tate, T.S. 138 ha watershed evaluated in this study. across observers. It is recommended that Koopmann, J. Cullor, and T. Huff. 1998. Mean time to collect all fecal material when comparisons are to be made, a single Developing and implementing a HACCP- from each transect was approximately 5 observer or consistent set of observers be based program to control C r y p t o s p o r i d i u m minutes. Mean time for a person to con- utilized to reduce potential bias from and other waterborne pathogens in Alameda observer differences. Creek Watershed: a case study, p.57–69. In: duct the comparative fecal load method Proc. of Amer. Water Works Assoc. Annu. per transect was approximately 2 minutes. Conf. Dallas, Tex. At 54 plots, that was 4.50 and 2.25 hours Dixon, W.J. 1992. BMDP Statistical software of transect sampling time for the hand col- Conclusion manual. University of California Press. lection and comparative methods, respec- Berkeley, Calif. tively. Regardless of the number of tran- Estimating livestock fecal loading on Frost, W.E., N.K. McDougald, and M.R. sects, the comparative fecal load method rangelands via hand collection of fecal George. 1990. Herbaceous plant measure- required an hour to select a reference fecal material from plots is cost prohibitive and ments, p.3–6. I n: W.J. Clawson (ed.), limits sample size in a system where large Monitoring California’s annual rangeland deposit set and sample the calibration vegetation. Univ. of California, Div. of Agr. fecal deposit set. Travel time between numbers of samples are preferable. We were able to successfully adapt the com- and Natur. Resources. Leaflet 21486. Davis, transects was faster for the comparative Calif. fecal load method because technicians parative herbage yield methodology to estimate livestock fecal loading on range- Gillen, R.L. and E.L. Smith. 1986. Evaluation were not carrying sacks of fecal material. of the dry-weight-rank method for determin- lands. This method provided a rapid, sim- ing species composition in tallgrass prairie. J. Travel time among transects accounted for ple method for collecting plot-based, site- about half of the time spent in the field Range Manage. 39:283–285. specific estimates of spatial and temporal Haydock, K.P. and N.H. Shaw. 1975. T h e each day. By employing the comparative livestock fecal loading on rangelands. fecal load method, sample size was comparative yield method for estimating dry When used in conjunction with a system- matter yield of pasture. Australian J. Exp. increased 45%, from 54 to 78 transects per atic sampling scheme of plots across a Agr. and Anim. Husb. 15:663–670. day. Also, only 1 observer rather than the pasture or watershed, fecal loading Mannetje, L.’t and K.P. Haydock. 1963. The team of 2 employed in the hand collection through time can be estimated at the ripar- dry-weight-rank method for the botanical method was required. ian, upland, pasture, or watershed scale. analysis of pasture. J. Brit. Grassl. Soc. If transects are permanently established Quantification of fecal loading through 27:268–275. for estimation of loading over time, all space and time is a critical first step in Nader, G.A., K.W. Tate, E.R. Atwill, and J. fecal deposits should be cleared from the evaluating the risk to water quality under Bushnell. 1998. Water quality effect of transect following ranking. This is not various grazing management scenarios. rangeland beef cattle excrement. Rangelands 20:19–25. necessary for temporary transects where Reese, G.A., R.L. Bayn, and N.E. West. repeated sampling is not an objective. Literature Cited 1980. Evaluation of double-sampling estima- Loading estimates for specific areas tors of subalpine herbage production. J. (flood-plain, uplands, etc.) of the water- Range Manage. 33:300–306. shed can be developed if transect locations Atwill, E.R. 1996. Assessing the link between USDI-BLM. 1996. Comparative yield method, are stratified or targeted to adequately rep- rangeland cattle and waterborne C r y p t o - p.116–122. In : Sampling vegetation attributes. resent the area of interest. If multiple sporidium parvum infections in humans. Interagency Tech. Ref. BLM/ST-96/002+1730. Rangelands 18:48–51. Denver, Colo.

510 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 J. Range Manage. 53: 511–517 September 2000 Early establishment of Douglas-fir and ponderosa pine in grassland seedbeds

YUGUANG BAI, DON THOMPSON, AND KLAAS BROERSMA

Authors are range ecologist, range plant physiologist, and soil scientist, Kamloops Range Research Unit, Agriculture and Agri-Food Canada, Kamloops, BC V2B 8A9 Canada. Funding for this research was provided by Beef Cattle Industry Development Fund, Matching Investment Initiative Fund, and Natural Science and Engineering Research Council (Canada).

Abstract Resumen

Grasslands of interior British Columbia are being encroached Los pastizales del interior de Columbia Británica están siendo upon by Douglas-fir (Pseudotsuga menziesii var. glauca (Beissn.) invadidos por "Douglas-fir" (Pseudotsuga menziesii var. Glauca Franco) and ponderosa pine (Pinus ponderosa Dougl.). A pot (Beissn.) Franco) y "Ponderosa Pine"(Pinus ponderosa D o u g l . ) . experiment placed in the field was conducted to determine the Se realizó un experimento en macetas, ubicado en el campo, con effect of forest and grassland seedbeds on seedling emergence el objetivo de determinar el efecto de la cama de siembra and early establishment of the 2 species with 2 seed collections (bosque y pastizal) y el origen de la semilla (dos colecciones) en la each. For these seedbeds, structural characteristics were evaluat- emergencia y establecimiento de plántulas de estas especies ed and the effect of seedbeds water extracts on seed germination arbóreas. Se evaluaron las características estructurales de la was determined. Seedling emergence of both species was signifi- cama de siembra y extractos de agua de ellas en la germinación cantly reduced by Douglas-fir needles and enhanced by fescue lit- de la semilla. Las hojas de "Douglas-fir" redujeron significativa- ter and cattle manure compared to mineral soil. The rate of mente la emergencia de plántulas de ambas especies, en cambio emergence was reduced by Douglas-fir needles and sagebrush lit- en el mantillo de "fescue" y el estiércol de ganado la aumen- ter, and for some collections, by ponderosa pine needles, but was taron, esto en comparación con el suelo mineral. La tasa de always enhanced by manure compared to mineral soil. Seedling emergencia se redujo por las hojas de ADouglas-fir" y el mantil- survival was generally not affected by seedbeds. Douglas-fir lo de "Sagebrush"; en algunas colecciones de semilla las hojas de seedlings emerging earlier in the season survived better, and both "Ponderosa pine" también redujeron la tasa de emergencia, sin Douglas-fir and ponderosa pine seedlings emerging earlier lived embargo, la tasa de emergencia siempre mejoró con el estiércol longer than these emerging later. Seed germination of ponderosa comparado con el suelo mineral. En general, las camas de siem- pine was not affected by the water extract while that of Douglas- bra no afectaron la sobrevivencia de las plántulas. Las plántulas fir was reduced by the water extract from sagebrush litter. de "Douglas-fir" emergieron mas temprano y sobrevivieron Therefore, differences in seedling emergence of the 2 species mejor, las plántulas que emergieron primero, tanto de "Douglas- among seedbeds were related more to structural than to chemical fir" como de "Ponderosa pine", vivieron mas tiempo que las que characteristics of seedbeds. Successful establishment of the 2 nacieron al último. La germinación de semillas de "Ponderosa species in grasslands within this region likely relies on the ability pine" no fue afectada por el extracto de agua mientras que la de of seeds to germinate early in the growing season on seedbeds in semillas "Douglas-fir" si se redujo con el extracto de agua de which soil moisture is conserved, as summer droughts are severe. mantillo de "Sagebrush". Por lo tanto, la diferencia en germi- nación de plántulas de las dos especies se relaciono mas a las car- acterísticas estructurales que a las características químicas de la Key Word: allelopathy, litter structure, seedbed ecology, P i n u s cama de siembra. El establecimiento exitoso de las dos especies p o n d e r o s a Dougl., Pseudotsuga menziesii var. glauca ( B e i s s n . ) en los pastizales de esta región, se basa, probablemente, en la Franco capacidad de las semilla para germinar a inicios de la estación de crecimiento en camas de siembra en las que la humedad del suelo se conserva en el verano cuando la sequía es severa. The encroachment of woody plants is a major threat to range resources at a global scale, reducing grassland area and carrying capacity, and inhibiting livestock movement (Burkhardt and suppression, disturbance, climatic variation, and interactions of Tisdale 1976, Strang and Parminter 1980, Gruell 1983, these factors in the last 100 to 150 years are thought to be respon- MacDonald and Wissel 1991, Richardson and Bond 1991). Fire sible for tree encroachment (Tisdale 1950, Parminter 1978, Strang and Parminter 1980, Arno and Gruell 1986, Mast et al. The authors wish to thank Dr. Walter Majak for advice on the water extract 1998). In the southern interior of British Columbia, conifer experiment, Ms. Barbara Brooke and Christine Kelly for technical assistance, and forests are the dominant vegetation; natural grasslands are limited Drs. J.T. Romo and Walter Willms for reviewing this manuscript. resources and are susceptible to tree encroachment. Tree Manuscript accepted 1 Jan. 2000. encroachment was first reported about 80 years ago (Whitford and Craig 1918) and the loss of grasslands to encroachment in recent years was more than 30% over a 30-year period in certain areas (Ross 1997).

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 511 The dominant conifer species in the Table 1. Descriptions of Douglas-fir and ponderosa pine seed collections from the Interior of grassland/forest ecotone, interior Douglas- British Columbia. fir (Pseudotsuga menziesii var. g l a u c a (Beissn.) Franco) and ponderosa pine Species/collection Location Elevation Date Seed (Pinus ponderosa Dougl.), regenerate (m) (No.g-1) from seeds. The invasion by trees requires Douglas-fir 1 Nesbitt Lake, 50°44'N 121°06'W 900 Sep. 93 75 a sequence of favorable events from seed Douglas-fir 2 Pher Lane, 50°43'N 120°52'W 750 Aug. 93 80 production and dispersal to seedling emer- Ponderosa pine 1 Sugarloaf Hill, 50°38'N 120°27'W 875 Aug. 94 19 gence and survival (Hill et al. 1995). Ponderosa pine 2 Indian Gardens, 50°44'N 120°49'W 725 Aug. 95 21 Therefore, understanding the eco-physio- logical aspects of seed germination, seedling emergence, and seedling estab- subjected to field stratification. Seeds Pots measuring 12 x 12 x 12 cm were lishment of these tree species under grass- were soaked in pre-cooled distilled water filled with 8 cm of mineral soil and cov- land conditions is critical in explaining at 4°C for 2 hours and then placed in the ered with seedbed materials. The litter this phenomenon, and in providing a base- field in early December 1997. The field depths selected were similar to those line for management strategies aimed at site was located within a long-term exclo- observed in the field, 1 cm depth for big controlling tree encroachment. Numerous sure dominated by rough fescue (F e s t u c a sagebrush litter and 3 cm for Douglas-fir researchers have studied the effects of for- campestris Rydb.) within the Lac du Bois and ponderosa pine needles, fescue litter est floors on seedling emergence and Grasslands near Kamloops at an elevation and cattle manure. Mineral soil treatment establishment of conifers (Bramble and of 900 m. Approximately 60 seeds were was not covered by any litter. Forty seeds Goddard 1942, Chrosciewicz 1974, placed in each custom made, metal screen were placed on the litter surface in each Williams et al. 1990, McLaren and Janke envelope (6 x 8 cm). These envelopes pot and shaken by hand. Seeds in the min- 1996). Douglas-fir litter inhibited seedling were randomly placed at the bottom of eral soil treatment remained visible on the emergence of itself in a monocultural each of the 4 metal screen boxes (60 x 60 surface; those in fescue litter and pon- plantation by reducing water availability x 5 cm). These boxes were fixed to the derosa pine needle treatments fell to the and by acting as a mechanical barrier ground with their bottoms 1 cm below the bottom of the seedbed materials within 1 (Caccia and Ballaré 1998). Seedling emer- soil surface and covered with grass litter. cm from mineral soils. Seeds in the gence of ponderosa pine was generally The lids of boxes were then sealed with Douglas-fir needle, sagebrush litter and enhanced by mineral or ash covered soils metal wire to minimize seed predation and cattle manure treatments remained at the compared to undisturbed soils covered covered with grass litter. Seeds were top 0.5 cm of the seedbed materials but with plant litter (Roe and Squillance 1950, retrieved from the field on 1 May 1998. invisible from the surface. Pots were then Foiles and Curtis 1965). No study has Seeds used in the water extract experiment placed outside under tree canopies and compared seedling establishment of were not stratified. covered with mesh metal screens to mini- conifers, particularly Douglas-fir and pon- mize seed predation by birds and other derosa in forest and grassland seedbeds. Seedbed experiment animals. The first trial was initiated on 12 Objectives of this study were to deter- Seedbed experiment was conducted in a May 1998 and the second trial 2 weeks mine: (1) effects of mineral soil, litter of randomized complete block design with 6 later. The duration of each trial was 90 grass and big sagebrush (Artemisia triden - treatments, including mineral soil, days. Air temperatures during the experi- t a t a Nutt. ssp. w y o m i n g e n s i s Beetle & Douglas-fir needle, ponderosa pine needle, mental period were obtained from the Young), needles of Douglas-fir and pon- sagebrush litter, fescue litter and cattle weather station at Kamloops Airport, derosa pine, and cattle manure on seedling manure. This was a pot experiment placed which is located about 1 km south of the emergence and establishment of Douglas- in the field with 4 replicates and the exper- experimental site. fir and ponderosa pine, and (2) if signifi- iment was repeated. Mineral soil from the Four hundred ml of distilled water were cant differences were found in seedling top 15 cm of the soil profile was collected applied to each pot at the beginning of the emergence and establishment among along a roadside in the Lac du Bois experiment to ensure that the soil reached seedbeds, whether they were attributed to Grasslands. The soil was then sifted field capacity. Four additional pots filled structural or chemical characteristics of through a 1.0 cm sieve and heated at 80°C with mineral soil were used as controls to these materials. for 48 hours to eliminate viable seeds. monitor soil water change and to ensure Other litter materials, including needles of no seeds survived the oven-drying Douglas-fir and ponderosa pine from process. Control pots were weighed every Materials and Methods ground, dry leaves of big sagebrush from 2 days before watering for soil water con- standing plants, standing dead leaves of tent determination. One hundred ml of dis- Seed collection and treatment rough fescue, and dry cattle manure more tilled water were added to each pot of the Two collections of Douglas-fir and 2 than 1-year old, were collected from the experiment on each of the first 3 checking collections of ponderosa pine seeds were Lac du Bois Grasslands. Douglas-fir nee- dates. Fifty ml of distilled water were obtained from British Columbia Ministry dles were separated from twigs, cones, and added each time when the soil moisture of Forests Tree Seed Centre and were rocks with a 3.5 mm sieve and then sorted content in the control pots was below 35% stored at –18°C before use (Table 1). on a 2.0 mm sieve to remove seeds and d.w., a pre-determined level about 80% of These seeds were collected from low ele- other small objects. Ponderosa pine nee- the soil field capacity. Seedling emergence vation sites (<1,000 m above sea level) in dles and the rough fescue litter were hand was checked every 2 days. Seedlings that the grassland/forest ecotone, within 50 km sorted and cut to fit into pots. Cattle reached at least 1 cm above the soil sur- from Kamloops. manure was broken into fragments of face (mineral soil treatment) or 3 mm In the seedbed experiment, seeds were about 1 cm in diameter. above litter materials (all other treatments) were considered emerged. Day of emer-

512 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 gence for each seedling was recorded and seed was considered germinated when the pine 2 (data not shown), which resulted in seedlings were marked with colored paper radicle was ³ 2 mm and germination was a significant interaction between treatment clips. Dead seedlings, determined by visu- checked every 2 days. Small amounts of and trial. The mean daily temperature in al inspection, were recorded and removed. distilled water were added to the filter the first 4 weeks was 1.5°C higher for trial The total length of each surviving seedling paper when needed. 2 than 1, and the soil water content in the was measured at the end of the experi- control pots was 2.4% lower between ment; seedlings were washed, oven-dried Data analysis weeks 2 and 4 for trial 2 than 1 (data not at 80°C for 48 hours, and weighed. Data for each seed collection were ana- shown). Bulk density of mineral soil and other lyzed separately. Seedling emergence rate Seedling emergence rate, measured by seedbed materials was determined as dry index (ERI)/seed germination rate index ERI, was significantly higher in manure weight per gross volume after oven drying (GRI) were calculated as the summation treatment and lower in Douglas-fir needle at 80°C for 48 hours. Net volume was esti- of percent seedling emergence/seed germi- and big sagebrush litter treatments than in mated by placing materials into beakers nation on each checking date divided by the mineral soil treatment (Table 2). The filled with a pre-determined amount of days, then multiplied by 100 (Evetts and emergence rate index (ERI) did not signif- distilled water. The increase in the volume Burnside 1972). Total seedling mortality icantly differ between mineral soil and of distilled water after adding these mate- was the percentage of dead seedlings over fescue litter treatments, but was similar or rials was determined after 10 seconds. The total emerging seedlings. Data of total lower in ponderosa pine needle compared relative volume of seedbed materials was emergence, mortality, and germination to mineral soil treatment. The interaction calculated as the percentage of net volume percentage were transformed with arcsine between treatment and trial was significant in gross volume. There were 8 replicates before being subjected to Analysis of for Douglas-fir 1 and ponderosa pine 2, for each seedbed type using different sam- Variance (Snedecor and Cochran 1980). A apparently caused by the lower ERI for the ples from the above seedbed experiment. 3-Way Analysis of Variance was used to mineral soil treatment in trial 2. analyze data of the seedbed experiment Water extract experiment (treatment, block and trial) and a 2-Way Seedling mortality and longevity as The 6 types of seedbed materials used in Analysis of Variance was used for the affected by seedbed type the above experiment as well as soils from water extract experiment (treatment and Total seedling mortality over the experi- beneath Douglas-fir needle, ponderosa trial). When interactions between treatment mental period ranged from 30 to 88% of pine needle, big sagebrush litter, fescue and trial were significant (P £ 0.05), data total emerging seedlings, but was general- litter and cattle manure were used in the were further analyzed within each trial. ly not affected by seedbed type (Table 2). water extract experiment. With distilled Means were separated with LSD test. Only in Douglas-fir 2 was seedling mor- water being the control, this experiment Seedling emergence data were pooled tality significantly higher in the Douglas- had 12 treatments arranged in a complete- according to species and collections and fir needle treatment than others. Longevity ly randomized design with 4 replicates and regression analysis was used to determine of seedlings which did not survive to the was repeated once. The amount of materi- relationships between seedling mortality end of the experiment was highest in min- als used for water extract was determined and day of emergence, and between eral soil, manure and fescue litter, lowest on a volume base: 1 x 12 x 12 cm for big longevity (difference between day of emer- in Douglas-fir needle, and intermediate in sagebrush litter, mineral soil and soils (top gence and day of death) and day of emer- ponderosa pine needle and sagebrush litter 15 cm) from beneath Douglas-fir needle, gence. Days in which less than 4 seedlings in 3 of the 4 collections. Seedling longevi- ponderosa pine needle, big sagebrush lit- emerged were excluded from regression ty of ponderosa pine 2 was not affected by ter, fescue litter and cattle manure; 3 x 12 analysis for mortality; best-fit equations seedbed type. x 12 cm for Douglas-fir needle, ponderosa were selected. Statistical significance for pine needle, fescue litter and cattle all tests was assumed at P £ 0.05. Seedling length and weight as affected manure. These materials were soaked in by seedbed type 400 ml of distilled water, shaken for 30 Results The length of seedlings that survived to min, and allowed to sit for 24 hours at the end of experiments was significantly room temperature (22 ± 1°C). The water affected by seedbed type for Douglas-fir, extract was obtained by filtering the solu- Effect of seedbed types on seedling but not ponderosa pine (Table 2). tion through 1 layer of Whatman No. 1 fil- emergence Seedlings were longest in mineral soil and ter paper then 1 layer of Whatman 0.2 µm Percent emergence and seedling emer- manure, shortest in sagebrush litter, and membrane filter paper. Seeds of 1 gence rate index (ERI) were significantly intermediate in Douglas-fir needle, pon- Douglas-fir collection and 1 ponderosa affected by seedbed types in both species, derosa pine needle, and fescue litter. pine collection were used in germination ranging from 2.5 to 84.6%, and from 0.06 Seedling dry weight was greatest in miner- tests with 50 seeds in each petri dish (9.5 to 5.15% day- 1, respectively (Table 2). al soil and manure, lowest in Douglas-fir cm in diameter). Five ml of water extract Compared to mineral soil treatment, emer- needle, and intermediate in others in all were added to seeds on top of 2 layers of gence from fescue litter and manure was seed collections except for ponderosa pine Whatman No. 1 filter paper and covered higher, and that from Douglas-fir needle 1, in which seedling weight was not with 1 layer of filter paper to ensure full and big sagebrush litter was mostly lower. affected by seedbed type. contact between seeds and water extract. Percent seedling emergence in the pon- Petri dishes were placed in clear plastic derosa pine needle treatment was not sig- Relation between seedling mortality, bags and incubated in a germinator at nificantly different from that in the miner- longevity, and day of emergence 30/20°C for 4 weeks, with 16-hour light/8- al soil treatment for both species. Mineral Pooling seedling mortality data accord- hour darkness. The top filter paper was soil treatment had lower seedling emer- ing to day of emergence showed that removed after 2 days of incubation. A gence in trial 2 than trial 1 for ponderosa

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 513 Table 2. Percent of seedling emergence, emergence rate index (ERI), mortality, longevity, and the length and weight of Douglas-fir and ponderosa pine seedlings as affected by seedbed type.

Species/Seedbed collection Emergence ERI Mortality Longevity Length Weight (%) (% day-1) (%) (day) (cm) (mg seedling-1) Douglas-fir 1 Mineral soil 48.4 ± 7.6 bc1 3.23 ± 0.47 b 49.5 ± 5.8 a 31.5 ± 1.7 ab 12.5 ± 1.2 a 19.5 ± 2.5 a Douglas-fir 12.5 ± 2.8 d 0.29 ± 0.07 e 61.9 ± 13.2 a 15.9 ± 4.0 c 10.5 ± 1.0 b 13.7 ± 2.6 b Ponderosa pine 55.6 ± 2.6 b 2.28 ± 0.29 c 58.5 ± 10.6 a 26.8 ± 2.7 b 10.3 ± 0.9 b 17.1 ± 1.0 ab Sagebrush 41.3 ± 5.9 c 1.31 ± 0.22 d 37.3 ± 4.7 a 11.9 ± 2.0 c 8.2 ± 0.6 c 17.1 ± 1.5 ab Fescue 75.9 ± 4.3 a 3.19 ± 0.22 b 43.1 ± 7.2 a 26.8 ± 2.1 b 10.5 ± 0.4 b 17.5 ± 1.8 ab Manure 82.2 ± 2.3 a 4.62 ± 0.23 a 29.3 ± 4.7 a 35.6 ± 5.2 a 12.4 ± 0.3 a 20.0 ± 0.7 a Douglas-fir 2 Mineral soil 47.8 ± 4.1 b 3.12 ± 0.29 bc 50.1 ± 5.6 b 29.4 ± 3.0 ab 12.4 ± 0.8 a 19.0 ± 1.3 a Douglas-fir 6.6 ± 2.1 d 0.14 ± 0.04 e 87.5 ± 6.6 a 14.0 ± 3.3 c 8.6 ± 0.4 cd 12.7 ± 3.0 cd Ponderosa pine 59.1 ± 2.6 b 2.42 ± 0.15 c 60.0 ± 5.6 b 24.7 ± 2.2 b 10.1 ± 0.7 bc 15.8 ± 0.8 bc Sagebrush 31.6 ± 7.0 c 1.04 ± 0.21 d 57.8 ± 10.2 b 16.5 ± 2.0 c 7.9 ± 0.9 d 11.4 ± 2.1 d Fescue 77.2 ± 2.1 a 3.50 ± 0.16 b 47.6 ± 3.8 b 27.2 ± 2.3 b 9.7 ± 0.4 cd 15.9 ± 0.7 abc Manure 84.6 ± 2.4 a 5.15 ± 0.17 a 34.7 ± 5.8 b 32.6 ± 2.0 a 11.8 ± 0.6 ab 18.8 ± 0.4 ab Ponderosa pine 1 Mineral soil 36.0 ± 9.0 b 2.00 ± 0.50 bc 50.6 ± 9.4 a 31.6 ± 3.3 a 19.1 ± 2.2 a 51.0 ± 8.1 a Douglas-fir 2.5 ± 1.3 c 0.06 ± 0.03 e 50.0 ± 0.0 a 2.0 ± 0.0 c 10.6 ± 1.2 a 37.0 ± 5.5 a Ponderosa pine 42.2 ± 4.8 b 1.34 ± 0.16 cd 37.7 ± 7.1 a 20.0 ± 3.0 b 16.2 ± 0.6 a 41.6 ± 1.9 a Sagebrush 23.6 ± 6.6 b 0.76 ± 0.22 d 60.0 ± 6.1 a 26.3 ± 6.6 ab 12.7 ± 1.5 a 40.0 ± 2.7 a Fescue 67.8 ± 5.0 a 2.22 ± 0.16 ab 61.5 ± 9.9 a 24.9 ± 1.6 ab 14.0 ± 1.5 a 42.1 ± 4.2 a Manure 67.2 ± 9.5 a 3.00 ± 0.42 a 57.0 ± 9.4 a 27.9 ± 2.9 ab 14.5 ± 1.3 a 47.8 ± 2.1 a Ponderosa pine 2 Mineral soil 41.1 ± 10.0 cd 2.30 ± 0.42 b 55.0 ± 4.6 a 32.3 ± 5.5 a 16.3 ± 2.0 a 44.9 ± 4.2 a Douglas-fir 3.1 ± 0.6 e 0.08 ± 0.02 e 42.9 ± 17.0 a 15.7 ± 1.7 a 12.6 ± 2.1 a 21.2 ± 2.7 c Ponderosa pine 49.7 ± 3.4 bc 1.48 ± 0.14 c 30.2 ± 5.9 a 18.2 ± 3.4 a 13.7 ± 1.1 a 34.9 ± 3.0 b Sagebrush 27.8 ± 5.1 d 0.91 ± 0.18 d 41.4 ± 8.7 a 18.2 ± 3.9 a 12.6 ± 1.2 a 36.1 ± 2.0 b Fescue 61.6 ± 3.5 b 2.16 ± 0.14 b 42.9 ± 7.8 a 21.7 ± 1.9 a 14.6 ± 1.1 a 38.8 ± 4.2 ab Manure 82.5 ± 4.3 a 3.48 ± 0.32 a 53.5 ± 9.2 a 25.7 ± 1.9 a 13.6 ± 0.7 a 38.0 ± 2.5 ab 1Values are mean ± SE of 8 replicates and data from 2 trials were pooled. Means with the same letter within a collection and a parameter are not significantly different at P £ 0.05. seedling mortality increased by nearly 1% Effect of water extract on seed germi- from soil beneath ponderosa pine needle day-1 with time of emergence in Douglas- nation enhanced germination by approximately fir, indicating that seedlings emerging Germination percentage and germina- 5%, and that from big sagebrush litter early in the season survived better than tion rate index (GRI) of ponderosa pine reduced germination by over 20%. Seed those emerging later (Fig. 1). Time of seeds were not affected by the 12 water germination in all other water extracts was emergence had no significant effect on extract treatments, averaging 89.8 ± 0.4% not significantly different from that in dis- seedling mortality in ponderosa pine as and 7.70 ± 0.17 % day -1, respectively (data tilled water and water extract from mineral indicated by the low r2 value (although for not shown). The treatment effect was sig- soil. The rate of germination, as measured ponderosa pine 2, P was less than 0.05). nificant for Douglas-fir (Table 3). by GRI, was lower in water extract from Longevity of those seedlings decreased Compared to distilled water and water big sagebrush litter and higher in water with time of emergence at a rate of extract from mineral soil, water extract extract of manure and mineral soil than in approximately 0.5 day- 1 for both species, distilled water. indicating that seedlings emerging early in Table 3. Seed germination percentage and germination rate index (GRI) of Douglas-fir as affected the season lived longer than those emerg- by water extract of seedbed materials and of soil beneath them. ing later (Fig. 2). Water extract Germination GRI Structural characteristics of seedbed (%) (% day-1) Both bulk density and relative volume Distilled water 90.7 ± 1.7 b1 9.69 ± 0.20 cd varied significantly among the 6 seedbed Douglas-fir 93.3 ± 1.1 ab 9.84 ± 0.22 c types (Fig. 3). Bulk density decreased in Ponderosa pine 90.5 ± 1.7 b 9.13 ± 0.33 d the order from mineral soil, cattle manure, Sagebrush 69.4 ± 3.5 c 7.13 ± 0.21 e Douglas-fir needle, big sagebrush litter, Fescue 90.3 ± 1.8 b 10.15 ± 0.27 bc ponderosa pine needle, to fescue litter. Manure 93.2 ± 1.1 ab 10.74 ± 0.22 ab Mineral soil 90.7 ± 1.8 b 11.04 ± 0.18 a Relative volume decreased in the order of Soil beneath Douglas-fir 91.8 ± 1.0 b 9.98 ± 0.28 c mineral soil, cattle manure, big sagebrush Soil beneath ponderosa pine 95.5 ± 1.1 a 10.31 ± 0.21 bc litter, Douglas-fir needle, ponderosa pine Soil beneath sagebrush 89.0 ± 1.6 b 10.35 ± 0.19 abc needle, and fescue litter, but there was no Soil beneath fescue 91.3 ± 1.6 b 9.85 ± 0.32 c significant difference between the latter 2 Soil beneath manure 91.2 ± 0.9 b 10.14 ± 0.22 bc types. 1Values are mean ± SE of 8 replicates. Means with the same letter within a collection and a parameter are not signifi- cantly different at P £ 0.05.

514 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Fig. 1. Seedling mortality of Douglas-fir and ponderosa pine as a function of day of emergence. Data for all treatments were pooled according to species and collection.

Fig. 2. Seedling longevity of Douglas-fir and ponderosa pine in relation to day of emergence. Data for all treatments were pooled according to species and collection.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 515 duced a different architecture, modifying temperatures under the litter layer (Bristow 1988). Mineral soil and manure fragments had the highest bulk density and relative volume among the 6 types of materials in our study, providing maxi- mum contact with seeds and subsequently optimal water supply for seed germina- tion. On the other hand, ponderosa pine needle and fescue litter were characterized by low bulk density and low relative vol- ume. Therefore, seeds are not likely trapped in their loose structure, but fall through, contacting mineral soil. Douglas- fir needle and big sagebrush litter had moderate bulk density and relative volume and seeds are more likely to be trapped within these materials, resulting in low seedling emergence. The 2 factors, namely structural and chemical characteristics, may interactively affect seed germination Fig. 3. Structural characteristics of the 6 seedbeds. Data are mean ± SE of 8 replicates. MS: as reported in the effect of oak litter on mineral soil, DF: Douglas-fir needle, PP: ponderosa pine needle, SB: big sagebrush litter, seedling emergence of table mountain pine F: rough fescue litter, M: cattle manure. (Pinus pungens Lam.) (Williams et al. 1990). Discussion seedling emergence observed in the field. The effect of structural characteristics of Allelopathy has been reported in a wide seedbeds is related to moisture availability, range of species and the effect was species which is highly variable in the field and may Substrate condition or seedbed is a specific (Grant 1964, Ballester et al. 1979, result in different results from controlled major determinant in the regeneration Kil and Yun 1992). The water extracts of experiments such as the current study. niche for seedlings (Grubb 1977). The 6 the 6 types of seedbed materials as well as Regardless of seedbed types, over 50% of seedbed types tested in the present study, of soils beneath them had no effect on seedlings of Douglas-fir and ponderosa pine including Douglas-fir and ponderosa pine seed germination of ponderosa pine and emerged in the first 4 weeks (data not needles, are representative of the forest- only a marginal effect on Douglas-fir. The shown). Seedlings emerging early in the sea- grassland ecotone in British Columbia, water extract of big sagebrush litter, but son have survival advantage (Miller 1987, because conifer trees are common in the not of Douglas-fir needle as previously Streng et al. 1989). Grasslands of British grasslands as a result of encroachment. reported (Daniel and Schmidt 1972), Columbia are characterized by low annual Bare soil surface is generally favorable for reduced seed germination in Douglas-fir. precipitation and extremely dry summers conifer seedling emergence because litter It is possible that allelopathy, even if it (Spilsbury and Tisdale 1944, Tisdale and accumulations on forest floors restrict seed exists in some of the materials studied, McLean 1957) and conifer seedling estab- germination and seedling establishment of could have been reduced during decompo- lishment usually occurrs when moisture con- conifer trees (Bramble and Goddard 1942, sition, since fresh needles or litter were ditions were close to or above the long-term Chrosciewicz 1974, Van Wagner 1983), mostly absent on the soil surface when average (Arno and Gruell 1983, 1986). including Douglas-fir (Caccia and Ballaré samples were collected in the spring. Successful encroachment of these 2 tree 1998) and ponderosa pine (Roe and Nevertheless, structural characteristics of species may therefore depend on the ability Squillance 1950, Foiles and Curtis 1965). these materials may provide a better of seeds to germinate early in the season In the present study, both Douglas-fir nee- explanation of the observed differences when soil moisture conditions are favorable dle and big sagebrush litter restricted among seedbeds. and on seedbeds that can provide sufficient seedling emergence of Douglas-fir and The effect of structural characteristics of mo i s t u r e . ponderosa pine compared to mineral soil, seedbed materials on seed germination and In conclusion, seedbeds in grasslands of but fescue litter and dry cattle manure seedling emergence has not received as British Columbia, such as mineral soil, fragment actually enhanced the emergence much attention as allelopathy, possibly fescue litter and cattle manure fragment, of the 2 species. Interestingly, ponderosa due to the difficulty of quantitative com- favor seed germination and seedling estab- pine needles did not restrict seedling parison among seedbeds. In general, litter lishment of Douglas-fir and ponderosa emergence of its own or Douglas-fir, pos- materials buffer soil temperatures and help pine compared to those on forest floors, sibly because the water availability in our conserve soil moisture (Sydes and Grime such as Douglas-fir needle. Differences experiment was less limited than in other 1981, Fowler 1986, Williams et al. 1990). among seedbeds are most likely due to field studies (Roe and Squillance 1950, On the other hand, seeds can be entrapped their structural, but not chemical charac- Foiles and Curtis 1965). between layers of litter (Burbidge 1945), teristics. Fire suppression, grazing and tree The effect of litter on seed germination preventing germination by reducing the and seedling emergence depends on litter logging will modify the soil surface, and contact between seeds and surrounding type and species studied; both the physical other factors such as weather conditions materials. Even for the same type of litter, structure and chemical composition of lit- and competition will interactively influ- horizontal or vertical arrangement pro- ter may contribute to differences in ence seedling establishment of the 2

516 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 species. Field studies are therefore neces- Gruell, G.E. 1983. Fire and vegetation trends Tisdale E.W. and A. McLean. 1957. T h e sary for isolating factors causing tree in the Northern Rockies: Interpretations from Douglas fir zone of southern British encroachment for specific areas. 1871-1982 photos. USDA Forest Serv. Gen. Columbia. Ecol. Monogr. 27:247–266. Tech. Rep. INT-158. Ogden, Ut. Van Wagner, C.E. 1983. Fire behaviour in Hill, J.D., C.D. Canham, and D.M. Wood. northern conifer forests and shrublands, p. 1 9 9 5 . Patterns and causes of resistance to 65-80. I n: R.W.Wein and D.A. MacLean Literature Cited tree invasion in rights of way. Ecol. Appl. (eds.), The role of fire in northern circumpo- 5:459–470. lar ecosystems. Wiley, New York, N.Y. Arno, S.F. and G.E. Gruell. 1983. Fire history Kil, B.S. and K.W. Yun. 1992. A l l e l o p a t h i c Whitford, H.N. and R.D. Craig. 1918. Forest at the forest-grassland ecotone in southwest- effects of water extracts of Artemisia prin - of British Columbia. Canada Comm. on ern Montana. J. Range Manage. 36:332–336. ceps var. orientalis on selected plant species. Forests, Ottawa, ON. Arno, S.F. and G.E. Gruell. 1986. Douglas-fir J. Chem. Ecol. 18:39–51. Williams, C.E., M.V. Lipscomb, W.C. encroachment into mountain grasslands in MacDonald, I.A.W. and C. Wissel. 1991. Johnson, and E.T. Nilsen. 1990. I n f l u e n c e southwestern Montana. J. Range Manage. Determining optimal clearing treatments for of leaf litter and soil moisture regime on 39:272–276. the alien invasive shrub A c a c i a s a l i g i n a i n early establishment of Pinus pungens. Amer. the Southwestern Cape, South-Africa. Agr. Ballester, A., A.M. Vietez, and E. Vietez. Midland Natur. 124:142–152. Ecosystems and Environ. 39:169–186. 1979. The allelopathic potential of Erica aus - Mast, J.N., T.T. Veblen, and Y.B. Linhart. t r a l i s L. and E. arborea. Bot. Gazette 1998. Disturbance and climatic influences on 140:433–436. age structure of ponderosa pine at the pine- Bramble, W.C. and M.K. Goddard. 1942. grassland ecotone, Colorado Front Range. J. Effect of animal coaction and seedbed condi- Biogeography 25:743–755. tion on regeneration of pitch pine in the bar- McLaren, B.E. and R.A. Janke. 1996. rens of central Pennsylvania. Ecol. Seedbed and canopy cover effects on balsam 23:330–335. fir seedling establishment in Isle Royale Bristow, K.L. 1988. The role of mulch and its National Park. Can. J. Forest Res. architecture in modifying soil temperature. 26:782–793. Australian J. Soil Res. 26:269–280. Miller, T.E. 1987. Effects of emergence time Burbidge, N.T. 1945. Germination studies of on survival and growth in an early old-field Australian Chenopodiaceae with special ref- plant community. Oecologia 72:272–278. erence to the conditions necessary for germi- Parminter, J.V. 1978. Forest encroachment nation. I. Atriplex vesicaria H. Trans. Royal upon grassland range in the Chilcotin region Soc. of South Australia 69:73–85. of British Columbia. M.Sc. Thesis, Univ. of Burkhardt, J.W. and E.W. Tisdale. 1976. British Columbia. Vancouver, BC. Causes of juniper invasion in south-western Richardson, D.M. and W.J. Bond. 1991. Idaho. Ecol. 57:472–484. Determinants of plant distribution: Evidence Caccia, F.D. and C.L. Ballaré. 1998. E f f e c t s from pine invasion. Amer. Natur. of tree cover, understory vegetation, and lit- 173:639–668. ter on regeneration of Douglas-fir Roe, A.L. and A.E. Squillance. 1950. Can we induce prompt regeneration in selective-cut (P s e u d o t s u g a m e n z i e s i i) in southwestern ponderosa pine stands? USDA For. Serv., Argentina. Can. J. Forest Res. 28:638–692. North Rocky Mountain Forest and Range Chrosciewicz, Z. 1974. Evaluation of fire-pro- Exp. Sta, Res. Note 81. Missoula, Mont. duced seedbeds for jack pine regeneration in Ross, T.J. 1997. Forest ingrowth and forest central Ontario. Can. J. Forest Res. encroachment on Bald Mountain and Becher 4:455–457. Prairie between 1962 and 1993/95. Prepared Daniel, T.W. and J. Schmidt. 1972. L e t h a l for: BC Ministry of Agr., Cariboo/Chilcotin and nonlethal effects of the organic horizons Grazing Enhancement Fund. Ross Range and of the forested soils on the germination of Reclamation Services, Cranbrook, BC. seeds from several associated conifer species Snedecor, G.W. and W.C. Cochran. 1980. on the Rocky Mountains. Can. J. Forest Res. Statistical methods (7th ed.). Iowa State Univ. 2:179–184. Press. Ames, Iowa. Evetts, L.L. and O.C. Burnside. 1972. Spilsbury, R.H. and E.W. Tisdale. 1944. Soil- Germination and seedling development of plant relationships and vertical zonation in common milkweed and other species. Weed the southern interior of British Columbia. Sci. 20:371–378. Sci. Agr. 24:395-436. Foiles, M.W. and J.D. Curtis. 1965. N a t u r a l Strang, R.M. and J.V. Parminter. 1980. regeneration of ponderosa pine on scarified Conifer encroachment on the Chilcotin grass- group cuttings in central Idaho. J. Forest. lands of British Columbia. Forest. Chronicle 63:530–535. 56:13–18. Fowler, N.L. 1986. Microsite requirements for Streng, D.R., J.S. Glitzenstein, and P.A. germination and establishment of three grass Harcombe. 1989. Woody seedling dynamics species. Amer. Midland Natur. 115:131–145. in an East Texas floodplain forest. Ecol. Grant, E.A. 1964. Influence of plant extracts Monogr. 59:177–204. Sydes, C. and J.P. Grime. 1981. Effects of on germination and growth of eight forage tree leaf litter on herbaceous vegetation in species. J. Brit. Grassl. Soc. 19:191–197. decidous woodland. II. An experimental Grubb, P.J. 1977. The maintenance of species- investigation. J. Ecol. 69:249–262. richness in plant communities: The impor- Tisdale, E.W. 1950. Grazing of forest lands in tance of the regeneration niche. Biol. Rev. interior British Columbia. J. Forest. 52:107–145. 48:856–860.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 517 J. Range Manage. 53: 518–528 September 2000 Grassland biomass dynamics along an altitudinal gradient in the Pampa

CAROLINA A. PÉREZ AND JORGE L. FRANGI

Authors are plant ecologists of the Laboratorio de Investigación de Sistemas Ecológicos y Ambientales (LISEA), Universidad Nacional de La Plata, CC 31, 1900 La Plata, Argentina. E-mail: [email protected]

Abstract Resumen

Above and below-ground biomass and necromass dynamics Se evaluó la dinámica de la biomasa y necromasa, aéreas y were assessed for 3 grassland sites located at 550, 850, and 1,025 subterráneas, de tres sitios de pastizal ubicados a 550, 850, y m elevation in Sierra de la Ventana range (38˚1'S 62˚2'W) in 1,025 m de elevación en la Sierra de la Ventana (38˚1'S 62˚2'O), Argentina. The objective was to determine if differences existed Argentina. El objetivo fue determinar si existían diferencias en la in dry matter structure, mycorrhizae infection, net primary pro- estructura de la materia seca, la micorrización, la asignación de ductivity (NPP) partitioning to aboveground and belowground la producción primaria neta (NPP) a tejidos aéreos y subterrá- tissues, senescence and litter fall, and seasonal patterns of dry neos, la senescencia y caída al mantillo, y los patrones de los flu- matter fluxes with altitude. Soil properties, water budgets and jos de materia seca con la altitud. También se consideraron las temperature at the sites were also assessed. Biomass plus necro- propiedades del suelo, los balances de agua y la temperatura de mass (without litter) was 1,184 ± 41, 1,208 ± 70, and 1,507 ± 63 cada sitio. La biomasa más la necromasa (sin broza) fue de 1,184 gDM m-2 for the lower, intermediate and upper sites, respectively. ± 41, 1,208 ± 70 y 1,507 ± 63 gMS m- 2 para los sitios inferior, The below:aboveground biomass ratio increased with elevation. intermedio y superior, respectivamente. El cociente biomasa Total NPP was 1,131, 1,280, and 1,157 gDM m- 2 y e a r- 1, respec- subterránea:aérea se incrementó con la elevación. La NPP total tively, for the 3 grassland sites. belowground allocation of net fue 1,131, 1,280 y 1,157 gMS m-2 año-1, respectivamente, para los productivity increased with altitude. Both mass and proportion tres sitios de pastizal. La asignación subterránea de la produc- of thin roots increased with elevation, and so did mycorrhizae tividad primaria neta se incrementó con la altitud. La masa y la infection. The aboveground and belowground turnover rates proporción de raíces finas, como también el porcentaje de micor- decreased with altitude, but rates were faster for aboveground rización, aumentaron con la elevación. Las tasas de renovación tissues. We found different temporal patterns in productivity, aérea y subterránea decrecieron con el aumento de la altitud, senescence and disappearance among grassland sites despite sim- pero las tasas fueron más rápidas para los tejidos aéreos. ilar total NPP. Water holding capacity of soils and temperature Nosotros encontramos distintos patrones temporales en la pro- were important factors related to several of the observed trends ductividad, senescencia y desaparición entre pastizales a pesar in structure and function. Differences in grassland structure and de su similar NPP. La capacidad de retención de agua del suelo y fluxes are discussed as related to soils and local climate at each la temperatura fueron importantes factores relacionados a site. varias de las tendencias observadas en la estructura y fun- cionamiento de los pastizales. Las diferencias en la estructura y los flujos de los pastizales son comentadas en relación a los suelos Key Words: aboveground, belowground, litter, phytomass, pro- y clima local de cada sitio. ductivity, roots

Natural grasslands dominate the landscape of the Sierras peak Tres Picos (1,243 m elevation) (Harrington 1947, Suero Australes (Buenos Aires, Argentina). The grass genera with more 1972). The macroclimate at this area has been described else- native species in the area are S t i p a and P i p t o c h a e t i u m; other where (Burgos 1968, Cappanini et al. 1971, Frangi and Bottino important species because of their abundance include B r i z a, 1995). According to Thornthwaite (1948) the climate is C2 B'2 r S o r g h a s t r u m and F e s t u c a. These hills (spanish s i e r r a s) are a a' -humid-subhumid, mesothermal, with a small to null water gondwanic folding system extending NE to SW, 170 km in length deficit, and a summer thermal concentration <48 %. Because of x 65 km maximum width. The highest point in the Pampas is the low elevation of the hills, and being perpendicular to the atmosphere general circulation, they are not an effective con- denser of atmospheric humidity but are effective in temperature Research was funded in part by Comisión de Investigaciones Científicas de la reduction. Precipitation diminishes from NE to SW. Sierra de la Provincia de Buenos Aires, Universidad Nacional de La Plata and the Cooperadora Ventana (249 m asl), the nearest town to the study area, has an Parque E Tornquist, and developed in cooperation with the International Institute annual-long term mean temperature of 14.5°C and 809 mm pre- of Tropical Forestry-USDA Forest Service, Puerto Rico. Authors thank Marcelo cipitation. In the Pampas, the interannual variation of rains is ca. Arturi, Martha Cabello, Ricardo Varela and Walter Canónica for statistical, mycor- rhizal, geological and field assistance, respectively. We also thank Marcelo ± 50% of the mean (Vervoorst 1967, Cappaninni et al. 1971). Barrera, Richard Joost, Joe Trlica and an anonymous reviewer whose comments The local geology is characterized by siliciclastic sedimentites helped to improve earlier versions. of the Ventana Group, Lower Devonic age, containing fossils of Manuscript Accepted 17 Jan. 2000. Brachyopoda (Andreis et al. 1989, Von Gosen et al.1990). They

518 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 mostly consist of cross-bedded quartzites shrub. Climatic conditions are more ty, senescence and disappearance at 3 and sandstones, with grain-size diminish- extreme; they are colder and windy with grassland sites located at 550, 850, and ing upward cycle. Group thickness is the highest evaporative air capacity (Piche 1,025 m asl. Soil properties, water bud- about 1,400 m. Wide and shallow plat- method) that is reduced in the winter gets, and temperature at these sites were forms and littoral zones constituted the because of clouds and mist. also determined. paleoenvironment. Rocks were deformed Vegetation differs in physiognomy, by folding and internal faulting. The species composition, and phytogeographic resulting sierras profile was asymmetric, relationships among plant species (de la Methods with the eastern flank more extended and Sota 1967, Frangi and Bottino 1995). formed by steps motivated in the structural Biomass structure changes as well with Study sites repetition by folding of more resistant elevation, slope and aspect (Barrera and The 3 sites were located on an eastern lithological levels. The soils are lithic Frangi 1994). aboveground biomass Hapludols, typic and lithic Argiudols and secondary divide at peak El Destierro I decreases and belowground biomass (38°1'S 62°2'W), Ernesto Tornquist lithic Haplumbrets developed on quater- increases with altitude. aboveground net nary loessoid sediments sometimes mixed Nature Reserve. Slope was about 3% at all primary productivity (ANPP) has only sites. Plant communities (Frangi and with fragments of base rock or intercalated been measured in low altitude grassland CaCO rock layers (Cappanini et al. 1971, Bottino 1995) were described as short 3 sites (Frangi et al. 1980a, 1980b). grassland with Piptochaetium hackelii Vargas Gil and Scoppa 1973, Frangi et al. Mountain grasslands are used as grazing 1980a, 1980b, Pérez 1996). (Arechavaleta) Parodi and P. napostaense lands (Ricci 1992). Overgrazing encour- (Spegazzini) Hackel, with Briza subarista - Grassland communities have been ages bush encroachment (Barrera 1991, linked to differences in biotope and meso- t a Lamarck dominating the low site (550 Barrera and Frangi 1996). Non-disturbed m). Short grassland with Sorghastrum pel - climate (Frangi and Bottino 1995, grasslands are important in erosion control Kristensen and Frangi 1995). The greatest litum (Hackel) Parodi and Stipa filiculmis and water quality of tributaries of the Delile dominate the intermediate and high dissimilarity in local climates was between Sauce Grande river, a source of consump- lowland and mountain grassland sites. The sites (850 and 1,025 m). In this paper we tion water for the city of Bahía Blanca refer to the sites as lower (LS) 550 m ele- former are warm, anisothermic with high- (Dascanio and Bianchi pers. com.). Exotic er diurnal evaporative capacity. They are vation, intermediate (IS) 850 m elevation, tree plantations are assumed by local peo- and upper (US) 1,025 m elevation. covered by medium-sized feather grasses ple to protect steep slopes but symptoms (Stipa and Piptochaetium species) in the of overland erosion under forest canopy piedmont and medium-sized tussocks of are common (Frangi and Bottino 1995, Soils Stipa caudata Trinius or S. ambigua Barrera and Frangi 1996). Six to 9 soil cores (0–20 cm depth) were Spegazzini and S. caudata Trinius grass- obtained at random from a 0.5 ha area, in 1 The compositional, physiognomical and lands in deep soils of the valley bottoms . structural changes in biomass with topog- spring, fall and winter, from each site. Steep SW slopes differ markedly in local raphy, soil and local climate caused us to These samples were immediately cleaned climate and grass flora compared with select an altitudinal gradient with a more of roots and air-dried. A mixed composite gentle slopes (Frangi and Bottino 1995, limited complexity to answer the follow- soil sample for each site and date was pre- Kristensen and Frangi 1995). Steep SW ing questions: Are differences in dry mat- pared for physico-chemical analysis (n = 3 per site). Traditional methods were slopes are cold, isothermic, and very ter structure correlated with net primary employed for texture using a sieving humid, receive low solar radiation during productivity (NPP) changes or only with machine and densimeter (Duchaufour most of the year and are covered by partitioning among above- and below- Festuca pampeana Spegazzini tussock 1970). Soil pH was determined in a 1:1 ground tissues? Do seasonal patterns in (soil:water) paste using a pH meter with a grasslands. dry matter fluxes change along the altitu- The NE and SW gentle slopes are cov- combination electrode. Total soil nitrogen dinal gradient? The hypotheses were: (1) (N) and carbon (C) were determined using ered by Sorghastrum pellitum ( H a c k e l ) Above- and belowground community bio- Parodi and Stipa filiculmis Delile short the CNS LECO 2000 procedure mass structure changed with elevation to (Tabatabai and Bremner 1991). Available grassland. They show intermediate condi- reflect adaptations to improve available tions between the upper and lowland sites. P was analyzed with the Bray & Kurtz #1 resource absorption; (2) NPP diminished method (Page 1982). The cation exchange The thermal gradient supported the exis- with altitude because of less favorable tence of a montane bioclimatic belt above capacity (CEC) was assessed using 1N local climate and soil conditions; (3) NPP neutral ammonium acetate extraction 750 m. allocation changed with elevation toward a (Page 1982). Exchangeable Ca, Mg, and Mountain summits and high ridges are greater proportion of roots at higher eleva- Na were determined using 1N KCl extrac- covered with short grasses and cushion tions; and, (4) Community differences in tion method and exchangeable K with plants. Fine textured soils give rise to high temporal patterns of dry matter fluxes Olsen EDTA procedure (Hunter 1982). sierra meadows with Briza subaristata indicated differences in plant strategies to Additionally, during the spring sampling, Lamarck, B. brizoides (Lamarck) O. changes in temperature and water avail- apparent density (AD) was estimated for 3 Kuntze and Festuca ventanicola ability along the elevational gradient. soil cores of known volume (6.5 cm diam. Spegazzini in concave to flat more humid The objectives were to determine the x 10 cm height) at 0–10 and 10–20 cm sites, with Sorghastrum pellitum and Stipa biomass and necromass structure includ- depth per site and dried at 70°C to con- filiculmis short grassland on upper slopes. ing root sizes and mycorrhizae infection, stant weight. An average apparent density Stony soils are covered by G r i n d e l i a net primary productivity partitioning to for 0–20 cm soil depth was calculated for c h i l o e n s i s Cabrera an evergreen dwarf above- and belowground tissues, senes- each site (n = 3) discarding weight of peb- cence and litter fall, and seasonal patterns bles. Soils were classified according to the 1Species nomenclature follows Cabrera (1963, of above- and belowground net productivi- Soil Survey Staff (1992). 1970).

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 519 Water budgets roots. Plastic bags containing sampled Statistics Thornthwaite’s (Thornthwaite and plants, litter, or soils were stored in a freez- The differences in soil parameters Mather 1957) water budgets were calculat- er (–12°C) until they were separated, clas- among sites were compared with a one- ed for each site using a computer program. sified and dried to constant weight at 70°C. way ANOVA. The plant compartment dif- Water retention maximum capacity ferences among sites and dates were eval- (WRM) at each site was estimated accord- Mycorrhizae uated using a repeated measure ANOVA. ing to: We estimated arbuscular (A-) mycor- Previously we corroborated the variance WRM (mm) = (0.023 x sand fraction rhizae fungal infection of roots in each homogeneity. To compare percentages of + 0.25 x silt + 0.61 x clay) x AD (g grassland in December 1988. Six soil sam- aboveground and belowground biomass cm-3) X depth (cm) x 10. ples were taken from each soil depth. compartments with total biomass we trans- formed the data to arc sin (%1 / 2). We Monthly temperatures were estimated Roots were extracted, clarified with KOH (7.5%) at 90°C for 20-minutes and stained employed the F-test to determine signifi- with data from the Sierra de la Ventana cant differences between dates and sites weather station (SMN 1981), altitudinal employing the procedure of Phillips and Hayman (1970). Infection was measured and the studentized rank test of Tukey (P thermal gradients for each season < 0.05) for mean comparisons (Sokal and (Kristensen and Frangi 1995) and site ele- with a line-intercept method under a 40X stereomicroscope (Giovannetti and Mosse Rohlf 1969). Multiple regression analysis vation. Weekly precipitation during the was applied to explain relationships study period was measured with 1 stan- 1980). A-mycorrhizal spores were extract- ed with the wet-sieving and decant tech- among plant mass compartments or ratios dard US Weather Bureau type rain gauge and independent variables. This analysis per site. nique (Gerdemann and Nicolson 1963) and counted under a 10X stereomicroscope. was performed with the pooled data of the 3 sites. The partial correlation of this vari- Biomass and necromass able was calculated after fitting other inde- We used the following definitions: bio- Fluxes pendent variables into an equation. When mass was dry weight of live plant materi- For aboveground net primary products a non-significant partial correlation was al; necromass was dead plant mass (stand- (ANPP), senescence, litterfall, and disap- found, it was interpreted that the other ing dead, litter, and dead roots); above- pearance calculations we used the sum of independent variables accounted for most ground dry matter was aboveground bio- differences between live and dead com- of the differences among sites. mass plus standing dead; belowground dry partments (Frangi et al. 1980 a). Likewise matter was the sum of live and dead roots we modified the equations to calculate and, total dry matter was biomass plus belowground fluxes (Table 1). We used Results necromass. only the significant differences (Student t, Sampling was done in 0.5 ha grazing P < 0.05) between the extremes of inter- Soils and water budget vals with increasing or decreasing values exclosures, at 9 dates between July 1988 Soils at all sites were classified as Lithic trends of each compartment. and July 1989. A stratified randomized Hapludolls, notwithstanding a lower base Total net primary products (NPP) was design was used for sampling above- and saturation of some samples that indicated the sum of ANPP and belowground net belowground live and dead mass compart- at intermediate site (IS) and upper site primary products (BNPP). Aboveground ments. The aboveground dry matter was (US) there were Lithic Haplumbrepts also and belowground turnover rates were cal- estimated by the harvest method (Milner present. The texture was clay loam to culated as the ratio between productivity and Hughes 1970). At each grassland site loam at lower site (LS), loam at IS and and the annual mean of biomass of above and on each date we collected ten, 0.25 x 1 sandy loam at US. Clay decreased by 50% and belowground compartments, respec- m pseudoreplicates (according to Hurlbert between LS and US (Table 2). The organ- tively (Frissel 1981). Dry matter turnover 1984). Plant matter and litter were collect- ic carbon values were very high and the rates were calculated employing the equiv- ed in separate bags. To avoid resampling a C:N ratios are similar and common to alent dry matter values. plot, we put a stake in each clipped Pampa grassland soils. Soil pH decreased quadrat. Plant matter was hand-separated with elevation and was moderately acid at into categories: green grasses, green dicotyledonous herbs, standing dead mate- rial, and reproductive tissue. Table 1. Formula to calculate aboveground and belowground dry matter fluxes, net primary pro- ductivity (ANPP = aboveground; BNPP = belowground), senescence (M), litterfall (C) and disap- The belowground dry matter in 0–10 pearance (D). and 10–20 cm soil depth was estimated using a 6.5 cm diam. x 10 cm height soil Aboveground core. On each date and site we extracted ANPP = DB+ + (M - DB-) = DB+ + m 10 cores from each soil depth at each - - + M = DB + m being m = DS – DB + DH (whenever m > 0) aboveground harvest unit. Cores were C = DH+ + c being c = DB- – DS – DH+ (whenever c > 0) 2 processed by wet sieving over a 1-mm D = d – DH being d = DB- – DS (whenever D > 0 and d >0) mesh screen (Böhm 1979). Collected roots Where: DB+: Positive increment in biomass, DB-: Decrease in biomass, DS: Change in standing dead were washed free of soil, visually classi- stock, DH+: Positive increment of litter, DH: Change in litter stock. fied, and separated with tweezers into live Belowground (soft, translucent, white to brownish) or BNPP = DRV+ + (M – DRV-) =DRV+ + m dead (brittle, opaque, and grayish) accord- M = DRV- + m being m = DRM - DRV- (whenever m > 0) ing to Böhm (1979) and McClaugherty et D = DRV- – DRM (whenever D > 0) al. (1982) and subdivided according to Where: DRV+: Positive increment in live roots, DRV-: Decrease of live roots, DRM: Change in dead roots diameter as thin (<1mm) or thick (>1mm) stock.

520 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 LS and moderately to very acid at both other sites. The CEC was similar at all sites, but its variation increased with ele- vation. Base saturation was high at low altitude and moderate at high elevation (Table 2). A traditional fertility data inter- pretation indicated that these soils were very well supplied in total N, moderately supplied in P, and adequately supplied in exchangeable K. Exchangeable Ca was well supplied at LS, but its CEC-saturation was significantly lower at IS and US. Magnesium was high at all sites and showed the same trend as Ca. Low values of apparent soil bulk density can be related to the high organic matter content, with inter-site differences attribut- able to texture changes and also to a greater fraction of discarded pebbles at IS and US (Table 2). Annual precipitation (P) increased with elevation, with 745, 786, and 828 mm at the LS, IS, and US, respectively. Days with rains were 29, 30, and 25 at LS, IS, and US, respectively, and maximum con- secutive days without rains during the sampling intervals were about 17 days. Potential (PET) and actual evapotranspira- tion (AET) decreased with elevation whilst P/PET and P/AET ratios increased with elevation. The P/PET ratio was 1.1, 1.2, and 1.4 for LS, IS, and US, respec- tively, and the P/AET ratio was 1.41, 1.56, and 1.78 at the same site sequence. Maximum water retention capacity of soil diminished with elevation, with upper site 50% of lower site. The values observed were influenced by the shallow soil depth considered (20 cm) that corresponded to the rooting zone but also to the total profile at Fig. 1. Thornthwaite's water budgets in grasslands sites located at different altitudes from IS and US. The water budget showed clear- July 1988 to July 1989. WRM = maximum water retention capacity; PET = potential ly the alternation of humid and dry periods evapotranspiration; AET = actual evapotranspiration. in the soils (Fig. 1). The humid periods

Table 2. Physico-chemical characteristics of soils at lower (550 m asl), intermediate (850 m asl) and occurred during early spring and fall at all upper (1,025 m asl) grassland sites. Values are ranges or means for composite samples (n = 3) sites. The late-spring water deficit was per site. Different letters indicate significant differences among the 3 sites (Tukey, P < 0.05). greater at both IS and US. Summer is clear- ly the main water consumption period but Soil parameter Lower Intermediate Upper some recharge also happened; however, Apparent density (g cm-3) 0.850 a 0.649 b 0.606 b most recharge occurred in the autumn. pH (soil/water 1:1) 5.4 to 5.9 a 4.3 to 5.4 a 4.1 to 5.2 a Temperatures favored water consumption Texture (%) during winter at all sites but at US there Sand 40.50 b 44.40 b 58.40 a was a water surplus. Silt 30.70 a 39.00 a 28.00 a Clay 28.90 a 16.60 b 13.60 b Organic Carbon (%) 4.91 a 5.64 a 5.90 a Plant compartments mass Total Nitrogen (%) 0.44 a 0.48 a 0.55 a Total dry matter was greater at the upper C:N ratio 10.9 a 11.8 a 10.9 a -2 P available (ppm) 12.17 a 10.50 a 11.45 a site (US) (1,552 g m ) by 20% more than -1 -2 CEC (cmolc kg ) 22.20 a 22.30 a 20.15 a either of the other 2 sites (1,260 g m at IS Exchangeable bases (cmolc kg-1) and 1,225 g m-2 at LS). The belowground Ca+2 15.43 a 9.07 b 6.19 c +2 dry matter increased with altitude and was Mg 6.56 a 4.31 b 2.78 c 52, 62, and 73% of total dry matter at the Na+ 0.44 a 0.29 a 0.28 a + lower site (LS), intermediate site (IS), and K 0.84 a 0.53 a 0.43 a Base saturation (%) 90.0 to 98.6 a 50.0 to 92.2 a 39.0 to 68.3 b upper site (US), respectively. Litter

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 521 68% at US; the last 2 were not significant- ly different. Normalized proportions of each live compartment with respect to total biomass changed significantly among sites and dates. The green fraction decreased with increasing elevation. Thin roots were the main contributing root compartments to belowground biomass and were responsi- ble for the increase in total live roots with increasing elevation (Fig. 2b). Roots colo- nized by endophytes forming A-mycor- rhizae changed significantly from 40 to 57% with elevation (Fig. 3). Spores showed a non-significant decreasing trend in the same direction (20, 19, and 14, respectively). Inter-site differences in biomass and dry matter compartments over time resume a distinct balance of functional processes at each site and date, as indicated by statisti- cally significant interactions between loca- tion and sampling dates (Fig. 4). The US showed smooth changes in plant compart- ments over time, whereas both other sites manifested more noticeable stock peaks at different months.

Plant mass and the physical environ- ment The multiple regression analyses per- formed for aboveground dry matter, belowground dry matter, belowground biomass, or above-/belowground biomass ratio with independent variables of month- Fig. 2. Annual mean dry weight for different plant compartments in grasslands sites: a) ly mean minimum temperature, monthly aboveground (A-), belowground (B-) and total dry matter, all in (g m-2); b) Aboveground (A-) biomass and live thin and thick roots (%) with respect to total biomass (g m- 2) . available soil water, and monthly actual Different letters indicate significant differences among compartment mean for the 3 sites evapotranspiration were significant (Table (Tukey, P < 0.05). Least squares trend lines are indicated with broken line. 3). The regression coefficients of monthly attained only about 3% of total dry matter at all sites. Above- and belowground dry matter dif- fered according to location and sampling dates. aboveground dry matter decreased with elevation, while belowground dry matter increased with elevation (Fig. 2a). The rate of decrease in aboveground dry matter was more pronounced (43 g m-2 per 100 m elevation) between LS and IS than between IS and US (21 g m-2 per 100 m). The rate of increase in belowground dry matter was greater between IS and US (199 g m- 2 per 100 m) than between LS and IS (47 g m-2 per 100 m). Although aboveground dry matter was mainly standing dead and aboveground biomass was similar at all sites, the rela- tive contribution of live tissue increased with elevation (27, 34, and 36% at LS, IS, and US, respectively). belowground dry matter was mainly contained in the 0-10 Fig. 3. Thin-/thick live roots ratio and mycorhizae infection (%) in grasslands sites at differ- cm soil layer: 71% at IS, 64% at LS and ent altitudes from July 1988 to July 1989. Vertical lines indicate ± SE.

522 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 tively) (Table 4). Total NPP was bimodal, with one peak during the winter and spring and, another, in summer and autumn. The winter-spring NPP peak was more notice- able at IS and US (73 and 64% of annual NPP, respectively), since productivity was more evenly distributed at the LS (46% in w i n ter-spring and 54% in summer- autumn). Reproductive tissues showed a major period of production during spring and lesser production at the beginning of autumn. The spring period showed 2 peaks of reproductive productivity at LS; one at the end of October comprised mainly by the winter exotic annual Vulpia dertonen - s i s (Allioni) Gola, and another at the beginning of December composed of feathers (S t i p a and P i p t o c h a e t i u m) . Annual grasses were unimportant at both other sites. Total senescence was greater and simi- lar at LS and US. The root contribution to total senescence increased with site eleva- tion (61, 64, and 68%). Litterfall increased slightly with elevation and represented about 10% of total senescence. Total dis- appearance of above- and belowground plant parts was greater in both elevational extreme sites (LS and US), where the main contribution came from below- ground mass (88 and 68%, respectively). Total disappearance at IS was nearly 50% of both other sites and equally distributed between below- and aboveground tissue Fig. 4. Above- and belowground dry weight (g m-2) for various compartments at 3 grasslands (Table 4). As a consequence of differences sites located at different altitudes from July 1988 to July 1989. in litterfall and aboveground disappear- ance dynamics among sites the LS accu- mulated a maximum of litter at the end of actual evapotranspiration were non-signif- belowground tissues changed among sites February, whereas both the other sites had icant for aboveground dry matter and (41–59% at lower site, 46–54% at interme- maximum accumulation at the end of above/belowground biomass. diate site and 33–67% at upper site, respec- December (Fig. 5). The variable site was significantly cor- related with all the dependent variables. When these correlations were partialed out Table 3. a) Multiple regression analysis of dry matter and biomass (g m- 2) in grassland sites to by the independent variables of the regres- water, temperature and evapotranspiration. Only significant equations are given; the non-signifi- sion models, no significant correlations cant contribution (P > 0.05) of individual variables are indicated as (ns). b) Correlation and par- tial correlation of the site variable with the dependent variables. W = monthly available water were obtained with aboveground dry mat- (mm); T = minimum monthly temperature (°C); ET = monthly actual evapotranspiration (mm); ter and above/belowground biomass. A- = aboveground; B- = belowground. Monthly available soil water and mean minimum temperature (in that order of a) importance) mainly explained the differ- 2 ences in dependent variables among sites. Dependent variables Equation Adjusted R P n The similar range of aboveground biomass Above-ground dry matter Y = 275 + 8 W + 19 T - 1.4 ET(ns) 0.55 <0.0001 27 at all sites, together with asinchronic inter- Above-ground biomass –0.075 0.759 27 Below-ground dry matter Y = 989 – 14 W – 38 T + 8 ET 0.35 <0.009 24 site variation, appears to explain the Below-ground biomass Y = 661 – 13 W – 64 T + 14 ET 0.42 <0.002 24 absence of correlation with the monthly B/A biomass ratio Y = 6 + 0.14 W – 0.8 T– 0.2 ET(ns) 0.54 <0.0004 24 independent variables (Fig. 4). b) Dependent variable Fluxes and turnover rates A-dry matter B-dry matter B-biomass A/B biomass Annual net primary products (NPP) was Site variable r P r P r P r P similar at all sites (ca. 1,131 to 1,280 g m-2 ) Correlation 0.59 <0.001 –0.80 <0.001 –0.77 <0.001 –0.67 <0.001 but the partitioning between above- and Partial correlation 0.006 0.98 -0.81 <0.001 -0.84 <0.001 -0.21 0.36

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 523 Table 4. Above- and belowground biomass, productivity, senescence, litterfall, disappearance and increased with elevation. Sims et al. turnover rates at lower (550 m asl), intermediate (850 m asl) and upper (1,025 m asl) grassland (1978), in their studies of North American sites. NPP = net primary productivity. grasslands, indicated that it was not possi- ble to assume a general relationship Parameter Site between aridity and high below- Lower Intermediate Upper ground/aboveground ratios. This ratio Biomass (g m-2) could be more related to life forms (Noy- Aboveground 146 a 138 a 131 a Meir 1973, Liang et al. 1989) or tempera- Belowground 410 c 615 b 904 a ture regime than with aridity (Noy-Meir Total biomass 556 c 753 b 1035 a -2 -1 1973, Sims et al. 1978). Net primary productivity (g m y ) In Sierra de la Ventana altitudinal differ- Aboveground 462 585 378 ences in belowground/aboveground bio- Belowground 669 695 779 Total NPP 1131 1280 1157 mass ratio could not be related to aridity Senescence (g m-2 y-1) as indicated by general climate and P/AET Aboveground 434 374 319 and P/PET ratios at each site. Lower soil Belowground 685 368 691 water retention with elevation probably Total senescence 1119 742 1010 played an important role in magnifying the Litterfall (g m-2 y-1) 107 210 232 -2 -1 temporary water shortage derived from Disappearance (g m y ) Aboveground 103 200 210 periods with several consecutive days Belowground 756 275 731 without rain. Total disappearance 859 475 941 The relationship of the below- Dry matter turnover rate (y-1) ground/aboveground biomass ratio with Aboveground 0.9 1.4 1.0 life form has been recognized in tussock Belowground 1.0 0.9 0.7 -1 and non-tussock grasses of these sierra Biomass turnover rate (y ) Aboveground 3.2 4.2 2.9 grasslands (Barrera and Frangi 1994); Belowground 1.6 1.1 0.9 however, within each life form the ratio increased mainly as temperature decreased. Turnover rates for aboveground biomass crest grasslands are colder and more The belowground/aboveground dry mat- were higher than for belowground biomass humid with the lowest air saturation ter ratio also increased with elevation. at all sites (Table 4). The US presented the deficit, stronger winds provoke higher air This ratio was 1.2 for the lower, 1.8 for slowest turnover rates of live tissues. evaporative capacity than found at low- the intermediate, and 3.0 for the upper lands (Kristensen and Frangi 1995). The site. These values are within the total wind effect should be important in range of physiognomic and altitudinal Discussion increasing water vapor loss from soil sur- grassland types of Sierra de la Ventana, faces and aboveground biomass at the where the ratio varies between 0.13 (at Soils and water budget upper site. 350 m with tussocks) to 4 (in the crests at The reduction of the mineral colloidal 1,100 m elevation) (Barrera and Frangi fraction with elevation was buffered by Biomass and necromass structure 1994). the high content of soil organic matter and Barrera and Frangi (1994) demonstrated Körner and Renhardt (1987) have indi- this explains similar CEC's at all sites. The that, between 350 and 1,100 m elevation, cated for the Central Alps that the lower base saturation and pH at higher ele- belowground dry matter increased linearly increased dicotyledonous herb below- vation indicated that even though these and aboveground dry matter decreased ground dry matter fraction at higher eleva- were fertile soils, exchangeable nutrients exponentially with elevation. Our sites tion resulted from shoot reductions and a were more limited at higher elevations. were located within the elevational range proportional increase in thin roots. An The relief heterogeneity and increased pre- studied by Barrera and Frangi (1994) and extended thin root system at higher eleva- cipitation and runoff at the higher eleva- did not include the lowland tussock grass- tions could be a functional substitute for tion site also influenced soil characteristics lands and high sierra meadows at the mycorrhizae that are more common at low and helped to explain the increased varia- extremes of the gradient. Both indepen- elevations. Although mycorrhizal coloniza- tion in soil parameters. Thornthwaite's dent studies showed similar trends in bio- tion directly with increases in soil tempera- water budgets, estimation of runoff, and mass with temperature change (Fig. 6). ture (Azcón-Aguilar et al. 1984), the actual evapotranspiration gave comparable The results were consistent with the degree of fungal colonization forming A- results to those empirically derived during inverse relationship between belowground mycorrhizae is also dependent on root sys- the same period from the 158 ha water- dry matter and temperature demonstrated tem morphology, since thin roots are more shed where the research sites were located by Sims et al. (1978). susceptible to colonization (Hetrick et al. (Bianchi and Dascanio pers. com.). The Biomass distribution within plants is 1988, Reinhardt and Miller 1990). In sierra runoff at upper site and intermediate site related to historic selective pressures and grasslands under a milder climate than the were 44 and 36% of precipitation, respec- the present environment (Piper 1989). In a alpine, our results suggested a greater myc- tively, vs 39–40% from the entire water- similar manner, the belowground/above- orrhizal activity coincidently with more shed above 550 m. Thornthwaite's water ground biomass ratio is an indicator of the thin roots rather than with temperature, as budget calculation did not include the allocation of energy to belowground and temperature decreased with elevation. wind effect in the evapotranspiration esti- aboveground organs (Sims et al. 1978). In belowground biomass and mycorrhizae mation. Although the upper slopes and the sierra grasslands, this quotient abundance indicated greater soil volume

524 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 in the upper grassland than in the lower one; in the intermediate grassland the rest period was dur- ing summer drought, not in winter. At the upper site, summer was not a true rest period but one with reduced production rates (Fig. 5). This behavior was consistent with seasonal changes in temperature and evaporative capacity that have been reported for this eastern slope (Kristensen and Frangi 1995). Roots in the upper soil layer were the most dynamic component of belowground dry matter. The lesser variability in the deeper root mass has been attributed to the more constant thermal and hydric conditions of this layer relative to the surface layer (Dahlman and Kucera 1965, Fernández and Caldwell 1975, Ares 1976). Although live thin roots increased during favorable thermic and hydric periods, periods of high senescence followed during dry consecutive days. The ecological differences among sites were not expressed by total net primary production (NPP), which was similar at all sites, but by its partitioning in space and time. The higher eleva- tion grassland community invested a higher proportion of annual NPP in roots, independent of season. The allocation pattern changed seasonally at the other 2-sites. At the end of summer and beginning of autumn, which was more humid than in spring, above ground net primary production (ANPP) showed a decreasing trend with Fig. 5. Aboveground fluxes: net primary productivity (ANPP), senescence (ASEN), litterfall (FALL) increase in altitude. The proportion and disappearance (ADIS); and belowground fluxes: net primary productivity (BNPP), senescence of NPP allocated to photosynthetic (BSEN) and disappearance (BDIS) in grassland sites at different altitudes from July, 1988 to July, tissues increased from the summits 1989. to the piedmont. occupation when availability of soil Summer and winter were the seasons Turnover rates resources decreased. Chapin (1991) has when necromass increased and biomass Faster turnover rates for aboveground proposed that a more intense occupation of decreased. The aboveground dry matter compartments than for belowground com- soil volume is a general response of plants increased during the spring as temperature partments have been reported in other to stress when resources are limiting. and rainfall increased, but maximum grassland systems (Kucera et al. 1967, stocks were found at different dates vary- Sims and Singh 1978, Aerts et al. 1989, Soriano 1992). The differences in photo- Dry matter dynamics and fluxes ing with elevation. The spring peak in growth occurred earlier in the lower grass- syntate partitioning and biomass distribu- The 3 sites showed similar trends in dry tion are shown by the different turnover matter and litter dynamics throughout the land. Temperatures more favorable to aboveground growth were attained later at rates (productivity/biomass) of above- year. aboveground compartments and live ground and belowground live tissues at roots increased during spring and early higher altitudes. We found differences among altitudinal each site. Since mean aboveground bio- autumn. Root growth during winter-spring mass and total net productivity were simi- preceded aboveground biomass increases. sites in the length of quiescence of vegeta- tion. The cold period was more extended lar at all sites, the lower turnover rate of

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 525 than at Pampa sites despite the partial over- lap in the quotient range (61 to 75% vs 54 to 67%, respectively). Since sites like Hays and Osage (Sims et al. 1978) were the most comparable to Sierra de la Ventana in latitude, mean temperature, precipitation, and vegetation parameters, we suggest that the harder seasonality (226 to 272 days length of thermal potential growth season, respectively) in the Northern Hemisphere compared to our sites, located in more oceanic climates (more than 300 days of estimated length of thermal potential growth season), influenced the structural and functional differences. With respect to productivity, Laurenroth (1979) considered that the most productive grassland sites of the world were those with around 800 mm annual precipitation and a mean temperature >15°C. His grass- land group 6 included a wide set of sites characterized by a lack of a significant drought period at any time of the year and an ANPP in the range of 160—1,387 g m-2 y e a r- 1 (mean = 525 g m - 2 y e a r - 1). The ANPP of the Sierra de la Ventana grass- lands were ± 12% to –28% of the grassland Fig. 6. Above (A-) (circle) and belowground (B-) (square) dry matter in sites located within group 6 mean. The main difference the elevational range from 350 to 1,100 m asl. Sites from Barrera and Frangi (1994) (solid appeared for the higher elevation site, symbols) include the lowland tussock grasslands and high sierra meadows at the extremes where productivity was composed mainly of the gradient. Open symbols indicate data from our sites. Vertical lines indicate ±SE. of belowground tissues. Gómez and Estimates of sites normal annual mean temperature (calculated with standard weather Gallopín (1991) modeled the productivity reports data for 1961–1970 period and 0.7˚C thermal gradient) is indicated as a second x- of Latin American biomes and the range axis. for Pampean grasslands was 400–600 g m-2 ye a r -1 . The sierra grasslands ANPP (378 to aboveground biomass at higher elevation the elevational gradient indicated a higher 585 g m- 2 y e a r- 1) fit within that range, as reflected the greater allocation of photo- belowground senescence rate and decom- did those from Pampa lowland ungrazed syntates to roots. In contrast, the lower position rate per mass unit of root, under sites like Salado River basin grasslands turnover of roots at higher elevations indi- soil conditions that prevailed at the lower with 532 g m-2 ye a r -1 (Sala et al. 1981) and cated that although belowground produc- site. Inversely, the markedly greater annu- La Pampa grasslands with 410 g m-2 year-1 tivity is greater, root replacement is slower al absolute aboveground disappearance (de Wysiecki 1993). The greater above- because of its higher biomass compared to suggested that aboveground disappearance ground productivity (700 g m-2 year-1) esti- lower grasslands. The higher turnover rate per unit of aboveground dry mass was mated for the lower site a decade before rates and lower biomass of roots at the higher at the upper site, probably (Frangi et al. 1980 a) suggested that annu- lower site suggested that more fertile soils enhanced by higher humidity and lower al productivity variation in the sierra were faster exploited during the humid saturation deficits. grasslands is important. periods, prolonged because of more loamy textures, while periods with scarce water Comparison with other temperate availability may have stimulated root grasslands Conclusions replacement. A texture that holds more aboveground dry matter and net primary water facilitates this strategy for sustaining production (NPP) values for these sierra The similar mean annual aboveground a similar aboveground mass of live tissue grasslands fell within the range reported biomass at all sierra grassland sites sug- to produce more aboveground tissues. for USA ungrazed temperate grasslands gested that a common threshold carrying Probably high temperatures and evapo- (Sims et al. 1978, Sims and Singh 1978). capacity of green tissues had been attained. transpiration at lower sierra slopes trigger belowground dry matter was lower than in As total production was also similar among a higher turnover rate of aboveground live USA comparable mixed-grass and tall- 3 elevational sites this suggested that pho- tissue. As a consequence of higher produc- grass prairie sites. Although NPP of sierra tosynthetic efficiency was equivalent at all tivity and senescence and lower disappear- grasslands was near the lower value of sites. The differences among sites observed ance of aboveground tissues, standing mixed-grass prairie, the above ground net in belowground biomass and proportion of dead material accumulated more at the primary production (ANPP) at sierra sites fine roots that were linked to temperature lower site and turnover rate of above- was higher than at USA sites reported by and soil water holding capacity appeared to ground dry matter was slower. A similar Sims and Singh (1978). Consequently, reflect an adaptive trend for improved annual absolute mass senescence and dis- there was greater BNPP/NPP ratio at USA water and nutrient absorption with eleva- appearance of roots at both extremes of

526 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 tion. As there were no differences in Cabrera, A.L. (ed.) 1970. Flora de la Hetrick, B.A., D.G. Kitt and G.T. Wilson. aboveground biomass, grass life forms or Provincia de Buenos Aires. Parte II 1 9 8 8 . Mycorrhizal dependence and growth NPP among sites, and evapotranspiration Gramíneas. Colección Científica INTA, 4. habit of warm season and cool-season tall- could be similar at all sites, the below- Buenos Aires grass prairie plants. Can.J.Bot. 66:1376–1380. ground biomass is probably satisfying a Cappanini, D., C.O. Scoppa, and J.R. Hunter, A.H. 1982. International soil fertility Vargas Gil. 1971. Suelos de las Sierras similar water demand. Because of a lower evaluation and improvement: Laboratory Australes de la Provincia de Buenos Aires. p. procedures. I n: Department of Soil Science water holding capacity with elevation, a 203-234. I n: Reunión Geología Sierras (ed.) North Carolina State University. more intense occupation of soil is neces- Australes. Comisión Investigaciones Raleigh, N.C. sary to accomplish water absorption under Científicas, La Plata (Argentina). Hurlbert, S.H. 1984. Pseudoreplication and colder conditions. Chapin, F.S III. 1991. Integrated Responses of the design of ecological field experiments. Plants to Stress. BioScience 41:29–36. Ecol. Monogr. 54:187–211. Dahlman, R.C. and C.L. Kucera. 1965. Root Körner, Ch. and V. Renhardt. 1987. D r y Literature Cited productivity and turnover in native prairie. matter partitioning and root length/leaf area Ecol. 46:84–89. ratios in herbaceous perennial plants with de la Sota, E.R. 1967. Composición, origen y diverse altitudinal distribution. Oecologia Aerts, R., F. Berendse, N.M. Klerk, and C. vinculaciones de la flora pteridológica de las (Berlin) 74:411–418. Bakker. 1989. Root production and root Sierras de Buenos Aires (Argentina). Bull. Kristensen, M.J. and J.L. Frangi. 1995. turnover in two dominant species of wet Soc. Argentina Botánica 11:105–128. Mesoclimas de pastizales de la Sierra de la heathlands. Oecologia (Berlin) 81:374–378. de Wysiecki, M.L. 1993. Productividad pri- Ventana. Ecología Austral (Argentina) Ares, J. 1976. Dynamics of the root system of maria neta aérea de un pastizal natural de la blue grama. J. Range Manage. 29:208–213. 5:55–64. provincia de La Pampa, Argentina. Rev. Andreis, R., A.M. Iñiguez, L.L. Lluch and S. Kucera, C.L., R.C. Dahlman, and M.R. Facultad Agronomía, La Plata (Argentina) Rodriguez. 1989. Cuenca paleozoica de Koelling. 1967. Total net productivity and 69:23–29. Ventania, Sierras Australes, Provincia de turnover on an energy basis for tallgrass Buenos Aires, p. 265-298. In: Chebli, G. and Duchaufour, P. 1970. Precis de Pedologie. prairie. Ecol. 48:536–541 L. Spalletti (eds.) Cuencas Sedimentarias Masson et Cie. Paris. Lauenroth, W.K. 1979. Grassland Primary Argentinas. Serie Correlación Geológica 6. Fernández, O.A. and M.N. Caldwell. 1975. Production: North American grassland in San Miguel de Tucumán (Argentina). Phenology and dynamics of root growth of perspective, p. 1–24. I n: N.R. French (ed.) Azcón-Aguilar, C., J.M. Barea and B. three cool semi-desert shrubs under field Perspectives in grassland ecology. Ecol. Roldán-Fajardo. 1984. Avances recientes conditions. J. Ecol. 63:703–714. Studies 32. Springer Verlag. Berlín. en el estudio de las micorrizas VA. II. Frangi, J.L. and O. Bottino. 1995. Las comu- Liang, Y.M., D.L. Hazlett, and W.K. Factores que afectan su formación y función nidades vegetales de la Sierra de la Ventana, Lauenroth. 1989. Biomass dynamics and y aplicaciones prácticas en agricultura. Provincia de Buenos Aires, Argentina. Rev. water use efficiencies of five plant communi- Anales Edafología y Agrobiología (Spain) Facultad Agronomía, La Plata (Argentina) ties in the shortgrass steppe. Oecologia 42:943–958. 71:93–133. (Berlin) 80:148–153. Barrera, M.D. 1991. Ecología de las comu- Frangi, J.L., N.E. Sánchez, M.G. Ronco, G. McClaugherty, C.A., J.D. Aber, and J.R. nidades con arbustos de la Sierra de la Rovetta, and R. Vicari. 1980 a. D i n á m i c a Melillo. 1982. The role of fine roots in the Ventana. Ph.D. Thesis, Univ. Nacional La de la biomasa y productividad primaria aérea organic matter and nitrogen budgets of two Plata, La Plata (Argentina). neta de un pastizal de "flechillas" de Sierra forested ecosystems. Ecol. 64:1481–1490. Barrera, M.D. and J.L. Frangi. 1994. de la Ventana (Buenos Aires, Argentina). Milner, C. and R. E. Hughes. 1970. Methods Estructura de la biomasa de pastizales en Bull. Soc. Argentina Botánica 19:203–228. for the measurement of the primary produc- Sierra de la Ventana (Buenos Aires, Frangi, J.L., M.G. Ronco, N.E. Sánchez, G. tion of grasslands. IBP Handb. 6. Blackwell Argentina). Rev. Museo La Plata (NS) Rovetta, and R. Vicari. 1980 b. Efecto del Sci. Publications, Oxford and Edinburgh. 14:243–262. fuego sobre la composición y dinámica de la Noy-Meir, I. 1973. Desert Ecosystems: Barrera, M.D. and J.L. Frangi. 1996. biomasa de un pastizal de Sierra de la Environment and Producers. Annu. Rev. Estructura poblacional de los arbustos Ventana (Buenos Aires, Argentina). Ecol. Syst. 4:25–51. Eupatorium buniifolium y Discaria ameri- Darwiniana (Argentina) 22:565–585. Page, L. A. 1982. Methods of soil analysis. cana con relación a la pedregosidad del suelo Frissel, M.J. 1981. The definition of residence Part 2, N˚9. Amer. Soc. of Agron., Madison, y el pastoreo en la Sierra de la Ventana, time in ecological models, p. 117–122. I n: Wis. Buenos Aires, Argentina. Rev. Facultad F.E. Clark and T. Roswall (eds.) Terrestrial Pérez, C. 1996. Productividad en pastizales Agronomía, La Plata (Argentina) Nitrogen Cycles. Ecol. Bull.33. Stockholm. serranos a lo largo del gradiente altitudinal. 101:113–125. Gerdemann, J.W. and T.H. Nicolson. 1963. Ph.D. Thesis. Univ. Nacional La Plata. La Böhm, W. 1979. Methods of studying root sys- Spores of mycorrhizal endogone species Plata, Argentina. tems. Ecological Studies 33. Springer- extracted from soil by wet sieving and Phillips, J.M. and D.S. Hayman. 1970. Verlag, Berlín. decanting. Trans. Brit. Mycol. Soc. Improved procedures for clearing roots and Burgos, J.J. 1968. El clima de la Provincia de 46:235–244. staining parasitic and VA-mycorrhizal fungi Buenos Aires en relación con la vegetación Giovannetti, M. and B. Mosse. 1980. An eval- for rapid assessment of infection. Trans. Brit. natural y el suelo, p. 34-97. In: A.L. Cabrera uation of techniques for measuring vesicular- Mycol. Soc. 55:158–161. (ed.) Flora de la Provincia de Buenos Aires. arbuscular mycorrhizae infection in roots. Piper, J.K. 1989. Distribution of dry mass Colección Científica INTA 4(1). Buenos New Phytol. 84:489–500. between shoot and root in nine understory Aires. Gómez, I. A. and G.C. Gallopín. 1991. species. Amer. Midland Natur. 122:114–119. Cabrera, A.L. (ed.) 1963. Flora de la Estimación de la Productividad primaria neta Reinhardt, D.R. and R.M. Miller. 1990. Size Provincia de Buenos Aires. Parte VI de ecosistemas del mundo en relación a fac- classes of root diameter and mycorrhizal fun- Compuestas. Colección Científica INTA, 4. tores ambientales. Ecología Austral gal colonization in two temperate grassland Buenos Aires (Argentina) 1:24–40. communities. New Phytol. 116:129–136. Cabrera, A.L. 1968. La vegetación de la Harrington, H. 1947. Hojas geológicas 33 m y Ricci, S. 1992. Relaciones entre la vegetación y Provincia de Buenos Aires, p. 101–126. I n: 34 m, Sierras de Curamalal y de la Ventana, la actividad agropecuaria en el área de Sierra A.L. Cabrera (ed.), Flora de la Provincia de Provincia de Buenos Aires. Serv. Nac. de la Ventana. Ph.D. Thesis, Univ. Nacional Buenos Aires. Colección Científica INTA, Minero Geológico. Bull.61. Buenos Aires. 4(1).Buenos Aires. La Plata, La Plata (Argentina).

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 527 Sala, O.E., V.A. Deregibus, T. Schlichter, Sokal, R.R. and F.J. Rohlf. 1969. B i o m e t r y . Thornthwaite, C.W. 1948. An approach and H. Alippe. 1981. Productivity dynamic The principles and practice of statistics in toward a rational classification of climate. of a native temperate grassland in Argentina. biological research. Freeman and Co., San The Geograph. Review 38:55-94. J. Range Manage. 34:48–51. Francisco, Calif. Thornthwaite, C.W. and J.R. Mather. 1957. Sims, P.L. and S. Singh. 1978. The structure Soriano, A. 1992. Rio de la Plata Grasslands, Instructions and tables for computing poten- and function of ten western North American p. 367–407. In: R.T. Coupland (ed.) Natural tial evapotranspiration and the water balance. grasslands (III). J. Ecol. 66:573–597. grasslands: Introduction and Western Publ. Climatol. 10(3), Laboratory of Sims, P.L., J.S. Singh, and W.K. Lauenroth. Hemisphere. Ecosystems of the World 8 A. Climatology, Drexel Institute of Technology, 1978. The structure and function of ten west- Elsevier. Amsterdam. Centerton, N.J. ern North American grasslands (I). J. Ecol. Suero, T. 1972. Compilación geológica de las Vargas Gil, J.R. and C.O. Scoppa. 1973. 66:251–285. Sierras Australes de la Provincia de Buenos Suelos de las Sierras de la Provincia de Buenos Aires. Rev. Investig. Agropecuarias SMN. 1981. Estadística Climatológica 1961- Aires. LEMIT, La Plata (Argentina) (Argentina), Ser. 3 Clima y Suelo 10:57–79. 1970. Serie B, 35. Fuerza Aérea Argentina, Ser.II(116). Vervoorst, F. 1967. La vegetación de la Servicio Meteorológico Nacional. Buenos Tabatabai, A.M. and J.M. Bremner. 1991. República Argentina. VII. Las comunidades Aires. Automated Instruments for Determination of vegetales de la depresión del Salado Soil Survey Staff. 1992. Keys to Soil total Carbon, Nitrogen and Sulfur in Soils by (Provincia de Buenos Aires). INTA, Serie Taxonomy. Agency Int. Develop-USDA Soil Combustion Techniques, p. 261–286. In: Soil Fitogeog. 7. Buenos Aires. Conservation Service, SMSS Technical Analysis, Modern Instrumental Techniques, Von Gosen, W.W., W. Buggisch and L.V. Monograph 19. Pocahontas Press Inc, 2nd Ed. Marcel Dekker. Inc. New York, Dimieri. 1990. Structural and metamorphic Blacksburg, Virg. N.Y. evolution of the Sierras Australes (Buenos Aires Province, Argentina). Geologische Runtschau 79:797–821.

528 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 J. Range Manage. 53: 529–536 September 2000 Estimating grazing index values for plants from arid regions

PIÉRRE C. V. DU TOIT

Author is principal agricultural scientist, Pasture Science Department, Grootfontein Agricultural Development Institute, Private Bag X529, 5900, Middleburg, Republic of South Africa.

Abstract Resumen

The Ecological Index Method impacted directly on the estima- El método de Indice Ecológico impactó directamente en la esti- tion of grazing capacities in the Karoo, an arid region in South mación de la capacidad de apacentamiento del Karoo, una Africa. Due to inherent deficiencies in the ecological index región árida de Sudáfrica. Debido a las deficiencias inherentes method, index values formed a disjunct series, 10, 7, 4, and 1 and del método de Indice Ecológico los valores de los indices for- the fact that different value systems were employed to score plant maron un serie disyuntiva, 10, 7, 4 y 1, y el hecho de que se species, necessitated that index values of species commonly emplearon diferentes sistemas de valores para calificar las plan- encountered during botanical surveys be subjectively adjusted by tas, exige que los valores de los indices para las especies común- means of a species by species comparison, forming a continuous mente encontradas durante los reconocimientos botánicos sean series of index values. These index values were still subjective subjetivamente ajustados mediante la comparación de especies value judgements of the agronomic value of the Karoo plant por especie formando series continuas de valores de los indices. species. The author felt that a method should be developed to Estos valores de los indices todavía son valores subjetivos del objectively estimate grazing index values from certain plant vari- valor agronómicos de las especies de plantas del Karoo. El autor ables, i.e. size, animal available dry matter production, and sintió que se debería desarrollar un método para estimar objeti- chemical properties of the species. The use of these properties vamente los valores del indice de apacentamiento a partir de would describe the agronomic value of the plant species, which ciertas variables de la planta, por ejemplo, el tamaño, la disponi- would lead to agronomically sound current grazing capacities bilidad de materia seca para el animal y las propiedades quími- being estimated. The following 2 models were proposed to deal cas de las especies. El uso de estas propiedades describirá el with the karoo subshrubs and the grasses of the karoo separate- valor agronómico de las especies vegetales, el cual conduciría a la ly. Grazing index value for the karoo subshrubs = {(canopy estimación sensata de la capacidad de apacentamiento. Los sigu- spread cover + available forage + TDN + [K ÷ (Ca + Mg)]) ÷ ientes 2 modelos fueron propuestos para tratar por separado los ether extract} ÷ 100, and grazing index value for the grasses = arbustos y zacates del Karoo. El valor del indice de apacen- {(canopy spread cover + available forage + TDN + [K ÷ (Ca + tamiento para arbustos fue igual a: {(cobertura de la copa exten- Mg)]) x ether extract} ÷ 100. dida + forraje disponible + NTD + [K ÷ (Ca + Mg)] ) ÷ extracto etéreo} ÷ 100 y el valor del indice de apacentamiento para zacates fue igual a: {cobertura de la copa extendida + forraje Key Words: available dry matter production, canopy spread disponible + NTD + [K ÷ (Ca + Mg)] ) x extracto etéreo} ÷ 100. cover, current grazing capacity, K÷(Ca+Mg) ratio, total digestible nutrients Vorster (1982) scored the grasses on ecological attributes and Publication of the Ecological Index Method by Vorster (1982) karoo subshrubs on palatability, an agronomic attribute, the range directly influenced the estimation of grazing capacities in the condition score computed for the sample site represents the eco- Karoo. He allocated index values to the karoo plants on the basis: logical state of health of the grass sward on one hand, but the 1) perennial climax grasses, index value 10; in close agreement agronomic state of health of the karoo subshrubs on the other (see with Tainton et al. (1980); Tainton 1981). Since the score is based partly on ecological prin- 2) perennial sub-climax grasses and palatable karoo subshrubs, ciples and partly on palatability, it is inappropriate to estimate index value 7; grazing capacities, which are agronomic values, from these range 3) perennial pioneer grasses and the less palatable karoo sub- condition scores which are assumed to indicate the agronomic shrubs, index value 4; and potential of the site to support livestock. Several other factors, for 4) annual pioneer grasses, unpalatable karoo subshrubs and instance dry matter production (Botha et al. 1993) need to be con- invader plants, index value 1. sidered when a realistic grazing capacity is to be estimated. The ecological index method is largely dependent upon classi- This method has been used throughout the Karoo, with largely cal succession theory, similar to many methods currently acceptable results. The ecological index values are used in the employed elsewhere in South Africa (Foran et al. 1978, Tainton computation of range condition scores (see Tainton 1981), fol- et al. 1980, Mentis et al. 1980, Hardy and Hurt 1989, Hurt and lowing surveys of sample sites. Grazing capacities for these sam- Bosch 1991). Another of the deficiencies was that some of the ple sites are then estimated from the range condition scores. Since late developmental stage grasses, previously referred to as climax grasses, received too high an index value to be used in the com- Manuscript accepted 25 Nov. 1999. putations when estimating the current grazing capacity (Tainton

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 529 et al. 1980, Edwards 1981, Vorster 1982). as unrelated species of known similar of canopy spread cover. The canopy Research revealed that not all the late grazing value. The phases of the opera- spread cover of 50 plants per species developmental stage grasses should have tional research study (Wilkes 1989) were obtained this way included a representa- an index value of 10 (Du Toit and Botha followed. tive sample of sizes, from small plants to 1993). Because of the above m e n t i o n e d These index values, however, still repre- large plants. The 50 measurements gave a deficiencies, a committee investigated the sent subjective value judgements of the reliable average value (Roux 1963). The apparent shortcomings in order to correct actual agronomic value of the Karoo plant plants were harvested in toto at a height of them in an effort to move away from the species. The author felt that a method approximately 20 mm to 50 mm above the classical succession theory. The index val- should be developed to estimate grazing soil surface. The plants were selected at ues of all the species commonly encoun- index values more objectively from certain random over an area of 1 hectare (Vlok tered during botanical surveys, were agronomic plant variables; size, dry matter 1963), to minimize variations in plant eco- adjusted subjectively by means of a production, and chemical properties of the typic differences. The harvested material species by species comparison. Each species. These properties were studied for was separated into potentially grazable species being assessed on the basis of the a period of 3 years, to largely exclude and ungrazable material on the basis of the following 6 variables: extreme variations in productivity which is 2 mm rule of thumb (Botha 1981, Du Toit 1) The ability to produce forage, i.e. the so characteristic of vegetation from semi- 1996a). The 2 mm rule of thumb states amount of grazable dry matter pro- arid range types (Sneva and Hyder 1962). that stems thicker than 2 mm are not duced per year: Low producers score Species from distinct a g r o - e c o l o g i c a l grazed. While stems thicker than 2 mm 1, high producers score 10. regions were studied. It was attempted to were regarded as ungrazable, thinner 2) The nutritional value during the grow- study related species in the different stems contributed to the available grazable ing season: Species with a low nutri- regions, to ascertain differences in grazing forage (Botha 1981, Du Toit 1993). Grass tional value score 1, species with a value between different species in an area tufts were harvested by clipping to a stub- high nutritional value score 10. and between related species of different ble height of 20–50 mm and the whole 3) The nutritional value during the dor- areas. These objective grazing index val- fraction was regarded as being grazable. mant season, as in No. 2. ues will replace the ecological index val- Plant material was dried in a forced draft 4) Relative ease with which the species ues. To avoid future confusion as to which oven between 60°C to 80°C, for 24 hours. can be grazed i.e. presence or absence index values have been used, and that an of spines: spiny species score 1, agronomic method of grazing capacity Relation between canopy spread species without spines score 10. The estimation was used, the ecological index cover and mass presence of resins and aromatic oils method is replaced by the Grazing Index The canopy spread cover : total mass downgraded the score due to anti-her- Method (Du Toit 1995). relation of specific species were deter- bivory. Grazing index values estimated by the mined by regression to determine the rela- 5) Perenniality: Annual species score 1, model, should reflect agronomic values of tion between plant cover and the available strong perennial sodforming grasses plant species more objectively and when forage (Payne 1974). score 10. used in conjunction with the Grazing 6) Ability of the plant to protect the soil Index Method, estimate more acceptable against surface soil erosion: Upright current grazing capacities. These grazing Chemical analyses karoo subshrubs have a low score, index values should furthermore verify the The grazable material was chemically decumbent bushes score fairly high, relative position of the species in the analyzed for N according to the Kjeldahl annual tufted grasses score 1, tufted series, when compared to the subjective method (AOAC 1964), Ca, P, Na, Mg, K perennial grasses, depending on pro- grazing index values developed earlier were analyzed by atomic adsorption spec- ductivity and habit, have an interme- (Du Toit et al. 1995). trophotometer, (Homer and Parker 1961) diate score due to the erosive chan- and acid detergent fibre was analyzed neling effect they may have on according to a modified Weende method runoff water, sod forming grasses, Methodology (Van Soest 1963). where infiltration of rainwater is heightened, score 10. Canopy spread cover Model development These 6 agronomic properties of the The canopy spread cover of a number of The proposed model attempts to objec- species were chosen to ensure that individ- grasses and bushes was measured at ran- tively allocate a grazing index value to ual species scores were comparable and dom at each locality from 1990 to 1993. each of the studied species, based on the the estimate free from bias. Each variable Fifty plants per species were selected from agronomic variables; size of the species scored a maximum of 10. Individual vari- small to large plants, including the range i.e. canopy spread cover, the cornerstone able scores were summed and the final of sizes normally found within the species. of the range condition assessment tech- species score calculated to occupy a posi- A frame 1 m x 1 m, subdivided into blocks nique, animal available dry matter, and tion between 0 and 10. The average of the of 50 mm x 50 mm for the large species certain chemical properties. Stepwise scores submitted by the members, and and 25 mm x 25 mm for the smaller regressions (Statistical Graphics standard deviation was calculated for each species, was used for this purpose. The Corporation 1991) of the subjective graz- species score. Values falling outside the canopy spread cover was estimated by ing index values on the values obtained for range x ± SD were discarded. A new mean counting the blocks of which half or more the different parameters of the studied was calculated for the remaining values. of each block was engaged by the perpen- species indicated the most important ele- The resultant score called the subjective dicular projection of the crown (Goebel et ments to be included in the model. grazing index value was vetted taking into al. 1958). The number of blocks counted account scores of related species, as well were then expressed as square centimeters

530 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 Taxonomy Nomenclature follows Arnold and De Wet Table 1. Species names and their authorities used in this paper and in Tables 1 to 3. (1993). Species names used in this paper and in the tables with their authorities are Chrysocoma ciliata L. given in Table 1. Eberlanzia ferox (L. Bol.) L. Bol. Eriocephalus aspalathoides DC. Eriocephalus ericoides (L.f.) Druce Results and discussion Eriocephalus spinescens Burch. Felicia fascicularis DC. Felicia filifolia (Vent.) Burtt Davy Available forage and canopy spread Felicia macrorrhiza (Thunb.) DC. cover Felicia muricata (Thunb.) Nees The relation between available forage Galenia procumbens L.f. and canopy spread cover, for all values, Galenia secunda (L.f.) Sond. 2 Helichrysum dregeanum Sond. & Harv. has a coefficient of determination, r = Helichrysum lucilioides Less. 0.62. The slope of the regression line is Hermannia desertorum Eckl. & Zeyh. less than 45°. This means that as the plants Hertia pallens (DC.) Kuntze increase in size, there is not the expected Lycium cinerium Thunb. (Sens. Lat.) corresponding increase in available forage. Monechma incanum (Nees) C.B.Cl. With the available forage being over-esti- Nenax microphylla (Sond.) Salter Osteospermum microphyllum DC. mated by about 25% (Du Toit 1993, Osteospermum spinescens Thunb. 1996a), the regression relation should Pentzia globosa Less. have a slope of much less than 45°. The Pentzia incana (Thunb.) Kuntze available mass recovered from a plant Pentzia spinescens Less. with a certain canopy spread cover varies Phymaspermum parvifolium (DC.) Benth. & considerably. Therefore, cover on its own Hook. ex Jackson Plinthus cryptocarpus Fenzl does not provide a reliable estimate of Plinthus karooicus Verdoorn available forage, in contrast to the finding Protasparagus suaveolens (Burch.) Oberm. of Mueller Dombois and Ellenberg (1974). Pteronia adenocarpa Harv. Uresk (1990) proposed that frequency and Pteronia glauca Thunb. canopy spread cover be multiplied to ren- Pteronia glomerata L.f. der an index of forage mass, which would Pteronia sordida N.E.Br. Pteronia staehelinoides DC. probably be a reliable indication of avail- Pterothrix spinescens DC. able forage. In this case the relation Rhigozum obovatum Burch. between plant cover, available forage and Rosenia humilis (Less.) Bremer forage value was to be used to calculate a Rosenia oppositifolia (DC.) Bremer grazing index value for each species. Salsola calluna Fenzl ex C.H.Wr. Floristic composition measured by the Salsola rabieana Verdoorn Salsola tuberculata (Moq.) Fenzl grazing index method, together with the Walafrida geniculata (L.f.) Rolfe grazing index values of the species would Walafrida saxatilis (E.Mey.) Rolfe then describe the grazing capacity. Zygophyllum gilfillanii N.E.Br. Zygophyllum lichtensteinianum Cham. & Schlechtd. Nutrient requirements of grazing ani- Zygophyllum microphyllum L.f. mals According to Du Toit et al. (1940), Grasses growing sheep needs approximately Aristida adscensionis L. 0.16% calcium and 0.14% phosphorus in Aristida congesta Roem. & Schult. subsp. congesta Aristida diffusa Trin. subsp. diffusa their diet, while Louw (1969) recommends Cymbopogon plurinodis (Stapf) Stapf ex Burtt Davy that the diet must contain between 0.2% to Cynodon dactylon (L.) Pers. 0.5% calcium and 0.2% to 0.46% phos- Digitaria eriantha Steud. phorus, which agrees with values quoted Enneapogon desvauxii Beauv. by Woods (1959). The calcium : phospho- Eragrostis bergiana (Kunth) Trin. rus ratio must be in the region of 1.0 for Eragrostis chloromelas Steud. Eragrostis curvula (Schrad.) Nees var. conferta normal growth and development. Eragrostis lehmanniana Nees var. lehmanniana Potassium, calcium, and magnesium are Eragrostis obtusa Munro ex Fical. & Hiern included as the K ÷ (Ca + Mg) ratio. Fingerhuthia africana Lehm. These elements have a direct influence on Heteropogon contortus (L.) Roem. & Schult. the grazing value since unfavourable ratios Hyparrhenia hirta (L.) Stapf lead to tetany (Kemp and T'Hart 1957, Merxmuellera disticha (Nees) Conert Sporobolus fimbriatus (Trin.) Nees Kidambi et al. 1989). Values of these Stipagrostis ciliata (Desf.) De Winter var. ca p e n s i s macronutrients are included in the formu- Stipagrostis obtusa (Del.) Nees lae for the objective estimation of the Themeda triandra Forssk. grazing index values. Low or high values and unfavourable ratios have detrimental

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 531 Table 2. Objective grazing index values (OGIV) calculated for species from the False Upper Karoo (Acocks 1988), compared to their corresponding subjective grazing index values (SGIV), illustrating the annual and seasonal variation.

OGIV SGIV 1990/91 May Aug. Nov. Jan. Mean Chrysocoma ciliata 1.22 0.31 1.29 0.51 0.83 1.50 Eriocephalus ericoides 2.44 1.67 3.44 4.16 2.93 5.00 Pentzia incana 1.64 1.33 2.01 1.94 1.73 5.70 Aristida congesta 3.36 1.92 0.94 1.07 1.82 1.30 Digitaria eriantha 19.59 7.72 13.51 6.46 11.82 8.90 Themeda triandra 15.04 10.19 8.31 6.69 10.06 9.30 1991/92 May Aug Nov Jan Mean Eriocephalus spinescens 3.09 2.75 0.29 2.73 2.22 4.50 Pentzia incana 1.67 1.17 1.42 1.93 1.55 5.70 Rosenia humilis 3.28 2.90 3.99 3.03 3.30 3.50 Heteropogon contortus 4.17 5.36 4.37 3.77 4.42 7.20 Stipagrostis ciliata 3.78 4.48 2.38 3.61 3.56 7.20 Stipagrostis obtusa 3.49 4.46 1.01 2.34 2.83 6.60 1992/93 May Aug Nov Jan Mean Eriocephalus ericoides 1.48 1.78 2.85 2.63 2.18 5.00 Pentzia incana 1.19 0.97 1.16 1.57 1.22 5.70 Rosenia humilis 3.16 4.56 4.60 3.65 3.99 3.50 Aristida congesta 1.82 1.58 0.78 0.85 1.26 1.30 Heteropogon contortus 3.77 5.06 4.15 3.98 4.24 7.20 Themeda triandra 6.64 6.26 5.65 4.70 5.81 9.30 effects on animal production (Van Hoven 75.1 + [(6.25 x log % crude protein)– the ether extract values of certain shrubs, and Ebedes 1988). Low values lead to 0.75 x % crude fibre)] and; the higher the loss of energy rich esters, serious deficiencies in animal feeding 2) Du Toit (1996b) where : Crude fibre = ethers and aldehydes through the urine of (Woods 1959). Unfavourable values nega- – 4.32 + 0.92 x % acid detergent sheep (Cook et al. 1952). The usable ener- tively affect index values. fibre. gy of these forages is not as high as their In addition to the chemical properties \ TDN = 75.1 + {(6.25 x log % N) – calculated TDN values indicate. mentioned above, the % ash and ether (0.75 x [– 4.32 + (0.92 x % acid Consequently the sum of the calculated extract of the species were obtained from detergent fibre)])} (Du Toit 1996b). values of the different variables is divided the literature for analyses previously car- by the ether extract value, a lower index ried out on plantspecies occurring in the value for karoo subshrubs with high ether same areas (Louw et al. 1968a, 1968b, Proposed model, used to estimate extract values are estimated, in accordance 1968c, Steenkamp and Hayward 1979, grazing index values with their relatively low nutritional value. Botha and Nash 1990, Botha et al. 1990a, The model takes into account the size of 1990b, 1990c). the plant i.e. canopy spread cover, which is the cornerstone of the method of botani- Model a, grazing index value for cal survey for estimating grazing capacity, karoo subshrubs Total digestible nutrients (TDN) through the Grazing Index Method. Two Grazing index value for the karoo sub- Because of the prominence that total approaches were followed with the formu- shrubs = {(canopy spread cover + avail- digestible nutrients receives in animal lation of the model. In the case of the able forage + TDN + [K ÷ (Ca + Mg)]) ÷ feeding ration formulations (Maynard and karoo subshrubs the ether extract value ether extract} ÷ 100 (Du Toit 1996b) Loosli 1962), in the comparison of differ- negatively affects the grazing value, on (Table 3). ent feedstuffs (Swift 1957) and in this account of it's contribution to the smell In the case of grasses, ether extract values study, the comparison of different natural and taste of the karoo subshrubs. The positively contributes to the grazing value, forages, it is included in these equations as higher the ether extract value of the karoo through the carotene content (Van Der a primary variable. The 2 variables incor- subshrub, the higher the resin and aromat- Merwe 1985). Grasses with high ether porated in the determination of TDN ic oil content. Experience teaches that extract values are late developmental stage (Glover et al. 1960, Bartholomew 1985, plants with high resin and aromatic oil species, they are more productive and Bredon and Meaker undated), i.e. the per- contents are the more unpalatable karoo assumed to be more nutritious, than early centages of nitrogen and acid detergent subshrubs. Compare the unpalatable karoo developmental stage grasses with normally fibre, can be excluded from the model, subshrub Chrysocoma ciliata with ether low ether extract percentages and low since their influence on the index value is extract values of 8.9 in summer and 9.2 in carotene contents. High ether extract values already taken into account. winter to the more palatable karoo sub- indicate high carotene contents of the grass- The formula for the estimation of TDN shrub Pentzia incana with ether extract es and a more favourable vitamin A, B, E, using acid detergent fibre, is composed of values of 2.3 and 3.9 in summer and win- and probably D content (McDonald et al. the formulae developed by: ter respectively (Botha et al. 1990c) (Table 1973). Late developmental stage grasses 1) Glover et al. (1960) and Bredon and 2). High ether extract contents act as Digitaria eriantha and Themeda triandra Meaker (undated) where : TDN = deterrents to herbivores, while the higher have the highest ether extract values, 2.02

532 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Table 3. Mean objective grazing index values (OGIV) calculated according to the model, com- ing from the use of the objectively esti- pared to the subjective grazing index values (SGIV) mated grazing index values are closer to the accepted grazing capacity norms than Karoobushes OGIV SGIV where the subjectively estimated grazing Min Max Mean Std dev index values have been used. Assuming Chrysocoma ciliata 0.31 1.29 0.83 0.43 1.50 that the grazing capacity norms set for the Eberlanzia ferox 1.07 3.68 1.96 0.67 2.70 various areas are correct, the conclusion Eriocephalus ericoides 0.57 7.67 2.54 1.67 5.00 can be reached that it is more appropriate Eriocephalus spinescens 0.29 5.77 3.65 1.20 4.50 to use the objectively estimated grazing Felicia fascicularis 0.22 0.95 0.48 0.23 6.20 index values in the estimate of the grazing Felicia filifolia 1.94 4.23 2.63 0.93 5.90 Felicia macrorrhiza 1.70 2.37 2.05 0.24 5.70 capacity, than either the ecological index Galenia secunda 2.53 6.21 4.92 1.46 4.70 values or the subjectively estimated graz- Helichrysum dregeanum 0.43 0.53 0.47 0.04 6.30 ing index values. It follows that the use of Helichrysum lucilioides 0.94 3.09 1.75 0.57 5.20 the objectively estimated grazing index Hermannia desertorum 1.09 1.35 1.21 0.09 5.90 values will result in more realistic esti- Monechma incanum 6.01 18.69 9.76 5.20 5.40 mates of the current grazing capacities of Nenax microphylla 0.50 4.07 1.16 0.83 7.00 Osteospermum microphyllum 1.51 1.79 1.66 0.10 7.00 rangeland. Furthermore, the objectively Osteospermum spinescens 2.92 6.03 4.63 1.34 6.00 estimated grazing index values verify the Pentzia globosa 0.71 1.73 1.26 0.30 4.80 relative position of the subjectively esti- Pentzia incana 0.97 2.01 1.50 0.33 5.70 mated grazing index values of the differ- Pentzia spinescens 0.44 6.16 2.43 1.12 4.80 ent plant species on the grazing index Phymaspermum parvifolium 0.39 3.71 1.60 1.12 6.20 value scale. Plinthus cryptocarpus 0.54 4.70 1.95 1.35 6.70 Plinthus karooicus 0.81 3.06 2.06 0.72 6.40 With this model, the grazing index value Pteronia adenocarpa 1.19 3.19 1.82 0.59 3.90 of any species can be estimated with a cer- Pteronia glauca 1.89 3.81 2.77 0.80 3.20 tain degree of confidence. Quality parame- Pteronia glomerata 1.15 2.63 2.06 0.55 3.90 ters can be gleaned from the literature, or Pteronia staehelinoides 0.14 0.22 0.18 0.03 4.00 be determined empirically. It may be pos- Pterothrix spinescens 0.11 3.03 1.91 1.91 2.00 sible to refine this model, by incorporating Rosenia humilis 2.44 6.06 4.00 0.89 3.50 Rosenia oppositifolia 0.50 1.24 0.82 0.30 3.10 some of the parameters excluded at the Salsola calluna 4.24 5.71 5.11 0.62 7.20 moment. However, different grazing index Salsola rabieana 2.79 4.09 3.39 0.42 6.70 values are estimated for species for the Salsola tuberculata 1.64 6.88 4.21 2.04 6.90 different seasons (Table 2) which in turn Walafrida geniculata 2.01 4.29 3.33 0.75 7.00 sheds more light on the concept of palat- Walafrida saxatilis 0.54 0.86 0.69 0.12 2.00 able and unpalatable species, as well as Zygophyllum lichtensteinianum 1.20 1.62 1.43 0.17 4.00 Zygophyllum microphyllum 2.91 4.52 3.79 0.57 4.00 the seasonal palatability of a species. Grasses The index value scale Aristida congesta 0.78 3.36 1.54 0.80 1.30 Aristida diffusa 2.56 11.00 5.75 2.36 5.10 It is questionable whether the top of the Digitaria eriantha 4.82 19.59 8.58 4.47 8.90 grazing index value scale should be 10. The Eragrostis curvula conferta 0.95 4.98 2.76 1.40 6.90 use of a scale with values from 1 to 10 has Eragrostis lehmanniana 1.87 5.64 2.70 0.89 5.40 become customary and is often proposed in Fingerhuthia africana 2.07 5.19 3.41 1.11 6.60 biological work (cf. Curtis and McIntosh Heteropogon contortus 3.77 5.36 4.33 0.55 7.20 1951, Brown and Curtis 1952, Vorster Hyparrhenia hirta 7.74 10.78 8.71 1.21 6.30 Merxmuellera disticha 4.91 12.24 7.46 2.82 5.00 1982, Hurt and Bosch 1991). It is clear that Sporobolus fimbriatus 5.22 10.02 7.08 1.80 9.50 unbeknown to the estimators of the subjec- Stipagrostis ciliata 1.87 12.93 5.82 3.12 7.20 tively estimated grazing index values, the Stipagrostis obtusa 0.62 5.65 2.41 1.28 6.60 old ecological index method scale (Vorster Themeda triandra 2.32 15.04 6.10 3.09 9.30 1982) and the climax adaptation values (Curtis and McIntosh 1951, Brown and Curtis 1952) played an important role in (summer) and 1.92 (winter) as opposed to Model b, grazing index value for positioning the subjective estimates of the the early developmental stage grasses grasses grazing index values. From the foregoing it Aristida congesta and A. adscensionis Grazing index value for the grasses = then becomes clear that all the subjectively with a content of 1.34 in both summer and {(canopy spread cover + available forage estimated grazing index values will have to winter (Botha et al. 1990c) (Table 2) + TDN + [K ÷ (Ca + Mg)]) x ether be adjusted, mainly downwards, so as to (Acocks 1988). Therefore the sum of the extract} ÷ 100 (Du Toit 1996b) (Table 3). fall into line with the objectively estimated calculated values of the different variables grazing index values. Once this action is is multiplied by the ether extract value. The estimated grazing index values accomplished, current grazing capacities This action favours grasses with high ether Although not all the plant species pre- can be estimated much more reliably and extract values. sented in the botanical surveys (Table 4) realistically and different homogeneous have objectively estimated grazing index areas can be directly compared to each values, the range condition scores and the other on a realistic and agriculturally resultant current grazing capacities result- sound scientific basis.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 533 Table 4. Botanical composition, percentage canopy spread cover (CS), range condition scores (VCS) and the respective estimated current grazing capacities (CGC) presented for 7 sample sites from 4 reasonably homogeneous areas in the Karoo, i.e. the Bushmanland, the Eastern Mixed Karoo, the Great Karoo and the Central Upper Karoo. Sites follow a gradient from the arid northwest through the south and central Karoo to the relative- ly moist north eastern Karoo. The ecological, subjective grazing index and objective grazing index values of the different species are compared, as well as the results of their respective products with the percentage strikes on the species. Where no OGIV was available for a plant species, the SGIV was included as the OGIV score, this is indicated by an asterisk next to the species name. CGC is given in ha.LSU-1. EIV = ecological index value, SGIV = subjective grazing index value and OGIV = objective grazing index value.

Grappies Farm (29° 25'S, 19° 57'E), Pofadder district; June 1992, median rainfall 82 mm.a-1, grazing capacity norm 39 ha.LSU-1, Bushmanland.

% EIV Score SGIV Score OGIV Score Aristida congesta 1 1 1 1.3 1.3 1.5 1.5 Eriocephalus spinescens 2 4 8 4.5 9.0 3.7 7.4 Stipagrostis ciliata 39 10 390 7.2 280.8 5.8 226.2 Stipagrostis obtusa 25 10 250 6.6 165.0 2.4 60.0 CS 67 VCS 649 VCS 456 VCS 295.1 CGC 7.2 10.2 15.7

Knolepark Farm (30° 38'S, 26° 20'E), Burgersdorp district; February 1992, median rainfall 450 mm.a-1, grazing capacity norm 10 ha.LSU-1, Eastern Mixed Karoo. % EIV Score SGIV Score OGIV Score Cymbopogon plurinodis* 78 10 780 7.6 592.8 7.6 592.8 Digitaria eriantha 2 10 20 8.9 17.8 8.3 16.6 Pentzia globosa 1 4 4 4.8 4.8 1.3 1.3 Themeda triandra 9 10 90 9.3 83.7 6.1 54.9 CS 90 VCS 894 VCS 699 VCS 665.6 CGC 5.2 6.6 7.0

Diephoek Farm (30°07'S, 24° 15'E), Petrusville district; July 1992, median rainfall 312 mm.a-1, grazing capacity norm 20 ha.LSU-1, Eastern Mixed Karoo.

% EIV Score SGIV Score OGIV Score Eragrostis lehmanniana 30 7 210 5.4 162.0 2.7 81.0 Pentzia globosa 15 4 60 4.8 72.0 1.3 19.5 Aristida congesta 4 1 4 1.3 5.2 1.5 6.0 Plinthus karooicus 8 7 56 6.4 51.2 2.1 16.8 Enneapogon desvauxii* 1 1 1 1.0 1.0 1.0 1.0 Hertia pallens* 2 1 2 1.2 2.4 1.2 2.4 Lycium cinereum* 1 1 1 3.0 3.0 3.0 3.0 Protasparagus suaveolens* 4 1 4 1.0 4.0 1.0 4.0 Eriocephalus aspalathoides* 1 4 4 4.0 4.0 4.0 4.0 Galenia procumbens* 5 4 20 4.3 21.5 4.3 21.5 Eragrostis bergiana* 3 4 12 2.8 8.4 2.8 8.4 CS 74 VCS 374 VCS 3.35 VCS 167.7 CGC 12.4 13.9 27.7

Hillstone Farm (31° 20'S, 25° 31'E), Middelburg district; July 1992, median rainfall 346 mm.a-1, grazing capacity norm 16 ha.LSU-1, Eastern Mixed Karoo.

% EIV Score SGIV Score OGIV Score Sporobolus fimbriatus 14 10 140 9.5 133.0 7.1 99.4 Eragrostis obtusa* 9 7 63 4.0 36.0 4.0 36.0 Eragrostis lehmanniana 1 7 7 5.4 5.4 2.7 2.7 Eriocephalus ericoides 15 4 60 5.0 75.0 2.5 37.5 Pentzia globosa 6 4 24 4.8 28.8 1.3 7.8 Aristida congesta 20 1 20 1.3 26.0 1.5 30.0 Cynodon dactylon* 4 4 16 4.5 18.0 4.5 18.0 Lycium cinereum* 8 1 8 3.0 24.0 3.0 24.0 Eragrostis chloromelas* 1 7 7 5.5 5.5 5.5 5.5 Eriocephalus spinescens 1 4 4 4.5 4.5 3.7 3.7 Chrysocoma cilata 1 1 1 1.5 1.5 0.8 0.8 CS 80 VCS 350 VCS 358 VCS 265.4 CGC 13.3 13.0 17.5

(Continued on Page 535)

534 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 (Table 4. Continued) Swartgrond Farm (33° 04'S, 22° 56'E), Beaufort West district; August 1992, median rainfall 199 mm. a-1, grazing capacity norm 32 ha. LSU-1, Great Karoo.

% EIV Score SGIV Score OGIV Score Aristida congesta 2 1 2 1.3 2.6 1.5 3.0 Protasparagus suaveolens* 1 1 1 1.0 1.0 1.0 1.0 Pentzia incana 14 4 56 5.7 79.8 1.5 21.0 Rhigozum obovatum* 2 7 14 6.6 13.2 6.6 13.2 Eberlanzia ferox 1 1 1 2.7 2.7 2.0 2.0 Hermannia desertorum 5 7 35 5.9 29.5 1.2 6.0 Stipagrostis obtusa 2 10 20 6.6 13.2 2.4 4.8 Eriocephalus ericoides 1 4 4 5.0 5.0 2.5 2.5 Felicia muricata* 3 7 21 6.5 19.5 6.5 19.5 Lycium cinereum* 1 1 1 3.0 3.0 3.0 3.0 Pteronia sordida* 1 4 4 4.5 4.5 4.5 4.5 Eriocephalus spinescens 1 4 4 4.5 4.5 3.7 3.7 Enneapogon desvauxii* 1 1 1 1.0 1.0 1.0 1.0 Stipagrostis ciliata 2 10 20 7.2 14.4 5.8 11.6 Felicia filifolia 1 7 7 5.9 5.9 2.6 2.6 Zygophyllum microphyllum 1 4 4 4.0 4.0 3.8 3.8 CS 39 VCS 195 VCS 204 VCS 103.2 CGC 23.8 22.8 45.0

Abrahamskraal Farm (31° 46'S, 22° 40'E), Victoria West district; February 1992, median rainfall 227 mm.a-1 grazing capacity norm 26 ha.LSU-1,Central Upper Karoo.

Grass/karoo bush range % EIV Score SGIV Score OGIV Score

Eberlanzia ferox 1 1 1 2.9 2.9 2.0 2.0 Eriocephalus spinescens 1 4 4 4.5 4.5 3.7 3.7 Felicia filifolia 2 7 14 5.9 11.8 2.6 5.2 Pentzia incana 34 4 136 5.7 193.8 1.5 51.0 Pteronia glauca 1 4 4 3.2 3.2 2.8 2.8 Pteronia sordida* 4 4 16 4.5 18.0 4.5 18.0 Rosenia humilis 1 1 1 3.5 3.5 4.0 4.0 Salsola tuberculata 1 7 7 6.9 6.9 4.2 4.2 Stipagrostis obtusa 20 10 200 6.6 132.0 2.4 4.8 Walafrida geniculata 1 7 7 7.0 7.0 3.3 3.3 Zygophyllum gilfillanii* 2 7 14 5.9 11.8 5.9 11.8

CS 68 VCS 404 VCS 395 VCS 154 CGC 11.5 11.7 30.0 Mean current grazing capacity of Abrahamskraal farm 17.2 14.1 29.3

Conclusions of range condition scores and in the calcu- Association of Official Agricultural lation of current grazing capacities from Chemists (AOAC) 1964. Official methods these range condition scores (Table 4). of analysis of the association of official agri- The objectively estimated grazing index cultural chemists, 9th Ed, edited by W. These current grazing capacities have been Horwitz. Washington, DC. values verifies the relative position of the estimated for widely differing areas. It is species in the series and they can be com- Bartholomew, P.E. 1985. Beef production clear that they closely approximate the from kikuyu and Italian ryegrass. Ph.D. pared to the subjective grazing index val- long-term grazing capacity norms pre- Thesis. Univ. of Natal, Pietermaritzburg. ues developed earlier at the Grootfontein scribed for these areas by the South Botha, P. 1981. The influence of species selec- Agricultural Development Institute (Du African Department of Agriculture. tion by sheep, cattle and goats on the floristic Toit et al. 1995). composition of mixed Karoorange. (In It is possible to objectively estimate Afrikaans: Die invloed van spesieseleksie grazing index values for plant species of deur skape, beeste en bokke op die floristiese Literature Cited samestelling van gemengde Karooveld.) the arid areas by means of a suitable D.Sc. thesis. P.U. for C.H.E., Potchefstroom. model. In the estimates used in the model, Acocks, J.P.H. 1988. Veld Types of South Botha, P. and C.B. Nash. 1990. Mean phy- certain plant variables such as the canopy tomass and chemical composition of a num- Africa. Memoirs of the Botanical Survey of ber of plant species in the Karoo Midlands. spread cover, the available forage on the South Africa 57:1–146. Gov. Printer, plant and the chemical content of the Tech. Comm. no. 228. Gov. Printer, Pretoria. Pretoria. Botha, P., H.R. Becker, and I.J. Van Der available forage at that instant, play Arnold, T.H. and B.C. De Wet. 1993. Plants prominent roles. Merwe. 1990a. Mean phytomass and chemi- of Southern Africa: Names and Distribution. cal composition of a number of plant species The objective grazing index values have Memoirs of the Bot. Survey of So. Afr. 62:1- in the Great Karoo. Tech. Comm. no. 226. been successfully used in the computation 825. Nat. Bot. Institute, Pretoria Gov. Printer, Pretoria.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 535 Botha, P., C.H. Erasmus, and S.C. Theron. Foran, B.D., N.M. Tainton, and P. De V. Roux, P.W. 1963. The descending-point 1990b. Mean phytomass and chemical com- Booysen. 1978. The development of a method of vegetation survey. A point sam- position of a number of plant species in the method for assessing veld condition in three pling method for the measurement of semi- Nortwestern Karoo. Tech. Comm. no. 227. grassveld types in Natal. Proc. Grass. Soc. open grasslands and Karoo vegetation in Gov. Printer, Pretoria. So. Africa 13:27–33. South Africa. So. African J. Agr.. Sci. Botha, P., W.H. Van Staden, and J.D. Blom. Glover, J., D.W. Duthie, and H.W. Dougall, 6:273–288. 1990c. Mean phytomass and chemical com- 1960. The total digestible nutrients and gross Sneva, F.A. and D.N. Hyder. 1962. position of a number of plant species in the digestible energy of ruminant feeds. J. of Estimating herbage production on semiarid Northeastern Karoo. Tech. Comm. no. 229. Agr. Sci.. 55:403–408. ranges in the intermountain region. J. Range Gov. Printer, Pretoria. Goebel, C.J., L. Debano, and R.D. Lloyd. Manage. 15:88–93. Botha, W. Van D, P.C.V. Du Toit. C.D. 1958. A new method of determining cover Statistical Graphics Corporation 1991. Blom. H.R. Becker. D.J. Olivier. E.M. and production on desert shrub vegetation. J. Statgraphics. STSC Inc, Maryland. Meyer, and G.Z.J. Barnard. 1993. A tech- Range Manage. 11:244–246. Steenkamp, C.W.P. and F.C. Hayward. nique for estimating the grazing index values Hardy, M.B. and C.R. Hurt. 1989. An evalua- 1979. Chemical composition of the most of Karoo plants. Karoo Agr. 5(2):10–14. tion of range assessment techniques in important plant species in the south-central Bredon, R.M. and H.J. Meaker. Undated. Highland Sourveld. J. Grass. Soc. So. Africa sector of the Great Escarpment. Tech. How to calculate TDN and DCP of feeds. 6: 5 1 – 5 8 . Comm. no. 149. Gov. Printer, Pretoria. Leaflet no. 54. Cedara Agricultural Homer, D.C. and F.B. Parker. 1961. Methods Swift, R.W. 1957. The caloric value of TDN. J. Development Institute, Pietermaritzburg. of analysis for soil, plants and water. Univ. Anim. Sci. 16:753–755. Brown, R.T. and J.T. Curtis. 1952. T h e of California, Berkeley, Calif. Tainton, N.M. 1981. The ecology of the main upland conifer-hardwood forests of northern Hurt, C.R. and O.J.H. Bosch. 1991. A com- grazing lands of South Africa. I n : N . M . Wisconsin. Ecol. Mono. 22:217–234. parison of some range condition assessment Tainton (ed.), Range and Pasture Cook, C.W., L.A. Stoddard, and L.E. Harris. techniques used in southern African grass- Management in South Africa, pp 25–56. lands. J. Grass. Soc. So. Africa 8:131–137. 1952. Determining the digestibility and Shuter and Shooter, Pietermaritzburg. Kemp, A. and M.L. T'Hart. 1957. G r a s s metabolizable energy of winter range plants Tainton, N.M., P.J. Edwards, and M.T. tetany in grazing milking cows. Netherlands by sheep. J. Anim. Sci. 11:578–590. J. Agr.. Sci. 5:4–17. Mentis. 1980. A revised method for assess- Curtis, J.T. and R.P. McIntosh. 1951. A n Kidambi, S.P., A.G. Matches, and T.C. ing veld condition. Proc. Grass. Soc. So. upland forest continuum in the prairie-forest Griggs. 1989. Variability for Ca, Mg, K, Cu, Africa 15:37–42. border region of Wisconsin. Ecol. Zn and K÷(Ca+Mg) ratio among 3 wheat- Uresk, D.W. 1990. Using multivariate tech- 32:476–496. grasses and sainfoin on the southern high niques to quantitatively estimate ecological Du Toit, P.C.V. 1993. Stocking rate vs defolia- plains. J. Range Manage. 42:316–322. stages in a mixed grass prairie. J. Range tion of karoobushes. Karoo Agr. 5(1):17–21. Louw, G.N. 1969. The nutritive value of natur- Manage. 43:282–285. Du Toit, P.C.V. 1995. Grazing index method al grazings in South Africa. Proc. So. African Van Der Merwe, F.J. 1985. Animal Nutrition of range condition assessment and compari- Soc. Anim. Prod. 8:57–61. (In Afrikaans: Dierevoeding.) Kosmo pub- son of techniques. African J. Range For. Sci. Louw, G.N., C.W.P. Steenkamp, and E.L. lishing house, Stellenbosch, pp. 1–260. 12(2):61–67. Steenkamp. 1968a. Chemical composition Van Hoven, W. and H. Ebedes. 1988. Du Toit, P.C.V. 1996a. Karoobush defoliation of the most important plant species in the Nutrition, carrying capacity and care of in the Arid Karoo. J. Range Manage. Noorsveld. Tech. Comm. no. 77. Gov. young animals (In Afrikaans: Voeding, 49:105–111. Printer, Pretoria. drakrag en versorging van jong diere.) I n : Du Toit, P.C.V. 1996b. Development of a Louw, G.N., C.W.P. Steenkamp, and E.L. Bothma, J. Du P. (ed.) Wildplaasbestuur, model to estimate grazing index values for Steenkamp. 1968b. Chemical composition of Van Schaiks, Pretoria, pp 1–616. Karoo plant species. Ph. D. Thesis. Univ. of the most important plant species in the Van Soest, P.J. 1963. The use of detergents in Pretoria, Pretoria. Western Mountain Karoo in the Fraserburg the analysis of fibrous feeds: II A rapid Du Toit, P.C.V. and W. Van D. Botha. 1993. District. Tech. Comm. no. 78. Gov. Printer, method for the determination of fibre and Grazing Index Values of Karoo Plants. Paper Pr e t o r i a . lignin. J. Assoc. Agr. Chem. 46:829. read at the Arid Zone Ecology Forum held at Louw, G.N., C.W.P. Steenkamp, and E.L. Vlok, J.J.A.L. 1963. The nutritional value of Middelburg, 28 to 30 September 1993. Abst. Steenkamp. 1968c. Chemical composition the natural forages in the border region Proc. Arid Zone Ecol. Forum 1993:30–31. of the most important plant species in the between the winter and summer rainfall Du Toit, P.J,, J.G. Louw, and A.I. Malan. Arid, Semi-Arid, False Succulent and Central regions in the district of Calvinia. (In 1940. The feeding value of the natural graz- Upper Karoo. Tech. Comm. no. 79. Gov. Afrikaans: Die voedingswaarde van die natu- ings of the Union of South Africa (In Printer, Pretoria. urlike weidingsgewasse in die grensgebied Afrikaans: Die voed ingswaarde van die McDonald, P., R.A. Edwards, and J.F.D. tussen die winter-en somerreënstreke in die natuurlike weivelde van die Unie van Suid- Greenhalgh. 1973. Animal nutrition. Oliver distrik Calvinia.) M.Sci. Agr. Diss. Univ. of Afrika.) Farming in So. Africa 15:229–232. and Boyd, Edinburgh. Stellenbosch, Stellenbosch. Du Toit, P.C.V., W. Van D. Botha. C.D. Maynard, A.L. and J.K. Loosli. 1962. Animal Vorster, M. 1982. The development of the Blom. H.R. Becker. D.J. Olivier. E.M. nutrition. McGraw-Hill, New York. Ecological Index Method for assessing veld Meyer, and G.Z.J. Barnard. 1995. Mentis, M.T., R.F.H. Collinson, and M.G. condition in the Karoo. Proc. Grass. Soc. So. Estimating grazing index values for Karoo Wright. 1980. The precision of assessing Africa 17:84–89. plants. Tech. Comm. no. 239. Gov. Printer, components of the condition of moist tall Wilkes, M. 1989. Operational research : Pretoria. grassveld. Proc. Grass. Soc. So. Africa Analysis and applications. McGraw-Hill Edwards, P.J. 1981. Grazing management. In: 15 : 4 3 – 4 6 . Book Co., London. Tainton, N.M. (ed.) Veld and Pasture Mueller-Dombois, D. and H. Ellenberg. Woods, F.W. 1959. Nutritional aspects of Management in South Africa, pp 315–336. 1974. Aims and methods of vegetation ecolo- Wiregrass from West Florida Sandhills. J. Shuter and Shooter, Pietermaritzburg. gy. John Wiley and Sons, New York. Range Manage. 12:141. Payne, G.F. 1974. Cover-weight relationships. J. Range Manage. 27:403–404.

536 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 J. Range Manage. 53: 537–543 September 2000 Tiller recruitment patterns and biennial tiller production in prairie sandreed

J. R. HENDRICKSON, L. E. MOSER, AND P.E. REECE

Authors are rangeland scientist, Northern Great Plains Research Lab, USDA-ARS, Mandan, N. D. 58554; professor, Agronomy Department, University of Nebraska, Lincoln, Nebr., 68583; and associate professor, Panhandle Research and Extension Center, University of Nebraska, Scottsbluff, Nebr. 69361-4939 respectively. At the time of the research, the senior author was a research assistant, Agronomy Department, University of Nebraska, Lincoln, Nebr.

Abstract Resumen

Tiller recruitment is an essential process for ensuring the El renuevo de las macollas (número de macollas que brotan perenniality of grasses. The timing and extent of tiller recruit- durante todo el periodo de crecimiento) es un proceso esencial ment and the role of biennial tillers must be documented for key para asegurar la perennidad de los pastos. La regulación del range species. Prairie sandreed [Calamovilfa longifolia ( H o o k ) tiempo, la cantidad de renovación de las macollas, y el papel de Scribn.] is an important grass in the Nebraska Sandhills for both las macollas bienales deben ser documentados para las especies ecological functioning and as a forage. The objective of this study claves de las praderas. Prairie sandreed [Calamovilfa longifolia was to document tiller recruitment patterns and the occurrence (Hook) Scribn.] es un pasto importante en las lomas arenosas and contribution of current year and biennial tillers to biomass (sandhills) al oeste de Nebraska, no solo por su contribución production in prairie sandreed at 2 locations in Nebraska. Tiller ecológica sino también como forraje. Los objetivos de este estu- recruitment was monitored at 2-week periods throughout the dio han sido el de documentar el patrón de renuevo de las macol- growing season during a 2-year period. Newly emerged tillers las, la ocurrencia y la contribución de las macollas durante el were classified as intravaginal, extravaginal, or rhizomatous presente año y las macollas bienales del año anterior en la pro- tillers and marked with colored wire. Prairie sandreed has an ducción de la biomasa del pasto prairie sandreed en dos locali- unimodal pattern of tiller recruitment and over 50% of the cur- dades de Nebraska. El renuevo de las macollas fué determinado rent year tillers emerged by mid-May and 80% by mid-June. cada dos semanas durante la época de crecimiento del pasto y Rate of tiller emergence and absolute number of emerged tillers por un período de dos años. A medida que las macollas iban were poorly correlated with short- and long-term precipitation brotando fueron clasificadas en intravaginadas (crecimiento totals (r < 0.3 P > 0.20). The year after new tillers were marked, desde adentro de la vaina de la hoja) y extravaginadas (pene- biennial tillers and tillers initiated during the current-year were tración de la vaina de la hoja) o rizomas e identificadas con counted and clipped in September for biomass determination. alambres de colores. El pasto prairie sandreed posee un patrón Biennial tillers made up only 6 and 20% of the total tiller emer- unimodal en su renovación de las macollas; durante el presente gence at these locations and were generally only 30% as large as año mas del 50% de sus macollas, brotaron a mediados de Mayo the new tillers. Extravaginal tillers composed over 78% of the y el 80% a mediados de Junio. La proporción de la brotación de biennial tiller population as a result of both their dominance in las macollas y el número absoluto de las macollas que brotaron emerging populations and the higher percentage of tillers that estuvo pobremente correlacionado (r < 0.3 P > 0.20) con la pre- survived the winter. Current year tillers contributed the most to cipitación total a corto y a largo plazo. En el segundo año, las prairie sandreed forage production and their emergence was macollas bienales y las macollas que brotaron durante el pre- largely completed by mid-June. The lack of a relationship sente año fueron contadas y cortadas para determinar la bio- between tiller recruitment and precipitation patterns, combined masa. Las macollas bienales representaron de un 6 al 20% del with previous studies of prairie sandreed, indicates that tiller total de las macollas en las dos localidades , y fueron por lo gen- recruitment involves a process that begins the previous growing eral 30% tan grandes como las macollas nuevas. Las macollas season. extravaginadas formaron mas del 78% de la población de macol- las bienales, debido a su caracter dominante para brotar y tam- bién a su alto porcentaje de sobrevivencia durante el invierno. Key Words: Warm-season grass, tiller demography, Calamovilfa Las macollas del presente año fueron las que mas contribuyeron l o n g i f o l i a (Hook.) Scribn., grassland ecology, population ecolo- a la producción del pasto forrajero sandreed, completando la gy/biology mayor parte de su brotación a mediados de Junio. La carencia de relación entre el renuevo de las macollas y los patrones de la precipitación, combinado con resultados obtenidos en años ante- riores, indican que el renuevo de las macollas está involucrado The authors would like to thank Drs. D. Briske, R. Heitschmidt, J. Volesky and en un proceso el cual se inicia en el anterior período de crec- J. Stubbendieck for their comments on earlier versions of this manuscript, and N. imiento del pasto. Mason for providing the Spanish translation of the abstract. All programs and services of the U.S. Department of Agriculture are offered on a nondiscriminatory basis without regard to race, color, national origin, religion, sex, age, martial status, or handicap. This manuscript was published as the University of Nebraska Agricultural Research Division Journal Series No. 12350. Prairie sandreed [Calamovilfa longifolia (Hook) Scribn.] is a Manuscript accepted 7 Dec. 1999. widespread rhizomatous perennial grass on sandy soils in the Northern Great Plains of the United States. It is one of the most

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 537 common grasses of the Nebraska Sandhills often have a greater dry weight, leaf num- Average annual precipitation at the (Tolstead 1942) and together with sand ber, and seed yield than tillers recruited Gudmundsen Sandhills Laboratory is 514 bluestem [Andropogon gerardii var. pau - during the growing season because they mm and 70% falls during the growing sea- c i p i l u s (Nash) Fern.] and little bluestem have more time for growth and develop- son (April through September). At the [Schizachyrium scoparium (Michx.) Nash] ment (Langer 1956). Panhandle Experimental Range, the aver- provides 50–60% of the available forage Tiller type influences plant architecture age annual precipitation is 393 mm with in most years (Northup 1993). Tolstead (Briske 1991) and plant development the same distribution pattern. Precipitation (1942) and Hendrickson et al. (1998) (White 1977). Bunchgrasses are character- from 1 April to 30 September was 76, 119, described the developmental morphology ized by intravaginal tillers or tillers which and 90% of the 30-year average for the of prairie sandreed. arise from within the subtending leaf Gudmundsen Sandhills Laboratory in Knowledge of tiller dynamics con- while sod forming grasses have more 1990, 1991, and 1992 respectively. tributes to greater ecological understand- extravaginal and rhizomatous tillers Precipitation during the same years, at the ing and more effective management of (Briske 1991). Prairie sandreed has Panhandle Experimental Range, was 98, grasslands (Briske and Silvertown 1993, rhizamotous, intra- and extravaginal tillers 138, and 96% of the 30-year average. McKenzie 1997). Understanding tiller similar to switchgrass (Panicum virgatum Precipitation data were collected electroni- dynamics is critical in developing appro- L.) (Brejda et al. 1988). Rhizomatous cally during the growing season from a priate grazing strategies for rangelands tillers in prairie sandreed have an weather station at the Gudmundsen (Cullan et al. 1999). Most perennial grass- increased probability of producing seed- Sandhills Laboratory headquarters and es have bimodal recruitment patterns with heads (White 1977) which has an adverse flushes of new tillers emerging during the affect on nutritive value and palatability from a rain gauge located 1.25 km south spring and fall (Langer 1956, Briske and (Reece et al. 1999). of the Panhandle Experimental Range Butler 1989, Briske and Richards 1995). Despite the importance of prairie san- study site. Thirty-year averages were However, crested wheatgrass [A g r o p y r o n dreed to the Nebraska Sandhills in particu- taken from the closest U.S. Weather desertorum (Fisch. Ex Link) Schult.], lar and the Northern Great Plains in gener- Service stations located 20 km northeast bluebunch wheatgrass [P s e u d o r o e g n e r i a al, relatively little is known about its basic of the Gudmundsen Sandhills Laboratory s p i c a t u m (Pursh) A. Löve] (Mueller and demographic patterns and the role of bien- and 10.5 km southeast of the Panhandle Richards 1986), and big bluestem nial tillers in biomass determination. The Experimental Range. [Andropogon gerardii Vitmann var. g e r - objective of this study was to evaluate Study sites at the Gudmundsen a r d i i] (McKendrick et al. 1 9 7 5 ) , o n l y tiller recruitment patterns in prairie san- Sandhills Laboratory were excluded from produced 1 annual cohort. These estab- dreed, estimate the contributions of intrav- livestock grazing for 5 years previous to lished patterns of tiller recruitment may be aginal, extravaginal, and rhizomatous the initiation of the study. At the affected by defoliation which extends the tillers to overall tiller recruitment and Panhandle Experimental Range, study recruitment period (Butler and Briske determine the importance and contribution sites were grazed as part of a deferred 1988), promotes additional tiller cohorts of biennial tillers to forage production. rotation for 15 years previous to the study (Olson and Richards 1988a), and changes at a stocking rate of 1.01 AUM/ha. the timing of peak recruitment (Bullock et However, livestock were excluded from al 1994). Tiller recruitment is also a major Materials and Methods the sites during the investigation. method of perennial grass persistence Twelve permanent 0.5 m2 quadrats (70.5 (Hendrickson and Briske 1997) and annual cm x 70.5 cm) were randomly located tiller replacement is necessary to maintain Research was conducted in 1990 and within the study sites during the first week tiller density (Olson and Richards 1988b). 1991 at 2 University of Nebraska research of May 1990 at each location. There were Reduction in tiller density reduces both locations. The Gudmundsen Sandhills a total of 4 quadrats within each of the 3 current and future productivity since tiller Laboratory (42° 07'N, 101° 43'W) is cen- study areas at the Gudmundsen Sandhills density represents a pool of meristematic trally located in the Nebraska Sandhills Laboratory and 2 quadrats within each of tissue for future growth (Murphy and near Whitman, Nebr. and the Panhandle the 6 study areas at the Panhandle Briske 1992). Experimental Range (42° 08'N, 103° Experimental Range. Quadrats were moni- Timing of tiller recruitment has a direct 43'W) is located near Mitchell in the tored for new tiller emergence at 2-week effect on tiller longevity and tiller yield. Nebraska Panhandle approximately 240 intervals, from mid-May through mid- Tillers recruited in the fall often overwin- km west of the Gudmundsen Sandhills September during the 1990 growing sea- ter in the vegetative stage and resume Laboratory. There were 3 and 6 study sites son. Newly emerged tillers were classified growth the following spring (Briske located at the Gudmundsen Sandhills into 3 categories (intravaginal, extravagi- 1991). These “biennial tillers” have been Laboratory and the Panhandle Experimental nal, or rhizomatous) based on the position reported for certain grasses in the plains Range respectively. Study sites were located of the emerging tiller. Tillers that emerged states. In the Kansas Flint Hills, a majority on Valentine fine sands (mixed, mesic typic from within the subtending leaf sheath of indiangrass tillers [Sorghastrum nutans Ustipsamments) at the Gudmundsen were considered intravaginal tillers while L. (Nash)] were biennial but big bluestem Sandhills Laboratory and Valent fine sands extravaginal tillers emerged through the tillers only lived for 1 year (McKendrick (mixed, mesic, ustic Torripsamment) at the subtending leaf sheath (Briske 1991). et al. 1975). In the Nebraska Sandhills, Panhandle Experimental Range in areas Tillers that emerged more than 3 cm from biennial tillers were reported in prairie dominated or co-dominated by prairie san- the nearest existing tiller were considered sandreed but not its co-dominant sand dreed (³ 50% of the cover). Dominance or to be rhizomatous. Colored wires were bluestem (Bredja et al. 1988, Cullan et al. co-dominance by prairie sandreed is typical placed on the base of each new tiller to 1999) and prairie sandreed in Montana did indicate category and date of emergence. not produce biennial tillers (White 1977). of a majority of Sandhills vegetation that is in mid to high seral stage. In 1991, these same 12 quadrats at each Biennial tillers are important because they location were evaluated to determine the

538 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 winter survival of the 1990 tillers by day intervals prior to the sampling period. Experimental Range so total tiller emer- counting the number of live tillers. Tiller emergence was also correlated with gence was very similar between locations Surviving tillers were classified as bienni- the 30-year average precipitation on a (P > 0.75) (Fig. 2). Over 50% of the total al. Counts were done in mid-April, mid- monthly basis. Correlations were done tiller emergence for the season occurred May and at the end of September in 1991. using the PROC CORR procedure in SAS by mid-May for both years and locations In September, biennial and new tillers (SAS 1989). Significance was determined at (Fig. 1 and 2). After early June, tiller were clipped separately, separated into P £ 0.05 unless otherwise noted. Standard emergence generally declined and vegetative and reproductive tillers, dried errors were calculated averaged across sam- remained low for the remainder of the for 3 days at 55°C, and weighed. pling dates. growing season. In 1991, 12 new 0.5 m2 quadrats were The Panhandle Experimental Range had Time, but not location by time, had a sig- randomly located in the same study areas relatively few biennial tillers. Therefore, nificant effect on tiller recruitment in 1990, at each location and monitoring was con- most of the data used in evaluating the thus tiller emergence was pooled over ducted in the same manner as in 1990 contribution of biennial tillers came from location (Fig. 1). Each of the first 3 sam- except that at the Panhandle Experimental the Gudmundsen Sandhills Laboratory. pling periods in 1990 had significantly Range, monitoring of tiller emergence Standard errors used in the comparison of greater tiller recruitment than the subse- began in mid-April rather than mid-May. the contributions of new and biennial quent period (Fig. 1). In 1991, there was a Evaluation of these 12 quadrats for bienni- tillers to biomass and tiller numbers were location by time interaction (Fig. 2) and al tillers was conducted in April and calculated across study sites and quadrats. tiller recruitment patterns were analyzed September of 1992. Standard errors, used in evaluating the separately at each location. The pattern of Daily tiller initiation rate was calculated number and percent of biennial tillers tiller emergence at the Gudmundsen as follows: from each sampling date, were calculated Sandhills Laboratory was similar to 1990 across study sites and quadrats within although there were no significant differ- Daily tiller initiation rate = sampling dates. ences between the mid and late June time number of new tillers 1.0 m-2 periods (Fig. 2). At the Panhandle number of days since previous sampling Experimental Range, tiller emergence was The number of tillers emerged was con- Results similar in early and mid June and those verted from 0.5m- 2 to the more recogniz- time periods were significantly greater than able 1.0 m- 2. Based on the 1991 tiller Emergence date and category: late June (Fig. 2). emergence patterns at the Panhandle There were no significant differences in The Gudmundsen Sandhills Laboratory Experimental Range, 1 April was selected tiller recruitment between locations in had significantly more extravaginal and as the beginning date for calculating rate 1990 (P > 0.10); however, tiller emer- rhizomatous tillers than the Panhandle of tiller initiation for the first sampling gence during the 1990 growing season Experimental Range in 1990. However, period each year. was 160% greater at Gudmundsen intravaginal tiller number was influenced Tiller emergence was analyzed as a split Sandhills Laboratory than at Panhandle more by site within location than by loca- plot in time with time being a sub-plot fac- Experimental Range (Fig. 1). In 1991, tion. In 1991, the Gudmundsen Sandhills tor to allow for a greater resolution of time tiller emergence declined by 25% at the Laboratory had significantly more intrav- (Briske and Hendrickson 1998). Differences Gudmundsen Sandhills Laboratory but aginal tillers than the Panhandle between locations were analyzed using nearly doubled at the Panhandle Experimental Range, but the number of study site nested within location as an error term and differences between time and location by time were analyzed using time by study site nested within location as an error term. A majority of tiller emergence occurred early in the growing season and so only the first 4 time periods were included in the data analysis. Tiller emer- gence data were log transformed to correct for a non-normal distribution. Results are presented using non-transformed data for clarity of presentation. Standard errors were calculated across the first 4 time periods. Tillers within each type category were pooled over time periods, because of small sample sizes in some time periods and categories, and analyzed for differ- ences between locations and study sites within location. Analysis was conducted using the SAS PROC GLM module (SAS -2 1989). Means separation was done using Fig. 1. Number of tillers that emerged (m ) on each monitoring date in 1990 pooled across locations for the Gudmundsen Sandhills Laboratory (GSL) and the Panhandle the Student-Newman-Keuls’ test. Experimental Range (PER). Newly emerged tillers were classified as either intravaginal, The absolute number of tillers that extravaginal or rhizomatous at each monitoring date. Numbers in parentheses are the total emerged and the rates of tiller emergence number of tillers that emerged (tillers m-2). Data was pooled over locations for each emer- were correlated with the sum of the pre- gence date. Bars with different letters represent significant differences (P < 0.05) in total cipitation received during the 14 and 30 - tiller emergence among the first 4 dates.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 539 tively. Intravaginal tillers contributed the least number of tillers at both locations with only 7% and 16% of the emerged tillers being intravaginal at the Panhandle Experimental Range and the Gudmundsen Sandhills Laboratory respectively.

Tiller initiation correlations Absolute tiller recruitment and tiller ini- tiation rate were poorly correlated with 14-day and 30-day precipitation amounts (Fig. 3). Correlation coefficients were less than 0.3 (P > 0.20). An exception was tiller initiation rate and 30-day precipita- tion at Gudmundsen Sandhills Laboratory in 1991 (r = 0.43, P = 0.07). Correlations between tiller emergence and longer-term precipitation data (monthly totals over 30 years) was also low (r < 0.30, P > 0.20).

Biennial tiller yield: Most tillers did not survive the follow- ing winter. Only 22% of the tillers, which emerged during the 2-year study, became biennial tillers at the Gudmundsen Sandhills Laboratory and only 6% at the Panhandle Experimental Range. In 1991, less than 1% of all the tillers that emerged during the growing season at the Panhandle Experimental Range became biennial tillers. Because of their minimal contribution to numbers and biomass at the Panhandle Experimental Range, those biennial tillers were not included in any of the data analysis regarding biennial tillers. Approximately 30% of tillers harvested in 1991 from plots established in 1990 at the Gudmundsen Sandhills Laboratory were biennial tillers but this had declined to Fig. 2. Number of tillers that emerged (m- 2) on each monitoring date in 1991 at the Gudmundsen Sandhills Laboratory (GSL) and the Panhandle Experimental Range (PER). 20% in 1992 (Fig. 4). Biennial tillers made Newly emerged tillers were classified as either intravaginal, extravaginal or rhizomatous up 12% of the yield from all live prairie tillers at each monitoring date. Number in parentheses is the total number of tillers that sandreed tillers in 1991 and 6% of the emerged (tillers m- 2) during the growing season at each location. Data was not pooled 1992 biomass at the Gudmundsen because of a location by time interaction. Bars with different letters represent significant Sandhills Laboratory (Fig. 4). differences (P < 0.05) in total tiller emergence between the first 4 dates for each location. Yields of biennial tillers may have decreased because of harvesting in extravaginal tillers was similar between Range. Rhizomatous tillers were the next September rather than earlier in the grow- locations. The number of rhizomatous largest category and they were 20 and ing season. However, there were no signif- tillers, in 1991, was affected more by site 18% of total tiller emergence at the icant differences (P > 0.05) at the within location than by location. Gudmundsen Sandhills Laboratory and the Gudmundsen Sandhills Laboratory when Over the 2-year period, extravaginal Panhandle Experimental Range, respec- the number of biennial tillers recorded tillers made up more than 63% of all tillers recruited at the Gudmundsen Table 1. Mean individual tiller weights for tillers surviving from the previous year (biennial) and Sandhills Laboratory and over 75% of all tillers that emerged during the current year at the Panhandle Experimental Range (PER) and tillers recruited at the Panhandle Gudmundsen Sandhills Laboratory (GSL). Biennial tillers recorded in 1991 and 1992 had Experimental Range. Extravaginal tillers emerged in 1990 and 1991 respectively. comprised more than 50% of the emerged tillers at any single sample date (Figs. 1 PER GSL and 2) and they were the only tiller cate- Harvest Date Biennial Tillers Current Year Tillers Biennial Tillers Current Year Tillers gory to emerge from 24 July to 28 August ------(mg.) ------1990 and from 6 August to 19 August 1991 78 ± 4 759 ± 48 131 ± 3 443 ± 33 1992 321 ± 63 414 ± 21 93 ± 3 334 ± 8 1991 at the Panhandle Experimental

540 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 early in the season was compared to bien- nial tillers recorded later in the growing season. Tillers that emerged during the current growing season were 30 to 880% heavier then the biennial tillers (Table 1). These extremes were both recorded at the Panhandle Experimental Range but at the Gudmundsen Sandhills Laboratory, cur- rent year tillers were 230 to 260% heavier than the biennial tillers. At the Gudmundsen Sandhills Laboratory, tiller cohorts that emerged early in the growing season the previous year made up the largest percentage of biennial tillers (Fig. 5). However, tillers that emerged later in the growing season were more likely to become biennial tillers (Fig. 5). May tiller cohorts made up approximately 33% of the biennial tillers in 1991 and 60% of the biennial tillers in 1992. The 1991 biennial tiller population had a large contribution from tiller cohorts that emerged later in the season in 1990 because of increased tiller emergence later in the season combined with increased potential for these tillers to become bienni- al. Over 70% of the biennial tillers at Gudmundsen Sandhills Laboratory were from the extravaginal tiller category (Table 2) because of the greater number of extrav- aginal tillers. Extravaginal tillers had the highest survival rate in the 1990 cohort but Fig. 3. Daily tiller initiation rates (tillers m-2 day-1) and precipitation events (cm) during the survival rates were similar between the growing season at the Gudmundsen Sandhills Laboratory and the Panhandle extravaginal and intravaginal categories in Experimental Range during the 1990 and 1991 growing seasons. The numbers in paren- the 1991 tiller cohort (Table 2). thesis represent total precipitation received during the growing season at each location and each year.

Discussion bimodal tiller recruitment pattern while T o r r . ] (Hendrickson 1996). Moreover, big bluestem, had a unimodal pattern although both locations received more pre- (McKendrick et al. 1975). cipitation during the 1991 than in the 1990 The unimodal pattern of tiller emer- Regionally, patterns of tiller emergence growing season, total tiller emergence gence in prairie sandreed differs from the may be linked to precipitation patterns or decreased from 1990 to 1991 at more commonly reported bimodal tillering the length of the growing season (Briske Gudmundsen Sandhills Laboratory while pattern of other perennial grasses in tem- 1991). In our more site specific study, increasing by 100% at Panhandle perate zones (Langer 1963, Butler and tiller emergence during the growing sea- Experimental Range. However, Cullan et Briske 1988, Briske and Richards 1995). son and the rate of daily tiller emergence al. (1999) suggested that soil moisture and However, unimodal tillering patterns have were poorly correlated with either yearly the ability of plants to absorb soil moisture been reported for the C3 grasses, crested or 30-year average precipitation events may be critical factors in determining wheatgrass, and bluebunch wheatgrass, which was similar to reports from sideoats recruitment in prairie sandreed. The sandy (Mueller and Richards 1986). In a Kansas grama [Bouteloua curtipendula ( M i c h x . ) substrate at both locations complicates the study of 2, C4 grasses, indiangrass had a response to precipitation because its coarse texture allows for deep percolation Table 2. Percentage of tillers that emerged during the preceding year and survived to become of even minor rainfall events (Barnes and biennial tillers from each tiller category and the percentage that each tiller category con- tributed to the composition of the biennial tiller population in 1991, 1992 and averaged over Harrison 1982). Thus, frequency of pre- both years at the Gudmunsen Sandhills Laboratory. Data from the Panhandle Research Range cipitation events may be more important not presented. that total precipitation. Biennial tillers did not make a large con- Composition of the tribution to tiller recruitment (6 to 22%) or Survival Rate Biennial Tiller Population aboveground biomass, which was surpris- Tiller Category 1990 1991 Average 1990 1991 Average ing since the longer season of growth and ------(%) ------development is often considered to give a Extravaginal 27 ± 2.2 27 ± 3.8 27 ± 2.0 86 ± 2.7 70 ± 1.3 78 ± 3.9 growth advantage to biennial tillers Intravaginal 8 ± 2.7 25 ± 2.2 16 ± 4.2 6 ± 2.2 15 ± 1.3 10 ± 2.3 (Langer 1956, Briske 1991). However in Rhizomatous 9 ± 4.7 17 ± 2.3 13 ± 2.9 8 ± 2.7 15 ± 2.6 12 ± 2.4 our study, new tillers were 200% heavier

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 541 prairie sandreed with an alternative popu- lation maintenance mechanism. This observation is difficult to assess because of the limited time frame of the experi- ment but a majority of biennial tillers remained vegetative even at the end of their second growing season. Although the contribution of axillary buds and tiller recruitment to population persistence has been explored (Hendrickson and Briske 1997), there is limited information regard- ing the role of tiller longevity in this regard. There have been reports of tiller ages of up to 5 years in northern wheat- grass [Agropyron dasystachyum ( H o o k ) Scribn.) in Saskatchewan, Canada (Zhang and Romo 1995) and 3-year old western Fig. 4. Contribution of current year tillers and biennial tillers to biomass (g m-2) and tiller wheatgrass (Agropyron smithii R y d b . ) numbers m-2 at the Gudmundsen Sandhills Laboratory in 1991 and 1992. Biennial tillers tillers were observed in Montana (White had emerged during the previous growing season (1990 and 1991 respectively) and sur- 1977). These reports are from areas with vived overwinter. Current year tillers were tillers that emerged in the 1991 and 1992 grow- more severe winter climates and contrast ing seasons. Current year tiller biomass was separated into vegetative and reproductive with lifespan of 1–2 years in more moder- components but biennial tiller biomass was pooled over both components. ate climates (Langer 1956, Briske 1991, than biennial tillers thus early emergence did not seem to give a growth advantage to biennial tillers. If biennial tillers died prior to the harvesting date in September, their numbers and biomass could have been underestimated. However, biennial tillers numbers were generally similar between the early (April and May) and September monitoring dates. The largest percentage of biennial tillers were from tiller cohorts that emerged early in the previous growing season because of the larger tiller recruitment. However, later emerging tillers were more likely to become biennial. Although, tillers which emerge early in the growing season gener- ally have the greatest probability of becoming reproductive (Briske 1991), in our study a majority of biennial tillers were still vegetative when they were har- vested at the end of the second growing sea- son. A study of morphological development in prairie sandreed conducted during the same time period at Gudmundsen Sandhills Laboratory also indicated a majority of tillers remained vegetative throughout the growing season (Hendrickson et al. 1998). This suggests that the limited number of biennial tillers in prairie sandreed may not be a result of the tillers completing their life cycle but rather because of the small num- ber of tillers recruited late in the growing season when they were more likely to become biennial. In Montana, C3 g r a s s tillers took 2–3 years to flower and C4 gr a s s tillers took 1–2 years although the results varied by species (White 1977). -2 Biennial tillers may represent a small Fig. 5. The number and percent of new tillers m marked on each monitoring date in 1990 and 1991 that survived overwinter and became biennial tillers the following year at the pool of long-lived tillers that provide Gudmundsen Sandhills Laboratory.

542 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Briske and Richards 1995). From a practi- Literature Cited Hendrickson, J.R., L.E. Moser, K.J. Moore cal point of view, a majority of the prairie and S.S. Waller. 1998. M o r p h o l o g i c a l sandreed tillers appeared to function as development of two warm-season grasses in Barnes, P. W. and A. T. Harrison. 1982. annual tillers similar to reports of prairie the Nebraska Sandhills. J. Range Manage. Species distribution and community organi- 51:456–462. sandreed tillers in Montana (White 1977). zation in a Nebraska Sandhills mixed prairie Langer, R.H.M. 1956. Growth and nutrition of The limited contribution of biennial as influenced by plant/soil-water relation- timothy (Phleum pratense)I. The life history tillers in our study emphasized the impor- ships. Oecologia 52: 192–201. of individual tillers. Annu. of Applied Biol. tance of tillers initiated during the current Brejda, J. J., L. E. Moser, and S. S. Waller. 44:166–187. year to tiller density and productivity. 1 9 8 8 . Rhizome and tiller development of Langer, R.H.M. 1963. Tillering in herbage However tillers, which emerge during the three Nebraska Sandhills warm-season grass- grasses. Herbage Abstr. 33:141–148. growing season, are often initiated early es. p. 52. I n: Thomas B. Bragg and James McKendrick, J.D., C.E. Owensby and R.M. Hyde. 1975. Big bluestem and indiangrass and make some limited subterranean Stubbendieck (eds.), Proc. 11th North American Prairie Conference, Univ. of Nebr. vegetative reproduction and annual reserve growth in the previous growing season Printing, Lincoln, Nebr. carbohydrate and nitrogen cycles. Agro- (Bredja et al. 1988). Therefore, stress that Briske, D. D. 1991. Developmental morpholo- Ecosystems 2:75–93. occurs during this crucial time period gy and physiology of grasses. Grazing McKenzie, F.R. 1997. Influence of grazing fre- early in the growing season may affect Management: an ecological perspective. pp. quency on intensity of tiller appearance and tiller numbers into the following growing 85–108. (Eds. R.K. Heitschmidt & J.W. death rates of Lolium perenne L. under sub- season. For example, tiller and axillary Stuth). Timber Press, Portland, Ore. tropical conditions. Australian J. Agr. Res. Briske, D. D. and J.L. Butler. 1989. Density- 48:337–342. bud numbers of prairie sandreed per unit Mueller, R.J. and J.H. Richards. 1986. area decreased when prairie sandreed and dependent regulation of ramet populations within the bunchgrass Schizachrium scopari - Morphological analysis of tillering in its associated species were clipped early in Agropyron spicatum and Agropyron deserto - u m: interclonal versus intraclonal interfer- the growing season (June) over a 3-year rum. Annu. Bot. 58:911–921. ence. J. of Ecol. 77:963–974. period (Mullahey et al. 1991). Organic Mullahey, J. J., S. S. Waller, and L. E. Briske, D. D. and J.R. Hendrickson. 1998. Moser. 1991. Defoliation effects on yield reserves of prairie sandreed were also Does selective defoliation mediate competi- adversely affected by a single grazing and bud and tiller numbers of two Sandhills tive interactions in a semiarid savanna? A grasses. J. Range Manage. 44:241–245. event in June or July but not August demographic evaluation. J. of Veg. Sci. Murphy, J.S. and D.D. Briske. 1992. (Reece et al. 1996). These responses col- 9:611–622. Regulation of tillering by apical dominance: lectively indicate that tiller numbers dur- Briske, D. D. and J.H. Richards. 1995. Plant Chronology, interpretive value, and current ing the year may be determined relatively responses to defoliation: A physiological, perspectives. J. Range Manage. 45:419–429. early in the previous growing season. morphological and demographic evaluation. Northup, B.K. 1993. Utilization of native for- Consideration of the effects of environ- In: (eds. D.J. Bedunah and R.E. Sosebee), pp. ages of the Nebraska Sandhills by yearling 635–710. Wildland Plants: Physiological mental variables and management over cattle. Ph.D. Diss. University of Nebraska, ecology and developmental morphology. Lincoln, Nebr. multiple years needs to be incorporated Soc. of Range Manage., Denver, Colo. Olson, B.E. and J.H. Richards. 1988a. into current management decisions. Briske, D.D. and J.W. Silvertown. 1993. Annual replacement of the tillers of Plant demography and grassland community Agropyron desertorum following grazing. balance: the contribution of popuation regu- Oecologia 76:1–6. Conclusions lation mechanisms. I n: Proceedings of the Olson, B.E. and J.H. Richards. 1988b. XVII Internat. Grassl. Congr., pp. 291–298. Tussock regrowth after grazing: intercalary New Zealand. meristem and axillary bud activity of tiller of In prairie sandreed, the tillering pattern Bullock, J. M., B. Clear Hill, and J. Agropyron desertorum. Oikos 51: 374-382. and role of biennial tillers contrasts with Silvertown. 1994. Tiller dynamics of two Reece, P.E., T.L. Holman, and K.J. Moore. patterns observed in other species (Langer grasses- responses to grazing, density and 1999. Late-summer forage on prairie san- 1956, Briske 1991). Prairie sandreed has a weather. J. of Ecol. 82:331–340. dreed dominated rangeland after spring defo- unimodal tillering pattern and a majority Butler, J.L. and D.D. Briske. 1988. liation. J. Range Manage. 52:228–234. Population structure and tiller demography of Reece, P.E., J.E. Brummer, R.K. Engel, B.K. of tillers do not survive over winter. The Northup, and J.T. Nichols. 1996. G r a z i n g few biennial tillers that are produced are the bunchgrass Schizachyrium scoparium i n response to herbivory. Oikos 51:306–312. date and frequency effects on prairie san- small and make a minor contribution to dreed and sand bluestem. J. Range Manage. biomass production as opposed to tillers Cullan, A.P., P.E. Reece, and W.H. Schacht. 1999. Early summer grazing effects on defo- 49:112–116. SAS Institute. 1989. SAS user’s guide: statis- that emerge during the growing season. liation and demography of prairie sandreed. tics (Version 6 1s t ed.) SAS Institute Inc. The developmental morphology of prairie J. Range Manage. 52:447–453. Cary, N.C sandreed and the low correlations with Hendrickson, J.R. 1996. Mechanisms of per- precipitation indicate current-year tiller Tolstead, W.L. 1942. Vegetation of the north- sistence and species replacement in popula- ern Cherry County, Nebraska. Ecol. Monogr. production may be linked to conditions in tion of Bouteloua curtipendula with contrast- 12:255–292. previous years. Given the importance of ing grazing histories. Ph.D. Diss., Texas White, L. M. 1977. Perenniality and develop- prairie sandreed for ecological functions A&M Univ., College Station, Tex. ment of shoots of 12 forage species in and as a forage resource, quantifying the Hendrickson, J.R. and D.D. Briske. 1997. Montana. J. Range Manage. 30:107–110. effects of preceding and current-year man- Axillary bud banks of two semiarid perennial Zhang, J. and J. T. Romo. 1995. Impacts of grasses: occurrence, longevity, and contribu- defoliation on tiller production and survival agement as well as environmental vari- tion to population persistence. Oecologia ables on tiller emergence is warranted. in northern wheatgrass. J. Range Manage. 110:584–591. 48:115–120.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 543 J. Range Manage. 53: 544–549 September 2000 Seed biology of rush skeletonweed in sagebrush steppe

JULIA D. LIAO, STEPHEN B. MONSEN, VAL JO ANDERSON, AND NANCY L. SHAW

Authors are Ph.D. student, Department of Rangeland Ecology and Management, Texas A&M University, College Station, Tex.. 77843; botanist, USDA-FS, Rocky Mountain Research Station, Provo, Ut. 84606; associate professor, Botany and Range Science Department, Brigham Young University, Provo, Ut. 84602; and botanist, USDA-FS, Rocky Mountain Research Station, Boise, Ida. 83702. At the time of the research, the senior author was graduate student, Department of Botany and Range Science, Brigham Young University, Provo, Ut. 84602.

Abstract Resumen

Rush skeletonweed (Chondrilla juncea L.) is an invasive, herba- Chondrilla juncea L. (nombre vulgar: yuyo esqueletico) es una ceous, long-lived perennial species of Eurasian or Mediterranean especie herbácea perenne e invasora de larga vida y origen origin now occurring in many locations throughout the world. In eurasiático o mediterraneo que habita actualmente en muchos the United States, it occupies over 2.5 million ha of rangeland in lugares del mundo. En los Estados Unidos, ocupa más de 2.5 mil- the Pacific Northwest and California. Despite the ecological and lones de hectáreas de tierras ganaderas en el Noroeste Pacífico y economic significance of this species, little is known of the ecolo- California. A pesar del significado ecológico y económico de esta gy and life history characteristics of North American popula- especie, poco se sabe sobre la ecología y las características de la tions. The purpose of this study was to examine seed germination historia de vida de las poblaciones norteamericanas. El objetivo characteristics of 2 populations of rush skeletonweed in Idaho. del presente trabajo fue examinar las características de la germi- Seeds from rush skeletonweed plants in southwestern Idaho were nación de semillas de dos poblaciones de C. juncea en Idaho. Se collected during the 1994 and 1995 growing seasons. Mature colectaron semillas de plantas en el suroeste de Idaho durante la seeds were harvested on 6 dates between early July and early estación de crecimiento en 1994 y 1995. Se coleccionaron semillas October 1994, and on 5 dates between early July and late maduras en 6 ocasiones entre principios de julio y principios de September 1995. Fresh seeds from each harvest period were octubre de 1994, y en 5 ocasiones entre principios de julio y fines measured to determine seed weight, total germination, rate of de septiembre de 1995. Se midieron semillas frescas de cada germination, and viability (tetrazolium staining [TZ]) of non-ger- cosecha para determinar el peso de las semillas, la germinación minating seeds. An aliquot of seeds collected in 1994 was also total, la tasa de germinación, y la viabilidad (teñido con tetrazo- stored for 1 year to examine the effects of seed storage on germi- lio [TZ]) de semillas que no germinaron. Una alícuota de semillas nation. In southwestern Idaho, rush skeletonweed produces seeds colectadas en 1994 se almacenó por un año para examinar los continuously from mid-July through October. Seeds were capa- efectos del almacenamiento sobre la germinación. En el suroeste ble of immediate germination without scarification or wet de Idaho, C. juncea produce semillas continuamente desde medi- prechilling. Total germination generally ranged from 60 to 100% ados de julio hasta octubre. Las semillas fueron capaces de ger- throughout the entire seed production period. Germination was minar inmediatamente sin escarificación o mojado y enfriado also rapid, reaching 50% of total germination in less than 12 previos. La germinación total varió generalmente del 60 al 100% days. In general, germination was higher at the lower incubation durante el período completo de producción de semillas. La ger- temperature regime (20/10˚C), perhaps reflecting origins of this minación también fue rápida, alcanzando el 50% de la germi- species in Mediterranean winter rainfall regions. The TZ testing nación total en menos de 12 días. En general, la germinación fue indicated that 30 to 60% of non-germinating seeds were viable, mayor en el régimen de temperatura de incubación más bajo suggesting that seeds may persist in the soil seed bank. Up to (20/10°C), lo cual quizás refleja el origen de esta especie en 60% of seeds remained viable following 1 year of storage. Stored regiones mediterraneas de lluvias invernales. Las pruebas con seeds generally exhibited higher germination rates (x– = 90%) TZ indicaron que del 30 al 60% de las semillas que no germi- than fresh seeds (x– = 67%), indicating possible dormancy and naron eran viables, sugeriendo que las semillas podrían persistir afterripening effects. Germination characteristics of this species en el banco de semillas. Hasta el 60 % de las semillas seguían are consistent with those of other invasive alien species, and viables después de un año de almacenamiento. Las semillas favor rapid population growth leading to community dominance. almacenadas generalmente mostraron proporciones de germi- nación mayores (x– = 90%) que las frescas (x– = 67%), indicando posibles efectos de la dormancia y postmaduración. Key Words: Chondrilla juncea L., Pacific Northwest, rangeland Características de germinación de esta especie son consistentes weed, cereal crop weed, invasive species, germination con las de otras especies exóticas invasoras, y favorecen un crec- imiento rápido de las poblaciones que la lleva a dominar en la The authors wish to thank Beth-Anee Johnson, Mering Hurd, Danielle Scholten, comunidad. Kristi Worth, and Fred Galen for their help in data collection. We would like to thank Dr. Michael Longnecker for assistance with statistical analysis and Dr. Thomas W. Boutton and Dr. Marshall Haferkamp for manuscript review. We thank the USDA-FS, Rocky Mountain Research Station, Shrub Sciences Laboratory, and USDI-BLM, Idaho State Office for funding, support, and use of Rush skeletonweed (Chondrilla juncea L., Asteraceae: facilities and equipment. Partial support for manuscript preparation was provided Cichorieae), an apomictic, perennial forb with Eurasian and by the Department of Rangeland Ecology and Management at Texas A&M Mediterranean origins, has been introduced to other regions of University. Manuscript accepted 5 Dec.1999. the world, including Australia, South America, and the United

544 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 States (Cullen and Groves 1977, McVean yields by 26 to 42% (Cheney et al. 1981). (Orchard) is located 32 km southeast of 1966, Panetta and Dodd 1987). In North It is one of the most serious weeds of cere- Boise, Ada Co., Ida. (116° 04' W 43° 33' America, it was first discovered in the al cropping in Australia where it has N), at an elevation of 955 m (NOAA western United States near Spokane, reduced wheat yields by as much as 80% 1993, Kitchen 1995). Mean annual tem- Wash. in 1938. By 1981, its rate of spread (Piper 1983, Sheley and Hudak 1995, perature is 10.5°C. (Fig. 1) (NOAA 1993). was approximately 40,000 ha year- 1 Sheley et al. 1999). Mean annual precipitation is 250 mm with (Cheney et al. 1981). It now dominates Current technology is largely ineffective 81% occurring between November and more than 2.5 million ha of rangeland in at curtailing the spread of rush skeleton- June (Fig. 1). The frost-free period ranges the Pacific Northwest and California weed. Although the weedy characteristics from 140 to 190 days. This site supports a (Sheley and Hudak 1995). of Australian populations have been degraded Wyoming big sagebrush Rush skeletonweed grows in dense described, a better understanding of the (Artemisia tridentata Nutt. var. w y o m i n - monocultures and displaces native plants, reproductive potential and germination gensis [Beetle & A. Young] Welsh)/mixed thereby reducing biodiversity and forage characteristics of North American popula- bunchgrass community. Soils are sandy, production for both domestic and native tions is essential if we are to identify the mixed, mesic Xeric Torriorthents with herbivores (Carroll 1980, Sheley and biological characteristics that contribute to deep, well-drained profiles. The rooting Hudak 1995). Seeds are wind-dispersed. the ecological success of this species, and zone extends to a depth of 1.5 m or more New infestations establish on coarse, well- develop more effective management meth- and available water capacity is high drained soils along roadways and on over- ods (Wapshere et al. 1974, Lee 1986, (Collett 1980) grazed rangelands, abandoned croplands, Sheley et al. 1999). The objective of this The Shrub Garden site is located near and other disturbed sites. The species study was to evaluate the effects of envi- the Boise River, about 27 km northeast of exhibits a wide range of adaptability, ronmental conditions, maturation date, and Boise, Ada Co., Ida. (116° 04' W 43° 33' occurring at elevations from less than 225 dry storage on seed germination and via- N), at an elevation of 1,033 m. Mean m to over 1,830 m that receive 250 to bility of 2 rush skeletonweed populations annual temperature is 9.2°C (Fig. 1) 1,500 mm annual precipitation (Moore in southwestern Idaho. (NOAA 1993). Mean annual precipitation 1964, Sheley and Hudak 1995). is 430 mm, and the frost-free period aver- Although rush skeletonweed is spread- ages 126 days (Fig. 1). The site supports ing primarily on rangelands, its potential Materials and Methods an antelope bitterbrush (Purshia tridentata threat to agricultural crops is also a major [Pursh] DC.)/mountain big sagebrush concern as it competes aggressively for Study Sites (Artemisia tridentata Nutt. var. v a s e y a n a light, water, and nutrients (Schirman and Populations of rush skeletonweed were [Rydb.] J. Boivin) community with a Robocker 1967, Zimdahl 1980). In some identified at 2 locations in southwestern diverse grass-forb understory (Shaw and parts of Oregon and northern Idaho, rush Idaho. The Orchard Research Site Monsen 1982). Soils are loamy-skeletal, skeletonweed has reduced annual wheat mixed, mesic Aridic Argixerolls that are moderately deep and well-drained, devel- oping from weathered granite (Collett 1980).

Seed Collection and Testing Sixty plants at each study location were randomly selected and marked for seed collection in spring 1994. Mature seeds were hand harvested at 2-week intervals from the time of earliest maturation in late July through early October in 1994. In 1995, seeds were harvested on days within the same 7-day periods as in 1994, with the exception that no collections were made in October 1995. Weeks of harvest were: late July (22 to 28), early August (6 to 12), late August (18 to 24), early September (3 to 9), late September (22 to 29), and early October (7 to 13). Seeds were pooled among plants within a loca- tion on each harvest date. Seeds were cleaned by hand-rubbing and sieving to remove the pappus and debris. Subsamples of each cleaned seed lot were used to determine seed weight, germinability, and viability. On each col- Fig. 1. Air temperature and precipitation profiles for the Orchard and Shrub Garden sites, lection date, 3 subsamples of 100 seeds 1993-1995. Climatic data were based on NOAA information from the Boise Airport from each location were weighed to esti- (Orchard) and Lucky Peak (Shrub Garden) Idaho weather stations. Long term averages mate average seed weight. Three addition- encompass 1961–1990. (NOAA 1993, 1994, 1995). al subsamples of 100 seeds from each

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 545 location were immediately placed in incu- Results bation regime (Fig. 3A) (Table 1). Seeds bation trials (1 to 2 days after collection). collected in early August exhibited lower Remaining seeds from each 1994 collec- total germination percentages than seeds Climatic Conditions tion date not used in germination trials collected on other harvest dates. Total ger- Total 1994 precipitation was 4% below were stored in sealed glass containers at mination percentage was not affected by average at Orchard and 33% below aver- room temperature in darkness for a period incubation regime in 1994. However, total age at the Shrub Garden (Fig. 1). By con- of about 1 year. Three subsamples of 100 germination percentage of seeds collected trast, 1995 was a relatively wet year. Total stored seeds from each location and col- on 1995 harvest dates was greater when precipitation was 43% above average at lection date were placed in incubation incubated at the 20/10°C compared to the Orchard and 12% above average at the about 1 year later in late August 1995 to 30/20°C regime. Shrub Garden. In 1994 and 1995, the 14- examine the effects of dry storage on seed Fresh seeds harvested in 1994 and 1995 day mean maximum air temperature prior germination characteristics. Because the to each harvest date ranged from 27 to germinated rapidly, generally reaching effects of storage on seed germination 32°C at both locations, while the mean 14- 50% germination within 2 to 8 days (Fig. were only examined for seed produced in day minimum air temperature ranged from 3B). Germination rate was affected by a a single growing season (1994), inferences 8 to16°C. significant 3-way interaction between col- are not broadly applicable to seed generat- lection year, week of harvest, and incuba- ed in another year. tion temperature (Table 1). Germination trials were conducted by Seed Weight Days to 50% germination was similar placing each subsample of 100 seeds on 2 Weight of individual rush skeletonweed for seeds harvested on the early dates in blue blotters saturated with distilled water seeds ranged from 0.1 to 0.5 mg (Fig. 2). 1994 and incubated at either 20/10°C or Seed weight was greater in 1994 than in in an 11 x 11 cm germination box. 30/20°C. In 1995, seeds collected on the 1995 for the early harvest dates (p < 0.01) Germination boxes were then incubated at late harvest dates germinated more rapidly either 20/10 or 30/20°C (12 hrs/12 hrs). (Fig. 2). In 1994, seed weight was positive- when incubated at 20/10°C compared to Seeds were exposed to light (PAR = 15 ly correlated with total precipitation 30/20°C. In both 1994 and 1995 rate of ger- µM m- 2 s- 1) during the high temperature received during the 4-week interval prior to alteration. Germination counts were made each harvest date (R2 = 0.75, p < 0.05) and mination was more rapid for seeds collect- at 2-day intervals for 14 days. with the mean air temperature during the 2- ed later in the season and incubated at Germination boxes were rearranged ran- week period preceding each harvest date 20/10°C. There were no significant rela- domly in the incubation chambers after (R 2 = 0.98, p <0.05). No significant correla- tionships between seed weight and total each count. Seeds were considered germi- tions between seed weight and precipitation germination percentage or germination rate. nated when the radicle had emerged and or air temperature were found in 1995. Viability of nongerminating seeds dif- the cotyledons were green and spreading. fered between years (Table 1). In 1994 less than 10% of the nongerminating seed were After 14 days, non-germinated seeds were Germination of Fresh 1994 and 1995 viable following incubation, while in 1995 tested for viability using a 1% solution of Seeds more than 30% were viable (Fig. 3C). 2,3,5-triphenyl tetrazolium chloride Germination of fresh seeds varied with (Moore 1973). Maximum germination was year of collection, harvest date, and incu- calculated as number of seeds germinated divided by the number of germinated seeds plus the number of viable nongermi- nated seeds determined by TZ testing. Rate of germination was expressed as days to 50% of 14-day germination.

Statistical Analyses Experimental design was a randomized, complete block with site as the blocking factor. An ANOVA was used to evaluate differences in germination, germination rate, and viability of nongerminated fresh and stored seeds attributable to the main effects of year of collection, harvest date, storage treatment, and incubation regime and all possible interactions (SAS Institute 1988). Where significant interactions pre- cluded direct interpretation of main effects, pairwise comparisons were made using the Bonferroni t-test (Milliken and Johnson 1992). All differences reported were significant at p £ 0.05. Regression analyses were used to examine the rela- tionship between seed weight and environ- Fig. 2. Mean weight of Chondrilla juncea L. seed (mg) collected on 1994 and 1995 harvest mental conditions preceding each harvest dates from the Orchard and Shrub Garden sites in southwestern Idaho. date, germination, and rate of germination.

546 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Table 1. Degrees of freedom (df), MS values, and probability levels (p) for ANOVA models used both temperatures tested. Most seeds were to statistically analyze total germination, days to 50% germination, and viability [TZ] of capable of immediate germination without nongerminating fresh rush skeletonweed seed from 1994 and 1995 harvest dates and for fresh pretreatment to relieve dormancy. In both and stored seed from 1994 harvest dates in southwestern Idaho. Seeds were stored in sealed containers at room temperature from harvest to August 1995. years, seeds reached total germination per- centages in excess of 60% regardless of Days to 50% incubation regime. Total Germination Germination Viability Grant-Lipp (1966) reported that rush Model df MS p MS p MS p skeletonweed seeds from Canberra, ------fresh 1994 vs fresh 1995 seed ------Australia reached 90% germination with- Collection Site (Block) 1 167.6 * 5.9 ns 562.6 * in 24 hours when incubated at tempera- Year of Collection (Year) 1 1879.9 *** 18.7 * 8444.5 ** * tures between 18 and 30°C. Germination Week of Collection (Date) 4 1025.0 *** 83.7 *** 117.8 ns was complete within 48 hours. Slower Germination Temperature (Temp) 1 894.3 *** 6.4 ns 101.2 ns rates reported in this study were attributed Year*Date 4 452.4 *** 12.3 ** 105.0 ns to differences in criteria used to define Year*Temp 1 980.1 *** 28.9 *** 160.6 ns successful germination. Radicle emer- Date*Temp 4 105.3 ns 0.9 ns 73.7 ns gence was used by Grant-Lipp; criteria Year*Date*Temp 4 134.4 * 6.1 * 25.4 ns used for this study were believed to pro------fresh 1994 vs stored 1994 seed ------vide a more realistic index of seedling sur- Collection Site (Block) 1 121.8 ns 1.6 ns 266.0 ns vival in the field. Week of Collection (Date) 5 480.8 * 31.1 *** 198.7 ns Even in the drier year (1994) when pre- Germination Temperature (Temp) 1 154.8 ns 0.6 ns 0.0 ns cipitation was considerably less than the Storage 1 150.7 ns 60.7 *** 351.4 ns 30 year average, rush skeletonweed pro- Date*Temp 5 31.2 ns 0.7 ns 198.2 ns duced seeds that were highly germinable, Date*Storage 5 229.2 ns 24.4 *** 210.7 ns reaching total germination percentages Temp*Storage 1 196.8 ns 22.3 * 10.6 ns generally in excess of 80% (Fig. 3A). Date*Temp*Storage 5 28.8 ns 3.9 ns 159.6 ns These seeds also exhibited more rapid ger- * p < 0.05 mination than seeds from 1995. Studies ** p < 0.01 have found that when subjected to mild ***p < 0.001 drought stress, some plant species display higher germinability (Hilhorst and Toorop 1997). Dodd and Panetta (1987) found that Germination of Fresh and Stored Discussion and Conclusions a large numbers of seeds are produced 1994 Seeds even under drought conditions. Rush Total germination percentage of seeds Many weedy species of Eurasian and skeletonweed may be producing viable harvested in 1994 and incubated immedi- Mediterranean origins are well-adapted to seeds even under hot, dry conditions ately or stored in sealed containers at room the climatic conditions of the interior because flowering and seed production are temperature until August 1995 varied with western United States (Mack 1986). These independent of summer rainfall (Panetta harvest date (Table 1). For stored seeds, invasive plants occupy broad niches and and Dodd 1987). Plants acquire soil water those collected mid-season exhibited exhibit a large degree of phenotypic plas- through deep roots that may penetrate sev- greater total germination percentages than ticity (Baker 1986, Bazazz 1986). They eral meters into the soil profile (Panetta collections harvested at earlier or later col- are generally capable of flowering and set- and Dodd 1984, 1987). Elevated tempera- lection periods (Fig. 3D). Stored seeds ting seed under a wide range of environ- tures during the seed production period in were as germinable as fresh seed, general- mental conditions (Pimentel 1986). 1994 may have increased the germinabili- ly exceeding 80% germination following 1 The adaptability of rush skeletonweed to ty of rush skeletonweed seeds in this year of dry storage. climatic conditions in areas of the western study. Days to 50% germination for 1994 seeds United States has been demonstrated by its In both years, across all dates, there varied with storage treatment and harvest rapid expansion. In addition, it shares appeared to be a trend toward greater ger- date (Table 1). Stored seeds generally ger- many characteristics of highly successful mination rates and faster germination at minated more rapidly than fresh seeds invader species. Its apomictic habit elimi- the lower incubation temperature regime with the exception of seeds collected at the nates dependence on insect pollinators and (Fig. 3). Germination of rush skeletonweed later harvest dates and incubated at promotes genetic stability. Plants are capa- seeds was generally rapid in both incuba- 30/20°C ( Fig. 3E). Days to 50% germina- ble of establishing, reproducing, and dis- tion regimes. Plants reached 50% of total tion also varied with storage treatment and persing seeds in areas where environmen- germination within 2 to 8 days for most incubation regime (Table 1). The germina- tal factors may be limiting for plant dates in both 1994 and 1995. For seeds tion rate was similar for fresh or stored growth and development (Duke 1985, produced in 1995, germination was slower seeds when incubated at 30/20°C. Bazazz 1986). for seeds incubated at the higher tempera- However, stored seeds germinated more The southwestern Idaho rush skeleton- ture regime. This trend may reflect origins rapidly than fresh seeds when incubated at weed populations flowered and produced of rush skeletonweed in Mediterranean 20/10°C. The viability of nongerminating seeds continuously from July to October winter rainfall regions where cool-season seeds was similar for fresh and stored during the driest and warmest portion of germination may ensure access to winter 1994 seed collections averaging 13% the season. Despite the environmental lim- precipitation. The ability to germinate dur- (Table 1). itations, rush skeletonweed produced ing the cooler portions of the growing sea- viable seeds that germinated rapidly at son would be advantageous throughout

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 547 (Grant-Lipp 1966, McVean 1966, Cuthbertson 1970, Panetta and Dodd 1987). Cuthbertson (1970) stated that there appeared to be a short-lived dorman- cy in some field samples, while others suggest that rush skeletonweed seeds lack primary dormancy (Grant-Lipp 1966, Coleman-Harrell et al. 1979). Rush skele- tonweed seeds may be exhibiting germina- tion polymorphism, a common character- istic of successful weeds. It may be a sur- vival strategy whereby the range of condi- tions favorable for germination varies among seeds. This mechanism permits some seeds to germinate immediately whereas others may germinate at a later date (Pimentel 1986). A small percentage of non-germinating stored seeds, especially seeds produced later in the season, were still viable, indi- cating that rush skeletonweed seeds may acquire a secondary dormancy which could potentially enable some seeds to persist in the soil seed bank. Lonsdale et al. (1988) stated that persistence of seeds in soil seed banks is a critical component for maintaining populations of weedy species. Grant-Lipp (1966) indicated that rush skeletonweed seeds may acquire a secondary dormancy when exposed to high temperatures. Other studies have found that in some plant species, dispersed seeds are readily germinable immediately after shedding, if conditions are favorable. However, when conditions for germina- tion are not met, seeds may acquire a sec- ondary dormancy (Hilhorst and Toorop 1997). Subjecting rush skeletonweed seeds collected in 1994 to storage conditions may have induced a secondary dormancy Fig. 3. Germination, germination rate, and viability of non-germinated rush skeletonweed in some seeds. The findings in this study seeds immediately following 1994 and 1995 harvest dates from the Orchard and Shrub suggest that a portion of rush skeleton- Garden sites in southwestern Idaho. The legend in figure 3E applies to the entire graph. weed seeds may have a primary dormancy Figures 3A, B, and C are results for fresh seed incubated in 1994 and 1995. The effects of and that others may be acquiring a sec- storage on germination were evaluated for seed collected in 1994 only. Figures 3D, E, and ondary dormancy, enabling seeds to per- F represent results of stored seed collected in 1994 and germinated in August 1995. sist in the soil. In addition to germination polymor- much of the northwestern United States have a detrimental effect on germination phism, rush skeletonweed seeds also where most precipitation occurs during the of rush skeletonweed seeds. Cuthbertson showed seed polymorphism in producing winter. Conversely, lower germination at (1970) reported that germination percent- seeds of different weights. Although seed higher temperatures may reduce the risk of ages of mature rush skeletonweed seeds weight is positively correlated with ger- germination during the warmest, driest remained unchanged after 12 months of minability in many plant species (Saner et portion of the season when soil moisture dry storage. al. 1995), this relationship was not evident may be insufficient for seedling survival, The increased germination total and rate with rush skeletonweed in this study. establishment, and growth (Baskin and for stored 1994 seed indicates that some Seed weight was positively correlated Baskin 1985). fresh seeds may indeed possess a primary with both temperature and rainfall during About 10% of non-germinating seeds dormancy. Seeds with primary dormancy the growing season in 1994, but not in from 1994 and more than 30% of non- may undergo an afterripening process dur- 1995. Because 1995 was a wetter year, germinating seeds from 1995 were viable. ing storage that releases dormancy and environmental conditions may have been Viability of non-germinating seeds may be results in increased germination rates and conducive to greater seedling establish- an indication of some seeds possessing a percentages under a broader range of incu- ment in the earlier portions of the growing primary dormancy. Germination results bation conditions (Beckstead et al. 1995). season. Rush skeletonweed density was from seeds collected in 1994 and stored Primary dormancy in rush skeletonweed obviously higher in 1995 than in 1994 for 1 year indicated that storage did not seeds ranges from 0 to 40% among studies

548 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 (personal observation). We speculate that Coleman-Harrell, M., D. Ehrensing, G. Lee, NOAA. 1995. Climatological data annual sum- plants establishing early in the 1995 grow- W. Belles, D. Issacson, and R. Schirman. mary Idaho. Vol. 98. National Oceanic and ing season may have allocated more car- 1979. Rush skeletonweed. Univ. Idaho, Atmospheric Administration, Asheville, N.C. bon to growth while conditions were College Agr., Cooperative Ext. Serv., Curr. Panetta, F.D. and J. Dodd. 1984. favorable. However, those plants may Information Ser. 468. Moscow, Ida. Skeletonweed: how serious a threat in Collett, R.A. 1980. Soil survey of Ada County Western Australia? West. Aust. J. Agr. have experienced greater intraspecific area, Idaho. USDA-SCS, Washington, D.C. 25:37–41. competition in the latter portion of the Cullen, J.M. and R.H. Groves. 1977. T h e Panetta, F.D. and J. Dodd. 1987. The biology growing season when hot, dry conditions population biology of Chondrilla juncea i n of Australian weeds 16. Chondrilla juncea L. ensued. This may have in turn reduced the Australia. Proc. Ecol. Soc. Aust. 10:121–134. J. Aust. Inst. Agr. Sci. 53:83–95. amount of carbon available for plants to Cuthbertson, E.G. 1970. Chondrilla juncea in Pimentel, D. 1986. Biological invasions of allocate to seed production. Australia 3. Seed maturity and other factors plants and animals in agriculture and In conclusion, we found that germina- affecting germination and establishment. forestry, p.149-160. I n: H.A. Mooney and tion characteristics of rush skeletonweed Aust. J. Exp. Agr. Anim. Husb. 10:62–66. J.A. Drake (eds.), Vol. 58. Ecology of bio- were consistent with those of other suc- Dodd, J. and F.D. Panetta. 1987. Seed pro- logical invasions of North America and duction by skeletonweed in Western Hawaii. Springer-Verlag N.Y., Inc., N.Y. cessful alien invaders. Rapid germination, Australia in relation to summer drought. Piper, G.L. 1983. Rush skeletonweed. Weeds without any afterripening requirements, Aust. J. Agr. Res. 38:689–705. Today 14:5–7. favors rapid explosive population growth Duke, S.O. 1985. Weed physiology. Vol. I. Saner, M.A., D.R. Clements, M.R. Hall, D.J. and allows rush skeletonweed to spread, Reproduction and ecophysiology. CRC Doohan, and C.W. Crompton. 1995. T h e persist, and dominate disturbed communi- Press, Inc. Boca Raton, Fla. biology of Canadian weeds. 105. Linaria vul - ties. Secondary dormancy may allow some Grant-Lipp, A.E. 1966. Some properties of garis Mill. Can. J. Plant Sci. 75:525–537. skeletonweed seed to persist in the soil the seeds of skeletonweed (C h o n d r i l l a SAS Institute. 1988. SAS/STAT User’s guide and possibly reestablish a stand if unfavor- j u n c e a L.). CSIRO Division of Plant (Release 6.03). SAS Institute, Inc., Cary, able environmental conditions or manage- Industry. 5:17–24. Canberra, Aust. N.C. ment activities eliminate an initial popula- Hilhorst, H.W.M. and P.E. Toorop. 1997. Schirman, R. and W.C. Robocker. 1967. Review on dormancy, germinability, and ger- Rush skeletonweed-threat to dryland agricul- tion. Due to these characteristics, the con- mination in crop and weed seeds. Adv. ture. Weeds 15: 310–312. trol or eradication of rush skeletonweed in Agron. 61:111–165. Shaw, N.L. and S.B. Monsen. 1982. most infested areas may prove difficult. Kitchen, S.G. 1995. Return of the native: a Phenology and growth habits of nine ante- look at select accessions of North American lope bitterbrush, desert bitterbrush, stansbury Lewis flax, p. 321-326. I n: B.A. Roundy, cliffrose, and Apache-plume accessions, Literature Cited E.D. McArthur, J.S. Haley, and D.K. Mann p.55–69. In:USDA For. Serv. INT-GTR-152. (compilers), Proc. Wildland shrub and arid Ogden, Ut. land restoration symposium. USDA For. Sheley, R.L. and J.M. Hudak. 1995. R u s h Baker, H.G. 1986. Patterns of plant invasion in Serv. INT-GTR-315. Ogden, Ut. skeletonweed: a threat to Montana’s agricul- North America, p. 44-55. I n: H.A. Mooney Lee, G.A. 1986. Integrated control of rush ture. EB 132. Mont. State Univ., Bozeman, and J.A. Drake (eds.), Ecology of biological skeletonweed in the western U.S. Weed Sci. Mont. invasions of North America and Hawaii. Vol. 34:2–6. Sheley, R.L., J.M. Hudak, and R.T. Grubb. 58. Springer-Verlag, N.Y., Inc., N.Y. Lonsdale, W.M., K.L.S. Harley, and J.D. 1999. Rush skeletonweed. I n: R.L. Sheley Baskin, J.M. and C.C. Baskin. 1985. T h e Gillett. 1988. Seed bank dynamics in and J.K. Petroff (eds.), Biology and manage- annual dormancy cycle in buried weed seeds: Mimosa pigra, an invasive tropical shrub. J. ment of noxious rangeland weeds. Ore. State a continuum. Bioscience 35:492–498. Appl. Ecol. 25:963–976. Univ. Press, Corvallis, Ore. Bazazz, F.A. 1986. Life history of colonizing Mack, R.N. 1986. Alien plant invasion into the Wapshere, A.J., S. Hasan, W.K. Wahba, and plants: some demographic, genetic, and Intermountain West; a case history, p.191- L. Caresche. 1974. The ecology of physiological features, p. 96–108. I n: H.A. 210. In: H.A. Mooney and J.A. Drake (eds.), Chondrilla juncea in the western Mooney and J.A. Drake (eds.), Ecology of Vol. 58. Ecology of biological invasions of Mediterranean. J. Appl. Ecol. 11: 783–800. biological invasions of North America and North America and Hawaii. Springer-Verlag, Zimdahl, R.L. 1980. Weed crop competition. Hawaii. Vol. 58. Springer-Verlag, N.Y., Inc., New York, Inc., New York. Int. Plant Protection Center, Corvallis, Ore. N.Y. McVean, D.N. 1966. Ecology of C h o n d r i l l a Beckstead, J., S.E. Meyer, and P.S. Allen. juncea L. in south-eastern Australia. J. Ecol. 1 9 9 5 . Effects of afterripening on cheatgrass 54:345–365. (Bromus tectorum) and squirreltail (E l y m u s Milliken, G.A. and D.T. Johnson 1992. elymoides) germination, p. 165–172. In: B.A. Analysis of messy data. Vol. 1: Designed Roundy, E.D. McArthur, J.S. Haley, and experiments. Chapman and Hall, N.Y. D.K. Mann (compilers), Proc. Wildland Moore, R.M. 1964. Chondrilla juncea L . shrub and arid land restoration symposium. (skeleton weed) in Australia. Proc. 7th USDA For. Serv. INT-GTR-315. Ogden, Ut. British Weed Control Conference, p. Carroll, P. 1980. Wiry skeletonweed threatens 563–568. Oregon agriculture, rangelands. Rangelands Moore, R.P. 1973. Tetrazolium staining for 2:21. assessing seed quality. I n: W. Heydecker Cheney, T.M., G.L. Piper, G.A. Lee, W.F. (ed.), Seed Ecology. Butterworths, London. Barr, D.C. Thill, R.B. Hawkes, R.F. Line, NOAA. 1993. Climatological data annual sum- R.R. Old, L.L. Craft Jr., and E.B. Adams. mary Idaho Vol. 96. National Oceanic and 1981. Rush skeleton weed biology and con- Atmospheric Administration, Asheville, N.C. trol in the Pacific Northwest. Univ. Idaho, NOAA. 1994. Climatological data annual sum- Coll. Agr., Cooperative Ext. Serv., Curr. mary Idaho. Vol. 97. National Oceanic and Information Ser. 585. Moscow, Ida. Atmospheric Administration, Asheville, N.C.

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 549 J. Range Manage. 53: 550–555 September 2000 Seed production in sideoats grama populations with differ- ent grazing histories

STEVEN E. SMITH, REBECCA MOSHER, AND DEBRA FENDENHEIM

Authors are associate professor, School of Renewable Natural Resources, University of Arizona, Tucson, Ariz. 85721; graduate student, University Program in Genetics, Duke University, Durham, N.C. 27706; and research specialist, School of Renewable Natural Resources, University of Arizona, Tucson, Ariz. 85721.

Abstract Resumen

Frequent and intense defoliation of grasses has been associated La frecuente e intensa defoliación de los zacates ha sido asocia- with the evolution of “grazing morphotypes” that exhibit a vari- da con la evolución de "morfotipos de apacentamiento" que pre- ety of vegetative traits correlated with improved grazing resis- sentan una variedad de características vegetativas correla- tance. While recovery from a seed bank is not considered an cionadas con una mejor resistencia al apacentamiento. Mientras important grazing resistance mechanism, relatively little is actu- que la recuperación a partir del banco de semillas no se consid- ally known regarding seed (caryopsis) production in grazing era un mecanismo importante de resistencia al apacentamiento, morphotypes of caespitose grasses. The goal of this research was actualmente poco se sabe respecto a la producción de semilla to compare components of seed production in 2 populations of (cariópside) de los "morfotipos de apacentamiento" de zacates sideoats grama (Bouteloua curtipendula var. caespitosa Gould & cespitosos. La meta de este estudio fue comparar los compo- Kapadia) from nearby sites with different histories of livestock nentes de la producción de semilla de dos poblaciones de grazing. This was done using vegetative propagules of genotypes "Sideoat grama" (Bouteloua curtipendula var. caespitosa Gould y from both populations in a greenhouse study. The study was con- Kapadia) de sitios cercanos con diferentes historiales de apacen- ducted in 2 flowering seasons under conditions considered favor- tamiento por ganado. Esto se realizó a través de en un estudio de able for seed production. The population exposed to livestock invernadero utilizando propágulos vegetativos de genotipos de grazing showed a genetically based decrease in seed production ambas poblaciones. El estudio se condujo durante 2 estaciones de relative to the ungrazed population. Lower seed production per floración bajo condiciones consideradas favorables para la pro- plant in the grazed population was at least partially due to ducción de semilla. La población expuesta al apacentamiento por reduced numbers of tillers and panicles per plant and spikes per ganado mostró una disminución (genéticamente basada) en la panicle that may be associated with selection for grazing toler- producción de semilla en relación a la población sin apacen- ance. The grazed population also exhibited lower average seed tamiento. La baja producción de semilla por planta de la production per spike indicating lower inherent floral fertility. población con apacentamiento se debió parcialmente a el reduci- Seed production was not closely correlated with vegetative traits do número de hijuelos y panículas por planta y espigas por associated with increased grazing tolerance, nor was there evi- panícula, lo que puede estar asociado con la selección por toler- dence of obvious physiological trade-offs related to decreased ancia al apacentamiento. La población apacentada también seed production in the grazed population. Lower seed production exhibió un promedio menor de producción de semilla por espiga, potential in populations of sideoats grama intensively grazed by indicando una menor fertilidad floral inherente. La producción livestock may lead to reduced potential for seedling colonization. de semilla no se correlaciono estrechamente con las característi- cas vegetativas asociadas con el aumento en la tolerancia al apacentamiento, ni hubo evidencia obvia de cambios fisiológicos Key Words: B o u t e l o u a, defoliation tolerance, evolution, grazing relacionados con la disminución de producción de semilla de la resistance población apacentada. La potencial baja producción de semilla de poblaciones de "Sideoat grama" apacentadas intensivamente puede conducir a una reducción potencial de colonización por Grasses that are exposed to herbivores often possess unique life plántulas. history, morphological and physiological traits (Carmen and Briske 1985, Briske and Richards 1995, Briske 1996). When defoliation has been frequent over long periods of time, these 1983, Aarssen and Turkington 1985, Carmen and Briske 1985, traits may be associated with the evolution of genetically based Painter et al. 1989, 1993, Smith 1998 ). Defoliation tolerance in defoliation tolerance and the emergence of “grazing morpho- caespitose grasses is commonly correlated with a series of vege- types” in the plant populations affected (e.g., Detling and Painter tative traits including reduced plant height, a less upright growth form, fewer and smaller leaves, and increased tiller number Research was supported by the University of Arizona Foundation and the (Briske 1996). Arizona Agricultural Experiment Station. Authors wish to thank Lyman Nyquist Previous research (Smith 1998) compared vegetative responses for assistance in statistical analysis, Virgil Mercer for access to collection sites and under varying defoliation regimes of genotypes from 2 popula- information on their history, and Sharon Biedenbender, Mitchel McClaran, Bruce Munda, Mark Pater, and David Williams for technical assistance. tions of sideoats grama (Bouteloua curtipendula var. caespitosa Manuscript accepted 1 Jan. 2000. Gould & Kapadia) with different livestock grazing histories. In

550 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 these studies, a population with long-term Materials and Methods that population and were isolated from the exposure to cattle grazing (approximately other main plot in that replication or from 100 yr) exhibited traits associated with other replications by at least 80 cm. This research utilized genotypes from 2 improved defoliation tolerance (primarily Mean daily temperature in the green- populations of sideoats grama from the increased tiller survival and reduced tiller house during the study was 32.7°C (mean semidesert grassland biotic community in mass) compared to a population that was low of 27.7°C, mean high of 35.6°C). southeastern Arizona (Brown 1994). One not exposed to grazing by livestock. Fans were used to circulate pollen during population (“ungrazed”) occurs atop a Evidence of reduced genetic variation anthesis. Ramets were irrigated individual- steep-walled butte (32° 42' N, 110°, 29' W within the grazed population was also ly when leaf curling and color indicated 1,655 m asl) near Mammoth, Ariz. that noted for various vegetative traits associat- drought stress. Soil moisture content was probably has never been grazed by large ed with defoliation tolerance. While the ~ 20 to 30% of field capacity at this point. (>100 kg) herbivores, including most environments the 2 populations inhabited Ramets were allowed to grow from June domestic livestock (Hadley et al. 1991). differed in ways other than the presence of through October in 1997 and 1998, a peri- The other (“grazed”) population occurs on livestock, the response of genotypes from od exceeding the entire flowering season a site (1,338 m asl) that has been regularly both populations was consistent with the for sideoats grama in southern Arizona. grazed by cattle since before 1900 (V. evolution of increased grazing tolerance in Inflorescences in sideoats grama are pani- Mercer, Mammoth, Ariz., personal com- the population from the grazed site. cles with up to 50 short, pendulous spikes. munication) that is located 4.8 km from the Increased seed production and expan- As panicles matured, at least 3 were har- ungrazed population. Deer (O d o c o i l e u s sion of a soil seed bank may represent an vested individually from each ramet in spp.) and small mammalian herbivores are additional defoliation resistance mecha- each main plot. Spikes were counted for able to reach both sites and may graze on nism (Briske 1996), but the effects of live- each panicle and individual seeds per pan- s i d e o a t s grama (Krausman et al. 1997). stock grazing on the evolution of this icle determined following threshing. Mean Prairie dogs (C y n o m y s spp.) occurred in resistance mechanism are not well charac- seed mass was determined from 50-seed southeastern Arizona (Hoffmeister 1986) terized. Reduced production of reproduc- samples dried at room temperature. The and may have also grazed on both sites tive tillers has been observed in popula- mean number of spikelets per spike was until about the late 1930's. No description tions of perennial grasses derived from determined by dissecting 5 spikes from of fire history or precipitation is available sites exposed to contrasting grazing each ramet. Spikelets were assumed to for either site. Both sites are on 20–35% regimes (e.g., Trlica and Orodho 1989, contain 1 floret (Cronquist et al. 1977). slopes with numerous rock outcrops and Painter et al. 1989, 1993). Nevertheless, Tiller number was recorded for each ramet shallow Haplargid soils (D. Robinett, little information is available regarding 8 weeks after defoliation in a regrowth USDA-NRCS, Tucson, Ariz., personal any differences in seed (caryopsis) pro- cycle that began 5 June 1998. Tillers that communication). Sideoats grama occurs duction between such populations. Scott showed any sign of a mature spike were sporadically in the area between the 2 and Whalley (1984) found that the mean counted as reproductive. sites, most commonly in drainages on the number of seeds produced per inflores- Some of the genotypes used in this slopes of the butte, and this may permit cence was significantly lower in genotypes study were also included in a greenhouse gene flow between the populations. from populations of 2 species of study that evaluated vegetative response to Twenty randomly selected genotypes Danthonia that had been exposed to heavy different defoliation frequencies (Smith (basal area >15 cm) were dug from along grazing by sheep compared to relatively 1998). Seed production per spike in ramets a 400-m transect at each site in 1994. less grazed populations. However, protec- from these genotypes (ungrazed: N = 10, Individual genotypes were separated by at tion from livestock grazing for more than grazed: N = 9) was rank correlated (Sokal least 3 m and were assumed to represent 50 years did not result in altered seed pro- and Rohlf 1995) with various traits associ- the products of separate established duction in Indian ricegrass (A c h n a t h e r u m ated with vegetative growth from the pre- seedlings. Genotypes from both popula- hymenoides (Roemer & J.A. Schultes) vious study where defoliation occurred tions are high polyploids with 2n = 75–92 Barkworth) genotypes taken from grazed every 14, 28, or 56 days over a 168-day (D. Showalter, University of Arizona, or ungrazed sites when these genotypes period. These traits were: plant height, unpublished data, Gould and Kapadia were grown in a common environment root dry weight, shoot dry weight, total 1964). These genotypes were grown in 4- (Orodho et al. 1998). A lack of additional plant dry weight, and tiller number and liter pots in a shadehouse in Tucson, Ariz. data on seed reproduction and the evolu- dry weight. for 6 mo before a minimum of 6 ramets tion of grazing morphotypes prevents a Mixed-model analysis of variance was more integrated understanding of demo- were established for each genotype from used to compare traits among genotypes of graphic processes that may be affected by single rooted tillers (Smith 1998). the 2 populations. Proc MIXED of SAS the presence of grazing animals. Individual ramets were grown in 4-liter (Littell et al. 1996) was used with replica- The primary objective of this research pots in an evaporatively cooled green- tions, years, and genotypes within popula- was to determine whether exposure to dif- house and treated uniformly until June tions considered random effects. For com- ferent livestock grazing regimes is associ- 1997 with defoliation 5 cm above the soil parisons of the components of seed pro- ated with genetically based changes in seed surface every 6 to 12 weeks. The resulting duction (Table 1), significance of differ- production in sideoats grama. This was ramets were placed on greenhouse bench- ences between populations were assessed done under conditions of minimal defolia- es using a modified split-plot arrangement using single-degree of freedom contrasts. tion stress using genotypes from 2 popula- with 4 replications with population as the Least-squares means are reported through- tions with contrasting livestock grazing main plot and all 20 genotypes from each out for these traits. Distribution of mean histories (Smith 1998). Secondary objec- population placed randomly as sub plots seed production per spike in the 2 popula- tives were to determine whether variation within the main plots. Ramets within each tions was evaluated for each year separate- in seed production was associated with main plot were grouped together within an ly using frequency histograms and two- other traits related to defoliation tolerance. area of 1 m2 to concentrate pollen from

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 551 Table 1. Least squares means for components of seed production by structural unit in ungrazed and grazed populations of sideoats grama.

Structural unit Population Spike Panicle Plant1 (Seeds (Seeds (Spikelets (Seeds (Spikes (Tillers (% (Panicles (Spikes (Seeds spikelet-1) spike-1) spike-1) panicle-1) panicle-1) ramet-1) reproductive ramet-1) ramet-1) ramet-1 tillers) Ungrazed 0.050* 0.200* 3.37 6.3* 31.3 33.7* 40.9* 9.2* 284* 55.8* Grazed 0.013* 0.140* 3.75 4.1* 28.4 29.2* 37.6* 6.8* 197* 35.1* 1Estimated from tiller counts made in 1998. *Population means significantly different by single degree of freedom contrast (P £ 0.05). tailed Fisher’s exact tests. The same the average number of spikelets per spike, in the grazed population (23.6 + 4.8 mg). approach was used to compare the number the number of seeds per spike may be con- Still, the mass of seeds produced by each of spikes per panicle but data from both sidered as the primary component explain- ramet would represent less than 2% of the years were combined since means were ing differences in floral fertility between aboveground biomass productions of these not significantly different between years the populations. This resulted in signifi- ramets based on results for the 2 popula- for this trait. Relationships between com- cantly higher mean seed production per tions from previous research (Smith ponents of seed output could be evaluated panicle in the ungrazed population. Using 1998). using rank correlation coefficients which tiller count data, it was possible to esti- Over all genotypes in both populations, do not assume a linear relationship mate additional components of seed output seed production per spike was significant- between the 2 variables examined (Sokal on a whole-plant basis. This showed that ly correlated between years (rs = 0.66, P £ and Rohlf 1995). The numbers of seeds in the absence of significant defoliation, 0.05) and the genotype x year, and popula- produced per spike and spikes produced ramets from the ungrazed population tion x year interactions were not signifi- per panicle were also rank correlated with would generally be expected to produce cant. However, the mean number of seeds various vegetative traits associated with significantly more tillers overall and produced per spike was significantly response to defoliation using data from a reproductive tillers, and therefore more greater in 1998 (0.26 + 0.02) than in 1997 previous study (Smith 1998) for genotypes panicles and spikes than ramets from the (0.12 + 0.01) based on a Wilcoxon signed that were included in that study and the grazed population (Table 1). Combination rank test. Analysis of frequency his- present one. Conservative rank correlation of tiller data with that from individual tograms of seeds produced per spike over procedures were also used in this case, spikes and panicles further revealed that genotypes (Fig. 1) in each year also sug- again because linear relationships could significantly more seeds would be expect- gested differing distributions of this vari- not be assumed but also because data orig- ed per ramet under these conditions in an able between the populations. This led to inated from separate experiments for some entire flowering season within the separate analyses of the distribution of pairs of variables. ungrazed population than within the seed production for the 2 years. These dis- Genetic analysis was done using Box- grazed population. tributions differed significantly between Cox transformed values for the number of There was no significant difference the 2 populations in both years with seed seeds per spike (Sokal and Rohlf 1995) and between the populations in average seed production in the ungrazed population untransformed values for the number of mass (ungrazed: 0.62 + 0.05 mg [least- consistently skewed toward higher values spikes per panicle. Broad-sense heritability squares mean + SE], grazed: 0.70 + 0 . 0 8 (Fig. 1). In 1997, the median number of was estimated for both traits within each mg). Because ramets from the ungrazed seeds produced per spike was 0.125 in the population using a split-plot in time design population would produce more seeds, the ungrazed population compared to 0.064 after Nyquist (1991). When the estimates mean total mass of seeds produced per seeds per spike in the grazed population. were significantly greater than 0, standard ramet was significantly higher in the In 1998, medians for the 2 populations errors were calculated using the method of ungrazed population (34.2 + 5.7 mg) than were 0.320 and 0.091 seeds per spike. The Gordon (1979). Phenotypic correlations were utilized throughout since the array of genotypes and experimental design utilized did not permit accurate estimation of addi- tive genetic variance which would be required for calculation of genetic correla- tions (Lynch and Walsh 1998). Statistical significance was assigned at P £ 0.05 in all ca s e s .

Results

Ramets from the ungrazed population produced significantly more seeds per spikelet and spike overall than did those Fig. 1. Distribution of mean seed production per spike for genotypes from ungrazed and from the grazed population (Table 1). grazed populations of sideoats grama. Population distributions significantly different (P Since the 2 populations did not differ in 0.05) based on Fisher’s exact test for experiments conducted in 1997 and 1998.

552 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 mean number of spikes per panicle did not 35.8 + 11.3%). Broad- differ between years, nor was there a sig- sense heritability of the nificant population x year interaction for number of spikes per pan- this trait. The distributions of this trait also icle was less than 2% and did not differ significantly between popu- did not differ significantly lations (Fig. 2). Medians for the ungrazed from 0 in both popula- and grazed populations were 31.6 and 28.6 tions. spikes per panicle, respectively. The number of seeds per spike was sig- nificantly and negatively correlated with Discussion and the number of tillers per ramet in the Conclusions ungrazed population as measured in this study (Table 2). The number of spikes per As was the case with panicle was also positively correlated with vegetative traits (Smith tiller number in this population. Overall 1998), this research shows seed production per ramet was positively that 2 sideoats grama pop- associated with the number of seeds pro- ulations from nearby sites duced per spike in both populations. In the with contrasting livestock Fig. 2. Distribution of mean spike production of spikes per grazed population, the number of seeds grazing histories differed panicle for genotypes from ungrazed and grazed popula- per ramet was also positively correlated in reproductive character- tions of sideoats grama. Population distributions are not sig- with both the number of tillers and the istics that may be associ- nificantly different (P > 0.05) based on Fisher’s exact test. number of spikes produced per ramet. This ated with plant demogra- would be expected given the lower mean phy (Table 1, Fig. 1). While the grazed given that sideoats grama occurs in the seed production per spike in this popula- population displayed evidence of selection area between the sites, that many native tion. The average mass of an individual for vegetative traits associated with herbivores are able to reach both sites, and seed was not significantly correlated with increased grazing tolerance in previous that genotypes from the grazed population any other primary reproductive measures research, the current study indicates that exhibit vegetative characteristics that are in either population. this population also exhibits decreased consistent with selection for tolerance to Combining data from this study with average seed production potential per defoliation (Briske 1996), it appears rea- those from Smith (1998), the number of plant. Observation of significant broad- sonable to speculate that exposure to live- seeds produced per spike was significantly sense heritability for the number of seeds stock may have affected reproductive correlated only with mean tiller weight per spike in both populations demonstrates characteristics of the grazed population. with defoliation every 28 days (rs = 0.52, a genetic basis for the variation among As in this research, Scott and Whalley P £ 0.05). No other significant phenotypic genotypes within these populations. (1984) found that plants from populations correlations were observed between any Because the 2 populations occupy envi- of Danthonia linkii Kunth and D. richard - measures of reproductive output and vege- ronments that may differ in ways other sonii Cashmore that had been exposed to tative growth under different defoliation than their livestock grazing history, it is intense grazing by sheep produced fewer regimes for the 19 genotypes common to not possible to definitely conclude that seeds per panicle on average when grown both studies. differences in reproductive biology are without defoliation stress compared to Overall variability in the number of solely the result of exposure to livestock plants from populations from sites with seeds per spike appeared greater in the grazing in 1 population. For example, fire lower grazing stress. These results are also grazed population than in the ungrazed history and weather may differ at the 2 consistent with the observation that population (CV = 84 vs 121%). Broad- sites and could explain the differences in increased seed production to expand the sense heritability of the number of seeds vegetative traits (Smith 1998) and repro- seed bank is apparently not a common per spike over years was significantly ductive biology between the populations. mechanism of defoliation resistance in greater than 0 in both populations, but did The 2 populations may have also diverged perennial caespitose grasses (Pyke 1990). not differ significantly between the popu- genetically before livestock were intro- Reduced average seed production poten- lations (ungrazed: 24.5 + 6.0%, grazed: duced to the grazed site. Nevertheless, tial per plant in the grazed population is due in part to a combination of traits that Table 2. Rank correlation coefficients between components of seed production ungrazed (upper lead to a reduction in overall plant size. values and grazed (lower) populations of sideoats grama. These include a reduced number of tillers and panicles compared to the ungrazed Spikes Tillers % reproductive Spikes Seeds Seed -1 -1 -1 -1 population (Table 1). This morphology and panicle ramet tillers ramet ramet mass its effects on seed production is consistent Seeds spike — – 0.50*/0.40 — — 0.68*/0.93* — with findings from previous research on Spikes panicle 0.56*/0.04 — 0.48*/0.22 — – the evolution of grazing-adapted morpho- Tillers ramet 0.68*/0.81* 0.04/0.56* — types in caespitose grasses (e.g., Detling % reproductive and Painter 1983, Painter et al. 1989, tillers 0.85*/0.58* — — 1993). These morphological adaptations -1 Spikes ramet 0.31/0.55* — affecting seed production may therefore -1 Seeds ramet — represent simply secondary consequences *Significant at P £ 0.05, of strong selection for vegetative traits — = P > 0.05; Ungrazed population/Grazed population, N = 20 for both. associated with defoliation tolerance.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 553 The observation of reduced floral fertili- associated with grazing resistance in the Cheplick, G.P. 1995. Life history trade-offs in ty (number of seeds per spike) in the grazed population in the roughly 100 years Amphibromus scabrivalis ( P o a c e a e ) : grazed population suggests that selection that it has been exposed to livestock. Allocation to clonal growth, storage, and under grazing stress may have affected the Reduced seed production would have cleistogamous reproduction. Amer. J. Bot. production of seeds within an individual negative consequences most noticeably in 82:621–629. Cronquist, A., A.H. Homgren, N.H. spikelet. This could be driven by competi- reducing propagules available for colo- Holmgren, J.L. Reveal, and P.A. tive advantages afforded to genotypes that nization of open sites (Richards 1990). Holmgren. 1977. Intermountain Flora. allocate additional photosynthetic Reduced seed production as demonstrated Vascular plants of the Intermountain West, resources to vegetative growth and less to here could also affect genetic structure U.S.A. Columbia University Press, New seed production (Bazzaz et al. 1987, within the grazed population since the pro- York, N.Y. Ashman 1994). Physiological trade-offs of duction of fewer seeds would result in Detling, J.K. and E.L. Painter. 1983. this sort (Stearns 1992) would be expected reduced opportunities for sexual reproduc- Defoliation responses of western wheatgrass to lead to negative correlations between tion should it occur. Indeed, previous populations with diverse histories of prairie the life-history traits affected although genetic analyses of these 2 populations dog grazing. Oecologia 57: 65–71. Freter, L.E. and W.V. Brown. 1955. A cyto- these have been difficult to document in (Smith 1998) suggests that the grazed pop- taxonomic study of Bouteloua curtipendula perennial grasses (Reekie 1991, Cheplick ulation is less variable genetically in vege- and B. uniflora. Bull. Torr. Bot. Club 1995). While this research was not tative traits associated with defoliation tol- 82:121–130. designed to identify physiological trade- erance. The grazed population may there- Gordon, I.L. 1979. Standard errors of heri- offs, there is some evidence for such a fore be more susceptible to local extinc- tabilities based on perennial observations, trade-off in the ungrazed population where tion (O’Conner 1991) due to either the with application to Yorkshire fog grass. the number of tillers per plant is negative- direct demographic effects of reduced con- Euphytica 28:81–88. ly correlated with the number of caryposes tribution to the soil seed bank, or to the Gould, F.W. 1951. Notes on apomixis in per spike (Table 2). However, this is not more indirect genetic effect of limiting sideoats grama. J. Range Manage. 12:25–28. conclusive since seed production traits are available genetic variability due to Gould, F. W. and Z. J. Kapadia. 1964. Biosystematic studies in the Bouteloua cur - not correlated with the number of spikes reduced potential for outcrossing. t i p e n d u l a complex II. Taxonomy. Brittonia produced either by individual panicles or 16:182–207. whole plants in this population as might Hadley, D., P. Warshall, and D. Bufkin. be expected with this trade-off. Literature Cited 1991. Environmental change in Aravaipa, A commonly discussed physiological 1870-1970: an ethnoecological survey. trade-off related to seed production Aarssen, L.W. and R. Turkington. 1985. Cultural Resource Series 7, Ariz. Bureau of involves linkage of decreases in seed num- Within-species diversity in natural popula- Land Management, Phoenix, Ariz. ber per plant with increases in seed mass, tions of Holcus lanatus, Lolium perenne and Harlan, J.R. 1949. Apomixis in side-oats and with increases in the latter trait pre- Trifolium repens from four different-aged grama. Am. J. Bot. 36:495–499. pastures. J. Ecol. 73: 869–886. Hoffmeister, D. F. 1986. Mammals of sumably leading to improved seedling Arizona. University of Arizona Press, establishment (Westoby et al. 1992). This Ashman, T.-L. 1994. A dynamic perspective on the physiological cost of reproduction in Tucson, Ariz. does not appear to be the case here as Krausman, P. R., A. J. Kuenzi, R. C. mean seed mass did not differ between the plants. Amer. Nat. 144:300–316. Bazzaz, F.A., N.R. Chiariello, P.D. Coley, Etchberger, J. R. Rautenstrauch, L. L. populations and this trait was not correlat- and L.F. Pitelka. 1987. Allocating resources Ordway, and J. J. Hervert. 1997. Diets of ed with any components of seed output to reproduction and defense. BioScience desert mule deer. J. Range Manage. (Table 2). 37:58–67. 50:513–522. Available evidence suggests that the Briske, D.D. 1996. Strategies of plant survival Littell, R.C., G.A. Milliken, W.W. Stroup, polyploid sideoats grama populations in grazed systems: A functional interpreta- and R.D. Wolfinger. 1996. SAS System for native to southern Arizona are facultative- tion, p. 37–67. I n: J. Hodgson (ed.), The Mixed Models. SAS Institute, Inc., Cary, ecology and management of grazing systems. N.C. ly apomictic (Harlan 1949, Freter and Lynch, M. and B. Walsh. 1998. Genetics and Brown 1955, Gould 1951). The existence Oxford Univ Press, New York, N.Y. Briske, D.D. and J.H. Richards. 1995. P l a n t analysis of quantitative traits. Sinauer of facultative apomixis and complex sexu- Associates, Inc., Sunderland, Mass. al/apomictic cycles (Richards 1986) in this response to defoliation: A physiological, mor- phological and demographic evaluation, p. Nyquist, W.E. 1991. Estimation of heritability taxon would seem likely given current 635–710. In : D.J. Bedunah and R.E. Sosebee and prediction of selection response in plant understanding of its reproductive biology. (eds.), Wildland plants: Physiological ecology populations. Critical Rev. in Plant Sci. Recent research (Overath and Asmussen and developmental morphology. Soc. Range 10:235–322. 1998) suggests that it may be incorrect to Manage., Denver, Colo. O’Connor, T.G. 1991. Local extinction in assume that populations of facultatively Brown, D.E. 1994. Semidesert grasslands, p. perennial grasslands: A life-history approach. 123–135. In: D. E. Brown (ed.), Biotic com- Amer. Natur. 137:753–773. apomictic plants are necessarily depauper- Orodho, A.B., R.L. Cuany, and M.J. Trlica. ate genetically and therefore less respon- munities: southwestern United States and northwestern Mexico, University of Utah 1998. Previous grazing or clipping affects sive to selection than non-apomictic seed of Indian ricegrass. J. Range Manage. species. Canfield (1957) showed that indi- Press, Salt Lake City, Ut. Canfield, R.H. 1957. Reproduction and life 51:37–41. vidual sideoats grama plants did not live span of some perennial grasses of southern Overath, R.D. and M.A. Asumssen. 1998. longer than 3 years on sites grazed by cat- Arizona. J. Range Manage. 10:199–203. Genetic diversity at a single locus under via- tle in southern Arizona. Together, these Carmen, J.G. and D.D. Briske. 1985. bility selection and facultative apomixis: findings indicate that significant genetic Morphologic and allozymic variation Equilibrium structure and deviations from variation may have existed and sufficient between long-term grazed and non-grazed Hardy-Weinberg frequencies. Genetics mortality and seedling establishment could populations of the bunchgrass Sc h i z a c h y r i u m 148:2029–2039. have occurred to permit selection of traits s c o p a r i u m var. f r e q u e n s . O e c o l o g i a 66 : 3 3 2 – 3 3 7 .

554 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Painter, E.L., J.K. Detling, and D.A. Richards, A.J. 1986. Plant Breeding Systems. Smith, S.E. 1998. Variation in response to Steingraeber. 1989. Grazing history, defoli- Allen & Unwin, London. defoliation between populations of Bouteloua ation, and frequency-dependent competition: Richards, A.J. 1990. The implications of curtipendula var. caespitosa (Poaceae) with effects on two North American grasses. reproductive versatility for the structure of different livestock grazing histories. Amer. J. Amer. J. Bot. 76: 1368–1379. grass populations. p. 131-153. I n: G.P. Bot. 85:1266–1272. Painter, E.L., J.K. Detling, and D.A. Chapman (ed.), Reproductive versatility in Stearns, S.C. 1992. The evolution of life histo- Steingraeber. 1993. Plant morphology and the grasses. Cambridge Univ. Press, ries. Oxford Univ. Press, Oxford. grazing history: Relationships between native Cambridge. Trlica, M.J. and A.B. Orodho. 1989. Effects grasses and herbivores. Vegetatio 106:37–62. Scott, A.W. and R.D.B. Whalley. 1984. T h e of protection from grazing on morphological Pyke, D.A. 1990. Comparative demography of influence of intensive sheep grazing on geno- and chemical characteristics of Indian rice- co-occurring introduced and native tussock typic differentiation in Danthonia linkii, D . grass, O r y z o p s i s h y m e n o i d e s . O k i o s grasses: persistence and potential expansion. richardsonii and D. racemosa on the New 56:299–308. Oecologia 82:537–543. England tablelands. Aust. J. Ecol. Westoby, M., E. Jurado, and M. Leishman. Reekie, E.G. 1991. Cost of seed versus rhi- 9:419–429. 1 9 9 2 . Comparative evolutionary ecology of zome production in Agropyron repens. C a n . Sokal, R.R. and F.J. Rohlf. 1995. B i o m e t r y . seed size. Trends in Ecol. and Evol. J. Bot. 69:2678–2683. 3rd ed. W.H. Freeman, New York, N.Y. 7:368–372.

Your one-stop supplier for: • Pasture and range grasses • Alfalfa, shrubs, & wildflowers • Custom seed blends • Erosion control materials • Environmental consulting s e r v i c e s Ask for our catalog featuring over 600 native species and named vari- e t i e s .

Call or fax for our Catalog (801) 768-4422, fax (801) 768-3967 Granite Seed Co., 1697 W. 2100 North, Lehi, UT 84043 w w w . g r a n i t e s e e d . c o m

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 555 J. Range Manage. 53: 556–559 September 2000 Hoary cress reproduction in a sagebrush ecosystem

LARRY LARSON, GARY KIEMNEC, AND TERESA SMERGUT

Authors are professor Department of Rangeland Resources, associate professor Department of Crop and Soil Science, and graduate research assistant at time of research Department of Rangeland Resources, Oregon State University, Corvallis, Ore. 97331.

Abstract Resumen

Field studies were undertaken to evaluate hoary cress Se condujeron estudios de campo para evaluar la reproducción (Cardaria draba (L.) Desv.) reproduction and spread in a sage- y diseminación de "Hoary cress" [Cardaria draba (L.) Desv.] en brush ecosystem. Hoary cress germination, emergence, and sur- un ecosistema de "Sagebrush". La germinación, emergencia y vival were restricted to moist environments. These conditions sobrevivencia de "Hoary cress" ocurrió solo en ambientes húme- occurred 2 out of 8 years. Hoary cress germination under field dos. Estas condiciones se presentaron en 2 de 8 años. La germi- conditions was greatest on toe slope positions and areas of soil nación de "Hoary cress" bajo condiciones de campo fue mayor disturbance. Number of shoots varied annually for established en posiciones de la punta de la pendiente y en áreas con suelo hoary cress populations. Shoot propagation was reduced when disturbado. En las poblaciones establecidas de "Hoary cress", el early spring growth was followed by frost or drought. Shoot número de tallos varió anualmente. La propagación de los tallos numbers were increased when spring growth was delayed and se redujo cuando el crecimiento temprano de primavera fue warm, moist growing conditions occurred in May. Seed repro- seguido de una sequía o helada. El número de tallos se incremen- duction did not increase plant density in monitored populations. to cuando el crecimiento de primavera se retrasó y en Mayo se Established populations relied upon vegetative reproduction to presentaron condiciones de calor y humedad. La reproducción sequester resources and increase plant density. por semilla no incrementó la densidad de plantas en las pobla- ciones establecidas. Las poblaciones establecidas dependieron mas de la reproducción vegetativa para adquirir los recursos Key Words: Noxious weeds, Cardaria draba (L.) Desv., rangeland necesarios e incrementar la densidad de plantas. weeds

Hoary cress (Cardaria draba (L.) Desv.), also known as white- of its spreading potential. The objective of this research was to top or heart-podded whitetop, was introduced to North America evaluate the contribution of vegetative and seed reproduction to in the 1800's (Mulligan and Franklin 1962). The earliest collec- hoary cress spread in a sagebrush ecosystem by measuring germi- tions of hoary cress in the United States were made in California nation, emergence and shoot density. in 1876 (Groh 1940). Hoary cress, a perennial weed of forage crops, has extended its range into the sagebrush ecosystem of the Northern Great Basin. Materials and Methods Hoary cress reproduces by seed and creeping rootstock (Whitson et al. 1992). Two seeds develop in each heart-shaped Field studies were conducted between 1986 and 1994 in north- capsule. Adult plants can produce 1,200 to 4,800 seed with over eastern Oregon (44° 43' N, 117° 45' W) in the foothill range of 80% viability (Selleck 1965). Seed rain occurs as individual the Wallowa mountains. Study sites lie within the Wyoming big seeds are released through ruptures in the capsule sidewall, as sagebrush (Artemisia tridentata ssp w y o m i n g e n s i s N u t t . ) — b l u e- individual capsules, and as capsule clusters which break away bunch wheatgrass (Agropyron spicatum (Pursh) Scribn. & Smith from the parent plant. Hoary cress seed exude mucilage when = Pseudoroegneria spicata (Pursh) Scribn. & Smith) habitat type moistened (Kiemnec and Larson 1991). Mucilage produced by (Daubenmire 1970). At each study site the sagebrush component other species increases germination under conditions of low was intact but the bluebunch wheatgrass had been lost and later osmotic potential (0 to –0.2 MPa) (Harper and Benton 1966). replaced by crested wheatgrass (A. cristatum (L.) Gaertn.) seeded Germination occurs in the fall or spring and non-germinated seed during the 1960's. The predominant soil in the study area is an can remain viable in the soil seed bank for 3 to 4 years (Bellue Encina gravelly silt loam (fine, montmorillonitic, mesic Aridic 1946, Pryor 1959). Shoots develop on adult plants at the root Calcic Argixeroll). crown and along creeping root stock (Mulligan and Findlay Temperature and precipitation data obtained from the Baker 1974). Shoot development begins in the fall or spring, as a loose City airport (10 k from study area, 1,050 m elevation) were used rosette, and culminates with the production of a flowering stalk. to calculate average precipitation (mm), maximum/minimum Understanding the expression of reproductive characteristics of temperature (C) and growing degree day (Miller and Donahue hoary cress in a sagebrush ecosystem would aid in the evaluation 1990) estimates (threshold: 4°C). Precipitation and temperature frequency estimates were obtained from the Baker County Soil Survey (Laird et al. 1988) and represent data recorded in the peri- Manuscript accepted 25 Nov. 1999. od 1951 to 1981.

556 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 Seed reproduction - germination randomized block design. Treatments Results and Discussion Hoary cress seed was collected at the were replicated 3 times within each com- study area during July 1988. Seed were munity at each site (2 years, 2 sites, 2 Germination separated from capsules, cleaned and treatments, 3 replications). Emerged Germination in buried seed packets in counted into lots of 50. Shriveled seed was seedlings were counted at weekly intervals 1989 was restricted to the time period discarded. Each lot (50 seed) was sealed in May 1987 and 1988. Emergence data between 1 February and 15 March. Air with 5 ml of silt loam soil in a nylon mesh were analyzed using analysis of variance temperatures during the months of packet (8 cm x 8 cm; mesh opening = procedures with mean separation by least October through April were near normal 0.015 cm) and placed in the field in significant difference (P £ 0.05). Data (avg. max./min.; 16/–1, 7/–4, 2/–8, 1/–9, October 1988. were blocked by years, and sites were 4/–6, 9/–3, 14/0 °C). Precipitation was Seed packets were located in 4 topo- combined after it was determined that below normal in October (13 mm) (fre- graphic positions (ridgetop, toe slope, there were no site differences. quency < 2 yr in 10), whereas precipita- mid-slope north aspect, and mid-slope tion was above normal in November (44 south aspect). Seed packets were placed mm), December (35 mm), January (33 on the surface and at 1- and 3-cm depths Vegetative reproduction Hoary cress shoot density was estimated mm), and March (35 mm) (frequency < 2 and were retrieved at 6-week intervals yr in 10). Precipitation amounts during by counting live shoots along 8 permanent beginning on 15 December 1988 and end- February and April were near normal. transects (30-m length) in 2 infested ing on 1 May 1989. Each treatment was Germination was affected by topograph- replicated 5 times (4 topographic posi- Wyoming big sagebrush-crested wheat- ic position and burial depth (Table 1). tions, 3 soil depths, 4 retrieval dates, 5 replicates = 240 seed packets/experiment). Table 1. Field germination of hoary cress seed associated with topographic position and soil depth. The experiment was conducted at 2 sites within the study area. Soil Depth (cm) Seed packets were opened on the Position 0 1 3 Mean retrieval date and number of germinated 1 seed (emerged radicle) recorded. ------(%) ------Remaining seeds were placed on filter Toe slope 56 c 43 b 34 a 45 C paper moistened with distilled water in a North Aspect 16 a 34 c 28 b 26 B sealed Petri dish and placed in an environ- Ridgetop 19 a 27 b 26 b 24 B mental chamber (16-hour dark period at South Aspect 6 a 16 b 20 c 14 A 15°C, 8-hour light period at 20°C). 1Horizontal letter changes indicate significance (LSD = 4; P £ 0.05)within a topographic position. Upper case letters Germination was evaluated for 16 days indicate significance of means across topographic positions (LSD = 6; P £ 0.05). with germination counts occurring at 4- day intervals. grass communities. Parallel transects were Germination was greatest on the toe slope Data analyses were performed on 2 cate- established at each infestation (diameter with minimal soil coverage. Germination gories of seed germination: (1) field germi- approximately 70 m); 2 transects dissected at the other topographic positions nated seed—seeds germinated in the seed the infestation and 2 transects (non-infest- improved with some seed burial. Micro- packet prior to the retrieval dates of 15 ed) were tangent to the infestation. Shoot and macro-topographic characteristics December, 1 February, 15 March, and 1 influence seed germination (Harper 1983). May and (2) total germination—seeds ger- densities were tallied mid-May from 1987 through 1994. Density counts were taken These influences largely reflect the degree minated in the field plus seeds germinated 2 of contact between the seed and soil sur- in the environmental chamber after each using 0.2-m quadrats placed at 1.5 m intervals along the transect. Data were not face and the matric potential of the soil retrieval date. Data from the 2 sites were water (Fenner 1985). Osmotic potential collected in 1991. combined after it was determined there can also be a factor, especially in arid areas Twenty excavations (0.25 m2 X 20 cm were no site differences. The experimental where salt influences are more common. design was a complete randomized block. deep) within hoary cress infestations were Kiemnec and Larson (1991) reported hoary Data analysis (Dowdy and Wearden 1983) made in 1988. Counts of vertical root cress germination rates of 82, 55, 11 and included Hartley's data normality test, extensions and shoots were taken. 0% at osmotic potentials of 0, –0.5, –1.0, analysis of variance, and mean separation Differences in shoot density were deter- and –1.5 MPa. The process of soil freezing by least significant difference (P £ 0.05). mined by analysis of variance with mean and thawing may have also benefited ger- separation by least significant difference mination through vertical redistribution of Seed reproduction - emergence (P £ 0.05). Precipitation and growing soil moisture upward toward the freezing Hoary cress emergence was studied at 2 degree day estimates were determined bi- layer (Heidmann and Thorud 1976). In sites in 1986 and 1987. Seed were collect- monthly for the months of March, April, finer textured soils, freezing and thawing ed in July 1986 and 1987 using the proce- and May during the period of 1987–1994. can increase moisture levels at the soil sur- dures described previously. Seed were Stepwise linear regressions (Statsgraphics, face to near saturation (Krumbach and broadcast (50 seed/plot) in November Cambridge, Mass.) were performed to White 1964, Harlan 1973). We speculate 1986 and 1987 onto 0.1-m2 plots in both select the periods of precipitation and that some or all of these processes were Wyoming sagebrush-crested wheatgrass growing degree day accumulation that present during our study and likely con- and annual grass communities (toe slope accounted for the variation in annual shoot tributed to the pattern of field germination. topographic position). Tilled (churned density. Seed germination (field + lab) achieved with a shovel) and undisturbed seedbed rates of 71 to 94%. Compared to field ger- treatments were compared in a complete mination these germination rates suggest

JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 557 Table 2. Hoary cress emergence in 2 plant com- Table 3. Regression model of shoot density nation, emergence, and survival to munities with 2 levels of seedbed prepara- association with precipitation amount and microsites where moist conditions are pre- tion. accumulated degree days in 16–30 April and sent. These conditions most likely occur 1–15 May during the period 1987–1994. on toe slope positions and soil disturbance Year Seedbed Community Variable1 Coefficient t-value P areas where direct seed-to-soil contact is Sagebrush Annual maximized. Reproduction by seed Grass (b) occurred 2 out of 8 years and was associ- - - - (shoots 0.2 m-2)1 - - - Intercept (a) 17.14 5.20 0.01 2 ated with warm, moist weather in May. 1987 Degree Day Shoot densities in established hoary tilled 0.68 a 14.68 b 16–30 April (X1) –0.12 –4.82 0.01 1–15 May (X ) 0.09 3.25 0.01 cress populations vary annually. Densities undisturbed 0.0 a 0.0 a 2 are increased by early cool, dry periods Precipitation; cm followed by warmer, wetter conditions 1988 16–30 April (X3) –2.68 –5.04 0.01 and decreased by frost or drought damage. tilled 3.32 a 9.32 b 1–15 May (X4) 1.22 4.01 0.01 1 Measured changes, however, did not result undisturbed 0.0 a 0.0 a Regression model: Y = a + b1X1 + b2X2 + b3X3 + b4X4 2 1 Growing degree day = max temp + min temp - threshold in a measurable expansion or contraction Horizontal letter changes within each year indicate sig- of hoary cress populations. nificance (P £ 0.05) 2 We speculate that the spread of hoary cress, controlled primarily by environmen- that moisture and/or temperature rather Precipitation and degree day accumula- tal conditions, is occurring at a relatively than seed viability or dormancy were lim- tion for 16–30 April were negatively asso- slow rate. Soil moisture conditions needed iting germination in this study. ciated with shoot density (Table 3). for seed germination occur relatively Precipitation and degree day accumulated infrequently. However, established plants Emergence during the first half of May were positive- appear to readily sequester resources via a Seedling emergence occurred only on ly associated with shoot density. These creeping root system thus reducing the plots that were tilled and was greatest in the relationships suggest cool, dry weather in risk of shoot mortality. Factors such as annual grass community (Table 2). None of the later half of April and/or warm, moist irrigation, disking established populations, the seedlings that emerged during the study conditions in the first half of May are foraging by pocket gophers, and other survived beyond May. This was likely due associated with years of increased hoary influences that increase soil moisture or to a lack of available soil moisture. cress density. fragmentation of existing root systems will We speculate that rapid above ground tend to increase the rate of spread. growth in April is susceptible to frost or Vegetative reproduction drought damage in May. In 1990, warm, Shoot density fluctuations occurred moist conditions in April were followed Literature Cited within monitored hoary cress populations. by frost damage from 5 nights below Shoot densities ranged from 12 shoots freezing in May. These cold conditions 0.2m-2 (1989 and 1993) to 5 shoots 0.2m-2 Bellue, M. K. 1946. Weed seed handbook. were associated with a 56% reduction in Series VI. Calif. Dept. Agr. Bull. 22:288. (1990 and 1994). These changes did not shoot density compared to shoot density in Daubenmire, R. 1970. Steppe vegetation of result in a measurable expansion or con- Washington. Washington Agr. Exp. Sta. traction of the area occupied by the hoary 1987. Low temperatures of 2°C or lower occur later than 10 May in 5 out of 10 Tech. Bull. No. 62. cress populations. Hoary cress seedlings Dowdy, S. and S. Wearden. 1983. S t a t i s t i c s were observed in the monitored popula- years and temperatures below 4°C occur for research. John Wiley & Sons. New York. tions in 1989 and 1993, but did not sur- later than 4 May in 2 out of 10 years. In Fenner, M. 1985. Germination. p 87–102. I n vive beyond their first growing season 1994, rainfall total was 30 mm and aver- Seed Ecology. Chapman & Hall. New York. (data not shown). age temperature (max/min) was 58/34°C Groh, H. 1940. Hoary cresses in Canada. Sci. Excavations of hoary cress showed the in April. These conditions were followed Agr. 20:750–756. creeping root system averaged 19 vertical by below normal (0.8 mm) early May pre- Harlan, R.L. 1973. Analysis of coupled heat- cipitation and 6 days in which maximum fluid transport in partially frozen soil. Water extensions arising from lateral roots and Resour. Res. 9:1314–1323. each extension averaged 2.5 shoots with temperatures ranged from 24 to 29°C. Harper, J.L. 1983. The natural dynamics of numerous dormant buds. These excava- These growing conditions were associated plant populations. p. 515-646. In: Population tions suggest that the creeping root system with a 36% reduction in shoot density. Biology of Plants. Academic Press Inc., is the primary source of the observed den- By contrast a cool, dry April in 1989 and London. sity fluctuations. 1993 was followed by several rains (> 6 Harper, J.L., and R.A. Benton. 1966. T h e Precipitation amounts and the number of mm) during the first 15 days of May. Mean behavior of seeds in soil, part 2. The germi- accumulated growing degree days during daily maximum temperatures in May both nation of seeds on the surface of a water sup- 16–30 April and 1–15 May accounted for years was near 18°C and lows were above plying substrata. J. Ecol. 54:151–166. 2 Heidmann, L.J. and D.B. Thorud. 1976. 76% (R ) of the observed variation in 1°C. These conditions were associated with Controlling frost heaving of ponderosa pine shoot density (Table 3). Precipitation 31 and 41% increases in shoot density. seedlings in Arizona. USDA Forest Serv. (44%) and accumulated degree days Res. Pap. RM-172. Rocky Mt. For. Range (22%) for April 16-30 contributed 66% of Implications Expt. Sta., Fort Collins. Colo. the R 2 value. Precipitation (22%) and Kiemnec, G. and L. Larson. 1991. accumulated degree days (12%) for May Hoary cress germination is limited by Germination and root growth of two noxious 1–15 contributed the remaining 34% of dry environmental conditions (Kiemnec weeds as affected by water and salt stresses. the R2 value. and Larson 1991). The xeric nature of the Weed Technol. 5:612–615. sagebrush ecosystem restricts seed germi-

558 JOURNAL OF RANGE MANAGEMENT 53(5), September 2000 Krumbach, A.R. Jr., and D. P. White. 1964. Miller, R.W. and R.L. Donahue. 1990. Soils: Pryor, M.R. 1959. Hoary Cress: new control Moisture, pore space, and bulk density An introduction to soils and plant growth. findings. Calif. Dept.Agr. Bull. 48:11–14. changes in frozen soil. Soil Sci. Soc. Amer. Prentice Hall. New Jersey. Selleck, G.W. 1965. An ecological study of Proc. 28:422–425. Mulligan, G.A. and J.N. Findlay. 1974. T h e lens and globe-podded hoary cresses in Laird, W.E., M.H. Fillmore, G.D. Macdonald, biology of Canadian weeds. 3. Cardaria Saskatchewan. Weeds. 13:1–5. and D.P. Christenson. 1988. Soil survey of draba, C. chalepensis, and C. pubescens. Whitson, T., L. Burrill, S. Dewey, D. Baker County area, Oregon. USDA, Natural Can. J. Plant Sci. 54:149–160. Cudney, B. Nelson, R. Lee, and R. Parker. Resources Conservation Service. US Gov. Mulligan, G.A. and C. Franklin. 1962. 1992. Weeds of the west. University of Print. Office, Washington, DC. Taxonomy of the genus C a r d a r i a with par- Wyoming. Laramie, Wyo. ticular reference to the species introduced into North America. Can. J. Bot. 40:1411–1425.

International Center for Agricultural Research in the Dry Areas (ICARDA) RANGE/FORAGE SCIENTIST/ECOLOGIST VACANCY ANNOUNCEMENT

Arabian Peninsula The Organization ICARDA is a non-profit international agricultural research center in a worldwide consortium of 16 centers, supported by the Consultative Group on International Agricultural Research (CGIAR). The CGIAR is co-sponsored by the World Bank, the Food and Agriculture Organization of the United Nations, the United Nations Development Programme and the United Nations Environment Programme. With a staff of over 400 scientists and support personnel, representing more than 40 nationalities, ICARDA addresses two key areas of agricultural research: germplasm improvement (cereals and legumes) and natural resource management. The Position ICARDA wishes to recruit a dynamic Range Ecology/Forage/Management Scientist to work as part of ICARDA’s Arabian Peninsula Regional Program (APRP) whose main office is located in Dubai, UAE. The overall objective of the position is the development of inte- grated range/forage/livestock production systems and management practices for rangeland regeneration, through conservation and sustain- able management of rangeland vegetation and the production of stress tolerant and water use efficient forage crops. In collaboration with the Range Management Scientist at ICARDA headquarters, the incumbent will be responsible for the execution of the research strategy of the rangeland program, which will focus on the following: collection and evaluation of indigenous and exotic forage and rangeland species for the purpose of rangeland regeneration (restoration and rehabilitation) and as sources for alternative forage crops; identification of the appropriate technical options for forage seed/crop production in different agroecological zones in the region; identification of the appropri- ate technical options for the regeneration of degraded rangelands; adoption and demonstration of rangeland management practices in target areas and pilot sites. Required Qualifications & Experience • PhD in Range Science/Ecology, or related discipline, with strong background and scientific experience in range plant ecology and forage production in arid and salt-affected environments. Knowledge of animal/plant interactions and ecophysiology is important. • Minimum of five years’ experience, preferably with knowledge of the region and with international research organizations, demonstrated field orientated skills, and ability to work in an inter-disciplinary team. • Statistical skills to analyze range dynamics—including overgrazing. • Ability to communicate in spoken and written English (knowledge of Arabic is desirable). For further information about ICARDA, the position, terms of appointment, living in UAE and application procedure, please see our website: http://www.cgiar.org/icarda Application Qualified applicants are invited to send: a cover letter of interest, including recent salary history; curriculum vitae; names, addresses, fax and e-mail information for three professional referees, to: Personnel Office, ICARDA, P.O. Box 5466, Aleppo, Syria Telephone: (963-21) 2213477, 2225112, or 2225012 Fax: (963-21) 2213490, 2225105, or 5744622 E-mail: [email protected] (If sending by e-mail, please send documents in Rich Text Format and do not include graphics or other large file attachments.)

REFERENCE: quote INT/0019/00 on the application (or in the subject line if applying by e-mail). APPLICATION DEADLINE: Applications must be received by 30 September 2000. ICARDA is an equal opportunity employer, and encourages applications from women.

JOURNAL OF RANGE MANAGEMENT53(5), September 2000 559 Sci. Bull. 26:725–731). Newcomers have preserved almost 80,000 Book Review acres of open land, which is almost half of the private land in the county. This has maintained the critical mass of ranches needed for their continued survival and minimized rough edge effects between Old Fences, New Neighbors. By Peter R. Decker. 1998. The ranches and developments (Huntsinger and Hopkinson 1996, J. University of Arizona Press, Tucson, Arizona Range Manage. 49:167–173). [www.uapress.arizona.edu]. 159 p. US$40.00 hardbound, Old Fences, New Neighbors contains no warning that the zoning US$20.00 paper. ISBN 0-8165-1771-1 hardbound, ISBN 0- that was instrumental in maintaining open spaces and ranches in 8165-1905-6 paper. Ouray County can be transitory, as the populace and politics change. On the cusp of middle age I find that change, especially change I Not all of the current landowners have made a commitment to pre- don’t like, begins to seem inevitable. Growing up in the Midwest I serve their ranches in perpetuity. Conservation easements offer per- saw farms become shopping malls; as an adult living in the manent protection for open spaces, but require an agency with large Southwest I see ranches becoming subdivisions. Old Fences, New amounts of cash, and, as Decker points out, are of limited use to Neighbors has given me hope that we can preserve our ranches as small ranchers with little income to offset. open space. Although the physical landscape has largely been preserved, Decker’s book, which he calls a partial biography of a place, is his Decker now sees a different human landscape in Ouray County. The attempt “...to bring some order and understanding to the chaotic present community is less unified than before, as the residents no c h a n g e s . . .” that have occurred in Ouray County, Colorado during longer share common goals. But this area has reinvented itself many the past 20 years. Rapid population growth in this ranching commu- times, as Decker documents; the transition periods are painful, but nity led to higher land prices, which drove many ranchers out of the community has survived. For Decker, Ouray County is “...losing business. Decker traces this area’s waves of invasions over the past its cows, but...gaining in human strength to sustain and comfort its 150 years, from the U.S. military driving out the Utes, through citizens”. Anglo miners, farmers, and ranchers, to the latest newcomers, begin- Decker’s love of the West is evident, as he describes being a camp ning in the 1980s. This last group was able to pay more for land than Jack, the lowest form of life on a Montana sheep ranch as a teenager, the land could produce from ranching, shifting the area from an agri- rediscovering the West as a young adult, and later, ranching in Ouray cultural to a largely residential one. County. He simply and richly describes the joy at the birth of a calf, Old Fences, New Neighbors weaves together several themes. First the camaraderie of neighbors at branding time, and his battles with is isolation, as the San Juan Mountains slowed settlement and made the Dry Creek meadow with its “...steep slope and thin soils...”. transportation difficult. Its corollary is the influence of outside In a book with many dichotomies, Decker himself is a gray area. forces, over which the residents have no control. Historian Patricia Neither old timer nor newcomer, he bought his ranch in 1974 but Limerick, in her 1987 book Legacy of Conquest: The Unknown Past continued to hold an academic position until he began ranching full- of the American West, described western immigrants as denizens of time in 1980. He seems to have been accepted by the ranchers and to an “Empire of Innocent” driven by heartfelt motives but the victims have been able to write, without rancor, from both sides. But ranch- of forces beyond their control. In Ouray County these forces have ing has never been his sole source of income. Few ranchers receive included the federal government, which instituted grazing fees on National Endowment for the Humanities or Rockefeller Foundation public land in the 1930s, and had planned to flood the town of grants, and their books are rarely published. Decker also served as Ridgway behind a dam in the 1960s. National and world economies Chair of the County Planning Commission and State Commission of have determined the prices of minerals, farm produce, and cattle; big Agriculture for Colorado. But these additional sources of income businesses, including mines and railroads, began and ended opera- could not protect him against falling cattle prices and soaring land tions. During the past 20 years low cattle prices and high land prices; Decker sold most of his ranch and moved to Nebraska in prices, driven by changing demographics and increasing affluence, 1993. have forced many ranchers to sell out. Decker also describes the An egregious oversight of the book is the lack of any mention of relationship that each wave of immigrants has had with the land. estate planning or trusts, other than those supporting some of the The Utes hunted and gathered and miners, farmers, and ranchers county’s new residents. Although Decker and the other ranchers made their living directly from the land. Most of the current resi- bemoan estate taxes and recite stories of cash-poor ranches sold to dents were drawn by the land’s beauty, but make their living outside pay the taxes, he never mentions the simple steps to avoid this the county. occurrence. This is difficult to bring up with one’s parents or to Decker argues that higher cattle prices would be the best protec- accuse a neighbor of having foolishly overlooked. The onus is on tion for ranchers. Although it is possible to receive above-market the cattle industry to encourage and educate ranchers in the use of prices with a specialty product, such as “natural beef” (Voynick, estate planning, as a recent article in R a n g e l a n d s does (McCann, 1998, Coleman, Saguache, Colorado), the vast majority of ranchers 1999:21:3–4). must accept market prices for their cattle. Although land values also Although not a blueprint for preserving cattle ranches and open are beyond the control of individual ranchers, the residents of Ouray spaces in the New West, Old Fences, New Neighbors is for anyone County have been able to preserve open space by regulating land concerned about ranching and open lands. This book is not a call to values indirectly. The county had instituted unusually strict zone caring, but is for those who care: ranchers, environmentalists, plan- regulations in the 1970s, which included low housing density limits, ners, resource managers, and the citizens of Ouray County. It tells visual impact specifications, and protection of irrigated hay mead- the story of a difficult place, hit hard by demographic and economic ows. In response to increased pressure for development in the 1980s, changes, and its ability to reinvent itself. The book gives us hope density limits were increased, but open space requirements were that open spaces can be preserved and that ranching can continue, added. This limited the return from subdividing ranches, reducing although perhaps in a different form.—Lucinda F. Salo, School of their appeal to developers and leaving wealthy individuals from out- Renewable Natural Resources, University of Arizona, Tucson, side the county as the main ranch buyers. Arizona. Cash from outsiders, drawn by the area’s outstanding scenic beau- ty, has protected the ranches of Ouray County, illustrating how stewardship of private lands and working landscapes can be an important part of preserving our open spaces (Knight, 1998, Wildl.

560 JOURNAL OF RANGE MANAGEMENT53(5), September 2000 INSTRUCTIONS FOR AUTHORS (from revised Handbook and Style Manual)

Eligibility received an unsatisfactory review may file an appeal with the Editor. The Editor then may select another Associate Editor to review the appeal. The Associate Editor reviewing the appeal will be provided with copies of an The Journal of Range Management is a publication for reporting and correspondence relating to the original review of the manuscript. If the documenting results of original research and selected invitational papers. appeal is sustained, a new review of the manuscript may be implemented Previously published papers are unacceptable and will not be considered at the discretion of the Editor. for publication. Exceptions to this criterion are research results that were Authors should feel free to contact the Associate Editor assigned to originally published as department research summaries, field station their manuscript at any stage of the review process: to find out where the reports, abstracts of presentations, and other obscure and nontechnical paper is in the process; to ask questions about reviewer comments; to ask handout publications. Manuscripts submitted to the JRM are the property for clarification or options if a paper has been rejected. of the J o u r n a l until published or released back to the author(s). Manuscripts may not be submitted elsewhere while they are being con- sidered for this journal. Papers not accepted for publication are automat- Page Proofs ically released to the authors. Page proofs are provided to give the author a final opportunity to make Kinds of Manuscripts corrections of errors caused by editing and production. Authors will be charged when extensive revision is required because of author changes, Journal Articles report original findings in Plant Physiology, Animal even if page charges are not assessed for the article. One author per paper Nutrition, Ecology, Economics, Hydrology, Wildlife Habitat, will receive page proofs. These are to be returned to the Production Methodology, Taxonomy, Grazing Management, Soils, Land Recla- Editor, 445 Union Blvd., Suite 230, Lakewood, Colorado 80228, with- mation (reseeding), and Range Improvement (fire, mechanical chemical). in 48 hours after being received. If a problem arises that makes this Technical Notes are short articles (usually less than 2 printed pages) impossible, authors or their designates are asked to contact the reporting unique apparatus and experimental techniques. By invitation of Production Editor immediately, or production and publication may pro- the Editorial Board, a Review or Synthesis Paper may be printed in the ceed without the author’s approval of his edited manuscript. journal. Viewpoint articles or Research Observations discussing opinion or philosophical concepts regarding topical material or observational data Page Charges and Reprint Orders are acceptable. Such articles are identified by the word viewpoint or observations in the title. Authors are expected to pay current page charges. Since most research is funded for publication, it will be assumed that the authors are able to Manuscript Submission pay page charges unless they indicate otherwise in writing, when submit- ting a manuscript. When funds are unavailable to an author, no page Contributions are addressed to the Editor, Gary Frasier, Journal of charges will be assessed. Only the Editor will have knowledge of fund Range Management, 7820 Stag Hollow Road, Loveland, Colorado status of page charges; the Associate Editors and reviewers will accept or 80538. Manuscripts are to be prepared according to the instructions in the reject a manuscript on content only. Journal’s Handbook and Style Manual. If the manuscript is to be one of An order form for reprints is sent to one author with the page proofs. a series, the Editor must be notified. Four copies of the complete manu- Information as to price and procedure are provided at that time. The min- script, typed on paper with numbered line spaces are required. Authors imum order is 100; no reprints are provided free of charge. may retain original tables and figures until the paper is accepted, and send good quality photocopies for the review process. Receipt of all manu- Basic Writing Style scripts is acknowledged at once, and authors are informed about subse- quent steps of review, approval or release, and publication. Every paper should be written accurately, clearly, and concisely. It Manuscripts that do not follow the directives and style in Journal hand- should lead the reader from a clear statement of purpose through materi- book will be returned to the authors before being reviewed. A manuscript als and methods, results, and to discussion. The data should be reported number and submission date will be assigned when the paper is received in coherent sequence, with a sufficient number of tables, drawings, and in the appropriate format. photographs to clarify the text and to reduce the amount of discussion. Tables, graphs, and narrative should not duplicate each other. Manuscript Review Authors should have manuscripts thoroughly reviewed by colleagues in their own institution and elsewhere before submitting them. Peer Manuscripts are forwarded to an Associate Editor, who usually obtains review before submission insures that publications will present signifi- 2 or more additional reviews. Reviewers remain anonymous. These cant new information or interpretation of previous data and will speed reviewers have the major responsibility for critical evaluation to deter- JRM review process. mine whether or not a manuscript meets scientific and literary standards. Particular attention should be given to literature cited: names of Where reviewers disagree, the Associate Editor, at his discretion, may authors, date of publication, abbreviations or names of journals, titles, obtain additional reviews before accepting or rejecting a manuscript. The volumes, and page numbers. Associate Editor may also elect to return to the author those manuscripts It is not the task of Associate Editors or Journal reviewers to edit poor- that require revision to meet format criteria before the Journal review. ly prepared papers or to correct readily detectable errors. Papers not prop- The Associate Editor sends approved manuscripts, with recommenda- erly prepared will be returned to the author. tions for publication, to the Editor, who notifies the author of a projected publication date. Manuscripts found inappropriate for the JRM are released to the author by the Associate Editor. Manuscripts returned to an author for revision are returned to the Associate Editor for final accept - ability of the revision. Revisions not returned within 6 months, are con- sidered terminated. Authors who consider that their manuscript has