TABLE of CONTENTS: Vol

TABLE OF CONTENTS: Vol. 42, No. 6, November 1969

ARTICLES Invited Synthesis Paper 442 Remote sensing technology for rangeland management applications by Paul T. Tueller

General Interest 454 Motivation of Colorado ranchers with federal grazing allotments by E.T. Bartlett, R.G. Taylor, J.R. McKean, and J.G. Hof

Grazing Management 458 A test of grazing compensation and optimization of crested wheatgrass using a simulation model by Bret E. Olson, Richard L. Senft, and James H. Richards 468 Effects of stocking rate on quantity and quality of available forage in a southern mixed grass prairie by Rodney K. Heitschmidt, Steven L. Dowhower, William E. Pinchak, and Stephen K. Canon 474 Diet and forage quality of intermediate wheatgrass managed under continuous and short-duration graxing by M.L. Nelson, J.W. Finley, D.L. Scamecchia, and S.M. Parish

Animal Ecology 480 Association of relative food availabilities and locations by cattle by D.W. Bailey, L.R. Rittenhouse, R.H. Hart, D.M. Swift, and R.W. Richards 483 Herbivore effects on seeded alfalfa at four phryon-juniper sites in central Utah by Steven S. Rosenstock and Richard Stevens

Plant Physiology 490 Emergence and root growth of three pregerminated cool-season grasses under salt and water stress by D.M. Mueller and R.A. Bowman 496 Morphological and physiological variation among ecotypes of sweetvetch (Hedysurum boreule Nutt.) by Douglas A. Johnson, Timothy M.J. Ford, Melvin D. Rumbaugh, and Bland Z. Richardson 502 The effects of intra-row spacings and cutting heights on the yield of Leucaena leucocephala in Adana, Turkey by Tuncay Tukel and Riistii Hatipoglu 504 Gambel oak root carbohydrate response to spring, summer, and fall prescribed burning by Michael G. Harrington Published bimonthly-January, March, May. July, September, November Plant Ecology 508 Predicting peak standing crop on annual range using weather variables by Copyright 1989 by the Society for Range Management Melvin R. George, William A. Williams, Neil K. McDougald, W. James Clawson, and Alfred H. Murphy INDIVIDUAL SUBSCRIPTION is by membership in the Society for Range Management. LIBRARY or other INSTITUTIONAL SUBSCRIP- TECHNICAL NOTES TIONS on a calendar year basis are $56.00 for the United States postpaid and $68.00 for other coun- 514 Direct effect of parasitism by Dinarmus acutus Thomson on seed preda- tries, postpaid. Payment from outside the United tion by Acanthoscelidesperforatus Horn in Canada milk-vetch by A. Boe, States should be remitted in US dollars by interna- B. McDaniel, and K. Robbins tional money order or draft on a New York bank. BUSINESS CORRESPONDENCE, concerning sub- scriptions, advertising, reprints, back issues, and BOOK REVIEWS related matters, should be addressed to the Manag- ing Editor, 1839 York Street, Denver, Colorado 516 Viable Populations for Conservation Michael E. Soule (Ed.); Mammals 60266. of Arizona by Donald F. Hoffmeister; Rangelands: A Resource Under EDITORIAL CORRESPONDENCE, concerning manu- Siege. Proceedings of the Second International Rangeland Congress scriptsor othereditorial matters, should beaddressed Edited by P.J. Joss, P.W. Lynch and O.B. Williams; To Feed the Earth: to the Editor, 1839 York Street, Denver, Colorado 80208. Agro-Ecology for Sustainable Development by Michael Dover and Lee INSTRUCTIONS FOR AUTHORS appear on the M. Talbot; Wyoming Shrubland Ecology by Herbert G. Fisser. inside back cover of most issues. A Style Manual is also available from the Society for Range Manage- 519 Index, Volume 42 mentat theaboveaddress@$2.00forsinglecopies; 524 Table of Contents, Volume 42,1989 $1.25 each for 2 or more. THE JOURNAL OF RANGE MANAGEMENT (ISSN 0622-499X) is published six times yearly for $56.00 per year by the Society for Range Management, 1839 York Street, Denver, Colorado 80206. SECOND CLASS POSTAGE paid at Denver, Colorado. POSTMASTER: Return mth journal with addrosa chenge-RETURN POSTAGE GUARANTEED-to Society for Range Management, 1839 York Street, Denver, Colorado 89298.

Mana~lng Editor ASSOCIATE EDITORS PETER V. JACKSON Ill DON C. ADAMS RICHARD H. HART BRUCE ROUNDY 1639 York Street USDA ARS USDA-ARS 325 Biological Sciences Denver, Colorado 80206 Livestock & Range Research 8498 Hildreth Rd. East Building, Univ. Arizona Rt 1. Box3 Cheyenne, Wyoming 82009 Tucson, Arizona 85721 Editor Miles City, Montana 59301 PATRICIA G. SMITH RODNEY HEITSCHMIDT DAVID M. SWIFT Society for Range Management WILL BLACKBURN Box 1658 Natural Resources Ecology 1839 York Street N.W. Watershed Res. Center Vernon, Texas 76384 Colorado State University Denver, Colorado 80266 270 South Orchard Ft. Collins, Colorado 80523 PETE W. JACOBY, JR. (303) 3557070 Boise, Idaho 83705 P.O. Box 1858 PAUL TUELLER Book Rovfaw Edltor CARLTON BRITTON Vernon, Texas 78384 Range Wildlife 8 Forestry GRANT A. HARRIS Range 8 Wildlife Mgmt UNR 1600 Valley Road HOWARD MORTON Forestry and Range Management Texas Tech University Reno. Nevada 89512 2ooO E. Allen Road Washington State University Lubbock, Texas 79409 Pullman, Washington 991846410 Tucson. Arizona 85719 STEVE WHISENANT TIMOTHY E. FULBRIGHT Texas A8M Univesity College of Agriculture Dept. of Range Science Texas A81 College Station, Texas 77843-2126 POB 156, Sta. 1 Kingsville, Texas 78383

Invitational Synthesis Paper Dr. Paul T. Tueller is Professor of Range Ecology in the Department of Range, Wildlife and Forestry, University of Nev- ada Rena Paul has had over 30 years as student and professor working in the general area of Range Ecology and Management. He has initiated numerous studies concerned with range ecology, vegetation-soil relationships, range condition and trend, big game habitat management, brush encroachment, ecology of the pinyon- juniper woodland, mine waste reclamation, and remote sensing with emphasis on arid lands. Other qualifications include: commercial instrument rated pilot with over 3,000 hours, currently Associate Editor, JoumolofRanae. _ Management, Certified Range Management Consultant.

Remote sensing technology for rangeland management applications

PAUL T. TUELLER

Tbe future of rangeland resmwces development and manage- photography WM first used. The first black-and-white aerial pho- ment in dependent upon increased scientific capability. Remote tography at the usual scale of about 1:2O,ooObecame available for sensing technology can contribute information for a variety of range-resources investigations in 1935 and 1936 (Moyer 1950). rangeland resource management applications. In future we can Poulton’s history further describes the development of remote expect to see an increased number of profnsion~l range managers sensing as a science in the mid-to-late 1960s. The launch of Landsat with expertise in remote sensing. This training will include, in 1 in 1972 ushered in a new era extending remote sensing beyond air addition to principles of aerial photo interpretation, digttnl image photo interpretation into the realm of digital analysis of multispec- analysis technology, increased use of geographic information sys- tral and multitemporal data. tems, airborne video remote sensing, and the use of newly develop The collective area of rangeland is large; if forested ranges and ing high resolution systems. The data will be obtained from both natural vegetation in tropical savanna and tundra areas are aircraft and spacecraft. Appltcations will include inventory, eval- included, the total land area of rangelands may be as high as 47% of uation, and monitoring of rangeland resources and the incorporn- the global land surface (Williams et al. 1968). Due to the extensive tion of remote senstng data to support and improve the decision nature of rangelands and the recognized need to manage them at processes on the use, development, and management of nngelnnd low cost, remote sensing is considered to have significant promise lesOurce area*. for the future. Remote sensing is more than high quality color and color Key Words: Remote sensing, aerial photography, color infnred, infrared photographs and digital image manipulation. Rather it is, in itself, an art and a science. The science is provided mostly by iig, range managements g&graphic infknskm~ystems, airborne engineers, physicists and computer specialists who have increased video. our abilities to exploit information inherent in various regions “I The art and science of range management is being pushed to new wavebands of the electromagnetic spectrum. Range scientists con- heights as practitioners and scientists approach the future. Changes tribute specialized knowledge and interpretation. The proper in range management are inevitable. Expectations of high technol- interpretation and application of remote sensing is an art. Inherent ogy will be realized only when range professionals are able to apply in this is the importance of developing an understanding of the new scientific developments to important rangeland resource ecology of the landscapes and of the vegetation-landform-soif management problems. Remote sensing, the acquisition of infor- relationships as a basis for image interpretation. mation concerning an object or phenomenon without physical The basic logic of remute sensing is the logic of inference. If contact, is one such scientific discipline and the subject of this cause Q exists, then effect E1 will be observed, and if effect E1 is paper. Rangeland resource management will become strongly observed, then cause C1 must exist. In other words remute sensing dependent upon increasingly sophisticated, holistic approaches. interpreters can study certain features directly and other features Remote sensing, along with Geographic Information Systems only indirectly by inference or association. We are using the con- (GIS), can provide a fresh approach to the “se, development, and cept of surrogates wherein we identify and measun easily observed management of rangelands throughout the world. (via remote sensing) features that are related to more complex Remote sensing has been recommended for at least 30 years for features or phenomena that a range scientist or manager wishes to assisting with rangeland resources development and management identify, measure, and judge the significance of. For example, it is on a worldwide basis (Tueller 1982). A history of remote sensing possible to measure crown cover of shrubs on B large-scale vertical for range management is found in Poulton et al. (1975) and Poul- aerial photograph, but it is quite difticult, if not impossible, to ton (1985). These historiesgo back wellin to the 1930s whenaerial accurately measure height or biomass. Research in remote sensing

442 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 Visible

Fig. 1. l’he visible. near injrared. mid infrared, thermal or far infrared andmicrowave regions of the rkctromagnetic spectrum provide potential remote sensing applications for range management. is required to establish useful and unique inferential relationships displayed in a picture-like form such as a laser scan. Also by that are relevant to range management decision processes. combining various wavebands of the spectrum, we can produce a product similar to an aerial photograph but derived in an alto- Remote Sensing Procedures or Systems gether different way. Remote sensing information is derived from measurements of Much of the remote sensing data we are concerned with is electromagnetic radiation by air- or satellite-borne cameras, video obtained with a scanning radiometer, which by the use of a rotating cameras, ultraviolet and infrared detection apparatus; radar and or oscillating plane mirror can scan a path normal to the movement radio frequency receivers; the measurement of acoustical energy by of the radiometer. The mirror passes radiant energy to one or more seismographs, sonar and microphones; the measurement of nuclear detectors that record the energy from various wavelengths of the or ionizing radiation; and the measurement of force fields by electromagnetic spectrum. Resolution for a scanning radiometer gravimeters and magnetometers (Willow Run Laboratories 1978). system is expressed in terms of an instantaneous field of view The latter 3 forms of remote sensing have not been used extensively (IFOV). This denotes a narrow field of view designed into the for rangeland resource management. This paper will emphasize detectors of a scanning system, so that while as much as 120” may imaging and digital remote sensing in the visible, near infrared, mid be under scan, only electromagnetic radiation from a small area is infrared, far or thermal infrared, and microwave sections of the being recorded at any one instant. This ‘refers to the physical electromagnetic spectrum (Fig. 1). dimension on the ground within which one datum point, or spectral reflection or emission value, is recorded by the sensor system. Aerial Photography This area is referred to as a pixel (picture element) which deter- The most used form of remote sensing has been aerial photo- mines the minimum feature size that is represented by a unique graphy. Aerial photographs still provide the highest resolution and signature in the spectral data set. Normally ground features of capture the spatial and textural essence of the scene with greater interest are represented by the data in a few to many pixels. fidelity than any other procedure. Disadvantages include the cost Scale is an important concept related to resolution. Representa- of repeated coverage for change detection, including the costs of tive fraction scales (the ratio between the distance on the photo- film and processing, and the limited spectral sensitivity of conven- graph to the actual ground distance) are used to determine the size tional photography. of objects on a photo or image. Scales of remote sensing image Photogrammetry, the making of measurements from photo- products can vary from as large as 1: 100 to as small as 1:5,000,000 graphs includes stereoscopic viewing, which permits the measure- or smaller, all of which can be used by range managers. Digital data ment of height and the evaluation and interpretation of terrain can be used to produce hard copy images at a variety of scales, the features. Photo interpretation includes the use and evaluation of largest of which are difficult to interpret because of their low several important factors often referred to as the photo interpreta- resolution. tion principles: color, tone, texture, pattern, shadow, size, shape, and convergence of evidence (several different but related charac- Spatial, Temporal, and Speetnl ReaoIutIon teristics when combined lead to correct interpretations). In photo In addition to spatial resolution (related to scale), remote sens- interpretation the human mind acts as the master manipulator and ing also has aspects of both temporal and spectral resolution. integrator of information leading to an accurate interpretation. Spatial resolution is used in resource management, includmgrange For many rangeland uses thii approach is still the most fruitful. management, in a multistage sampling approach (Langley 1971). In multistage sampling, the remote sensing user evaluates the Ralolution and SeaIc resource fast on small-scale imagery or photography giving a Resolution is the ability of an entire remote sensor system, synoptic or large-area view. This enables one to examine the eco- including lens, antennae, display, exposure, processing, and other logical and land use patterns over distances of IO’s of kilometers. factors, to render a sharply defined image (Colwell 1983). For Images at incrementally larger scales provide increasing detail aerial photography it is usually considered in 1 of 2 ways: first as about smaller areas. Inferences from representative sites within effective ground resolution, which can be defined as the size an large scale images are extrapolated back to the smaller scales giving object must be before it can just be identified as a discrete entity on a summation of the total resource within the area of interest, an aerial photograph or image, and secondly as the number of providing sampling has been adequate and all categories of interest contrasting black and white line pairs of a given size and spacing have been included. Often this is accomplished using statistical that can be differentiated per milliieter or meter as viewed on a techniques such as sampling with probabilities proportional to standard test target. The term image in remote sensing is a broader sample size (PPS). term than “photograph.” Photographs are images, but images also Temporal resolution can be referred to as multidate remote include radar scenes, thermograms, video pictures, and other data sensing and is used for change detection or monitoring. Several

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 443 ,. VISIBLE+ REFLEC?lVE INFRARED bl

6. _ ,. PHOTOGFWHIC NemiOTmRAPHlC H H

50 -

40 -

SPOT iands 30 - +iel N

20 - I Landsat MSS Bands 4,516, 7 I 1 I J ‘O-/\/ Landsat Thematic Mapper Spectral Bands www e w * 0- 0.4 12.5

I I I I I I I I I I I I I I I I I I I I I I I 0.5 1 .o 1.5 2.0 2.5 wavelength pm

Fig. 2. The spectrum of o typic01 green plant showing areas of high and low reflectance and the location of the visible and refictive infrmid, the photographic and non-photographic wavekngths, and the La&at TM and SPOT wovebands. change detection algorithm alternatives are available including flown on Landsat 4 and 5 which has 7 channels 3 in the visible part image differencing, image ratioing, classification comparison, of the spectrum, 1 in the near infrared, 2 in the mid infrared or comparison of preprocessed imagery, and change vector analysis water absorption region and a thermal channel: the SPOT (Sys- based on changes between sampling dates in spectral appearance of teme Pour 1‘ Observation de la Terre) which has 3 channels, 2 in the the features in the scene (Jensen 1981). Two good examples in visible and 1 in the near infrared (Fig. 2); and the AVHRR range management are the assessment of vegetation change (range (Advanced Very High Resolution Radiometer), with 5 channels, 1 trend) and the evaluation of utilization differences brought about in the visible, 1 in the near infrared, 1 in the mid infrared, and 2 by differential use of various pastures in a grazing management thermal channels. All are multispectral imaging sensors. The system. AVHRR is a multispectral imaging sensor that was designed to Somewhat paralleling the use of optical filters in conventional permit detection and discrimination between clouds, land, water, photography, acquisition of the multispectral scanner data allows snow, and ice. one to much more finely control the wave-length window, or MSS data has pixels approximately 80 m on a side and the TM amount of the electromagnetic spectrum, that is used to establish pixels are 30 m on a side. SPOT data from the French Satellite has one datum point. The width of this spectral window is referred to as 20 m pixels. In addition SPOT has a single panchromatic black and spectral resolution. They can be very broad, i.e., 1 band can white band with 10 m resolution. The imagery from these 2 systems encompass all the green light or reflective infrared energy in a can be co-registered, giving effective 10 m resolution of color and single band or measurement (thus a single datum point), or each color infrared products. These systems have relatively wide spec- can be broken down into 2 or more bands, each with its own set of tral bands compared with the high resolution remote sensing sys- data points. These readings can be made for large or small IFOV’s, tems currently under development. The orbital characteristics of i.e., ground resolution cells (pixels). Multispectral reconaissance these result in repeat coverage of the same ground area at regular or multispectral sampling, refers to the concept of looking at the intervals between 9 and 18 days. In the case of SPOT, repeat same feature or polygon with several of these discrete bands or coverage of selected sites can be accomplished more often because wavelengths of the electromagnetic spectrum. Each band or com- the scanning instrument can be pointed to areas not directly bination of bands provides new or unique information about a underneath the satellite. resource feature. The NOAA weather satellites provide remotely sensed data at a scale of 1: 10,000,000 and resolution cells or pixels that are approx- MSS, TM, SPOT and AVHRR imately 1 km on a side from AVHRR. AVHRR data have been Numerous spacecraft, both past and present, have been used to used to study desertification in the Sahel (Tucker and Justice 1983; acquire remotely sensed data. Some of the systems presently in use Tucker et al. 1985) and to study large grazing piospheres in Austra- are the Landsat Multispectral Scanner (MSS) flown on Landsats lia (Graetz and Ludwig 1978). One advantage to this system is the 1,2, and 3 which has 4 channels, 2 in the visible part of the spectrum capability of obtaining data on a daily basis. A disadvantage is the and 2 in the near infrared; the Landsat Thematic Mapper (TM) low resolution. These data are useful where daily synoptic views

444 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 with multispectral data are required. An example on rangelands at We&co, TX are pacesetters in airborne video applications, would be to examine range readiness and fire fuel hazard for an especially for range vegetation studies and range management. In a entire mountain range or several mountain ranges as such parame- recent review article, Everitt and Escobar (1989) point out that ters change on a daily basis. another advantage of video is the capabiity of obtaining imagery in very narrow spectral bands (0.05-0.10 pm) and in the near and RadarSystem mid-infrared bands. They state that the main disadvantage of video Microwave systems have potential for rangeland applications. is its low resolution relative to aerial photographs. Typical tape These include both Side Looking Airborne Radar (SLAR) and recorder resolution for color and black-and-white video is 240 and Synthetic Aperature Radar (SAR). SLAR allows a relatively wide 300 lines, respectively, across the format tield, whereas 35mm slide swath to be imaged. SAR allows the creation electronically of long film resolution has 720 lines across the long axis of the format. The aperatures capable of greater resolution. These systems are useful primary reason for the low resolution of video is the recorder. because they are not constrained by either night (darkness) or Video cameras are available with at least 1,400 lines of resolution cloud cover. They create their own signals that are reflected back and a new super VHS recorder with over 400 lines of resolution will from the earth’s surface giving imagery that has utility for vegeta- improve video resolution (Lusch 1988). tion mapping and strong capabilities for looking at terrain Airborne video imaging systems have been evaluated for a var- features. iety of rangeland management applications. These have included Green (1986) used Shuttle Imaging Radar-A (SIR-A) data and distinquishing among plant species and mapping rangeland vege- found that during periods when most of the vegetation in a shrub tation (Everitt and Nixon 1985) and assessing grass production rangeland in South Australia was non-vigorous and spectrally (Everitt et al. 1986). Video applications for rangeland management homogeneous, the SIR-A data, as a surrogate measure of shrub will almost certainly escalate in the near future. cover, allowed the reflectance due to shrubs in Landsat data to be separated from the reflectance due to the intervening ground. High Spectral Resolutiam System Radar has potential to measure and monitor soil moisture and may Physicists and engineers have actively improved remote sensing be used in conjunction with other forms of remote sensing as instrumentation, while application scientists have been researching supplemental information for rangeland monitoring. Several other their feasibility and use. Thii has resulted in a new area of remote range management applications are feasible with radar data sensing that we might refer to as high spectral resolution remote although the cost of most systems and the acquisition of the data sensing. Three systems can be mentioned here AIS, AVIRIS, and currently tends to be prohibitive. HIRIS. All 3 of these systems have been developed at NASA’s Jet Propulsion Laboratory. AIS (Airborne Imaging Spectrometer) is Alrborne Seamrem an instrument that images 32 cross-track pixels simultaneously, Several valuable instruments have been developed to be flown as each in 128 spectral bands in the 1.2-to 2.4 pm region (Vane et al. multispectral scanners from aircraft. One such instrument is the 1984). This system is somewhat lacking in that no data are acquired Thermal Infrared Multispectral Scanner (TIMS), an aircraft- in the visible part of the spectrum. Preliminary land vegetation borne scanner providing six-channel spectral capability in the arid studies with AIS indicated that brightness was reduced by the thermal infrared (8.2-l 1.7pm) or that portion of the infrared that vegetation primarily due to shadowing. At periods of peak growth, has thermal or emitted heat sensitivity. This instrument was deve- saltbush and Great Basin sagebrush communities were found to loped at NASA’s Jet Propulsion Laboratory. Other airborne scan- have unique spectral curves (Ustin et al. 1985). ning instruments include the Thematic Mapper Simulators, TMS , The AVIRIS (Airborne Visible and Infrared Imaging Speo and NSOOl. The added feature of measuring surface temperature trometer) instrument, has been designed as a successive step in the has been used to map soils in arid regions (Miller et al. 1986). There process of developing systems for high spectral resolution remote appears to be some potential for discriminating soil textural fea- sensing of the Earth. AVIRIS covers the spectral region from 0.41 tures because of the ability of the system to identify emmissivity to 2.45 pm using bands IO-nm wide. This instrument consists of 4 minima for silicate minerals (Kahle and Goetz 1983). Langran spectrometers that scan the test site from an aircraft. At any one (1985) used TIMS data to monitor vegetation regrowth in primary moment the spectrometers are viewing a 20-meter square pixel. and secondary disturbance zones on Mount St. Helens. Studies This pixel is viewed simultaneously in 224 spectral bands. The data with TIMS and similar data sets suggest a potential application are recorded on a tape recorder for later analysis. Computer pro- whereby unique temperature-related soil and vegetation type maps cessing of the data will produce an image of the test site in any one of rangelands may be created and updated. of the 224 spectral bands, the spectrum corresponding to any pixel Video Systems or group of pixels in the scene (Porter and Enmark 1987). It may seem somewhat strange to turn now to a relatively low- And into the futum, HIRIS (High-Resolution Imaging Spec- resolution system such as airborne video. However, it seems clear trometer) is planned for the 1990s. This instrument will acquire that the future of resource management will be strongly influenced simultaneous images in 192 spectral bands in the dominant wave- by video remote sensing systems. Video cameras differ from film lengths of the solar spectrum, 0.4 to 2.5 micrometers, at a spectral cameras. They contain no tape or film and the image is relayed sampling interval of 10 nanometers. The ground instantaneous electronically to a video cassette recorder (VCR) for tape storage field-of-view (GIFOV) will be 30 meters over a 3Okilometer track. or to a TV monitor for real-time display. In addition a pointing capability will allow image acquisition up to Video has several attractive attributes, the most prominent of +60 degrees/ -30 degrees down-track and plus or minus 24 degrees which is the near-real-time availability of imagery. Also video data cross-track (NASA 1987). This latter feature may someday be very are relatively inexpensive to obtain. These factors can be very important to range managers interested in evaluating rangeland important whenever the application is very time-sensitive or when vegetation for suitability, phenology, carrying capacity, and range film and film processing is not available such as in many develop readiness since the cross-track pointing will also allow multiple ing countries. Examples of time sensitivity in range management viewing opportunities during one orbital revisit cycle, nominally 16 include surveillance and action to ameliorate wildfires or floods, days. recording range readiness/ phenology, and measuring vegetation High resolution systems will enhance the ability of remote sens- changes (monitoring) generated by intensive grazing management. ing procedures to scrutinize the vegetation with much greater J.H. Everitt and his colleagues at the USDA/ARS laboratories

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 445 finesse than heretofore available. For example, Dr. Chris Elvidge approaches to remote sensing became available. Computer-aided of the Desert Research Institute, University of Nevada System, has interpretation of remotely sensed data requires (1) the availability been pioneering the analysis of vegetation spectra of plant canopies of digitixed spectral (or multispectral) data, (2) advanced computer including rangeland species (Elvidge 1989). He and others have technology, (3) algorithms applicable to remote sensing, and (4) hypothesized that high spectral resolution data from aircraft can methodologies or strategies for analysis. These strategies have been be used to measure lignin, nitrogen, and both green and non-green slow to develop for resource management and very few are avail- (dry leaves, dry reproductive structures, bark, and wood) plant able for use in range management. materials. For big sagebrush (Arzemisia rridenrutu) Elvidge’s data Think of your TV set at home. It currently has about 500 lines of show great variation in the spectral response of gray bark, gray resolution, meaning that the resolving power of the system is such wood, gray leaves, brown wood, brown flowers and senesced that 500 line pairs can be identified on this medium. Such a system leaves (Fig 3). is made up of 500 by 500 data elements corresponding to IFOVs The future may well allow high resolution remote sensing sys- each having both spatial and spectral aspects. The spatial aspect tems to routinely gather spectral data concerning several species on defines the apparent size of the resolution cell and the spectral a rangeland site and provide an instant analysis of the quality, aspect defines the intensity of the spectral response for that cell quantity, and condition of rangeland and wildlife habitat forage (pixel) in a particular waveband. Visual interpretation of a group and browse. of like pixels leads the eye to identify discrete clusters, blobs, or polygons and judge them to be recognizable features on a digitized image representing real and meaningful elements of the landscape. I I I I I I Bffi SAGEBRUSH Multispectral data derived from scanners provide radiance or lArtemiaia- brightness information in each of several discrete wavebands within the electromagnetic spectrum for each pixel. These brightness levels are divided into quantizing levels. Tucker (1979) found that the current 256 quantizing levels of most satellite scanning systems was sufficient for satellite monitoring of vegetation resources. When several of these wavebands are looked at simultaneously we refer to the result as an n-dimensional spectral signa- / ture where n represents the number of spectral bands used. The signatures vary with the scanning device and its sensitivity to the wavelength being measured. 7 In addition to multispectral scanners, it is possible and feasible to digitize both aerial photographs and video images. This paves the way for computer enhanced digital image processing and inter- / pretation of these digitized images or any available image products. A range management example would be to make canopy cover measurements from large-scale digitized aerial photographs or video images with a classification algorithm used with a PC driven image processing system.

Vegetdion Indices For live green vegetation there is a significant differential in reflectance and absorption of electromagnetic radiation when / going from visible to near or mid infrared wavelengths (Fig. 2). / These differences have led to the development of several multispectral band ratios and indices that involve both the red/infrared / differences and coefftcients derived from several bands. These ratios and indices are indicative of the quantity and quality of green I I I I I I 0.4 0.7 1.0 1.9 1.2 1.0 2.2 and senescent vegetation. On rangelands, especially arid and semi-

WAVELENQTN (mlcrona) arid rangelands, soil background conditions and shadows often influence the signal received by a multispectral scanner which acts Fig. 3. Sagebrush spectra from Ehddge (1989) to complicate the use of these indices for evaluating vegetation One of the several problems for the future is the prodigious rate (Itteller 1987a). at which data can be generated by these systems. Fortunately, it Jackson et al. (1983) pointed out that an ideal vegetation index appears that faster and faster computers with ever-growing speed would be highly sensitive to vegetation, insensitive to soil back- and capacity will allow scientists, including range scientists, and ground changes and only slightly influenced by atmospheric path managers to analyze the data sets coming from these instruments radiance. On rangelands the ideal index would have the capability and make the information readily available and helpful in the of sorting out the influence of shadow and the influence of the great making of rangeland management decisions. However, it must be variety of leaf reflectances among the many species and species cautioned that the range manager/scientist must be an integral groups found as well as the standing dead vegetation and litter part of the computer/data/information system before the poten- (Tuelkr 1987). tial of remote sensing can be real&d. The multispectral or n-space indices originated from Kauth and Thomas (1976), who proposed a transformation that used linear Image Proceaaing combinations of 4 Landsat MSS bands to produce 4 indices: brightness, greenness, yellowness, and nonsuch. The soil spectra in According to Mausel(l985), it was not until the mid-1960% that Qdimensional Landsat MSS signal space were found to be distrib- all the elements needed to develop digital image processing uted along a plane, known as the plane of soils (Fig. 4). It is

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 meadow and riparian snow and ice, . . -. .- dry grassland *a.. l J .;:.. :.a .:.I: green J A grassland ‘* l . . l;* . . . .*. .I .*.. . . .*.* .yy; .* $:,:~‘::::~. I *.. 0. . *.** . ..*. l.*: ‘.-: I:... 0:. .* l . ..Z” . l l.*..*... .- shrub dominated 8 dense woodland S l . l .’ . .- 5 .*.I. ..* :: l J a .* . .* . . .:g :’ G= l .‘5’. I . *. f* l ‘** .. . “..‘ .*-\,....“ . .l .*:..,. ’ I? _..** ‘*‘. . . l .*. l l .* l l A l:- A soil brightness line m . l .l : l**up. .: f -*‘..**. ‘...*’ l . 0. 92 . ‘0... . ._*. ..* . l E ..‘.‘.-. .*.’ :‘. 8. l * 5 ..* l.s.** _$.:.* .*l l :t’t.;*‘.’ l.**:. +a. :.* .‘.-.*. .*. . . - water region . . b Red Reflectance

Fig. 4. An infrared/red ratio plot showing the distribution, in spectral space of components of a rangeland scene: aknse woodland; green grassland; meadow andripark~ dry gra&nd;shrub dominatedcommunities; theplane of soils or soil brightness line: bright objects such as snow, ice, white sand, bright bare soil or dry playa’ and water. sometimes referred to as a soil baseline in 2dimensional digital the best-fit line for lands with low vegetation cover. The MND space. This observation led to the soil brightness index (SBI) which maximizes the contrast in “greenness” while assigning a value of established the data plane for soils. The SBI is measured as the zero for no vegetation or dry vegetation. Huete’s (1988) soil- vector distance in the direction of the soil baseline. The greenness adjusted vegetation index (SAVI) is a similar modification of the vegetation index (GVI) is defined as the measured distance per- NVDI. A constant of 0.5 was recommended for sites with interme- pendicular (orthogonal) to the soil baseline towards a point of all diate amounts of vegetation. Ratio-based indices are relatively vegetation (Kauth and Thomas 1976). Figure 4 shows the distribu- independent of illumination intensity and eliminate the irradiince tion in infrared/red spectral space of pixels from various range measurements required for the calculation of reflectances (Pinter landscape types. et al. 1983). The perpendicular vegetation index (PVI) is a combination of Spectral response from any land area (including rangelands) and infrared and red bands (Richardson and Wiegand 1977). Thii any sensor depends on numerous variables that must be evaluated 2dimensional n-space equivalent of the GVI has been defined as alone and in various combinations. These characteristics include the orthogonal distance of a given spectral point from the soil such things as: (a) vegetation cover by species, (b) total vegetation baseline. The PVI responds to changes in both the quality and cover, (c) species or vegetation structure or geometry, (d) leaf quantity of vegetation. The PVI, unlike ratio-based indices, min- geometry, (e) bare ground percentage, (f) amount of shadow, (g) imizes the influence of the soil background for the assessment of cryptogam cover on the bare soil, (h) lichen cover on soils and green biomass (Elvidge and Lyon 1985). rocks, (i) algal mats, (j) microbial desert crusts, (k) gravel and/ or Ratio-based indices for vegetation assessment typically use the pavement, (1) standing dead vegetation, and (m) topography. red and near-infrared (NIR) bands. They contrast the high chloro- These factors along with such factors as season, time of day, solar phyll absorption region in the red against the high reflectivity in the zenith angle (Singh 1988), soil moisture, soil and foliage color, and NIR. The ratio between the NIR and red radiations was found to vegetation maturity all contribute noise to vegetation indices. be a sensitive indicator of green biomass (Tucker 1979). These To understand these variables there is a recognized need to indices include the ratio vegetation index where RVI = NIR/ red carefully examine the spectral characteristics of various rangeland and the normalized difference vegetation index (NDVI) where ND scenes and components of those scenes. For example, Asrar et al. = (NIR + red)/(NIR - red). The ND increases as the vegetation (1985) found that burned and unburned tallgrass prairie grass becomes greener or more dense. A modified normalized difference canopies showed distinctly different, diurnal and seasonal, spectral (MND) is a ratio based index where MND = (NIR -(1.2*red> reflectance characteristics in the visible and infrared regions of the )/(NIR + red). Tucker (1979) and Jackson and others (1983) give spectrum. Musick (1984) found that green vegetation indices thorough descriptions of ratiebased indices. tended to remove the influence of shrubs in New Mexico mesquite/ - Paltridge and Barber (1988), in a study of arid lands in Australia, grass vegetation that were green in June, giving an estimate of grass produced the MND by modifying the NDVI to include the slope of cover. Tueller and Oleson (1989) described fluctuations in radiance

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 447 values of pixels in arid shrubland vegetation depending upon Rangeland Vegetation Mapping differences in ground target, latitude, time of year and time of day. Range managers are interested in the distribution and condition Pixel luodelllng of the vegetation and forage base as it occurs in space and time. Areas of rangeland vegetation can be mapped from space or air- Pixel modelling can be defiied as any procedure whereby the spectral mixture inherent in an individual pixel can be separated craft altitudes with reasonably high levels of accuracy using mul- into various known components. Several approaches have been tispectral data and image processing systems. Photo interpreters used. Gladwell (1982) used reflectance radiometry and the resul- with training in range management, plant ecology and soils, and tant wavelength spacing and shape of the absorption features of with field experience in the area in question, can map the various various clay minerals to separate component spectra from pixels plant communities or ecological sites. These interpreters, after gaining experience can effectively use vertical aerial photographs using a deconvolution procedure. Pech et al. (1986) used a class of at various scales and stero viewing to map homogeneous polygons statistical models called mixture models as a framework for the analysis of &lectance from multi-component surfaces. bounded by ecotones that represent vegetation communities. Adams et al. (1982) compared airborne or TM pixel samples In the Great Basin Tueller (1979a) found that a scale of 1: 10,000 with ground-obtained spectra in terms of their proportionate mix is optimum to map at the habitat type or ecological site level. However, many of these same features can be mapped at resource of ground level spectra that approximate the reflectance characteristics of pixels in the sample. Wilson and TueIler (1987) used a photographyscales(e.g., 1:24,OtXl).Photo interpmtation of l:l,OOO,OOO similar approach and found that the composite ground signatures scale Landsat 1 images allowed the mapping of the major vegetation zones in the Great Basin (Tueller et al. 1975). One can develop indicate that spectrally dark components exist in rangeland plant a map with good accuracy for a given area with remotely sensed communities which decrease the brightness of a scene measured data obtained at a particular time and date. lf, however, the data from the air. Shadow and litter are presumed to be the primary sourcea of darkening. are projected in either space or time the classifiitions or mapping accuracy is significantly reduced. This is because we lack under- Huete (1986) adapted a factor-analysis inversion model to standing of the spectral characteristics of our relatively complex decompose a data set of spectral mixtures into a sum of unique rangeland scenes (Tueller 1987a). reflecting components weighted by their corresponding amounts Machine processing techniques to map and evaluate range vege- Pech et al. (1986) studied how the landscape components, each tation communities begin with the creation of spectral class statis- covering a fraction of the total area, contribute to the measured tics of the pixels representing the area of interest (McGraw and reflectance on rangelands in Australia. Heihnan and Boyd (1986), Tueller 1983). Three basic methods used to create spectral class Richardson and Wiegand (1977), and Huete et al. (1984) all consi- statistics are supervised, unsupervised, and a mixed approach or dered the effect that soil background has on the vegetation component of the pixel as they attempted to discrimiite bare soil from guided clustering (Rohde 1978). In a supervised approach, training windows are designated in the low vegetation densities typical of arid rangelands. data set. These consist of a group of pixels, known to represent a Development of improved procedures for evaluating individual range plant community based on field observations, which have pixels should lead to improved applications of remote sensing to been selected and related to ground data. Statistics describing these rangeland management problems. An example is the potential to windows are generated by the computer (mean and standard devia- look at a rangeland scene on 2 dates and determine the change that tion) and then extrapolated over the entire area being mapped and has taken place, either in terms of changes in accumulated biomass, a classification and map are derived. A drawback to this approach level of productivity, degree of use or changes in the composition of is that a significant number of pixels are usually not classified dominant species that can be indicative of range trend. In this latter because it is difficult to locate and identify homogeneous training case the reference might be to modelling pixels to determine if the sites for all plant communities of interest. Even in highly homoge- spectral information can give a clear picture of the relative propor- nious plant communities there is still considerable pixel to pixel tions of green woody vegetation, green non-woody vegetation, variability or noise to contend with thus leading to a high percen- standing dead vegetation, litter, bareground, gravel, rock and tage of misclassified pixels. other parameters that are related to ecological conditions and Unsupervised classification approaches have worked best on trend. rangelands. Pixels are clustered into groups of pixels with similar Rangeland Applications spectral response. The clustering is controlled by setting maximum standard deviation, minimum distance between cluster centers, Rangeland applications have been described in several papers minimum number of pixels allowed in a cluster or the maximum (Cameggie 1968; Haas et al. 1975; Maxwell 1976; Poulton 1975; number of spectral classes that can be generated. This procedure Thompson 1977; Tueller 1979b, 1982, and 1983; Poulton 1985). classifies most pixels and can be repeated until the classification is Applications of remote sensing are best thought of as existing as satisfactory, i.e., can be demonstrated to match the features of part of a remote sensing cyck or a list of valid steps (Tueller 1983): interest on the real landscape. The idea is to create spectral classes that represent a specific polygon or set of polygons on the ground. I Define the probkm and seek infommtion. These polygons represent what are termed specrrul chases which II Determine the appropriate remote sensors. III Acquire the remotely sensed data. have unique spectral characteristics within the limits that have IV Correlate the data with ancillary ground data. been set for the wavebands used. The interpretation question is V Analyze the data for its information content. whether or not these spectral classes actually represent what may VI Report and/or publish the information. be termed information classes, i.e., discrete polygons on a range VII Interpret and use the information; landscape that represent specific plant communities or other fea- or as systematic models relating ancillary ground data to remotely tures with unique attributes. Information classes can represent sensed data (Maxwell 1976). Poulton et al. (1975) described the plant communities, grazing allotments, range seedings, succes- broad functions of resource allocation and management on range- sional communities (sites of a known range condition) or other lands and described training requirements and procedures for features. operational work flows involving remote sensing. Still the number of useful and fully applied procedures or techniques is appallingly low.

448 JOURNAL OF RANGE MANAGEMENT 42(6). November 1989 Other HangeIand Management Applkdom mentation changes as a result of upstream and rangeland watershed What else can remote sensing do for range managers? In range activities can be assessed using various spectral signatures. Remote management it is important to know what is happening to individ- sensing can provide spatial and temporal information on proper- ual species. For this reason considerable technology has developed ties of soils related to erosion, hydrology, and productivity (Ritchie around small-format cameras with either 35mm or 70mm format 1983). (Cameggie and Reppert 1969, Heintz et al. 1978, Cameggie et al. Wildlife habitat parameters in rangeland environments can 1971, Reppert and Driscolll970, Poulton etal. 1975, Tueller 1978, often be best evaluated and measured using remote sensing proce- Waller et al. 1978, Everitt et al. 1980, Meyer et al. 1982, Everitt dures (Tueller 1980). Both terrestrial and wetland habitats can be 1985 and Tueller 1987b and Tueller et al. 1988). These cameras so assessed. Examples of the former include mapping of deer provide photographic scales from 1500 to 1:2,000 or smaller and winter ranges with stands of bitterbrush, mountain mahogany are useful for identifying many species and making detailed mea- habitats or patches of aspen used for summer escape cover. Tueller surements. However, since the format is small and the scales large, et al. (1978) used Landsat digital maps to evaluate emu habitat in limited area coverage requires creative sampling to provide ade- West Australia. Asherin and associates (1978) used land cover quate data and make valid projections concerning rangeland areas maps emphasizing vegetation for predicting avian use of sections of interest. of land in southeastern Montana. Examples of the latter include Based on field experience we can use aerial photographs to color aerial photographs useful for marshland evaluations (Seher identify features, judge their significance, measure their number or and Tueller 1973) and for riparian area management (Hayes 1976; extent (invenrory), and determine over time using subsequent pho- Batson et al. 1987). tographs whether or not significant changes have taken place (monitoring). Both black-and white-panchromatic (photographs Geographic Information System taken where the emulsion is sensitive to the total range of the visible Geographic information system (GIS) technology is relatively part of the spectmm), color, and color infrared photographs are new but is growing extremely rapidly. A GIS is an information used. Color and color infrared film studied in concert usually technology system which stores, analyzes, and displays both spa- provides more total information than black-and-white, color, or tial and non-spatial data (Parker 1987). Furthermore, spatial data color infrared alone. The added information in color products occur in 3 forms, points, lines, and polygons or areas. All features justifies the extra cost for most uses on most rangelands. on a range landscape can be reduced to one of these categories. Grazing management can be evaluated and monitored using This technology has evolved as a result of the need to use mappable remote sensing imagery. Repeat photography or imagery either information to make decisions concerning preservation of land with film cameras, video cameras, or digital image products at resources based on a range of institutional, political, economic, appropriate scales can provide baseline information on range read- and environmental data concerns. GIS are powerful tools for iness, utilization, distribution of livestock, and other parameters of integrating and analyzing data derived from remotely sensed imag- interest. Pickup and Chewings (1988) used Landsat imagery along ery interpretations, soil surveys, vegetation maps, land ownership with animal distribution models to estimate the distribution of maps, utilities maps, water resources, geology, mining, and many grazing and patterns of cattle movement in large arid zone pad- other potential themes that can be presented spatially. These geo- docks in Australia. Applications to intensive grazing management graphically referenced data sets are spatially registered so that systems are exemplified by the ability to follow the influence of multiple themes of data can be quickly compared and analyzed changes in grazing of the vegetation. together. Vegetation productivity and biomass levels are of intense inter- Because many GIS applications are in their infancy, a large est to range managers. Using Landsat MSS data McDaniel and proportion of time is currently expended on data base creation Haas (1982) found that the GVI was highly correlated with wet (Johnston 1987). This is a modelling process wherein the manager green yield, dry green yield and cured vegetation cover on is creating digital, spatially oriented maps that are easily accessed. mesquite-grass vegetation. They concluded that quantitative mea- The focus is on the development of land management models that surement of rangeland vegetation condition can be made from can integrate several submodels created to analyze individual Landsat MSS data. management objectives, e.g., an assessment of rangeland monitor- Remote sensing has also been used to map and inventory range- ing data in relation to land ownership, class of livestock, wildlife land soils. Westin and Lemme (1978) successfully mapped soils populations, and associated vegetation changes warranting man- although they were acutely aware of the influence of the vegetation agement alterations. on the spectral signatures. Landsat has been used to distinguish For range management we might overlay, one-on-the-other, between eroding, stable, and depositional soil environments in spatial data on vegetation, soils and soil suitability interpretations, central Australia (Pickup and Nelson 1984). Their procedure water resources, roads, fences, improvements, wildlife summer and worked best when they there was a uniform level of greenness in the winter ranges, grazing management systems, use maps, range con- vegetation. Attempts have been made to predict soil loss from dition maps, carrying capacity or productivity data, ecological space-acquired remotely sensed data (Spanner and others 1983) sites, habitat types, and land ownership. Many other such themes but the techniques require additional research before they can be can be looked at individually or collectively to aid the range deemed successful. Gully erosion on rangelands is probably best manager to place all conflicting uses and data types quickly and measured with vertical stereo aerial photographs (Welch et al. efficiently into a proper perspective for rapid efficient interpreta- 1984). tion, evaluation and action. Watershed studies on rangelands can be enhanced by remote H. Dennison Parker, whose leadership and insight into GIS sensing. Zevenbergen (1985) found high correlations between systems and their utility for natural resources including range- rangeland runoff curve numbers and various reflectance indices lands, has quoted Light (1986) who analyzed the storage require- obtained from Landsat MSS data. The boundaries of watersheds ments for a digital database containing all 54,008 7.5’ quadrangle can be easily mapped on various scales of aerial photographs. maps required to cover the lower 49 states. Light concluded that Color infrared aerial photographs or digital spectral maps can be 1014bits of data would be required for a ground spatial resolution used to map riparian vegetation as well as areas of recharge and of 1.7 meters, and that all the resulting data could be stored on 4000 discharge in water balance studies. Water quality in streams, sedi- optical disks smaller than a phonograph record. ‘What’s the signif-

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 449 icance to global assessments? Well, if you compute storage tion overkill. However, smaller scales, such as Landsat images requirements at hectare ground resolution, a level which would be have been very useful. In most developing countries space or very useful in many developing countries, a similar spatial data satellite remote sensingdata is the only source of new or up-to-date base for the whole continent of Africa could be stored on 5 optical information about natural resources. In this sense the data is disks.” These data would include all the rangelands, and such invaluable for rangeland resource inventory, development, and storage levels would allow additional themes to be stored and monitoring. For the near future, satellite imagery, generally visu- retrieved as management levels increase. ally interpreted, will offer the best option for most developing Caution must be taken to see that only correct and necessary countries to document the condition and extent of their natural data sets are entered into the GIS and that the tendency to place resource base and to monitor their progress or failures in husband- everything available into the data base does not get out of hand. ing and developing range resources (personal communication, Also personnel retraining requirements present a sign&ant prob- C.E. Poulton 1989). lem. However, GIS systems are certainly going to be a large part of Remote sensing with both aerial photography and satellite remote sensing data analysis. Considerable rangeland manage- imagery, singly and in combination, is being applied to range ment data will be spatially displayed, overlayed, and importance management problems in many Third World countries such as given to the juxtaposition of the many data points as managers use Kenya (Isavwa 1988), Niger (Hiemaux 1988), India (Mehta 1988), this powerful tool to quickly and accurately make management Egypt (Gad and Daels 1986), Botswana (Ringrose and Matheson decisions. 1983), Sudan (Heilkema et al. 1986), Senegel (Tucker et al. 1985), Tanzania (Poulton 1979), Mauritania (Dalsted 1988); and many Economic Considerations other countries. Little has been written about the costs of remote sensing with the Poulton’s (1979) plan for the use of Landsat imagery in Tanzania exception of the contracting of aerial photography. Current costs, is excellent although its implementation is made difficult by social including aircraft mobilization costs, film processing and special and political constraints. His study included a 1: 1,ooO,OOO mosaic products can be agreed upon ahead of the job to be done by the of the region with climatic overlays and an interpretation, by Land numerous aerial photo firms around the world. Costs are usually Systems, of land use capability for both pastoralism and agricultu- quite reasonable unless there is a need for repeat photography ral cropping. Interpretation was based on knowledge of 550 Land which might be required for rangeland monitoring and even this Units (vegetation-landform complexes). Poulton concluded that can often be reasonably costed out. A recent consideration for little benefit was realized from the study for a number of reasons, aerial photo acquisition indicates that one firm would charge mostly not related to the professional quality of the work done. approximately $3,000 to provide 200 exposed and processed posi- The reasons were attributed to a lack of direct involvement of tive transparencies in a 9” format. These costs included aircraft country nationals, to the failure to provide follow-up continuity, mobilization costs, cameraman, pilot, aircraft (including climb and to the level of understanding that was possible to create in the and descent times), color films, and film processing. These 200 short time span of the study. frames could cover a large or small area depending upon the length Dalsted (1988) diicribed a Landsat/ GIS approach for Maurita- of the lens and flight altitude. nia but considered it a first step in assessing the natural resources of For repeat coverage digital multispectral information on com- the study area. Three important overlays were produced, forest puter compatible tapes may be less expensive. Much of this cost regeneration, soil erosion hazard, and rangeland carrying. Both was, previously, borne by NASA and NOAA. However, if one studies suggest that good field data and suitable resource classifica- were to amortize the cost of the spacecraft development, system tions must be developed within the country before the Landsat development, launch, etc., the cost would likely be prohibitive. imagery can be adequately interpreted. With appropriate institu- Since privatization of Landsat the costs of Landsat computer tional organization and efforts of newly trained scientists and compatible tapes have skyrocketed. managers, the applications of remote sensing will someday become The cost for a single scene for a single date (7 TM bands) is, as of commonplace throughout the world and, at the same time, serve to this writing, S3,600, which can become costly if multiple dates of improve the management of rangelands in developing countries. imagery are required for a range management application. Even a 512 by 512 pixel portion of a scene down-loaded to a floppy disk Training and Education for use on a PC is remarkably expensive at $600 for the 7 bands. It An impediment to rapid application of remote sensing science to seems clear that one must be careful in sampling and know pre- rangeland resourcc8 management is the lack of individual workers cisely the areas one needs to study or evaluate before purchasing educated in both remote sensing and range management. Very few either digital, photographic image, or satellite computer compati- range management professionals receive training in remote sensing ble tapes or disks. This leaves video as possibly the least costly (Tueller 1983). Those that have such training often do something remotely sensed data for rangeland management. The principal elseratherthanmmotesensingwhenemployed inrangemanagement. cost for a video system once the initial capital expenditure for There are 2 or 3 solutions; one is inservice education of range equipment, including camera, recorder, etc., has been met, is for conservationists or scientists already working in the field. The idea the plane and pilot. A multispectral video camera system can be would be to develop a skilled photointerpreter and remote sensing obtained for twenty to thirty thousand dollars (personal communi- analyst with expertise in landscape ecology and experience in cation, J.H. Eve&t). digital image processing and GIS. Such an analyst should be able Remote sensing, either aerial or satellite, has the potential to to support the information needs of his management colleagues. reduce the ratio of field to office time. For many range manage- He or she would conduct the surveys and analysis and often make ment requirements, this would constitute a real saving of man- inferences about successional status, or range condition and trend, power and field mobilization costs. and interpret the impacts of resource use and management (Poul- ton 1975). A second solution would be to provide range students Developing Countries with a working knowledge of photointerpretation, remote sensing, Surveys for management of the vast rangelands of the Third and GIS while they are still in school although this might entail a World should be done in scales between 1200,000 and 1:250,000 program going beyond the traditional 4 years. (ZoMeveld 1978). Larger scales tend to just produce an informa- A final and least desirable approach would be to have range

460 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 managers work closely with remote sensing spcciahsts. This latter CoIweIl, R.N. (edItor in CM@. 1963. Manual of remote seensing,Second approach has not worked well. Remote sensing spcciahsts with Ed.. Vol. I. Amer. Sot. Photogmmm.. Falla Church. Va. backgrounds in physics, computer science, or engineering, who are D&d, K.J. 1988. The use of a &dsst-based soil ai vegetation survey and graphic information system to evaluate sites for monitoring deserti- not knowledgeable of vegetation and soil science, end up making E&ion. Desert. Contr. Bull. 16:2&26. many of the image analysis decisions as well as the range manage- Elvidge, C.D., and R.J.P. Lyon. 1985. InfIucnce of rock-soil spectral ment interpretations of the data. This should be strenuously variation on the assessment of green bio-. Rem. Scns. of Environ. avoided. It just doesn’t work. Those who require the information 17:265-279. should use the remote sensing procedures themselves to obtain the BIvidge, C.D. 1989. Visible and near infrared refkctance characteristics of data which they will then interpret based on sound ecological and dry plant materials. Inter. J. Rem. Sens. (in Press). Eve&t, J.H., A.H. Gebemam, MA. Aladz, ud R.L. Bowen. 1989. Using range management experience and familiarity with the field prob- 70-mm aerie photography to identify rangeland sites. Photogramm. lems and locations involved. Eng. and Rem. Scns. 4731357-1362. The adaptation of remote sensing science to the solution of range Eve&t, J.H. 19gS. Using aerial photography for detecting blackbrush management problems will have to await the further training of (Acucicr rig&&la) on south Texas rangelands. J. Range Manage. individuals in both remote sensing and range management. It is 38:u8-231. important to provide continuity of personnel capable in remote Everitt, J.H., and P.R. Nixon. 1985. Video imagery: a new remote sensing tool for range management. J. Range Manage. 38:421424. sensing between the inventory and monitoring work and the pro- Eve&t, J.H., M.A. Huney, D.E. hobar, P.R. Nixon, and B. Pinkerton. cesses of information use and decision making. Emphasii must be 1986. Ameasment of grassland and phytomaas with airborne video imag- given to key training elements of range management, vegetation ery. Rem. Scns. of Environ. 21:299-306. science, plant ecology, soils, and vegetation-landform-soils rcla- Everttt, J.H., and D.E. kobar. 1989. The status of video systems for tionships. If understanding of these basic elements is lacking, the remote sensing applications. Proc. 12th Bienn. Workshop on the Use of interpretation job simply doesn’t get done. Color Photography and Viicography in the Plant Sciences and Related Fields. Reno, Nevada. (In Press). Conchwion and the Future Gad, A., and L. Daeb. 1986. Aanessment of soil degradation in an arid region using remote sensing. Proc. IGARSS’ 86 Symposium, Zurich. p. The future of rangeland remote sensing is a bit hazy. It seems fair 1241-1246. to conclude that research concerning the spectral characteristics of Gladwell, D.R. 1982. Application of reflectance spectrometry to clay min- heterogeneous rangeland scenes using high resolution systems, eral determination in geological materials using portabk radiometers. ground radiometers, and ancillary data will lead us closer and Proc. Inter. Symp. Remote Scns. Environ., Fort Worth. closer to an ability to use remote sensing to quickly and efficiently Graetr, R.D., and J.A. LudwIg. 1978. A method for the analysis of pio- sphere data applicable to ra& assessment. Aust. Range. J.-1:126-i36. measure many parameters of interest. Gretn. G.M. 1986. Use of SIR-A and Landsat MSS data in manoinn shrub Near real time video systems coupled with digital image analysis and’intershrub vegetation at Koonamorc, South Au&a. khoto- approaches and GIS data bases will become routine in the next few gramm. Eng. and Rem. Sena. 52659-670. years as range managers realize the great potential of these systems. Hau, R.H., D.W. Mg, J.W. Row, Jr., and J.A. S&U. 1975. Moni- Perhaps airborne video remote sensing technology will be the toring vegetation conditions for Landsat for use in range management. p. single most useful new technology for rangeland applications. As 507-568. In: Proc. NASA Earth Resources Symp. NASA. Houston, TCxaS. range managers become educated or trained in and feel comforta- Haam,R.H. 1986. A new view for resource managers. Rangelands. 8:99-102. ble with remote sensing data and PCs, then high-technology Hay- F. 1976. Application of color infrared 70 mm photography for remote sensing for range management will have come of age. assessing grazing impacts on stream-meadow ecosystems. Sta. Note 25, One can visual& range conservationists obtaining the majority Univ. Idaho, Forest, Wildl. and Range Exp. Sta.. of their required management information from remote sensing HeiIkema,J.U,S.D.Prince,andW.L.Astk.19(16.Rainfallandvegetation instruments within, say, 20 years. Poulton (198 1) reminded us that monitoring in the Savanna Zone of the Democratic Republic of Sudan using the NOAA advanced very high resolution radiometer. Inter. J. remote sensing is not a panacea and that its effective use increases Rem. Sena. 21499-1514. rather than decreases the demand that the analyst thoroughly Hellman, J.L., and W.E. Boyd. 1986. Soil background effects on the understand the ecology of the landscapes with which he or she spectral rcsponsc of a three-component rangeland scene. Rem. Sens. of works. And finally, Haas (1986) has cautioned us that to realixe Environ. 19127-137. these applications, there will have to be achieved support in the Helm, T.W., J.K. Lewis, and S.S. Walter. 1978. Low-level aerial photo- administrative and managerial echelons. graphy as a management and research tool for range inventory. J. Range Manage. 32:247-249. Literature cited Hiemaax, P. 19gg. Application of remote sensing techniqm for the quan- tification of range resources changes in the Sahel. ALtr. Vol. I, Thiid Adams,J.D., M. Smttb, and J.B. Adams. 19g2. UK. of laboratoryspectra Inter. Range. Congr., Range Manage. Sot. of India. Vigyan Bhavan, for determiningvegetation asscmblagca in Landsat images. Proc. Inter- New Delhi. p. 36. nat. Symp. Rem. Scns. of Environ., Fort Worth., Dec. Huctc, A.R. i986. Separation of soil-plant spectral mixtures by factor Ash&n, D.A., J.E. RoelIe, and ILL. Short. 1978. Characterization of analvtds. Rem. Scns. of Environ. 19:237-251. terrestrial vertebrate habitats using remotely sensed imagery. Procccd- Huete;A.R. 1988. A soil-adjusted vegetation index (SAVI). Rem. Scns. of ingn Pccora IV Symposium. Oct. Sioux Falls, South Dakota. pp 274-279. Environ. 25:295-309. Asrar, C., R.B. Myneni, Y. LI, and E.T. Kanemuu. 1985. Measuring and Huete, A.R., D.F. Post, and R.D. Jackson. 1984. Soil spectral effects on modeling spectral characteristica of a tall grass prairie. Remote Scns. &pace vegetation d -nation. Rem. Scm. of Environ. 15:155-165. Environ. 21:143-155. Isavwa, L.A. 1988. Inventory and monitoring of rangeland rcsounzca in Batum, F.T., P.E. Cuplh, and WA. Crbco. 1%7. The use of aerial Kenya using remote sensing techniques. Abstr. Vol. I, Third Inter. photography to inventory and monitor riparian areas. USDI-BLM, Range. Congr. Range Manage. Sot. of India. Vigyan Bhavan, New Riparian Area Management, Tech. Ref. 1737-2. Delhi p. 36. Carneggk, D.M. 1968. Analysis of remote sensing data for range resource Jackmon, R.D., P.N. Stater, P. Phter Jr., 1983. Discrimination of growth management. Annu. Prog. Rep. Officea of Space Sciences and Applica- and water stress in wheat by variousvegetation indices through clear and tion, NASA, Washington, D.C.. turbid atmospheres. Rem..Scns. of E&iron. 13:187-u)8. - Carwggie, D.M., and J.N. Reppert. 1969. Large scale 70 mm aerial color Jenwn, J.R. 1981. Urban change detection mapping using Landsat digital photography. Photogramm. Eng. 39249-257. data. Amer. Cartographer. 8:127-148. Cm D.M., D.G. WiIcox, and R.B. Hacker. 1971. The use of large Joiuuton, K.M. 1987. Natural resource modeling in the Geographic scale aerial photographs in the evaluation of western Australia range Information System Environment. Photogramm. Eng. and Rem. Scns, lands. Tech. Bull. 10. Dep. Agr. West. Aust., Perth, Wcat Au&.. 53:1411-1416.

JOURNAL OF RANGE MANAGEMENT 42(6). November 1989 451 Kehle, A.B., end A.F.H. Goetz. 19113.Mineralogic information from a new Poulton, C.E. 1979. Uses of Landsat imagery in resource analysis of the airborne thermal infrared multispectral scanner. Sci. 22224-27. Masai steppe region in northern Tanxania. Proc. of Rem. Scns. for Kautb, R.J., and G.S. Thomas. 1976. The tasseled-cap-a graphicdescrip- Natur. Res.; A Symp. Moscow,Idaho pp 361392. tion of the spectral-temporal development of agricultural crops as seen PouRon, C.E. lnl. The role of remote sensing in making decision about by Landsat. Proc. Machine Processing of Remotely Sensed Data. IEEE natural resources. Sixth Internat. Training Course in Rem. Sens. Appl. Cat. 76 CH 1103-l MPRSD. LARS. p. 4b-41~51. for Past. and Range. Dcv. Rome. May p. 81-102. LangBy, P.G. 1971. Multistage sampling of earth resources with aerial and Poulton, C.E. 19gS. Evolution of remote sensing in range management, space photography In: Colwell, R.N. (Ed) Monitoring Earth Resources Speculations on its future. Proc. of Pccora 10 Rem. Scns. in Forest and from Aircraft and Spacecmftz NASA SP-275, p 129-141. Range Res. Manage. Fort Collins, Colo., Aug. pp 2-17. Langran, KJ. 19lM. Monitoring vegetation recovery patterns on Mount St. Reppert, J.N., and R.S. DrIscolI. 1970.7~mm Aerial photogranhv-A Helens using thermal infrared multispectral data. The TIMS Data User’s remote sensing tool for wildland research and manag~ment.%t -USDA, Workshop Nat. Space Tech. Lab.. MississiDDi. JPL Pub. 86-38. Forest Service. Range and Wildlife Habitat Evaluation. a Rcs. Svma. Light, D.L: 1986. hsa storage estimates fb; the digital mapping era. pp. 190-193. USDAForcst Serv. Misc. Pub. 1147. ’ ~- or Photogramm. Eng. and Rem. SCM. LII:419-425. Rl&ardson, AJ., and C.L. Wiegend. 1977. Distinguishing vegetation from Lusch, D.P. 1988. Super VHS for improved airborne CIR videography. soil background information. Photogramm. Eng. and Rem. Scns. Proc. 1st Workshop on Videography._ _ _ Amer. Sot. Photogramm. and 43:1541-1552. Rem. Scns. p 23-29: RIngrose, S., and W. Mathon. 19(13.Descrtification in Botswana: pro- Mawl, Paul W. 1985. Characteristics and techniques of computer assisted gress towards a viable monitoring system. Desert. Contr. Bull. 96-8. processing of spectral data. DD. 292-309 In: The Surveillance Science. RIteIds+ J.C. 1)03. Remote sensing and erosion research. Proc. Natur. Rcs. Remote S&sing of Environ&tit. Robert K. Holx(cd). Second Ed. John Model. Symp. Pingrcc Park, Colo. p 449450. Wiley and Sons, Inc. Rokde, W.G. 197C.Digital image analysis techniques required for natural Maxwell, E.L. 1976. A remote rangeland analysis system. J. Range Man- resource inventories. AFIPS Conf. Proc. Vol. 47. USDI Gcol. Surv. age. 296672. EROS Data Center, Sioux Falls, S. Dak. McDau&I, K.C., sod B.H. Hau. 19g2. Assessing mesquite-grass vcgcta- SIngh, S.M. 1988. Lowest order correction for solar zenith angle to Global tion from Landsat. Photogramm. Eng. and Rem. Scns. 48:441450. Vegetation Index (GVI) data. Int. J. Rem. Scns. 9(10 & 11):1565-1572. McGraw, J.F., and P.T. TueUer. 1983. Landsat computer-aided analysis Saber, J.S., and P.T. Tuehr. 1973. Color aerial photos for marshland. techniques for range vegetation mapping. J. Range Manage. 36627-63 1. Photogmmm. Eng. 34489499. Met@ M.S. 1988.Thematic mapper data for evaluation and management Sparmar, M.A., A.H. StrahIer, and J.E. Eetee. 191)3.Soil loss prediction in of range ecosystems in part of Madhya Pradesh. Abstr. Vol. I. Third a geographic information system format. pp. 89-102 In: Proc. on Rem. Inter. Range. Congr., Range Manage. Sot. of India, Vigyan Bhavan, Sens. of Environ., Ann Arbor, Mich. May. New Delhi. p 3637. Thompson, M.D. 1977. Remote sensing of rangelands in western Canada. Meyer, M., F. B&on, and D. Whitmer. 191)2.Helicopter-borne 35 mm pp. 1,5061,514. In: Proc. Eleventh Internat. Symp. on Rem. Scns. of a&al photography applications to range and riparian studies. Remote Environ. Environ. Res. Inst. Michigan, Ann Arbor, Mich. Scnsinn Lab. Univ. Minnesota. St. Paul. IAFHERSL Rcs. Rcn. 82-l. Tucker, C J. 1979. Radiometric resolution for monitoring vegetation, how MiIkz, D’.A., G.W. Peterson, and A.B. Kahle. 19g6. Soil and geimorphic many bits are needed? NASA Tech. Mem. 80293. studies in arid region using the Thermal Infrared Multispectral Scanner Tucker, CJ., end C.O. Jwtice. 19g3. Satellite remotb sensing of desert (TIMS). Paper presented at the 3rd Arid Lands Remote Sensing Work- spatial extent. Deser. Contr. Bull. 92-5. shop, Univ. Arizona. Tucson, Apr. Tucker, CJ., C.L. VaaPreet, MJ. Shaman, and G. Van Itternum. 191)5. Moyer, R.H. 1958. Some uses of aerial photography inconnection with the Satellite &mote sensing of total herbaceous biomass production in the production and marketing programs of the U.S. Department of Agricul- Scncaalcsc Sahel: 1980-1984. Rem. Scns. of Envir. 17233-249. ture. Photogramm. Eng. 36305-307. TuellerrP.T., G. Loraln, R. Halvorwn, and J.M. RetUff. 1975. Mapping Musick, H.B. 19M. Assessment of Landsat multispectral scanner spectral Vegetation in the Great Basin from ERTS-I Imagery. Proc. Amer. SOC. indexes for monitoring arid rangeland. IEEE Trans. on Gcosci. and Photogramm. 41et Annu. Meeting. March Washington, D.C. pp. Rem. Scns. Ge-22512519. 338-370. NASA 19(17. HIRIS High-Resolution Imaging Spectromcter:Sciencc TueUer, P.T. 1978. -Large scale 70mm photography for range resources Opportunities for the 1990’s. Earth Observing System Volume 11~.Inatr. analysis in the western United States. Proc. 1lth Inter. Symp. on Rem. Panel Rep. Scns. of Environ. Ann Arbor, Mich.. p. 1057-1514. Paltridge, G.W., and J. Barber. 19811.Monitoring grassland dryness and Tueller, P.T., F.R. Honey, and I.L. TapIcy. 1978. Landsat and photogra- fire potential in Australia with NOAA/AVHRR data. Rem. Sens. of phic remote sensing for arid land applications in Australia. Proc. 12th Environ. 25381-394. Inter. Symp. on Rem. SCM. of Environ. Manila. p 2177-21291. Parker, H.D. 1987. What is a Geographic Information System? Proceed. TueUer, P.T. 1979a. Some aspects of the use of dichotomous keys to aid in Vol. I GIS ‘87 2nd. Annu. Inter. Conf. on Geograph. Informat. Sys. San the interpretation of color aerial photographs for vegetation mapping. Francisco. pp 72-80. 7th Bienn. Workshop of Color Aerial Photography in the Plant Sciences. Peck, R.P., R.D. Gmetx, A.W. Davis. 1986. Reflectance modeling and the Amer. Sot. Photogramm. Davis, Calif. p 189-200. derivation of vegetation indices for an Australian semi-arid shrubland. TueUer, P.T. 1979b. Rangeland remote sensing interpretation problems. Inter. J. Rem. Sens. 7:389-403. Proc. Rem. Scns. for Natur. Rcs.: A Symp. Moscow, Idaho pp. 450-465. Pickup, G, and D J. Nehon. 1984. Use of Landsat radiance parameters to TueUer, P.T. lsa0. Remote sensing applications for wildlife habitat man- distinquish soil erosion, stability, and deposition in Arid Central Austra- agement. Proc. California/Nevada !Icct. Wildlife Sot. Redding Calif. ha. Rem. Scns. of Environ. l&195-209. pp. 28-33. Pickup, G., amI V.H. Cbewlngs. 198(L. Estimating the distribution of TueUer, P.T. 19g2. Remote sensing for range management. Chapter 12. p. grazing and patterns of cattle movement in a large arid zone paddock: an 12514O.Zn:Johannscn and Sanders (ads). Remote Sensing for Resource approach using animal distribution models and Landsat imagery. Int. J. Management. Soil Conscrv. Sot. Amer. Rem. Scns. 91469-1490. Tuells, P.T. 1983. Rangeland remote sensing: an approach. Proc. RNRF Piuter, P J., Jr., R.D. Jackson, S.B. Idao,and R J. RegInato. 1983. Diurnal Symp. on the Application of Remote Sensing to Resource Management. patterns of wheat spectral rctlcctanccs. IEEE Trans. Gcosci. Rem. SCM. Amer. Sot. Photogramm. Seattle, Wash. p. 4245. GE-21:156163. TueUer, P.T. 1987a. Remote sensing science applications in arid mviron- Porter, W.A., and H.T. Emuark. 19111.A system overview of the airborne merits. Rem. Scns. of Environ. 23:143-154. visible/infrared imaging spectrometer (AVIRIS). p. 3-12. In: G. Vane Tuelkr, P.T. 1987b. Large-scale color aerial photography for monitoring (cd.). Airborne Visible/Infrared imaging spectrometer (AVIRIS), A shrub dominated rangelands. Proc. 1lth Bienn. Workshop on Color Description of the Sensor, Ground Data Processing Facility, Laboratory Aerial Photography and Videography in the Plant Sci. and Related Calibration, and First Results. NASA Jet Propulsion Laboratory. JP. Fields. Wcslaco, Texas p. 84-95. Pub. 87-38. TueUer, P.T., P.C. Lent, R.D. Stager, E.A. Jacobem, and K.A. Phtou. Poulton, C.E. 1975. Range resources: inventory, evaluation and monitor- 198g. Rangeland vegetation changes measured from helicopter-borne 35 ing. Chap. 18, p. 1427-1478. In: R.G. Reeves (Ed.) Manual of Remote mm aerial photography. Photogramm. Eng. and Rem. Sens. 54609-6 14. Sensing. Amer. Sot. Photogramm. Falls Church. Va, Tuelkr, P.T., end S.G. OIeaon. 19g9. Diurnal radiance and shadow fluctuations in a cold desert shrub plant community. Rem. Scns. of Environ. 27:(In Press).

452 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 Ustin, S.L., B.N. Rock, and R.A. Woodward. 19115.Analysis of substrate WiBiauu, R.E., B.E. ABred, R.M. Derdo, and HA. Pouken Jr. 196(1. and plant spectral features of semiarid shrub communities in the Owens Conservation. develoument and use of the world’s rangclands. J. Range Valley, California. Proc. Thematic Conf.: Rem. Sens. for Explor. Geol. Manage. 21:355-360.- San Francisco, Calif. pp 347-360. Willow Ruu Laboratorks. 1978. The University of Michigan Notes for a Vane, G., A.F.H. Goets, and J.B. Wellman. 19M. AirboNC imaging spcc- Program of Study in Remote Sensing of Earth Resources. Houston, trometer: a new tool for remote sensing. IEEE Trans. on Ckosci. and Texas. Rem. Scn. Vol. GE-223546449. Wilson, R.O., and P.T. Tuelkr. 19g7. Aerial and ground spectral character- Walkr, S.S., J.K. L&s, M.A. Brown, T.W. Heintx, RI. Butterfkld, and istics of rangeland plant communities in Nevada. Rem. Scns. of Environ. F.R. Gartner. 19’111.Use of 35 mm photography in vegetation sampling. 23:177-191. Proc. 1st Inter. Rangel. Congr. Denver, Colo. p. 5173520. Zevenbergen, A.W. 19(15.Runoff curve numbers for rangeland from Land- Wekh, R., T.R. Jordan, and A.W. Thomas. 19M. A photogrammatric sat data. USDA Hydrology Laboratory. Agr. Res. Serv. Tech. Rep. technique for measuring soil erosion. J. Soil and Water Conserv. HL85-1. 39191-194. Zonneveld, I.S. 197g. A critical review of survey methods for range man- Wuth, F.C., and CD. Lemme. 1978. Landsat spectral signatures: studies agement in the third world with special emphasis on remote sensing. with soil associations and vegetation. Photogramm. Eng. and Rem. Proc. 1st Internat. Range. Congr. Denver, Colo. p 510-513. scns. 44:315-325.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 453 Motivation of Colorado ranchers with federal grazing allotments

ET. BARTLETT, B.C. TAYLOR, J.R. MCKEAN, AND J.C. HOF

AbSdCt Net retumm to laveetment on weetern ranches are often low or found sign&ant, that we should implement questions on social negative. Ranchem who graze cattle on federal raoge during the and family ties, and that we should construct questions on per- summer in Colorado were eampled to determk the& wllllngnesa ceived mobility of resources,outside ranching. to wll their rmchee and to detemdne which factom were important The contingent market constructed in the questionnaire was ln their dedsloa to ranch. Cluster uulysia waa used to classify the specific to summer graxing for cattle. Because we had reason to rancher8 into 4 groups. Willingnenato rll the much wm the most believe that other seasons would each have a unique demand, and important factor in clessitylng groupa. Approsbnately 75% of the grazing for sheep also represents a different demand, those grazing federal pemdtteee would not coneider selling their ranches lo the markets were explicitly excluded. The sample was a simple random current market wblle over half rupooded that rate of return oo sample of the 1,530 ranchers who had summer federal graxing lnvestmeot wea of llttle or oo importance IOtheir dectslon to be in permits in Colorado in 1983, and who had cattle. the cattle bueineee. The groups also differed with respect to the Survey sampling and questionnaire design included a presurvey lmportmce of being near fmlly and Idends, and labor end amet which had several functions. Foremost among those functions was mobility. to set if the population of ranchers in Colorado understood and responded well to the questionnaire format and questions. The Key Wordaz publk la&, raocher behavior second function of the pm-survey was to estimate response var- Net returns to capital and management in western range live- iance and response rate to determine the required sample size. The stock operations are often low or negative. This fact has prompted Water Resources Council (WRC) (1979) recommends a pm-survey some economists to view a ranch as a consumption good to some sample of at least 200. We selected and mailed to a random sample degree rather than a pure production enterprise (Smith and Martin of 238 ranchers to test the contingent valuation questions. As a 1972). Consumptive components such as the value of a rural life- result of the presurvey, the survey was shortened considerably and style or a land ethic could raise ranch prices beyond that which some of the questions were made more explicit. However, ques- economic returns from livestock would justify. The degree to tions used in this study were unchanged in the final questionnaire. which these consumption attributes dominate ranch decisions will One thousand ranchers were sent the improved questionnaire with vary with the values of individual ranchers. one additional mailing to nonrespondents. For the questions used The purpose of this paper is to determine what motivates in this study, we pooled the presurvey and survey responses. Of the ranchers to stay in ranching. We grouped ranchers based on the 1,238questionnaires mailed, a total of 351 questionnaires of which importance they placed on various attributes including their will- 3 13 were useable in this study were returned. The low response rate ingness to sell their ranches. The results apply to ranchers using was at least partly due to the nature of the study to determine forage from federal land during the summer of 1983 in Colorado. forage value. Comments of respondents indicated that ranchers The research reported here was part of a study that applied the may have not responded because they felt the questionnaire or contingent valuation method to estimate the demand for summer study was a threat to them. The possibility of non-response bias forage from federal land in Colorado (McKean et al. 1986).There- cannot be ruled out, but, at the very least, the results apply to fore, Colorado ranchers not using federal forage were not included one-fifth of the population. in the population. Social parameters were measured and related to Seven variables relating to continuing in ranching were exam- demand for forage. ined in the survey. The variables were defined as: (1) desire to sell ranch (WOULD SELL), (2) strength of land ethic (LAND Methods ETHIC), (3) quality of family lie (FAMILY LIFE), (4) strength of A mail survey was used to elicit attitudes from a sample of social ties (SOCIAL TIES), (5) importance placed on mobility of ranchers. The questionnaire included Smith and Martin’s ques- the rancher’s labor (LABOR MOBILITY), (6) importance placed tions defining “land fundamentalism” as a starting point (Smith on difficulty of sale of ranch assets (ASSET MOBILITY), and (7) 1971). Smith defined land fundamentalism as “the attitude that a importance of investment return (PROFIT MOTIVE). The respond- greater amount of satisfaction can be received from associating ent rated variables on a scale of 1 (highly important) to 4 (not directly with the physical environment than with the ‘man-made*.” important) except the “desire to sell the ranch” question which Discussion with Rogers’ aided the design of the social parameter simply elicited a “yes” or -no” response. questions. We concluded that we should reduce the redundancy in Based on Smith and Martin’s work, a simple correlation analysis the “Land or Rural” preference questions that Smith and Martin did not appear to promise to identify relationships between rancher attitudes and the desire to be in ranching. Therefore, cluster analysis was used to determine sets of attitudes that affect whether a rancher is willing to sell or not. Cluster analysis is a technique that separates respondents into groups based upon sim- USDAForat Serviced ColoruloSt&c University with fundii from Colomdo ilarity of responses to a specified set of variables. The BMDP State University Agricultural Experiment Station (Projca CO-191). The usual dis- routine used was a K-mean clustering-a nonhierarchical tech- chimer applies. hdwwxipt acqted 24 April 1989. nique designed to group cases, not variables, for large amounts of data (Dixon 198 1). Nonhierarohical clustering was chosen because QJer& D. 1983. Personal communication. Formerly profeuor and department * Dept.of sociology, Colomdo state uniwmity. the initial partition of groups can be set by the researcher instead of

454 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 building a hierarchy from the most similar individuals in the sam- Some reasons for profit being assigned little importance can be ple (Johnson and Wichem 1982). found in the importance rankings of the lifestyle variables. A To make this analysis more statistically meaningful, the survey majority (56%) of ranchers assigned a high importance to a land data were divided by their standard deviations. This procedure had ethic (Fig. I), and 79% highly valued family life on the ranch (Fig. two effects on the composition of the final cluster groups. The first 2). effect was to standardize variables that were measured in different units. This was desired because the WOULD SELL question was 80 measured on a O-l scale, and the remainder of the questions on a m FAMILY LIFE m socuJ. TIE8 I scale of 1 to 4. A second effect from this procedure accounted for the different distributions of the responses for each variable. For example, a large proportion of the ranchers believed “quality of ranch-family life” to be of high importance, whereas the importance of investment return was much more dispersed. The standard deviation with the former would be very low, and with the latter, very high. By dividing the survey data by their standard deviations, the family-life variable was weighted relatively more because its standard deviation was so low. Conversely, the relative importance of the investment-return variable was decreased with its higher standard deviation. If a given rancher does not feel, as the over- whelming majority do, that family lie on a ranch is important, that dissenting opinion was weighted more heavily by the standardixa- HiGH LOW NONE tion procedure. Thus, in those questions that obtained a majority RESPONSE CATEGORY of responses in the “high” or the “none” categories, the dissenting Fig. 2. Percentage of ranchers’ responses to the questions of “quality of opinions were given more importance through the standardization family li/ on a ranch” (FAMIL Y LIFE) and “strength of social ties to process. frkndr and relatives on a ranch”(SOCIAL TIES). (A “none” response Cluster analysis can be used as an explanatory or descriptive indicates no importance.) technique to point out associations of variables that form natural In addition to the social attitudes of ranchers, there are also groups. To utilize the results of clustering, judgment concerning constraints on rancher behavior. About 49% of ranchers said that the level of clustering that best describes a consistent picture of it would be difficult to find an alternative job. Also, 50% felt that rancher behavior was necessary. The goal was to utilize the fewest the difficulty in selling their ranch was an important reason to possible groups without obscuring associations. To accomplish continue ranching (Fig. 3). this objective, the clustering routine was run for 2,3, and 4, and arbitrarily up to 8 clusters. “V I Resulta and Discussion Even a visual examination of the social and lifestyle variables revealed a great deal about rancher behavior. Approximately 75% of federal permittees in Colorado would not consider selling their ranches in the current market. Smith and Martin (1972) identified the desire to hold ranch properties as the principal reason for an historic acceptance of low return rates on ranchland investment. This conclusion is supported in our study, with over half of the ranchers responding that rate of return on investment was of little or no importance (Fig. 1).

md M ETHIC I I HiGH LbW NdNE RESPONSE CATEGORY Fig. 3. Percentage of ranchers’responses to the questions of ‘difficulty of obtaininga job outside the ranch”(LABOR MOBILITY)and”d@ctdty of selling the ranch in the current market” (ASSET MOBILITY). (A “none” response indicates no importance.) With cluster analysis, 4 groups provide traits observable in real-world rancher behavior. At greater than 4 clusters, the additional clusters did not increase clarification. For 2 or 3 clusters, coherent aggregate behavior was not discernible. Thus, ranchers in the sample were classified into one of four groups. The summary statistics of the clustering are given in Table 1 and Figure 4. HI’GH LbW NOi4E Certain variables were more important in differentiating the RESPONSE CATEGORY cluster groups than others. Because response to the SELL RANCH question may be related to many other variables as Smith and Fig. 1. Percentage of ranchers’responses to the questions of “owning a Martin suggested, it is not surprising that the desire to sell one’s ranch allows me tofeel&ser to theearth”(LAND ETHIC)ond”obtaining a good return in investment” (PROFIT MOTIVEA (A “none” ranch was the most significant variable in placement of ranchers response indicates no importance.) into groups.

JOURNAL OF RANGE MANAGEMENT 42(6). November 1999 455 Tdh 1. Clustermean, ~tandud deviation, and F-&IO from an analysis of vulmcc betweendusters for ercb aochl or rttItudbmlvukble w&binnch cluster group of ranchers.

Cluster WOULD SELL’ LAND2 FAMILY2 SOCIAL2 PROFIT2 JOB’ ASSET’ Number RANCH ETHIC LIFE TIES MOTIVE MOBILITY MOBILITY 1 mean 0.96 1.68 1.48 3.31 2.91 3.38 2.91 sd 0.20 0.91 0.80 0.78 1.01 0.93 1.10

2 mean 0.00 1.76 1.67 2.93 2.68 3.09 2.12 sd 0.00 1.01 1.00 1.08 1.12 1.11 1.04 3 mean ::: 1.54 1.38 2.18 2.04 1.62 1.38 sd 0.69 0.68 0.99 1.01 0.78 0.62

4 mean 0.97 1.61 1.21 1.50 2.81 3.19 3.45 sd 0.18 0.84 0.55 0.50 0.89 1.07 0.78

GRAND MEAN 0.75 1.65 1.45 2.62 2.64 2.87 2.48 FOR ALL CLUSTERS F ratio-+ 521.6 0.86 3.85 63.7 11.4 52.7 64.4 lRcsponsc wa8Would Sell = 0 and Would Not Sell = 1. from 1 to 4 with 1 = important, 2 = medium, 3 = low, and 4 = no importance for continuing in ranching. ‘tRcsponscA rewonse ““P o I mdicated mobilitv was dficult and an imoortant reason that thev continued in ranching, a response of 4 indicated difficulty in finding a job or in selling the ranch-was not important. 4Dcgrcc of freedom 3,309.

The significance of the remaining variables showed that the did the other groups. Thus, cluster 2 ranchers were the ranchers constraints a rancher faces in labor and asset mobility are just as that would be willing to sell but were otherwise similar to other important as the social and attitude factors for continuation in ranchers regarding social values. ranching. Asset mobility and social ties had the largest F-ratios Group I ranchers are not considering selling their ranches in the (Table 1) followed by job mobility, profit motivation, and the current market. They placed a relatively high importance on own- quality of family life on the ranch. The land ethic variable was not ing a ranch and on family life. Ties to friends and relatives, and significantly different between cluster groups (F = 0.86 in Table l), obtaining a good return on their investment were of the lowest although all groups placed high or medium importance to this importance of any group. Further, the opportunity to leave ranch- variable. ing, as indicated by the low importance placed on asset and labor mobility, was not a constraint. Group 4 was similar to group 1, but ranchers in this group had the highest degree of asset mobility and placed the highest importance on family related matters. They found that being near friends and relatives was a very important reason for staying in ranching. This would indicate that one-half of the ranchers in Colorado (Fig. 4) have other opportunities but choose to be in ranching because of land fundamentalism and social reasons. GROUP 1 (31. OX) GROUP 2 124.0%) Groups 3 and 4 were very similar in their behavior and social attitudes. Social ties, family, emphasis on profit, and land ethic for both groups were all more important than the overall average, and exhibited only slight difference in intensity of these variables. The major difference was that ranchers in group 3 felt that the lack of other job opportunities and the difficulty in selling their ranch were important reasons why they continued in ranching. These two clusters show that, even though ranchers can realistically face GROUP 4 (20.0%) GROUP 3 (25.0%) differing mobility of ranch assets and labor, this does not affect their desire to continue in ranching as long as the ranch provides the quality of family Me, some financial return, and social ties. Groups were analyzed with respect to ranch size and degree on dependency on federal forage during the summer. Ranch size was based on total animal units (AU’s) and averaged 335 AU’s Ranch size across the four cluster groups was not signiticantly different nor was the percent contribution of federal forage to summer Fig. 4. l%e percentage of ranchers designated in each of the fw cluster grazing. groups. Conclusions Beyond the ranking of significance of the variables, cluster analysis provided a description of each cluster group. Group 2 was the Most Colorado cattle ranchers using federal forage in the only cluster of ranchers that would sell their ranch at or near the summer were not willing to sell their ranches at 1983 prices. They current market price, although they cited the difficulty of selling valued owning a ranch regardless of other attitudes and also valued ranch assets in the current market as a fairly important reason for raising a family on a ranch. While the profit motive was significant continuing to ranch. This group placed slightly less importance on in classifying ranchers, job and asset mobility better differentiated the land ethic and the desirability of ranch life for their family than ranchers into groups.

450 JOURNAL OF RANGE MANAGEMENT 42(6). November 1989 Those who had the highest difficulty in obtaining other jobs or Literature Cited selling their ranches attached the highest importance to profit as Dixon, W.J., cd. 191)l. BMDP statistical software 1981. Univ. California the reason that they are in ranching. Perhaps because of the mobil- Press, Berkeley. ity constraints, these ranchers are more aware that a good return Johnson, R.A., sod O.W. Wicbern. 19112.Applied multivariatestatistical on investment is important to their survival even though they value analysis. Prentice-Hail Inc., Englewood Cliffs, NJ. ranching as a way of life. McKean, J.R., J.G. Hof, E.T. Bartlett, and R.G. Taylor. 1986. Contingent Ranchers who did not consider job and asset mobility as impor- value estimates for existing federal grazing. Colorado State Univ. Agr. Exp. Sta.. Fort Collins. Tech. Bull. TB8Cl. tant reasons for being in ranching (groups 1 and 4) ranked profit Smith, A.H. 1971. A socioeconomic analysis of the goals and attitudes of lower than the other groups. These individuals are likely to be the Arizona cattle ranchers. Unnub. Ph.D. Dim. Univ. of Aria.. Tucson. last to sell their ranches. Smith, AM., and W.E. Mar&. 1972. Socioeconomic b&a&or of cattle Ranchers who were willing to sell their ranches listed difficulty in ranchers with implications for rural community development in the selling them as a moderate reason for being in ranching, but were West. Amer. J. Agr. Econ. 54~217-225. not as profit oriented as those with the lowest labor and asset Water Resourcea Council. 1979. Procedures for evaluation of national economic development benefits and costs in water resources planning, mobility. This would suggest that even though those who would final rule. Fed. Reg. 8 CFR Part 713, Vol. 44, No. 242, Rules and sell at market prices thought that they would have difficulty in Regulations. p. 72892-72976. doing so. Thus, although economic factors may cause some to quit cattle ranching, the majority value ranching as a way of life.

NOWHERE will YOU find more information in one place on this increasingly important subject. A greater variety of uses for rangeland, a growing world population, plus deeper concerns for the environment make knowledge about insect grazers of rangeland ever more critical. Written to be understood by the average reader, RANGELAND ENTOMOLOGYstill contains the extensive bibliography, lists of both common and scientific names, and suggestions for needed research to make it a valuable addition to the scientist’s library. The second edition, edited by J. Gordon Watts, has been expanded to over 300 pages with state-of-the-art information on insect friends and foes of rangeland plants and animals and the role of integrated pest management. Range Science Series No. 2, Second Edition, will be available from the Society for Range Management by February 1999.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 457 A test of grazing compensation and optimization of crested wheatgrass using a simulation model

BRET E. OLSON, RICHARD L. SENFT, AND JAMES H. RICHARDS

AbStlWt

We developed 8 simuhtion model b8sed on tiller population recharge from winter snow is the primary source of water for spring processes to test gr8xing compens8tion 8nd optimhtion in created growth (Caldwell 1985). wbe8tgr8ss (Agropyron -0~ (Fiscb. es Link) Scbult.). Traditionally, these established grasses have been managed Model functions dacribing tiller dynamics 8nd growth were extensively, e.g., continuous season-long grazing, but there has derived from field observations in west-centml Ut8b. Predicted heen interest in intensifying grazing management of these species tiller growth 8nd new tiller production following defoli8tion were (Malecbek and Dwyer 1983). With extensive or minimal manage- verified 8g8inst 8ddition8l d8ti from tbe sune site; total produc- ment, grazing is highly variable in space and time, whereas inten- tion w8s v8lid8ted 8g8inst 8 30-ycu-old d8t8 set from 8 different sive management should increase control over the critical factors site. We then sinuhted 2 gr8xing experiments. First, @wing com- of the timing, intensity, and frequency of defoliation. pens8tiOn ~814determined 88 8 function Ofthe timing Of 8 single With extensive management most tussocks of crested wheat- defoliation during tbe growing se8son. Raponse vuirbles included grass are grazed only once, on average, during the growing season tiller density, plot growth n&s, shmling crop, 8nd se8sonrl pro- (Norton and Johnson 1983). Yet because pastures are grazed duction. Second, graxing optimizrrtion, 8 combin8tion of gr8xing incrementally, individual tussocks are defoliated at different phe- frquency 8nd intensity tb8t incre8sea primuy production rbove nological stages. The response of a tussock graxed in early spring tb8t of ungrmsed phnts, w8.s 8sseas4 by the systenmtic vuirtion of differs considerably from the response of a neighboring tussock these defolhtion puuncters under simuhted dry, 8vCnge, 8nd wet grazed in late spring (Richards and Caldwell 1985). McNaughton winters (SepMIbH-M8y). Results of the first experiment indi- (1983) suggested that plants defoliated early in the growing season uted tb8t compenmtion depended nminly on tbe timing of defolir- when soil moisture and nutrient levels are high can compensate for tiOn, presumrbly beauss of pbenologic8l conshints to regrowth foliage removed by grazers. We define compensation as the ability 8nd tbe short growing se8son in tbis cold-desert region. Overcom- of a grazed plant to maintain a similar level of production as an pen&On Only occurred when phnts were defoli8ted before the ungrazed plant (after Relsky 1986). Compensation equal to or tr8ditiolul st8rt of tbe gmxing season. Altbougb defolhtion greater than ungraxed plants has been observed after early spring incre8sed tiller growth rites,, tbe second experiment f8ikd to reved defoliations of crested wheatgrass (Miller et al. 1984, Olson and 8n optimum combhtion of defolhion frquency 8nd intensity Richards 1988a). Early spring defoliations resulting in compensa- resulting in m8ximUm bionmss prOdUCtiOn except after 8 dry win- tion occur just hefore the grasses begin their rapid growth phase. ter. Results from the second experiment indhted tb8t implement- * Thus, only relatively small increases in tiller growth rates are ing intensive rot8tiorul gmzing systems will seldom incre8se needed to achieve compensation (Hilhert et al. 1981). Compensa- crested wbe8tgr8ss production in these colddeaert systems. tion following grazing must he assessed carefully hecause graxing affects other ecosystem processes (Relsky 1987); for example, the Key Wordsz Agropyron dcsatorum, tiller, defolhtion, nmets, status of neighboring plants can alter the response of grazed plants r8met popuhtion model, production model to the physical loss of foliage (Mueggler 1972, Olson and Richards A short spring growing season constrains the extent and timing 1989). of primary productivity on colddesert regions of the world (West With intensive management grasses may be grazed several times 1983). In many of these regions the native tussoclc grasses have not within a growing season. To justify the costs of implementing an evolved tolerance to large mammalian herbivory (black and intensive system, the increased control over graxing intensity and Thompson 1982). Because many native grasses in western North frequency should maximize primary production and ultimately America are sensitive to spring livestock grazing, grazing-tolerant livestock production by enhancing regrowth. Evidence for and tussock grasses from Eurasia, primarily of the Triticeae, have heen against grazing compensation has often come from African G widely established. Most notably, crested wheatgrass (Agropyron sedges and grasses clipped or grazed at frequent, 2- to 14day, deserrorum (Fiscb. ex Link) Scbult.) has heen seeded to provide intervals (McNaughton 1979,198s; Relsky 1986). Whether the Cs early spring forage and to control soil erosion (Rogler and Lorenz crested wheatgrass can withstand such short graxing intervals is 1983). On cold-desert rangelands where it has heen established, unknown. Hilhert et al. (1981) hypothesized that frequent grazing growth is usually cued to short (2 to 3 month) periods of adequate is unlikely to result in compensation because large increases in temperature, nutrient availability, and soil moisture. Soil moisture relative growth rates would he needed after each grazing event. With one notable exception (Cook et al. 1958), time and cost Authors arc research umciate, Reclamation Raearch Unit, Montana State Uni- versity, Bozcman 59717; research ecologirt, USDA ARS, South Central Family Farm constraints have prohibited experiments to determine tussock Research Center, Boon&Be, Arkansas 7292%and amociate profewor, grass response to the numerous permutations of grazing intensity of Range Science and the Ecology Center, Utah State Uni~~ty, Logan Tmcnt84 224230. At the time of the -h, the fint author was -h assistant, Department of and frequency that occur in intensive systems. Because the growth Range Science, Utah State Univenity. The second author was &stant profmsor, and development of tussock grasses are fairly deterministic func- rtment of Animal, Dairy, and Veterinary Sciences, Utah State University. T e Utah Agricultural Experiment Station (Project 780) and the National Science tions of the number and size of tillers, both of which are affected by Foundation (BSR-8207171 and BSR-8705492) sup~rtcd our -h. This is Jour- grazing, simulation modeling of tiller population processes may nal Paper No. 3439 of the Utah Agricultural Ex nment Station, Utah State Univcr- sity, Logan. We thank J.L.K. Pcdcrscn for p”teld asbtance and M. Andrew, M. allow experiments combining grazing intensity and frequency at Coughenour, D. Pykc, H. ‘Ihorgckon, ad two anonymous reviewera for construe- minimal cost. tive comments on the manuscript. Manmript accepted 24 Aped 1989. We developed a model hased on weekly observations of the growth and population dynamics of grazed and ungraxed tillers in

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 456 a short-duration system (Olson and Richards 1988a). Although the structure of the model did not include other factors associated with -08 A. grazing, i.e., trampling, nutrient cycling, altered competitive backgrounds, these factors arc implicitly included in the model because 1 ____------, Maximum rate for the data used to develop the model were from a field study of .06 1 defoliated tillers crested wheatgrass response to gruzing by cattle (Olson and : Richards 1988a,b). z Defoliated :a, Once validated, WC used the model to assess compensation of \ tillers ; ‘t, :: crested wheatgrass following a single defoliation at various times ; : \ ’ : : : \ during the spring growing season. We also simulated a factorial : ‘\ \ I ; I : \ experiment (defoliation intensity and frequency, and precipita- ‘\ : \\ tion) to determine if an optimal combination of non-zero defolia- ‘\ : ‘\ A ‘\ I ‘\ ;: ‘\.\ \ tion parameters could be used in a grazing system on crested \ I ‘\ , . . \ wheatgrass. These results were compared with McNaughton’s Baseline (undefoliated) rate (1979) grazing optimization hypothesis. I 1 I Model Concqtudhation April MY Tussocks of crested wheatgrass consist of metapopulations of tillers (sensu White 1979). The emergence, growth, and death of Date tillers represent whole tussock response to the environment. Hence, a model of tiller population processes integrates the physiological and morphological attributes that are manifested in tussock growth. Our simulation model was based on observations describ “41 ing growth rates of different types of tillers and the transfer of B. tillers among population classes. Tillers were divided into 3 classes: undefoliated parent tillers, defoliated parent tillers, and new spring ll2 ------7 Maximum rate tillers. The latter were subdivided into cohorts (tiller pulses) pro- .lo : _\: ‘, for new tillers I , duced following individual defoliation events. I : : \ In a normal year, almost all tillers of crested wheatgrass present .08- I I New tillers i \ ::‘\ in early spring were produced the previous fall, overwinter with 1 \ : : ‘\ \ to 3 leaves, and resume growth in late March or early April. The .06- : ,:’ \ , :’ : \ growing season begins with the spring growth of overwintered ;’ : ,“, \ .04- 8, : ‘1, \ tillers. In the model, we started spring growth on 1 April. All tillers I :, \ are undefoliated parent tillers at thii time. Grazing converts these mo2 Baseline rate i i 1, ;’ 1, 1, tillers to defoliated parent tillers. Subsequent forage production : ; : 8’ : depends on the growth of the undefoliated tillers, the regrowth of I : \_- ,, defoliated tillers, or the emergence and growth of new spring tillers O- I I I (Olson and Richards 1988a). All tillers senesce by mid-July when April WY soil moisture is depleted. Date Model Development Fig. 1. Height growth rates of & defoiiated tiUers andb. new tillers. ‘%s Spring growth and regrowth of crested wheatgrass were simu- figures represent simukxtedgrowth ratesfora moderate& wet winter (350 lated as functions of winter precipitation (September-May), and mm). Plots were dcforiatrd 30. 50, and 70 doys after the initiation of the timing (relative to the initiation of growth) and intensity of a growth. Note wferent scaks.

single defoliation event. The model operated on a daily time step BW.I = O.OOO048mm d-’ l P (2) and was based on changes in the number and size (height and The realized growth rate declined linearly from the initial rate weight) of undefoliated, defoliated, and new spring tillers through throughout the active growth phase (Fig. la), depending on the time. multiplier ZMOD which integrated the growth of all tiller cohorts Functions describing the growth of tillers of crested wheatgrass (see equation 4). Height-growth rate varied among years with were derived from nondestructive measurements by Olson and precipitation. The height-growth function operated until a maxi- Richards (1988a). In summary, ungraxed tillers grew slowly until mum height based on the preceding winter precipitation was early May, then rapidly until mid-June when infIorescences were reached, and then growth stopped abruptly, similar to our field fully developed. Grazing before mid-May often enhanced tiller observations. Maximum tiller height (H-) was calculated from a growth, whereas grazing after mid-May usually reduced tiller correlation between observed heights and winter precipitation. growth. Grazing from mid-May to mid-June stimulated the production of new spring tillers. H,= 1.1 l P (3) The daily height growth increment (mm d”) of undefoliated tillers (AH) was simulated by: Height growth.. was further modified by negative feedback from AH=Ht+Bd (1) biomass growth: where Ht was tiller height (mm) at time-t, t was time since growth ZMOD = 1 - BMt l (1.5 l BM,,,.x)-’ initiation, and B- (mm mm-’ d-i) was actual height growth (4) rate. ti, baseline height growth rate (mm mm-’ d-i), was a function of winter precipitation (P in mm): where ZMOD was a modifier to height growth which varied between 0 and 1, BMt was current total biomass (kg/ha), and

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 1983

--..__.__.Predicted

I 0 I 1 April May June Intensity Date

B. 400 1984

E 300

., E .P, 200 _E $ F 100

/ 0 April May Jlne April May June Dale

1985

300 400 April May June Winter precipitation, mm Date

Fig. 3. Observed andpredicted growth (height mm) of undefoliated tillers Fig. 2. Three of thefactors which controlled the size of a filleringptdse in from I ApriI to 10 July. Observed data are means of I50 tillers in 1983 (10 the moakl: a defoliation intensity (I), b. akte of defoliation within the tillers/plant, 15plants)and 100 tillers (10 tillers/plant, IOphmts)in 1984 growing season (S), c. September-May precipitation (Y, Y= 0.002 * P”’ and 1985; aakptedfrom Oison 1986. where P is precipitation in mm). I, S. and Y are dimensionless scalars. BM, was the total biomass supported under a given precipitation During regrowth following defoliation, height growth rate (mm regimen. BM, was estimated by: mm-’ d- ) of defoliated parent tillers was immediately enhanced and then declined over a 3-week period to the baseline rate (Fig. la, BM,. = 1.11 + BM, l P l (P&-l (5) Olson and Richards 1988a). Peak height growth rate following where longterm average biomass production, BM, and winter defoliation declined as the season progressed. Defoliated tillers precipitation, P, were 900 kg/ ha and 300 mm, respectively (mod- ceased growth at the end of the spring growing season lo-15 days ified from Sneva and Hyder 1961). earlier than undefoliited tillers.

480 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 New spring tillers were 20 mm in length within the leaf sheaths of defoliated parent tillers. Thus, the size of a tillering pulse was their progenitors before defoliation (from Mueller and Richards estimated as: 1986). Their subsequent growth was initially equal to that of parent tillers, increased over a 10-l 1 day period (maximum height growth D-= I+S+Y*N (10) rate = 0.12 mm mm’ d-l), and then returned to rates comparable In simulations with multiple ddoliations, the cumulative number with parent tillers (Fig. lb). The peak growth rate of new spring of new tillers produced per parent was constrained to a maximum tillers, similar to defoliated tillers, declined as the season progressed. of 4.5 (from Mueller and Richards 1986). As the cumulative Temperature was not explicitly included in the model. However, number of new tillers approached this limit, tiller pulse size was the empirically derived growth parameters in Equations 3,4, and 5 modified from equation 10 by: implicitly included temperature effects. For example, high temperatures at our central Utah site probably depressed tiller growth TMOD = I - D,, l (4.5 l N-)-l (11) rates late in the growing season (Busso 1988). Temperature effects, where TMOD was a dimensionless factor varying between 0 and 1, as in most cold-desert regions, were probably confounded with soil and Ncym and Dcum were cumulative numbers of defoliated parent drying. Because spring growth was usually completed by mid-July, tillers and of new tillers, respectively. To scale up to plot-level mid- and late-summer rains had little effect on spring production responses, initial tiller density was 535 tillers/m* (from B. Olson and thus were not included in the model. 1986). For seasonal biomass production and peak standing crop, Standing biomass of undefoliated parent tillers on tussocks was these plot-level responses were scaled up to kg/ ha. determined using a cumulative biomass(C)/ cumulative height (A) formula developed for crested wheatgrass (Johnson 1987): Model VaIidation C = 1.88 A - 0.00888 A* (6) Methods C and A were expressed as proportions of their respective maxi- Field observations of crested wheatgrass in 1983 and 1985 were mum values achieved in the spring growing season. A was defined used to develop and verify the model whereas data collected from as the cumulative height from the base of a tussock and C was the same site in 1984 provided a validation (Olson and Richards cumulative live biomass. A and C were converted to mm and g by 1988a). Performance criteria were the quantitative correspondence scaling equations 3 and 5. of predicted height dynamics of defoliated and unddoliated tillers, For defoliated parent tillers and new spring tillers, allometric and size of new tiller pulses to observed responses (Grant 1986). height-weight relationships were determined from biweekly har- Growth and regrowth were simulated for the springs of 1983, vests of tillers at soil level during the spring growing seasons of 1984, and 1985. Actual September-May precipitation was 380 1983-1985 (B. Olson 1986). Weights (MS of defoliated parent mm, 351 mm, and 309 mm in 1983-1985. Initial parent tiller tillers and new spring tillers were estimated daily from tiller height heights were set to 75 mm (B. Olson 1986). Regrowth following (Ht) measurements using power functions: defoliation was simulated for 3 dates of defoliation in each growing season. These dates were 45,60, and 75 days after growth began Mt = 0.00032 l Ht” defoliated tillers and for (7) and corresponded to when grasses were grazed in the field, about Mt = 0.00006 l Ht’” for new tillers, (8) 15 May, 1 June, 15 June. Verification exercises were repeated with where the units of M and H were g/tiller and mm/ tiller, respe& defoliation dates varied f5 days because of the uncertainty in tively. Standing biomass (kg/ha) was estimated for defoliated observed defoliation dates with respect to the initiation of growth, parent and new tillers as the product of tiller number and average and of the strong effect of defoliation date on the size of tillering mass per tiller. Standing biomass was estimated for each cohort of pulses in the model (Fig. 2b). defoliated parent and new tillers. At each simulated defoliation, 55% of the nondefoliited tillers Few tillers present in early spring senesce before the end of the were “graxed” at heights ranging from 47 to 90 mm. This simulated growing season (unpublished data). Therefore, changes in total 50-70% removal of the current season’s herbage production. tiller numbers are primarily a function of a series of tillering pulses Plot-level biomass predictions were validated against the results (additions) following defoliation. The addition of new tillers (NT) of an independent clipping experiment (Cook et al. 1958). Plot- was calculated for each daily time step by: level predictions aggregated growth responses of tiller cohorts. In NT=Da*RT(D,.-Dd)+D--’ (9) the study by Cook et al., tussocks of crested wheatgrass were where Da was the number in a single pulse of new tillers at time t, defoliated 1 to 11 times during the spring and early summer of 5 D,. was the maximum number of new tillers produced for a given consecutive years. Data from the first year (1949) were used to pulse c, and RT was a growth rate constant. This logistic function validate our model because production in subsequent years was was based on weekly observations of new spring tiller emergence subject to cumulative treatment effects. September through May following biweekly graxing events during spring (Olson and precipitation in 1948-1949 (390 mm) was similar to our wet winter Richards 1988a). These observations indicated that the size of a of 1982-1983. tillering pulse (D-) was controlled by 3 factors: the intensity and To minimixe the differences between the experimental proce- date of defoliations within the growing season, and the previous dures that affected Cook et al.‘s observations and our model pre- September-May precipitation. On average, tillering pulses lasted dictions, the seasonal decline in peak growth rates of defoliated tillers was removed from the model (Fig. la). In Cook et al.‘s study, 21 days. soil moisture was maintained by periodic watering of plots whereas The intensity effect (I) on Dmu was a monotonic function of defoliation intensity (Fig. 2a). Seasonal tillering potential (sea- our model was developed from observations on a site that was not irrigated and has higher summer temperatures. Because Cook et al. sonal factor, S) rose from zero before 1 May, peaked on 20 May and declined to zero by late June (Fig. 2b). This factor was modi- combined spring and summer growth in a fall harvest, we also simulated a fall harvest. fied by winter precipitation. After wet winters defoliated parent tillers could produce new tillers earlier and later in the growing ResuIts and dIsewsion of model performance season, whereas after dry winters the tillering season was truncated The predicted growth (height) of nondefoliated tillers was sim- (Fig. 2b). The precipitation effect(Y) was defined as the maximum ilar to the observed (Fig. 3). The slightly lower predicted growth in number of new tillers produced per defoliated parent tiller as a 1984 could have been due to a dry May at the field site followed by function of winter precipitation (Fig. 2~). N was the number of

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 481 substantial rainfall (49 mm) in early June that may have contrib- 150 uted to the observed late season growth. The model did not include 1983 June precipitation. The average predicted sizes of the tillering pulses in the 3 years were similar to those observed (r = 0.94, n = 7, Pig. 4). _

Observed

Predicted ,/ 30 I I 0 / May June /i / l Date / / l 1983 $’ 0 1954 / / n 1985 I I I lo 30 50 70 Observed tillers/100 parents

Fig. 4. Observed and predictednew tiller pulses from 1983 - 1985.

The predicted heights of defoliated tillers were also similar to the 2 80 ___--- observed (Fig. S), and were neither consistently lower nor higher - i= than the observed. Tillers defoliated in early spring usually grew more in the following month than those defoliated later. Predicted and observed plot-level production are shown in Fig- June ure 6. After adjusting for differences in experimental procedures, May predicted production was similar to the observed (r = 0.80, n = 35, Date solid circles, Fig. 6), although predicted production of plots initially defoliated after 1 June was lower than the observed production (n = 8, open circles, Fig. 6). With the near optimal condi- 15’1 1985 tions in Cook et al.% study, defoliated tiller regrowth or new spring tiller growth probably continued into summer (see Caldwell et al. 1981). Midsummer growth under watered conditions probably explains why our model, which had minimal summer growth, under-predicted production for plots defoliated in early summer. On sites with higher annual precipitation or abundant summer precipitation and cooler temperatures, e.g., high elevation seedings (Currie and Smith 1970), growth could continue through summer. Since grazing of crested wheatgrass on Intermountain rangelands normally begins in late April or early May and ends before early July, the discrepancy between our model predictions and Cook et 30 -I al.% observations following late-spring defoliations does not limit June the validity of the model for applications in the Intermountain May region. Date Models have a valid domain, a range of input variables with Fig. 5. Observed andpredicted regrowth (height mm) of tillers defoliated which reasonably accurate model predictions can be expected. 15 May, 1 June, and 16 June. Regrowthperiods depictedare onemonth. Validation of a grazing system-level model, based on this model, Observed data are means of150 tilkrs in 1983 (10 tillers/plant, 15plants) indicated that the valid domain consisted of colddesert foothill and 100 tillers (10 tillers/plant. lOplants) in 1984 andl985;adaptedfrom pastures with 175 to 410 mm September-May precipitation (Senft Olson 1986. 1988). Below 175 mm, inadequate soil moisture recharge would (Rogler and Lorenz 1983). Thus, the model’s valid domain covers limit spring growth and tillering. Above 410 mm, especially where the range of precipitation regimes where it is normally seeded precipitation continues into summer, there would probably be Often models simulating grass growth and regrowth following considerable regrowth following grazing. However, our model is defoliation give adequate simulations of data used in developing based on observations from a semi-arid rangeland where late the model, but comparisons with an independent data set are not spring decreases in tiller pulses and growth rates limit regrowth. performed (Coughenour et al. 1984,1985a,b; Johnson and Parsons Finally, because in our model grass growth and regrowth is based 1985; Chen 1986). Whenvalidation is possible, predicted responses on winter precipitation, it is probably not applicable where spring- sometimes compare unfavorably with observed responses (Dayan early summer rains provide most of the annual precipitation, e.g., et al. 1981). The successful verification and validation of our Great Plains. Crested wheatgrass has seldom been seeded on range- tiller-based model by observations from 2 different sites and a lands with precipitation less than 230 mm or greater than 380 mm number of years indicates that crested wheatgrass is relatively

482 JOURNAL OF RANGE MANAGEMENT 42(8), November 1989 1400

I200

t a 1000

._5 tJ 800

-8 g 600

B t 400 .- 121 B. % 200

0 0 1000 2000 3000 4000 Observed production g/plot

Fig. 6. Predicted and observed seasonal plot-~1 production. Data for observedproduction are from Cook et al. (1958). Solid circks indicate plots that were initially &foliated before 1 June. @en clrcks indicate plots that were initially &foliated on or after I June. predictable in its response to defoliation, and the model can be used for plot-level simulation experiments within the valid domain.

Simulation Experiments MdhodS Grazing Compensation Simulation To assess the effect of “time" of grazing on the degree of compensation by crested wheatgrass, single defoliitions, from 5 to 95 days (at l-day intervals) after spring growth started, were simulated. The series of simulations was run for an average winter (300 mm September-May precipitation). A run with no defoliation was the control, In each trial, 50% of the standing crop was removed. A standard clipping height of 55 mm simulated the average observed 7001 grazing heights of crested wheatgrass by cattle at our field site (K. April WY Ane Olson 1986). To remove a constant 50% of standing crop, the model simulated a defoliation of 80 (early spring) to 62 (late) 900 D. percent of the tillers as plant heights increased during the growing season. F Model responses were assessed at the “plot” level and thus 8700 represented a summation of responses across all tiller populations. _ Response variables were: (1) end-of-season tiller density (tillers/m*); 8 (2) mean plot-level relative growth rate (RGR, mg g-l d-l); (3) ti I.--.---.-.:---.-.--.-.--.-.:-.- seasonal biomass production (kg/ ha) including biomass removed by defoliation; and (4) end-of-season standing crop (kg/ha). Mean plot-level RGR (G) was calculated by averaging day-today instantaneous growth rates:

3004 I I I G = ((B&I - BMt + BR) l BMi’) l L-’ l 1,000 mg/g (19 Ame where BMt was current total biomass, BR was biomass removed by April MY defoliation, and L was length of the growing season (L = 100 d). To Fig. 7. Effect of dclte of simulated defoliation on growth responses of determine the contribution of new spring tillers to production, the crested wheatgrass. a end-of-season tilkr density (tnitial density = 535 experiment was repeated without the tillering subroutine. tillersjms), b. mean plot-level relative growth rate (G), c. seasonal biomass production, and d end-of-season standing crop. Each point along Grazing Optimization Simulation the solid and dotted lines represents the growth or population response To assess whether grazing optimization (maximum production of a single plot &foliated at a particuhtr drrte. Responses of a serks of at some non-zero grazing frequency and intensity) occurs with defoliatedplots withandwithout normalspring tilleringare indicated by solidand dotted lines, respectively. The dt@ferencebetween the solidand crested wheatgrass, the frequency and intensity of defoliation were dotted lines indicates the contribution of new sprbtg tillers to production. varied in a factorial design. Simulated plots were defoliated 1 to 6 Dashed lines indicate the response of a singk unakfoliated plot (the times in a 60day grazing season, about 1 May to 30 June. Because control). JOURNAL OF RANGE MANAGEMENT 42(g), November 1999 463 the previous simulation emphasized the importance of defoliation defoliations after 1 July. This is consistent with our field observa- date, we averaged model responses for 3 runs at each frequency. tions (Olson and Richards 1988a) and probably resulted from low The first run assumed even spacing of defoliation events within the soil moisture availability. grazing season, whereas defoliation events in the other 2 runs were Mean plot-level RGRs (G) following different times of defolia- shifted 10 (frequency = 1) to days (frequency = 6) before and after tion were mediated by several factors. First, peak growth rates of the dates in the first run. defoliated parent tillers declined after mid-May (Fig. la). Second, Defoliation intensity was varied from lO-90% of standing her- tiller pulses added to mean plot-level RGR in mid-season (Fig. 2b), bage. Standard clipping height was again 55 mm. The proportion but this was only a minor contribution. Overall, mean plot-level of tillers defoliated was varied to obtain the desired graxing inten- RGR was highest in plots defoliated early in the growing season sity, which generally parallels the grazing pattern observed in the when regrowth was primarily from the continued growth of defol- field. At the heaviest grazing intensities, clipping heights were iated parent tillers (Fig. 7b). When defoliation occurred after early reduced to obtain the desired intensities. The lowest clipping height May, mean plot-level RGR declined because of the seasonal reduc- was 17 mm. This was above an estimated 5 mm minimum cropping tion in growth rates of defoliated parent tillers, and the negligible limit of domestic cattle (Olson et al. 1986). contribution from new spring tillers. Defoliation increased mean The simulation experiment was repeated for dry, normal and plot-level RGR above those of ungraxed plots, except when it wet winters (200, 300, 400 mm September-May precipitation). occurred after early June. Controls consisted of ungrazed simulations run under each winter Although mean plot-level RGR is increased simply by reducing type. In each run, tiller population, height and biomass dynamics biomass (the denominator in mg g-l) following defoliation, sea- were monitored. The model responses of interest were the same as sonal production was higher than the controls when grasses were in the first experiment. defoliated before early May (Fig. 7~). Production, including biomass removed by defoliation, varied between 85 and 110% of Results nondefoliated plots. Overall, new tillers contributed little to sea- Grazing Compensation sonal production (Fig. 7~). Recause tussocks compensated follow- The influence of defoliation date on end-of-season tiller density ing early spring defoliations, end-of-season standing crop was was mediated by the seasonal control of tillering pulse size (Fig. constant for plots defoliated before early May (Fig. 7d). For plots 2b). Defoliations before early May did not stimulate new spring defoliated later, standing crop was lower because grass regrowth tiller emergence (Fig. 7a). Predicted new tiller production, and thus was limited by declines in tiller growth rates. In addition, the pulse tiller density, was greatest when defoliations occurred in mid- to of new spring tillers contributed little to end-of-season biomass. late May. New spring tiller emergence was not stimulated by

Dry year Average year Wet year

Fig. 8. wfects of defokation frequency and intensity, ondprecipitation, on growth ond regrowth of crested wheotgrass. HT. Bui-of-seaon tiller den&y (initioldensity = 535 tillerslm~), and&$ meanplot-level RGR (G)for dry, overage andwet years. Thepoints associated with zero d~oiiationfrequency represent on ungrazed control.

464 JOURNAL OF RANGE MANAGEMENT 42(6), November lQ6Q Grazing Optimizarion depressed biomass production relative to undefoliated controls For all combinations of precipitation and grazing intensity, (Fig. 9b.c). maximum tillering occurred when plants were defoliated 3 times End-of-season standing crop declined with increasing defolii- (Fig. 8a-c), resulting in peak tiller densities of 680,813, and 931 tion frequency and intensity (Fig. W-f). Winter precipitation again tillers/m* for dry, average, and wet winters. Tillering tended to was a scaling variable. The nonlinear response to defoliation inten- increase with defoliation intensity; however, after a dry winter sity indicated that new spring tillers emerged and grew at moderate simulation, tillering was reduced at all defoliation intensities when grazing intensities without reducing the regrowth of defoliated plants were defoliated more than 3 times (Fig. 8a). This reduction tillers. in tillering with increased defoliation frequency was less evident after normal and wet winters (Fig. 8b,c). Discussion Predicted mean plot-level RGR increased monotonically with ConstraMs on Re~owth increasing frequency and intensity of defoliation (Fig. 8d-f). This The general agreement between predicted and observed responses response resulted primarily from the increased relative growth of verification (our 1984 data) and validation (data from Cook et rates of defoliated parent tillers; new spring tillers contr&uted little al. 1958 within the valid domain) simulations reflects the empirical to plot growth rates. Enhanced relative growth rates of defoliated basis of the model and the inherent biological constraints on tillers mainly reflect lower available biomass following defoliation, growth in the field. In many grass production models, grass however, and not high absolute growth rates. Lower remaining response to the timing of a defoliation often depends primarily on biomass makes relative growth rates large even when increases in environmental factors (Johnson and Thomley 1983, Johnson and absolute growth rates are small. Precipitation was a scalar to Parsons 1985, Chen 1986, Parsons et al. 1988), decmphasizing the herbage growth rates, particularly for baseline growth rates of potential importance of morphological or developmental con- undefoliited tillers (frequency = 0 in Figs. 8d-f). straints on regrowth. This approach is appropriate for species Predicted seasonal biomass production was generally stable at where developmental constraints are minor and thus the environ- moderate defoliation intensities, but declined at the heaviest inten- ment strongly affects regrowth, but for crested wheatgrass devel- sities (Fig. 9a-c). After wet and average winters, production opmental constraints seem to be important, at least at certain times declined with increasing frequency of defoliation whereas fre- of the year (Richards and Caldwell 1985, Olson and Richards quency had little effect on production after the dry winter. Follow- 1988a). Such constraints were included in this model; for example, ing dry winters, defoliated plots generally produced more than tillering occurred primarily when the grasses were defoliated dur- undefoliated plots (Fig. 9a). As precipitation increased, defoliation ing a time corresponding to the elevation of apical meristems (but

Dry year Average year Wet year

Fig. 9. E/fi?clsof dcfoliationfrequenqv and intensity, andprecipitation, ongrowth of crested wheatgrass. a-c. Seasonalproduction, and&J en&o/-sraron standing cropfor dry. average and wet years. lkpoints associated with zero defoliation frequency represent an ungrazed control. Note: To successfdly interpret rrsponse q&aces. the lowest point of the suflace shouldface the reaak Thus, X and Y axes are reversedfrom Fig. 8.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 465 see Richards et al. 1988). Tillering was not based explicitly on model nor was it observed (Olson and Richards 1988a), presum- assimilate supply (Dayan et al. 1981, Coughenour et al. 1984, bly because low soil moisture, high temperatures, or both limited 1985a). tillering and growth rates of defoliated tillers in late spring. These constraints resulted in a reduction in plot-level herbage Mid-spring defoliation reduced the seasonal production of growth rate following late spring defoliations (Fig. 7b). Models crested wheatgrass, as others have noted (Cook et al. 1958, Miller based on a plentiful supply of water and nutrients often do not et al. 1984). because of declining growth rates of defoliated tillers incorporate a seasonal decline in growth rates following defolia- and the negligible contribution from new spring tillers. Although tion (Dayan et al. 198 1, Coughenour et al. 1985a). In a previous new spring tillers contributed little to forage production and model of crested wheatgrass, Senft and Malechek (1985) assumed senesced in summer, they probably increased the number of axil- that there was a constant addition to available forage when it was lary buds available for fall tiller production and thus annual tiller grazed at any time in spring. In contrast, our model incorporates replacement (Olson and Richards 1988b). limited opportunity for regrowth with late-season defoliitions. Grazing Optlmlzation This limited response is more appropriate in cold-desert regions With extensive management, single early- or late-spring defolia- with a short spring growing season that depends on winter precipi- tions that result in overcompensation are possible, but these would tation. Grazing management decisions would differ considerably occur at the start and the end, respectively, of a typical spring depending on whether regrowth is constant or declines following grazing season in the Intermountain region. With intensive man- late spring defoliations. agement, individual tussocks could be grazed several times during Grazing Compematlon the spring growing season. In some grasses, grazing at some fre- The results of this simulation indicated that compensation after quency and intensity may increase net primary production above grazing in this tussock grass is not constant but varies. When that of ungrazed plots (i.e., grazing optimization, McNaughton crested wheatgrass was defoliated once before early May, there was 1979), yet there was little evidence that a particular grazing overcompensation in aboveground production (Fig 7~). This defo- frequency-intensity combination would increase primary produc- liation enhanced the growth rates of defoliated tillers relative to tion in crested wheatgrass. Only after a dry winter was simulated undefoliated tillers (Fig. lb, Olson and Richards 1988a). Cook et seasonal productivity slightly greater with “grazing” than without al. (1958) also observed overcompensation after early spring defo- (Fig. 9a). liation of crested wheatgrass. Furthermore, they observed over- For full compensation after defoliation to occur, mean plot-level compensation after late spring defoliations, but this probably RGRs should have been an exponential function of grazing inten- resulted from the supplemental watering of their grasses. sity (Hilbert et al. 1981). However, we found that mean plot-level Although defoliation enhanced mean plot-level RGR, over- RGRs were generally a linear function of grazing intensity, thus compensation in seasonal production occurred only after early full compensation would not have been expected following defolia- spring defoliations (Fig. 7c, compareMcNaughton et al. 1983, tion. This agrees with the limited conditions where the model McNaughton 1985, Relsky 1986). Early spring defoliations allowed predicted full compensation or where it has been observed in the enough time for enhanced growth rates to produce considerable field with crested wheatgrass (Cook et al. 1958, Olson and amounts of new foliage, thus increasing seasonal production. Richards 1988a). The positive relationship between grazing inten- Recause compensatory growth rates are often short-lived (Olson sity and mean plot-level RGR was not entirely linear. The highest and Richards 1988a), assessments of compensatory growth should grazing intensity (0.9) in our model was least, not most, likely to be based on seasonal or annual production, the integration of increase productivity above ungrazed tussocks, presumably because growth rates through time. the required increase in growth rate was not biologically possible. The potential enhancement of mean plot-level RGR (Fig. 7b) Moderate grazing intensities, in this and Hilbert et al’s (198 1) and available foliage declined as the season progressed, thereby model, generally increased production more than heavy or light lowering regrowth and seasonal production (Fig. 7~). Carbohy- defoliation intensities (Fig. 9a-c). drate reserves available for regrowth are highest when we observed Low defoliation frequencies should result in greater compensa- and simulated this seasonal decline in regrowth (Caldwell et al. tion than high frequencies because greater growth increases are 1981, Deregibus et al. 1982, Richards and Caldwell 1985, Olson needed at high frequencies (Hilbert et al. 1981). Our model and Richards 1988a). Our observations and simulations indicate responses tended to agree with this prediction when defoliation that regrowth depends primarily on the efficiency of carbon utiliza- intensity was light or heavy after average-and wet winters. In the tion for shoot production and photosynthetic area, but not on dry winter simulation, however, frequent moderate defoliations mobilizing large carbohydrate reserves. At least for crested wheat- increased production more than infrequent defoliations (Fig. 9a). grass, shoot regrowth efficiency depends on shoot meristem activ- This latter response reflects another postulate of Hilbert et al. ity, and concurrent photosynthesis contributes considerably more (1981) that smaller growth rate increases are needed to achieve carbon to regrowth than carbohydrate reserves contribute (Richards compensation when growth is limited, e.g., after a dry winter, than and Caldwell 1985). when grasses are rapidly growing under favorable conditions. Although the model is empirically based, the responses agree Our model indicated that grazing tolerance, for which crested with the theoretical constructs of Hilbert et al.‘s (1981) analysis of wheatgrass is well known (Caldwell et al. 198 1), does not necessar- relative growth rates. Overcompensation after early-spring defoli- ily translate into the potential for grazing optimization of net ation indicated that crested wheatgrass can increase relative primary productivity. McNaughton (1979) suggested that grazing growth rates of defoliated tillers considerably above the rates of optimization is not possible with domestic livestock as in our undefoliated tillers. After mid-spring defoliation, when undefol- model because the grasses and grazers have not evolved together. iated plants were growing rapidly, the large increase in relative Alternatively, created wheatgrass tillers are taller and less dense growth rates of defoliated tillers needed for compensation were not than those of KyIlinga nervosa (McNaughton et al. 1983). These real&d, possibly because of developmental constraints. After late- morphological differences may partially explain why defoliation spring defoliation, when undefoliated grasses were growing slowly, failed to consistently increase net primary productivity in crested only a small increase ingrowth rates of defoliated tillers would wheatgrass (Coughenour et al. 1985b). Even in Kyllinga, however, have been necessary to equal the productivity of ungrazed tillers grazing does not consistently maximize productivity (McNaughton (Hilbert et al. 1981). Yetsuch compensation did not occur in our 1979).

486 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 With our empirical simulation model, we developed rcsponsc Ma&, R.N., ad J.N. Thompson. 19g2. Evolution in steppe with few, large surfaces for several herbap variables to dctcrmine if compensation hooved mammals. Amer. Natur. 119757-773. might bc expected from graxcd tussocks of crested wheatgrass, and Makdwk, J.C., and D.D. Dwyar. 19g3. Short duration graxing doubles to test whether McNaughton’s grazing optimization hypothesis your livestock? Utah Sci. 44:32-37. McNaugbton, S.J. 1979. Grazing as an optimization process: grass- dcscribcs crcstcd wheatgrass regrowth following 3 types of winters. ungulate relationshipsin the Serengeti. Amer. Nat. 113~691-703. The time of defoliation, incorporating dcvclopmental and cnvir- McNanghton, S.J. 1%3. Compensatory plant growth as a response to omnental factors, dctcrmincd the extent of compensation. Further, herbivory. Gikos 403329-336. except after the driest winter there was no evidence for a combina- McNaught&, S.J. 1%5. Ecology of a graxing system: the Serengeti. Ecol. tion of non-zero graxing intensity and frequency that maximiis Mono. 553259294. seasonal hcrbagc production. McNaugbton, S.J., L.L. WaBaee, and M.B. Coughenour. 19113.Plant adaptation in an ecosystem context: effects of defoliation, nitrogen, and Literature Cited water on growth of an African G sedge. Ecology 64:307-318. bW9 Ad. 1986.DOCS herbivory benefitplants? A reviewof the evidence. MIBtr, R.F., M.R. Haferkamp, and R.F. AngaB. 19M. Defoliation and Amer. Natur. 127:870-892. growth of crested wheatgrass. I. Defoliation effects on forage growth, p. BaRQ, AJ. 1987. The effects of grazing: confounding of ecosystem, com- 12-17. Research in beef cattle nutrition and management. Oregon Agr. munity, and 0rghsm scales. Amer. Natur. 129:777-783. Exp. Sta. Spec. Rep. 714. Buno, C.A. 1988. Factors affecting recovery from defoliation during Mueggler, W.F. 1972. Intlucnce of competition on the response of blue- drought in two aridland tussock grasses. PhD. Dim. Utah State Univ., bunch wheatgrass to clipping. J. Range Manage. 2588-92. Mueller, R J., and J.H. Rkhards. 1986. Morphological analysis of tillering Logan CaIdwaB, MM. 191u. Cold desert. p. 198-212. In: B.F. Chabot and HA. of Agropyronsp~~ and Agropyron ak&orui. Ann. Bat. 5291 l-921. Mooney (eds.). Physiological ecology of North American plant com- Norton. B.E.. and P.S. Johnson. 1983. Pattern of defoliation by cattle munities. Chapman and Hall, New York. grazing c&ted wheatgrass pastures. p. 422424.Zn:J.A. Smithard V.W. CaIdweB, MM., J.H. Rkhuds, D.A. Johnson, B.S. Nowak, and R.S. Hayes (Cdr.), Proc. XIV Graasl. Cong., Westview Press, Boulder, Colo. Dzurec. 1981. Coping with herbivory: photosynthetic capacity and 0lm0,B.E. 19g6. Tiller dynamics of Agropyron desertorum in response to resource allocation in two semiarid Agropyron bunchgrasses. Geeologia grazing and resource manipulation. Ph.D. Diss., Utah State Univ., 50: 14-24. Logan. Chen, J. 1986. Optimal cutting frequency and intervals derived from John- Obon, B.E., and J.H. Rkhuds. 19Mla. Tussock regrowth after graxing: son and Thomky’s model of grass growth. Agr. Sys. 22305-314. intcrcalary meristem and axillary bud activity of tillers of Agropyron Cook, C.W., L.A. Stoddut, and F.E. KInsInger. 19%. Responses of desertorum. Oikos 51~374-382. cmsted wheatgrass to various clipping treatments. Ecol. Mono. 28237- Olson, B.E., and J.H. Richards. 1988b. Annual replacement of the tigers of 272. Agropyron akswtorum following grazing. Gecologia 76: Id. Cougbmour, M.B.,S.J. McNaagbton,andL.L. W4Baea. 19g4. Modellihg Ohon, B.E., and J.H. Rkhards. 1#9. Crested wheatgrass growth and primary production of perennial graminoids-uniting physiological replacement following fertihration, thinning and neighbor removal. J. processes and morphometric traits. Ecol. Model. 26101-134. Range Manage. 429397. Cougbenour, M.B., S.J. McNaaghton, and L.L. WaBaea. 1%5a. Siiula- Ohm, K.C. 1986. Animal nutritional response to sward structure under tion of East African perennial graminoid responses to defoliation. Ecol. short duration graxing management. Ph.D. Diss., Utah State Univ., Model. % 177-202. Logan. Cougbsnour, M.B., SJ. McNmghton, and L.L. WaIIaec. 1%5b. Shoot Olson, K.C., R.L. S&t, aod J.C. Makebek. 1986. A predictive model of growth and motphometric analyses of Senngeti graminoids. African J. cattle ingestive behavior in response to sward characteristics. p. 259-261. Ecol. 23:179-l%. In: Proc. West. Sect. Amer. Sot. Anim,Sci. 373259261. Cur&, P.O., and D.W. Smith. 1970. Response of seeded ranges to different Parsons, A.J., I.R. Johnson, and A. Harvey. 19gg. Use of a model to grazing intensities in the ponderosa pine zone of Colorado, USDA For. optimize the interaction between frequency and severity of intermittent Serv. Prod. Res. Rep. 112. defoliation and to provide a fundamental comparison of the continuous Dayan, E., H. Van Keulen, and A. Domt. 1981. Tier dynamics and and intermittent defoliation of grass, Grass and Forage Sci. 43:49-59. growth of Rhodes grass after defoliation: a model named TILDYN. Richards, J.H., and M.M. CaIdweU. l%S. Soluble carbohydrates, concur- Agro-Ecosys. 7:101-112. rent photosynthesis and eftlcicncy in regrowth following defoliation: a Dare&us, V.A., M.J. Trlh, and D.A. Junaon. 1962.Organic rcmvcs in tield study with Agropyron species. J. Appl. lkol. 22907-920. he&age plants: their relationship to grasslmd mana~&nt. p. 315-344. Riehude, J.H., R.J. Muetler, end J.J. Mott. 19g8. Tillering in tussock grasses in relation to defoliation and apical bud removal. AM. Bot. In: M. Recheixl (cd.). CRC Handbook of Axricuhural Productivitv._. Vol I, Plant Prod&&y. CRC Press, Boca I&on, Fla. 62173-179. Frkbknecht, N.C., modL.E. Harrb. 196g. Grazing intensities and systems Rogkr, G.A., and RJ. Loranx. 1983. Crested wheatgrass -early history in on crested wheatgrass in central Utah: response of vegetation and cattle. the United States. J. Rang&iIanage. 3691-93. USDA Forest Serv. Tech. Bull. 1388. S&t, R.L. 19g8. Short duration grazing on crested wheatgrass (Agro- Grant, W.E. 1986. Systems analysis and simulation in wildlife and fnheries pyron desertonrm): effects of grazing system variables on cattle and science. Wii-Interscience, New York. herbage production. Agr. Sys. 28189211. HBbert, D.W., D.M. SwBt, J.K. DetBng, and M.I. Dyss. 1981. Relative Senft, R.L., and J.C. Makebet 1985. Short duration grazingcell parame- growth rates and the grazing optimization hypothesis. Gecologia ters and cattle production: a low-resolution model. Proc. West. Sect. 51:14-18. Amer. Sot. Anim. Sci. x282-285. Johnson, I.R., aadA.& Parsons. 19M. Use of a model to analyse the effects Soeva, F.A., and D.N. Hydr. 1961. Estimating herbage production on of continuous grazing managementon seasonal patterns of grass produc- semiarid ranges of the Intermountain region. J. Range Manage. tion. Grassand Forage Sci. 4Oz449-458. 1588-93. Johnson, I.R., and J.H.M. Thomky. 19g3. Vegetative crop growth model West, N.E. 1983. Overview of North American temperate deserts and incorporating leaf area expansion and senescence, and applied to grass. semideserts. p. 321-330. In: N.E. West (cd). Temperate Deserts and Plant, Cell and Envir. 6721-729. Semideserts, Ehevier. Amsterdam. Johnson, P.S. 19g7. Utiliration of plants of crested wheatgrass by cattle Whita, J. 1979. The plant as a metapopulation. Amt. Rev. Ecol. Sys. under several grazing regimes. Ph.D. Dim. Utah State Univ., Logan. l&109-145.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 487 Effects of stocking rate on quantity and quality of available forage in a southern mixed grass prairie -

RODNEYK. HEITSCHMIDT,STEVEN L. DOWHOWER, WILLIAM E. PINCHAK,AND STEPHENK. CANON

AbShCt The objective of tbie study ~8s to qu8ntify the long-term (25 ye8rs) effecte of be8vy (HC) 8nd modcnte (MC) r8te!3of &ckin~ on qtmntity 8nd qtmlity of forage 8v8iLble. Study desip required frequent lmrvest of rt8nding crop on 5 r8np rite8 in twice rep& c8ted, 244 ha tre8tment p8eturee. Resulta from the 20montb study lowed 8boveground st8nding crop dynrrmkewere aimil8r in botb trerrtments,qu8nthy of8v8ikbk fonp ~8s gre8terin the MC tbrn HC tr#tment, qtmlity of lv8K8ble forye ~8s gre8ter generally ia tbe HC tb8n MC treatment, 8nd tb8t heavy etocking 18vored a domhunce of w8rm-ee8eon rhortgr8mea u opposed to 8 domi- arace of w8rm-se8son mid-. Averaged acroms drtea 8nd 8djusted for differenceaamong puturea in rrnp rite composftion, 8boveground berb8ceoue at8nding crop 8ver8ged 1,341 kg/b8 in the HC p8aturee M compared to 1,816 kg/al, in the MC tratment p8stures. Crude protein md 0-c matter digatibility 8venged 0 - 8.6% 8nd 49.396,respectively, in the HC p8atura md 7.7% md J A S 0 N OIJ F M A M J J A S 0 N.DIJ F M A M J 46.7% respectively, in the MC p8&um. It ia concluded tb8t the I 1985 ’ : 1986 1987 grerter vuiltion among yeur in cow/calf production in the HC tburhtbcMCtrcrrtmentbprimuilybcclrrwforrp8vrilrbllltyin Fig. 1. Monthly (bar graphs) and long-term (conrinrtous line)precipiration tbe HC treatment is lees tb8n in tbe MC treatment. (cm) at Texas Experimental Ranch. Key Words: speder dombunce, r8nge site, stmdding crop, crude precipitation (1960-1987) is 68 cm. Average maximum daily protein, o&c nutter digeetibility temperatures range from a high of 36’ C in July to a low of 1 lo C in Livestock production from rangeland is a function primarily of January the the quantity and quality of forage consumed (nutrient intake), Topography at the ranch is rollmg and ranges from broad val- which varies primarily as a function of the quantity and quality of leys (5%). Soils range from the deep, well-drained clay 1979), the major factors within a given climatic regime are the and clay loams (fine, mixed, thermic Calciustolls, Typic Ustocrepts inherent productivity potential of a site (range site) and the kinds and Cumulic Haplustolls) found on the uplands, flood plains and and number of plants present (range condition) (Ellison 1960, Sims gentle slopes, to the shallow, stony clay and clay loams (clayey, et al. 1978, Bartolome 1984). A basic premise in range management mixed, thermic Typic Ustochrepts) found on the steeper slopes. is that as range condition declines, livestock carrying capacity Dominant range sites are Clay Loam, Clayey Upland, Clay Slopes, declines also because quantity and/ or quality of forage produced Shallow Clay, Rocky Hills and Loamy Bottomland (SCS 1984). declines (Stoddard et al. 1975, Danckwerts and Aucamp 1984, Elevation ranges from 408 to 463 m. Malechek 1984). The objective of this study was to quantify the Vegetation is a mixed-grass complex within a light to moderate long-term (25 years) effects of moderate and heavy rates of stock- stand of honey mesquite [(Rosopisghdulosa (Torr.) var. grCmdu- ing on quantity and quality of forage available in a southern low)]. Dominant midgrasses are Texas wintergrass (St&w leuco- mixed-grass prairie. Subsequent papers will examine the on-going rricha Trin. & Rupr.), a cool-season perennial; sideoats grama effects of these treatments on quantity and quality of forage con- [(Boutelouu curripendulu (Michx.) Torr.], a warm-season peren- sumed and ultimately livestock production. nial, and Japanese brome (Bromus japonicus Thunb.), a cool- season annual. Dominant shortgrasses are buffalograsses [(B&z- Methods lee dacryloides (Nutt.) Engelm.] and common curlymesquite Research was conducted at the Texas Experimental Ranch [(Hiloria berlungeri (Steud.) Nash], both warm-season perennials. located (99’14’W, 33’2O’N) on the eastern edge of the Rolling The dominant forb is annual broomweed (Xunrhocephdum spp.). Plains. Climate is continental and semiarid. Precipitation is highly For a more complete description of the climate, soils, and vegeta- variable and bimodally distributed with peak rainfall occurring in tion at the ranch, see Heitschmidt et al. (1985). May (9.2 cm) and September (10.6) (Fig. 1). Average annual Grazing treatments were yearlong continuous stocked with mature Fl Angus/Hereford cows at heavy (HC) and moderate Authors arc profesw, research auc&te, amistsnt professor, and rmawch asso- (MC) rates. The twice replicated treatments were initiated in 1960 ciate, Texas Agricultural Experiment Station, P.O. Box 1658, Vernon, Texas 76384. Appreciation k expressed to the Swen R. Swcluon Cattk Co. for providing the in 4 pastures of similar condition (Heitschmidt et al. 1985) follow- land, livutock, and facilities for this study and the Texas Experimental Ranch ing subdivision of a single pasture. Size of pastures ranged from Committee for provid@ finurir a&tance. Report is ubli&d wth approval of the Director, Texas Agricultural Experiment 218 to 266 ha. Average rates of stocking from 1960 through 1984 Station, as #A-24071. were about 5 and 7 ha/ cow/ yr, respectively. Recommended rate of Manwcriptaaqtcd 16 March 1989.

466 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 stocking in this region for good condition range is about 8 assigned for relative rank in each quadrat. Assigned values were 3 ha/cow/ yr (SCS 1984). During this 2&month study, rates of stock- (rank I), 2 (rank 2). 1 (rank 3) and 0 (not ranked). Following ing were 5.0 and 6.4 halcowlyr. summarization, all data were subjected to least squares analysis of Range site composition varied among treatment pastures (Fig. variance utilizing a split-split plot model (Little and Hills 1978) to 2). The dominant range site in 3 of the 4 pastures was Clay Slopes test for the main effect of treatment (HC vs. MC), the split-plot (48%). In the fourth pasture (HC, Rep II), the dominant range site effect of range site, and the split-split plot effect of date. Pasture was Clay Loam (25%). summaries were analyzed using a split-plot model following adjustment for diierencs among pastures in range site composition. Percentage CP and OMD estimates were analyzed following arcsine transformations. Tukey Q (KO.05) procedures (Snedecor cl LB and Cochran 1967) were used for mean separation of significant

k8 RH (p

JOURNAL OF RANGE MANAGEMENT 42(6), November 1980 469 Table 1. Mean rank Uca for doodnent specke averaged acrom ssmpk data.

Ra ngc Site’ Treatment* ChY Clayey Clay Rocky Loamy Treatment3 Common Name Loam upland Slopes Hills Bottomland HC MC HC MC Gra!3%%3 Buffalograss- curlymesquite l.llY 1.4ab 1.3b 0.74 1.4a 1.4a l.Ob l&l l.lb Texas wintergrass l&i 1.8a 1Sb 0.4c 1.5b l.la 1.6b 1.2a 1.7b Sidcoatr 8rama 0.7b 0.5c 0.8b 1.4a 0.8b 0.8 0.8 0.8 0.6 Japanese bromc 0.4b OSb 0.3C O.ld 0.7a 0.3a 0.4b 0.3 0.4 Hairy grama CO.lb

Table 2. Meao rank iadica for dominant specke lvereged NKIUIIrange sitee end grazing treatments.

Date Species Get 85 Jan 86 Apr 86 July 86 Nov 86 Feb 87 Mar 87 June 87 Grasses Buffalograss~urlymesquitc 1.3al 1.3a O.!% 1.3a 1.2ab 1% l.Obc 1.2ab Texas wintergrass 1.3b 1.3b 1.4ab l/tab 1.3b 1.4ab 1.5a 1.5a Sideoats grama 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.7 Japanese brome 0.5ab 0.7a 0.6a

Table 3. Herbage standiog crop (kg/ha) averaged across sempk dates.

Ra ngc Site’ Treatment* Treatment3 Clay Clayey Clay Rocky Loamy Category Loam upland Slopes Hills Bottomland HC MC HC MC Live grass 629b’ 617bc 544c 269d 759a 505 622 496a 6Olb 628 966 606a Deadgrass ------991b 768c 776c 453d 997a 961b Total 162Ob 1385c 132Oc 722d 1756a 1133 1588 1102a 1562b

Forbs ------222 221 230 219 334 247 244 239 254 Total 1842b 1606c 155Oc 941d 209Oa 1380 1832 1341a 1816b ‘Averagedacross grazing treatments and dates. ZAveragedacross noge sites and dates. aAveragedacross dates and adjusted for diiereoces among treatmentgastures in range qite,compositipn. 4Meam within a mw within range sites and treatments followed by d erent letters arc s~gmticantlyd&rent at KO.05.

470 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 Table 4. Percentage crude protein, organ& matter di~atibbby and percentage Hve gnu atmdhg crop lvmged ~UOSSdata.

Ra nge Sit& Treatment* Treatment3 Clay Clayey ClaY Rocky Loamy Category Loam upland s10pca Hills Bottomland HC MC HC MC Crude Protein (%) 8.2b4 8.7a 8.Ob 7.4c 8.3b 8.5 8.6a Organic matter 48.3 41.2 48.7 49.3 47.3 49.5 2:: 49.3 23” digcsibiity (%) Live (%I) 37% 42Sa 38.9l.s 36.4c 41.9ab 41.2 37.7 41.6 36.9

~Averaged across grazing treatments and dates. *Averaged acro8s nngc &es and dates. aAveraged across dates and jWified for diiercnm among treatmentrun3 in range a+ ympositi?n. Weans within a row within range sites and treatmenta followed by d ercnt letter8 arc s~gmfiintly dlffercnt at KO.05.

tion effects but no treatment, treatment X range site, treatment X cant date X range site interactions that occurred when standing date, or treatment X range site X date interaction effects. Live, crops were averaged across treatments were considered to be of dead, total grass (live + dead) and total (grass + forbs) standing minor importance when F-values were examined. For example, crops were greatest on the Loamy Bottomland sites and least on the F-values for range site and date were 4- to 8-fold greater than Rocky Hills site (Table 3). Standing crop was greater generally on F-values for the date by range site interaction. Moreover, when the Clay Loam than the Clayey Upland and Clay Slopes sites. standing crops were adjusted for differences among treatment There were no significant differences among sites in forb standing pastures in range site composition, only the main effects of treat- crop. Although there were no significant treatment effects when ment and date were significant. Thus, at the pasture level, standing the data were averaged across dates and range sites, standing crops crop of live and dead grass, total grass, and total herbage were were significantly greater in the MC than HC treatment pastures significantly greater in the MC than HC treatment on all dates after adjustment for differences among pastures in range site com- (Fig. 3). There was no difference between treatments in forb stand- position (Table 3). ing crop. The significant date effects reflected the effect of clitic conditions on rates of herbage growth and disappearance. The sign& Percentage CP of the grass standing crop varied significantly among range sites and dates but not treatments when averaged across range sites (Table 4). Averaged across dates, percentage CP ranged from a high of 9.3% on the Clayey Upland site to a low of 7.4% on the Rocky Hills site. Averaged across range sites, percentage CP ranged from a high of 10.4% in March 1987 to a low of 5.8% in January 1986. Although there was a significant range site X date interaction effect, the effect was minor in that percentage CP was greatest on all dates, except June 1987, on the Clayey Upland site and least on all dates, except October 1985, on the Rocky Hills site. When the percentage CP data were adjusted for differences among treatment pastures in range site composition, percentage CP varied significantly between treatments (Table 4) and among dates (Fig. 4). There was no significant treatment X date interaction effect. Percentage OMD of the grass standing crop did not vary significantly among range sites or between treatments (Table 4) regard- 2.- less of method of summarixation. However, percentage OMD did vary significantly among dates (Fig. 4) in a pattern similar to a percentage CP. Although there was a significant range site X date e 8 interaction effect, no definitive pattern of variation was apparent in that percentage OMD ranged across dates from 49 to 65% on the E' Loamy Bottomland and Clay Slopes sites, 49 to 62% on the Clay Loam site, 50 to 64% on the Rocky Hill site, and from 51 to 65% on f the Clayey Upland site. Moreover, there was considerable variation among range sites when they were ranked within a date 0 N DIJ F Y A Y J J A 5 0 N DIJ F Y A Y J J 1966 1966 1967 according to percentage OMD. For example, ranks over the 8 dates ranged from 1 (greatest) to 5 (least) for the Clayey Upland Fig. 3. Grass and total (grass +forbs) standing crops (kgJha)forpastures in heavy (HC) and modhate (MC) continuous grasing treatments. and Rocky Hills sites, 2 to 5 for the Clay Loam and Loamy jrtcotment effects were signi#kant (X0.10) (QJ = 7 %51 for grass and Bottomland sites and from 1 to 4 for the Clay Slopes site. total standing crop, respectivelyJ Date effects were signifiant KO.05) There was a relatively strong relationship between percentage (a = 480 % 432, respectively). Zkeatment by &te interaction was not CP, OMD and live standing crops (Fig. 4, Table 4). Correlation signi@ant in either analysts.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1080 471 (Table 1); (2) aboveground standing crop dynamics were similar in both treatments (Fig. 3); (3) quantity of available forage was greater in the MC than HC treatment (Fig. 3, Table 3); and (4) forage quality was greater generally in the HC than MC treatment (Fig. 4, Table 4). Although species composition was not measured directly in this study, we believe the greater abundance of shortgrasses in the HC treatment was the result of a general shift in species composition from a Texas wintergrass/ warm-season midgrass complex (good range condition) to a Texas wintergrass/ warm-season shortgrass complex (fair range condition). Previous research in these same pastures (Heitschmidt et al. 1985) supports this conclusion in that it showed frequency of buffalograss and end-of-season standing crop of warm-season shortgrasses inside temporary exclosures was 1965 1966 1967 70 substantially greater in the HC than MC pastures on most range T sites. Standing crop dynamics were similar between treatments g 60 J because species composition by functional group (cool-season vs. warm-season species) was similar. Quantity of available forage was greater in the MC than HC treatment, probably because of the combined effects of both reduced forage demand (stocking rate) and greater aboveground net primary production (ANPP). Al- HC e-e though the vegetation sampling scheme utilized in this study was MC A-A not conducive for estimating ANPP, previous research in these same study pastures (Heitschmidt et al. 1985) has shown single year (1982) ANPP estimates of 3,285,2,640, and 1,850 kg/ha in the HC treatment and 3,110, and 2,940, and 2,505 kg/ ha in the MC treatment for Loamy Bottomland, Clay Loam, and Rocky Hills sites, respectively. Quality of forage was greater generally in the HC than MC treatment because the proportional amount of low quality 1965 1966 1987 707 (senesced) forage was greater in the MC (63%) than in the HC treatment (58%). 60-- The results of this study also provide strong evidence as to why cow-calf production is more variable across years in the HC treatment than in the MC treatment (Heitschmidt et al. 1988). For example, during the period from 1982 through 1987 production/ cow in the HC treatment averaged 211 kg and production/ha averaged 44 kg as compared to production estimates of 215 kg/cow and 34 kg/ ha in the MC treatment (Heitschmidt et al. 1988). These averages were attained at average stocking rates of 4.9 and 6.4 ha/ cowl yr, respectively, and the feeding of an average of 112 and 24 kg/ cow/ year of a 20% CP range cube during winter in the HC and MC treatments, respectively. However, production/cow, pro- OJ m 0 0 0 0 0 s 0 m a m a a 0 m n 1 n n 1 n 1 ONDJFYAYJJASONDJFMAYJJ duction/ ha, stocking rate, and amount of winter supplement fed in 1965 1966 1967 the HC treatment ranged from 169 to 248 kg/cow, 29 to 5 1 kg/ ha, 4.6 to 5.9 ha/ cow and 67 to 186 kg/ cow, respectively, as compared Fig. 4. Percentage CP, OMD andlive (green)standing cropforpastures in heavy (HC) and moderate (MC) continuous grasing treatments. neat- to respective ranges of 182 to 237 kg/cow, 29 to 38 kg/ ha, 6.2 to 6.4 ment effcs were significant (P

472 JOURNAL OF RANGE MANAGEMENT 42(8), November 1989 a constant herbivore density over time, ecosystem response has Iieitecbmidt,R.K.,S.K.Canon,W.E.Phbek,andS.L.D0rrb0w~.1~. been limited primarily to a shift in plant species composition. In Effects of ycarlong continuous, deferred rotation, and short duration this instance, the shit is towards a species complex that is generally grazing treatmentson cow/calf production at the Texas Experimental Ranch. J. Prod Agr. (In press). less palatable (i.e., annual broomweed), less productive (i.e., short- Lawnroth, W.K. 1979. Grassland primary production: North American grasses) and more grazing tolerant (i.e., short-grasses). grasslands in perspective. p. 3-24. In: Norman R. French (cd.) Penpcc- Literature Cited tivcs in grassland ecology. Springer-Vcrlag, New York. Little, TM., and F.J. IIRIs. 1978. Agricultural experimentation: Design Awe. Oft Anal. Chem. 1975. Gffcial methods of analysis (11th ed.). and analysis. John Wiley and Sons, New York. Assoc. Off. Anal. C&m., Washington, DC. M&&k, J.C. 1984. Impacts of graxing intensity and spccialixcd grazing Bartolome, J.W. 1984. Impactsof grazing intensity and grazing systemson systems on livestock response. p. 1129-l 158. In: Developing strategies vegetation composition and production. p. 917-215. In: Developing for rangeland management. Westview Press, Boulder, Colo. strategies for rangeland management. Westview Press, Boulder, Colo. Sime, P.L., J.S. Sir&, end W.K. Lauenrotb. 1978. The structure and Danckwerta, J.E., end A.J. Aucamp. 1986.The effect of rangecondition on function of ten w&cm North American grasslands. I. Abiotic and the grazing capacity of semi-aridSouth African savanna. p. 229-230.In: veaetational characteristics. J. Ecol. 66951-285. Proc. 2nd Int. Bangeland Congress. (eds. P.J. Joss, P.W. Lynch and Snehcar, G.W., and W.C. Co&ran. 1967. Statistical methods (6th cd.). O.B. Wims), Amt. Acsd. of Sci., Canberra, Australia. Iowa St. Univ. Press. Ames. KBlson, Lincoln. 1%8. Influence of graxing on plant succession of rangc- Soil Comervatlon Service. 1%4. Field offi technical guide. USDA. Abi- lands. Bot. Rev. 261-78. lene, Texas. Heitschmidt, R.K., R.A. Gordon, and Jd. Bluntxer. 1982. Short duration Stoddud, L.A., A.D. Smith, end T.W. Box. 1975. Range management. grazing at the Texas ExperimentalRanch: Effects on forage quality. J. McGraw-Hill Book Co.. New York. Range Manage. 35:372-374. TBley,J.M.A.,andR.A.Terry.1%3.Atwo-stagctechniquefortheinvitro Hcitechidt, R.K., S.L. Dowbower, R.A. Gordon, and D.L. Price. 1985. digestion of forage crops. J. Brit. Grassland Sot. 18:104-111. Response of vegetation to livestock grazing at the Texas Experimental Van Sosst. PJ., and R.H. Wine. 1967. Use of detergents in the analyses of Ranch. Texas Agr. Exp. Sta. Bull. 1515. fibrous feeds. IV. The determination of plant cell wall constituents. Heitechmidt, R.K., S.L. Dowbower, and J.W. Walker. 1987.1~vs. 42- Assoc. off. Anal. Chem. 5050. paddockrotationalgraxing:Foragequality.J.RangcManagc.40:315-317.

Moving in the near future? Please send us your change of address as soon as you can so you won’t miss any issues of the Journal or Rangelands. Please send your old address lable plus the following information to Society for Range Management: 1839 York Street, Denver, CO 80206. Name New Address C&Y Stete Zip

ATTACH OLD ADDRESS LABEL HERE

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 473 Diet and forage quality of intermediate wheatgrass managed under continuous and short-duration grazing

M.L. NELSON, J.W. FINLEY, D.L. SCARNECCHU, AND S.M. PARISH

AhStUCt Diet quality and foragequality were determined under lort- ment models to predict animal, plant, and economic performance dlw8tioa8ndcontinaou8~of iIlwmd8te WhaQlnlm of alternative grazing systems. Therefore, objectives of these trials (~brqpprouriidanrarlitlm)in~ygizlng~in19115~d were to determine seasonal changes in diet composition and in vivo lw16. lhe short-duration lllly waedivided into 8 e&units grazed organic matter digestibility under short duration and continuous eequentially for 3 dayr each. Ss crowbred heifers and 2 eaopba- grazing and determine effects of intermediate wheatgrass maturity geally fietuhhd steere were randomly aedgned to uch grazing on intake, digestibility, in situ rate of digestion and fluid and treatment. Aoimab were wdghed and fecal mmplm, pasture mm- particulate passage rates. plea, and diet (aopbageal masticate) samplea were collected in Materials and Methods each of tbe three 24day per&de. In vitro organic matter diup peuence (IVOMD) of steer dkta uoda shortduration grezing Grazing Trials declinedlineariyavoaperfodsofbothyeusendacroesdays Two 72day grazing trials were conducted on part of a 48-ha within periods in 1986. Crude protein content of steer diets under intermediate wheatgrass (Agropyron intermedium] pasture in an- ehortduration grazing dedined quadratically acroes periodn in tral Washington. The study area was described in detail by Pierson MM. Crude protein and IVOMD content of steer dieto under and Scamecchia (1987). The area was divided by ocular and strati- continuous grazing dedined linearly in 1985. The effects of 4 graphic estimates of equivalent standing crop into 2 units, 1 for mataritia of intermediate wbea@rase on d@tbUty and mminal SDG and 1 for continuous grazing. The SDG unit was further kbetice were compared in a 4 X 4 Latin quare dedgn with 4 subdivided into 8 subunits. Short duration grazed subunits aver- rumhully and abomasally w crossbred wethere. Organic aged .35 ha in both years and the continuously grazed unit was 3.76 matter intake and diptibility, in eku rate and extent of NDF and 3.54 ha in 1985 and 1986, respectively. digatiotb Uq&l passage rate and partkubte man fhbwing from Six crossbred pregnant heifers (average weight 356 kg) and 2 tbeN3ncndcereucdliawly~hrenucdfo~ge~~.Thac crossbred steers with esophageal fistulas (averageweight 326 kg in d8ta suggested that effects of forage maturity or period of grazing 1985, 273 kg in 1986) were randomly assigned to each of the 2 bad similar effects 011diet quality and forage quality. However, grazing systems. Animals in the continuous grazing treatment dictq~llndcrrbort_duntion~gIbodcelined~orrdlyr grazed the entire unit for 72 days. Initial stocking density on the withinsubpaib. continuous unit was .8 and .9 AU/ ha and on the SDG unit was 1. I and 1.1 AU/ha for 1985 and 1986, respectively. Animals in the Key Wards: Agropyron iidcnrsormum,fora~beef&tle,Weth~ SIX treatment grazed each subunit for 3 days; then, all animals dbatim, Fee r8tee were moved into the next subunit. Therefore, subunits were grazed Because the demand for grazing is projected to increase (Reid for 3 days followed by 21 days of rest in each of the 3 periods. Initial and Klopfcnstein 1983), management systems must be developed stocking density for the SDG subunits ranged from 6.9 to 9.9 to increase range and pasture productivity and cff%ziency.Maturity AU/ ha in 1985 and 6.2 to 11.7 AU/ha in 1986. Stocking variable of forage plants (Brady 1973)and type of grazing system (Pitt 1986) terminology was according to Scamecchia and Kothmann (1982) affect the chemical composition of forage. As forage plants and animal-unit-equivalents were based on the model of Scamec- mature, crude protein content and digestibility decrease and fiber chia and Gaskins (1987). The 72day studies were conducted from fraction concentration increases. Dietary chemical composition 18 May to 29 July 1985 and 23 May to 3 Aug. 1986. Mineral and increased forage maturity alter rate of passage (Bull et al. 1979) supplement (50% dicalcium phosphate and 50% trace mineralized and fermentation kinetics (Mertens 1977). salt’) was provided ad libitum. Short duration grazing (SDG) may allow greater stocking rates Twenty-five randomly distributed .5 m2 plots were clipped to than continuous grazing (Heitschmidt et al. 1982b, Jung et al. ground level before and after grazing in each subunit through the 3 1985). This could be due to increased forage quality and quantity rotations to estimate standing crop. In the continuously grazed or efficiency of harvest (Heitschmidt et al. 1982b). However, SDG unit, standing crop was estimated by clipping, to ground level, 40 has, in some studies, been shown to have no effect on animal weight randomly distributed .5 m2 plots at 7day intervals in 1985 and gain (Heitschmidt et al. 1982a.Jung et al. 1985).Chemical compo- 24day intervals in 1986. sition of diet (esophageal masticate) may be the most sensitive In 1985, heifers were weighed on day 1,2,7,8,9,31,32,33,55, measure of change due to grazing management. However, compo- 56,57,71, and 72. Esophageal masticate samples were collected on sition of forage selected by animals grazing a cool-season forage day 7 to 12 in each period alternating between a.m. and p.m. times under continuous and SDG systems as forage mature4 has not been of grazing without a previous fast to avoid fasting-induced, seIec- reported. These data are needed to adequately evaluate the grazing tive grazing (Siiahmed et al. 1977).Esophageal masticate samples system by forage maturity interactions and to be used in manage- were frozen (-20” C) until they were lyophilized. Esophageal masticate samples were composited (w/w) by steer, within day 7 to 9 AuthorsUC urirunt professor and pduate ashant, Department of Animal Scimaq associntc rofessor, Dcputmcnt of Natural Resource scicnca; and anao- and within day 10 to 12, representing subunits 3 and 4 under SDG, ciate p+amr, CcllL gc of VeteliMry Malick. Wuhington state univenity. Full- nun 99164. ‘97%NaCl, .002%Se, .014%Co, .OlS%I .032%Cu. 24% Fe, .31% Mn, ~56%Zn. Scicntifii paper 7845, a&Se of Apiculturr Md Home Emlomics. wMllingtoll State Univemitv. R-h cowhctd uder Proiect 0713. Muldpt iazptcd 9 Much 1989. -

474 JOURNAL OF RANGE MANAQEMENT 42(6), November 1989 in each period. Fecal grab samples from heifers were collected on ing because we did not harvest enough forage. Wethers were fed, in day 7 and 8 in each period, and frozen until subsequent analyses. amounts to allow 20% feed refusals (orts), twice daily at 0700 and Subsamples were taken from pasture plots. 1900. Mineral supplement (50% diilcium phosphate and 50% In 1986,heifers were weighed on day 1,2,7,8,3 1,32,55,56,71, trace mineralized salt’) was provided ad libitum. and 72. Esophageal masticate samples were collected on day 7,8, Animals were housed in a temperature-controlled, continuously and 9 representing subunit 3 under SDG, ineach period alternating lighted room. Periods were 11 days in duration, which included between a.m. and p.m. collections. Esophageal masticate samples day 1 through 7 for diet adaptations and day 8 through 11 for were froxen (-200 C) until they were lyophiied. Fecal grab sam- collection of ruminal, abomasal, and fecal samples. Feed intake ples from heifers were colIected on day 7 and 8 in each period, and was determined as feed offered, corrected for arts from day 6 frozen until subsequent analyses. Subsamples were taken from through 9. pasture plots. Rate and extent of digestion of neutral detergent fiber (NDF) of Fecal grab samples were ovendried at 50“ C and cornposited the 4 diets was determined using 5 X 10 cm dacron bags2(pore size (w/w) by animal. Pasture samples were oven-dried at 500 C. All 52 f 16 p m). About 2 g of forage, which was previously ground samples were ground through a l-mm screen in a Wiley mill. through a 1-mm screen in a Wiley Mii, was placed in each dacron Chemical analyses of pasture samples and esophageal masticate bag. Bags containing the same forage that each sheep was fed were samples included crude protein (CP) by macro-Kjeldahl (AOAC suspended in the rumen at 0708 on day 8. Duplicate bags were 1984), neutral detergent fiber (NDF), acid detergent fiber (ADF) removed from each sheep at 4,8,12,24,48,72, and 96 hours of and acid detergent lignin (ADL) according to Goering and Van incubation. Bags, after removal from the rumen, were frozen (-Up Soest (1970) and in vitro organic matter digestibility (IVOMD) by C) until subsequent analysis. Bags were thawed, rinsed with water, the Moore modiication (Harris 1970) of Tilley and Terry (1%3). and the residue remaining in each bag was quantitatively trans- Crude protein contents of esophageal masticate were corrected for ferred for NDF analysis. urea-N content. Urea-N in esophageal masticate was measured by Rate of liquid passage was determined using a 5-g dose of cobalt extracting soluble N in 10% Burrough’s buffer (Burroughs et al. lithium ethylenediaminetetraacetic acid (CoLiEDTA) synthesized 1950) at 37” C for 1 hour and measuring ammonia-N by the according to Uden et al. (1980). Rate of particulate passas was phenol-hypochlorite method (Weatherbum 1%7). All chemical determined using a 15-g dose of ytterbium (Yb) labeled forage analyses are reported as a percentage of organic matter. Ash-free prepared according to Goetsch and Galyean (1983). Ytterbium- indigestible acid detergent fiber (IADF) was measured in fecal and labeled forage contained an average of 8.4 mg Yb/ g. An aliquot of esophageal samples according to Nelson et al. (1985). A prelimi- forage of each maturity was ytterbium labeled so that labeled nary study (ML. Nelson, unpublished data) showed no effect forage and the diet of each sheep within a period were the same. (p>.l) of sampling day within collection period on fecal IADF Markers in gelatin capsules were administered into the rumen content of grazing animals. Therefore, animals were assumed to be simultaneously with the insertion of dacron bags. Ruminal con- in steady state conditions. In vivo forage organic matter digestibil- tents, samples at 0,4,8, 12,24,48,72, and 96 hours after dosing, ity (in vivo OMD) was calculated using IADF as the internal were obtained using a 15-mm diameter rubber tube with a wire marker (Harris 1970). running through the center which was attached to a rubber stopper. The tube was inserted through the ruminal strata and then MlatlollTrlaI Intermediate wheatgrass was harvested from an adjacent un- sealed. Each collection was a composite of samples from different grazed portion of the 48-ha seeded field used for the graxing sites (ventral sac, dorsal sac, and reticulum) in the reticula-rumen. studies. Forage was harvested with a rotary mower and sun-cured Composite samples were strained through cheesecloth to separate liquid from particulate fractions. Contents of the abomasum were at 4 stages of maturity (late boot stage, full head, mature and post ripe) on 26 May, 19 June, 13 July, and 6 Aug. 1985. sampled at 1000,1300,1600, and 1908 on days 8,9,18, and 11 in Four ruminally and abomasally fistulated crossbred wethers each period, respectively. (average weight 82 kg) were randomly allotted to a 4 X 4 Latin ‘RIO2 Muvclaii White, Erlmgcr, Slumgut & Co., Inc.. New York, NY. square design. Dietary treatments were the 4 stages of maturity of intermediate wheatgrass. Two cells of maturity 1 forage were miss-

Tabkl. Qmntityandqdtyofinte-whatgruaun&rsbortda~t~oeandeoounmmsgruing.

1985;Period 1986; Period 24!29 17-222 11-163 2b31 22-242 16183

Item MaY Jun Jul fiY JUn Jill SEW Shortduration grazing Crude protein, $ of organic matter 8.3 6.3 4.3 9.2 6.2 3.9 0.4 (OW

InStanding vitro OM crop, digcstibiitv. kg/ unitb %+j 350377.6 284772.8 173265.2 324070.3 318465.3 % &! Standing crop, g/ rn% 125.6 102.1 62.1 116.1 114.1 72.6 8.5 Continuour gmzing Crude protein, 9%of or@ 9.1 6.2 3.9 9.6 5.6 6:*3 0.4 In vitro OM diitibilitY, Bbd 75.8 74.3 68.6 66.7 63.8 1.1 Standing crop, kg/ unit 3504 2460 1352 3199 3397 30th 700 Standing crop, g/m% 93.2 65.4 36.0 90.3 95.9 84.5 18.4 staluhrd errorof the man. ulcar effect of period (K.05). %lau effect of riod (K.01). %fkct of year ( E .01).

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 475 Tabk 2. C~podtion ofsteer dkk (amphaged masticate)under shortduratIon and conthmo~~ grazing, 1WrS.

-Period I -Period 2 -Period ‘I 24-26’ 21-29 17-19 20-22 11-13 14-16 Item May WY JUn Jun Jill Jul SEMb % of Organic Matter

1; vitro OM digestibiitfl 72.113.4 72.512.3 70.212.0 70.910.9 61.08.5 7.5 0.5 60.1 2.3 Neutral detergent fibcf 52.2 55.6 52.2 53.5 59.1 58.8 1.6 Acid detergent fibef 30.9 31.6 29.6 30.9 38.0 36.5 Acid detergent lignin 6.9 5.9 4.5 6.6 9.6 9.1 & Continuous grazing CNdt protein’ 13.4 13.0 10.3 8.9 7.3 6.3 In vitro OM digestibility’ 69.3 74.6 66.4 61.7 55.5 58.0 k9 Neutral detergent fiber 53.9 52.4 60.1 57.7 60.1 60.9 6:2 Acid detergent libef 30.8 29.6 33.8 32.9 31.5 38.6 2.7 Acid detergent lignin 8.3 1.3 6.8 7.6 9.6 12.1 1.2 tUndcr shortduration grazing, fit data under each period WCICOII subunit 3 and second dates on subunit 4. Standard error of the mean cabxdated from &ccr by period by day; n=2. xiir cffcct of Lwiod IJK.09. %ii CtTectof -&ad @c.osj. ‘Quadratic CtTcctof period (fc.10). Quadmtic effect of period (X05).

Samples of feed, orts, feces, abomasal contents, and ruminal equations were calculated to relate diet quality with standing crop contents were dried at 500 C in a forced-air oven. Samples of feed, composition. orts, feces, and abomasal contents were ground through a‘ Wiley Digestion Trial Mill (l-mm screen). Samples were composited (w/w) by wether Data were analyzed as a 4 X 4 Latin square (Steel and Torrie within period. Chemical composition of feed, orts, abomasal con- 1980). Linear and quadratic orthogonal contrasts for forage tents and feces were determined by methods described for the maturity were calculated. grazing studies. Samples of ruminal liquid were thawed and centrifuged at Ramlts and Discussion 30,090 X g for 20 min. The supematant was decanted for subsequent Co analysis. Samples of ruminal particulates were prepared Grazing Trials for Yb analysis according to the methods described by Ellis et al. Numbers of animal-units (AU) per grazing treatment calculated (1980). Concentrations of Co and Yb were determined using a according to Scarnecchia and Gaskins (1987) were 2.9 and 3.2 for Perkin-Elmer model 2380 atomic absorption spectra-photometer the beginning and end of the 1985 study and 3.1 and 3.4 for the (Perkin-Elmer 1971). The natural log of Co or Yb concentration beginning and end of the 1986 study. Grazing pressures for the was regressed on hour post-dosing to determine rates of fluid or SDGsubunitsrangedfrom5.5 to20and6.7 to 19AU/torifor 1985 particulate outflow from the rumen (Grovum and Williams 1973). and 1986, respectively. Grazing pressures for the continuous unit Ruminal liquid volume was calculated from predicted initial con- ranged from .9 to 3.5 and 1.0 to 1.63 AU/ton for 1985 and 1986, centration of Co and amount of Co pulsedosed into the rumen. respectively. Mean stocking density, calculated from mean animal Ruminal particulate mass was calculated from predicted initial Yb weight, for SDG subunits ranged from 7.4 to 10.5 and 6.6 to 10.6 concentration and actual Yb dosage. Particulate mass flowing AU/ha for 1985 and 1986, respectively. Stocking density for the from the rumen was calculated as the product of ruminal particu- continuous unit ranged from .8 to .85 and .9 to .% AU/ ha for 1985 late passage rate and ruminal particulate mass (Grovum and Willi- and 1986, respectively. System stocking levels were 2.7,2.8,2.0,2.2 ams 1973). Content of IADF in feed, abomasal, and fecal samples AUM/ha for the SDG unit in 1985 and 1986 and the continuous was used as an internal marker to determine digestibility coefti- unit in 1985 and 1986, respectively. cients. Digestibilities in the rumen, lower tract, and total tract were Pasture forage (Table 1) under both grazing systems contained calculated according to Harris (1970). In situ rate and extent of similar crude protein (CP) content both years. However, in vitro NDF digestion, and discrete lag time were determined by fitting the organic matter disappearance (IVOMD) was greater in 1986 than model of Mertens and Loften (1980). 1985 for both grazing systems. Pasture forage CP and IVOMD under both grazing systems and standing crop under SDG declined Statisticd Analysis linearly across period. Grazing Trials A primary objective was to identify interactions of period with Esophageal masticate composition and in vivo organic matter subunits or days within period. These interactions were expected digestibility were analyzed with split block or repeated measures due to reduced selective grazing across days within an SDG sub- designs (Gill and Hafs 1971, Steel and Torrie 1980). The model unit and the possibility of differential effects of a grazing system included effects of animal, period, day, and the 2 and 3 way across period on plant regrowth. Diet chemical composition of interactions. Animal by period was the error term for animal and steers under SDG in 1985 (Table 2) was not affected by subunit period. Animal by period by day was the error term for day and the within period. Therefore, no differences due to subunit were 2 way interactions with day. Linear and quadratic orthogonal detected, Diet CP and IVOMD declined; NDF and ADF increased contrasts were calculated for period. Standing crop mass and across period as the plants matured. composition were analyzed with the previously described repeated In 1986, diet IVOMD of steers under SDG (Table 3) decreased measures designs. Linear and quadratic orthogonal contrasts were across day within subunit and period. Diet NDF and ADF calculated for period. Regression equations were calculated to increased across day within subunit. This indicated that degree of relate diet quality and in vivo OMD with day of grazing. Additional selective grazing was altered across day within subunit as grazing

476 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 Table 3. Compodtloa of steer d&u (esophageal masticate) umler sbortduration and cootimwa grazhg, WM.

-Period 1 -Period 2 -Period b 29 30 31 22 23 24 16 Item fiY MaY MaY JUn JUll JUU Jul ;: ;: SEW $bof Organic Matter Shortduration grazing crude protein 18.6 18.0 17.4 13.7 20.8 11.0 16.9 9.1 11.6 4.2 In vitro OM digestibilityi’C 74.5 71.1 73.0 67.2 65.1 60.7 65.8 52.4 5.7 Neutral detergenttibef 63.7 72.1 z: 55.7 65.2 79.9 66.0 76.3 74.2 9.8 Acid detergentfit&t 35.3 39.2 36:0 30.6 30.2 43.5 33.0 40.2 41.7 3.7 Acid detergentlignin 5.6 5.3 5.0 4.0 6.4 6.2 6.4 7.0 5.8 .9 continuous glazing crude protein 17.2 15.3 16.3 13.2 18.0 15.5 11.5 9.8 9.4 3.9 In vitro OM digestibility’ 75.9 77.8 75.7 73.5 57.0 60.2 58.6 62 1 2.9 Neutral detergenttlber 60.5 59.7 64.6 60.8 61.7 65.8 71.8 z.3 66’2 5.2 Acid detergentfiber 32.8 33.4 35.5 34.1 37.1 38.4 40.9 34:0 35:5 2.5 Acid detergentligninb 3.3 3.3 4.0 5.0 6.3 5.5 6.5 5.5 5.1 1.1

%tandard error of the mean calculated from steer by period by dar. n = 2. bLinear effect of period (Iy.10). %ncar &cct of day (K.10). kinr.ar effect of day (K.05). -Period by day interaction (K.05). under SDG reduced the availability of the preferred plant parts. herbage available per treatment in period 2 (Table 1) was similar Chemical composition of steer diets under continuous graxing in between grazing units. However, standing crop (g/m*) appeared 1985 (Table 2) was not affected by day within period similar to greater in the SDG subunits, which may have provided a greater steers under SDG. Esophageal masticate CP and IVOMD declined chance for selectivity. and ADF increased across period as the piants matured. Significant differences across days in period 1 would not be In 1986, diet ADL of steers under continuous grazing (Table 3) expected since animals in both graxing systems were graxing hom- increased across period. An exception to the general trend was an ogeneous new pastures. By period 2, regrowth effects allowed increase in IVOMD on day 2 of period 3. No effects of day within expression of differences in plant parts and chemical composition period were detected for diet CP, NDF, or ADF, which indicated of forage caused by differences in grazing. By period 3, standing that degree of selective grazing was not altered across the 3 sam- crop averaged only 1,549 kg/treatment in 1985 (Table 1); mostly pling days in a period under continuous grazing. stem remained in both grazing treatments and the homogeneous A quadratic period by year interaction was detected for in vivo forage offered little opportunity for selective grazing. organic matter digestibility by heifers under SDG (Table 4). This The decline in CP content of esophageal masticate across periods was consistent with repotted values (Sims et al. 1971, Karnstra Table 4. In vlvo organic matter dl@lblllty ollntermedlate whatgram by 1973, Rauzi 1975, Svejcar and Vavra 1985). There is wide variation bdfera under abort duration or continuom grazing. reported for IVOMD of forage varying in maturity, although values reported by Svejcar and Vavra (1985) and White (1983) were Year similar to the IVOMD of esophageal masticate collected in the present study. Further, Pitts and Bryant (1987) and Taylor et al. 1985 1986 (1980) in Texas reported similar rates of change in esophageal Period PCriod masticate contents of CP and IVOMD across collection date under 1 2 3 1 2 3 SDG, continuous, or Merrill grazing systems, as in the present 24-25 17-18 11-12 29-30 22-23 16-17 G=iW study. Greatest differences between in vivo OM digestibiities and Management May Jun Jul May Jun Jul SE’ IVOMD occurred in periods 1 and 3 with IVOMD being higher in -lo viva organic matterdigestibility, %- both periods. The lower in vivo OM digestibility in period 1 could Short durationb 66.4 71.0 58.8 70.0 61.0 54.3 2.0 have been due to a faster rate of particulate passage which would Continuousc 63.5 60.5 51.7 64.9 59.0 51.0 1.3 have reduced in vivo OM digestibility. Increased NDF, ADF, and %tanda~crro~ of the mea! calc+ed from period by heifer (year) for n = 5. ADL in esophageal masticate across periods was consistent with bQ$.ad=d~~~~&y~~~~~)tion (cU.05). reported values (Kamstra 1973, Cogwell and Kamstra 1976, Hart . . et al. 1983). interaction was apparently due to increased digestibility in period 2 Heifer average daily gain under SDG averaged .68 and .89 kg/d of 1985. However, standing crop IVOMD~in 1985 showed only a in 1985 and 1986, respectively. Heifer average daily gain under small reduction from period 1 to 2, which may indicate that stand- continuous grazing averaged .63 and .83 kg/d in 1985 and 1986, ing crop measurements were not good single point predictors of in respectively. vivo measurements. In vivo organic matter digestibiity (OMD) of Digestion Trial heifers under continuous grazing decrease d across period from Forage crude protein content declined and forage NDF and 64.2 to 51.4% similar to standing crop and esophageal masticate ADF content increased with increased forage maturity (Table 5). IVOMD. Forage ADL was not affected by forage maturity. The decline of Chemical composition of esophageal masticate is a result of the CP with increased maturity of forage was greater for this study interaction between chemical composition of forage and selection than the CP decline reported by Sims et al. (1971), Kamstra (1973), by the animal. The quantity of forage present can affect diet and Svejcar and Vavra (1985). selection and intake by animals (Hodgson 1981). In 1985, total Organic matter intake by wethers and digestion coeffkients for

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 477 TM!& RfktOfWtdQd hwmedfate wheatgraw (barvat date) on compaltlon oforgank mattmend vohmtary Intake, dlgdbmy, ntwtml detcrlcdfearkiodka,andramidpaaa~Linakrofwcthn.

Harvest date Item 26 May 19Jun 13Jul 6 Aug SEW Composition, % of organic matter

crudeNeutral protein datargent’ fibar (NDF)* 70.110.5 72.57.1 7:: 7::; :: Acid detergent f&r (ADF)’ 41.4 422 4414 43.0 0.43 Acid detergent lignin (ADL) 4.2 5.3 6.3 6.4 1.03 Intake and digestibility by w&era Organic matter (OM) intaka, g/day” 973.2 747.2 563.2 447.0 55.9 OM digeatibiity, sbb 70.5 64.2 z 58.5 2.5

NDFADF digestibiity,diitibiity, %96”’ 76.275.4 E 66:7 62.765.0 33 Nitrogen digaatibiity, % 83.2 rro.1 81.8 83.8 2:o In situ forage NDF kin&u in wethem Digestion la& houra -0.t 0.07 1.2 1.8 Rata of NDF disaPPaarance, %/ hourC 4.7 4.4 :*: Extant of NDF disappeerancc, %” 82.8 75.5 69:o 2:; ::3 Ruminal Passage rate kin&Y of wathern Ruminal liquid volume, liters 2.8 3.3 3.3 3.2 0.2 Ruminal liquid Passage, rata, %/bout)’ 8.2 7.5 5.9 5.2 0.7 RttmiMlp#iCtdatamall~kg 1.5 1.7 1.2 1.5 0.2 Ruminal particulate pamage rate, /hour 4.8 4.1 3.9 2.7 1.1 Particulate mass outflow. n/ho 3 65.8 57.7 45.8 39.8 11.5 %andud crfor ofthe - for n=4. bLinarcffectofhawestdate(F<.10). ziuar effectof llalvat date (X.05). ‘Liacardiea of lluvat date (K.01). ‘oUdratic c&t of harvcltdate (X.10). Table 6. Ragradon aquatlona predktlog maymith ~aopbycrl~teof~~~rbo~d~toaladeoafiew~pulnltnw&yd~l.

YC8r 1985 1986 Root Root Itam. Equation rr MSE P Equation rr MSE P Short-duration grazing CP = 14.52 -.10 Day .89 .84 .004 CP = 18.87 - .ll Day .61 2.17 IVOMD = 75.18 - 23 Day .65 4.06 .053 IVOMD = 74.62 - 25 Day .72 3.74 :g NDF = 49.88 + .I4 Day :Z 2.85 .071 NDF = 65.64 + . 10 Day .19 4.98 .392 ADF =28.09+.15Day 3.59 .120 ADF = 35.66 + .03 Day .03 4.0 .732 ADL = 5.24 + .06 Day 20 2.67 .380 ADL = 5.01+ .02 Day 23 1.0 .334 Continuous grazing CP = 14.42 - .13 Day 90 1.02 CP = 18.02 - .I2 Day .52 2.87 .105 IVOMD = 72.94 - .29 Day .74 4.07 :g IVOMD = 76.89 - 30 Day 69 4.76 .041 NDF = 53.95 + .13 Day .ll 8.88 .526 NDF = 60.53 + .09 Day :: 10.46 .701 ADF = 29.52 + .14 Day 48 3.49 .125 ADF = 33.79 + .06 Day 6.85 692 ADL = 7.40 + .03 Day .16 1.50 ,434 ADL = 3.49 + .05 Day .63 .83 .058

‘CP = crudeprotein. IVOMD = in vitro organic matter diippmrame, NDF = neutral detergent fiber, ADF = acid detergent fiber, ADL = acid detergentIi.@.

OM, NDF, and ADF decmascd with increased forage maturity. OM digestibility, % = 69.75 - .15 Day (rr =, PC.05). Forage NDF, Post-ruminal digestion coefIlcients (calculated by difference) were ADF, and ADL regressions were not signifiint and no regmssions not affected by forage maturity and averaged .8,1.5, and 3.1% for were improved by fitting quadratic regression lines. In vivo OMD OM, NDF, and ADF, respectively. Additionally, in situ rate and relationships, under SDG, were OMD, % = 61.14 + .78 Day - .02 extent of NDF digestion, rate of liquid passage and particulate Day2 (rr = .72, KOOl) and OMD, 9%= 71.91 - .32 Day (rr = 64, mass outtlow from the rumen decreased with increased forage F<.OOl) for 1985 and 1986, respectively. Under continuous graxing maturity. Rate of particulate passage tended to decrease with the relationships derived were OMD, $ = 66.41- 24 Day (r2 = .72, increased forage maturity. K.0001) and OMD, $ = 67.61 - .29 Day (rr = .77, K.0001). Regmdon Rela~mhipa Differences in potential prediction equations between studies arc In the graxing trials, rate of decline (Table 6) in diet CP across likely due to differences in animal species, amount of diet eelectiv- ity allowed, and forage conservation. days of grazing ranged from .lO to .13 percentage units/d. Diet Diet CP and IVOMD content could be adequately predicted IVOMD declined from .23 to 30 percentage units/d across days of grazing. Regression equations predicting forage CP and OM diges- from pasture composition. In 1985 under SDG, the relationships tibility by wethers from harvest day after initiation of the graxing derived were Diet CP, %b=3.65 + 1.21 Pasture CP(r2= .89, X.01) and Diet IVOMD, % = 1.51+ .92 Pasture IVOMD (rr = .70, K.05). trial were Forage CP, % q 10.23 - .08 Day (r2 = .88, PC.001) and

478 JOURNAL OF RANGE MANAGEMENT 42(6). November lB6B In 1986, relationships under SDG were Diet CP, $ = 8.45 + 1.04 H~~,B.H.,O.M.A~~~~B~,D.H.CI~~~,M.B.M~NIUB,M.H.H~~&I~~~ Pasture CP (r* = .61, PC. 1) and Diet IVOMD, $ = 1.21 Pasture CP J.W. Waggoner, Jr. 1%3. Duality of forage and cattle dii on the - 12.09 (r* = .72, X.05). In 1985 under continuous graxing, the Wyomiqg-high plains. J. Rat@ M&age. 36&51. relationships derived were Diet CP, % = 2.80 + 1.18 Pasture CP (rr = H&chmIdt.B.K,R.A.Corda.aadJ.S.Bbmtsu.1982s.Shortduration graxing at-the T&s Expcrim&l Ranch Effects on forage quality. J. 90, X.01) and Diet IVOMD, % = 1.91 Pasture IVOMD - 75.53 Range Manage. 35:372-374. (r* = .82, X.01). In 1986, relationships under continuous grazing Hdtach&R,R.K,J.R.Frnure,D.L.Rke,andL.R.BItta&ome.1~b. were Diet CP, % = 8.44 + .89 Pasture CP (r2 = .41, X.2) and Diet Short duration graxingat the Texas Experimental Ranch: Weight gains IVOMD, % = 2.89 Pasture IVOMD - 111.88 (r2 = .73, K.05). of growing heifers. J. Range Manage. 32375-379. Hodgeon, J. 1981. InfIuencc of swardcharacteristics on diet selection and Summary herbage intake by the xraximzanimal. D. 153-166. In: J.B. Hacker. (Ed.) This study involved 3 trials designed to assess the effects of Nutritional lim& to a&al production from pastures. Commonwealth graxing system and plant maturity on animal and plant responses. Agricultural Bureaux, Famham Royal, UK. In this study, forage quality declined with increased maturity as Juag, H.G., R.W. Rice, amI L J. Koong. 1985. Comparison of heifer weight gains and forage quality for continuous and short-duration grax- evidenced by decmased CP and increased fiber fraction contents. ing systems. J. Range Manage. 383144-148. In the grazing trials, this decrease in forage quality resulted in Kamstm, L.D. 1973. Seasonal changes in quality of some important range decreased quality of forage consumed by animals. In vivo and in grasses. J. Range Manage. 26289-291. vitro digestibility also decreased with increased forage maturity. In M&em, D.R. 1977. Diet& fiber components: Relationship to rate and the digestion trial, rate and extent of ruminal NDF digestion extent of rumhal digestion. P&l. Proc. 36187-192. declined with increased forage maturity. Poppi (1980) suggested Mertem, D.R., and J.R. L&en. 19%. The effects of starch on forage !Iber that decreased extent of digestion led to decreased rate of passage digestion kinetics in vitro. J. Dairy Sci. 631437-1446. Nelson, M.L., TJ. Klopfemtein, R.A. B&ton, and S.R. Lowry. 1985. and a subsequent decrease in intake. This suggestion is supported Protein supplementation of ammoniatcd roughages. III. Corncobs sup by data from the digestion trial in which intake and particulate plementcd with a blood meal-corn gluten meal mixtum-stccr studies. J. mass flowing from the rumen declined significantly with increased Anim. Sci. 61:1567-1575. forage maturity. Perkin-Elmer. 1971. Analytical methods for atomic absorption spectro- Grazing systems have been used in attempts to alter chemical photometry. Perkin-Elmer Corp.. Norwalk, Ct. composition of plants and, ultimately, intake by ruminants. Pierson, F.B., and D.L. Scameechh. 1967. Defoliation of intennuliate wheat grass under seasonal and short-duration graxing. J. Range Man- Although statistical comparisons were not appropriate, these data age. 4&U&232. suggest that effects of maturity were similar under SDG or contin- Pitt, M.D. 19%. Assessment of spring defoliation to improve fall forage uous grazing. It is likely that variables more fundamental than the quality of bluebunch wheatgrass (Agropyron spicatumJ J. Range Man- choice of grazing system are more important in determining sea- age. 39175-181. sonal change in forage and diet quality and animal production. Pit@ J.S., and F.C. Bryant. 1987. Steer and vegetation response to short Therefore, numerous field studies are needed to develop and vali- duration and continuous gmxing. J. Range I&age. 40&6-389. PODDI.D.P.. DJ. Mlmou. sod J.H. Ternouth. 1980. Studies of cattle and date sound empirical and/or theoretical models of the effects of sh&p eating leaf and S&I fractions of grasses. I. The voluntary intake, grazing management on seasonal changes in forage and diet qual- digestibility and retention time in the reticul~rmnen. Australian J. Agr. ity and animal production. Rcs. 3299-108. Rauxl, F. 1975. Seasonal yield and chemical composition of crested whcat- Literature Cited grass in southeastern Wyoming. J. Range Manage. 28219221. AOAC. 1984. Official methods of analysis (14th Ed.). Association of Reid, R.L., and TJ. KIopfeusteln. 1983. Forages and crop residues: Qual- Off&l Analytical Chemists. Washington, DC. ity evaluation and systems of utilization. J. Anim. !ki. 52534-554. Brady, C.J. 1973. Changes accompanyinggrowth and senescence and effect Scameccbia, D.L., and C.T. Caskins. 19117. Modeling animal-unit- of physiologicalstress. p. 317~349.In:G.W. Butlerand R.W. Baiiy(Ed.) cquivaknts for beef cattle. Agr. Sys. 23: 19-26. - Chemistry and biochemistry of herbage. Academic Press, London. Scarnecchla, D.L., awl M.M. Kotbmam. 1982. A dynamic approach to BuB,L.S.,W.V.Rumpler,T.F.Swaeney,andR.A.Ebm.1979.Influcnceof graxing management terminology. J. Range Manage. 35262464. rmninal turnover on site and extent of digestion. Fed. Proc. 3827132719. Sidabmed,A.E.,J.E.Morrb,W.C.WeIr,andD.T.TmeU.1977.Effectof Burroughs, W., N.A. Frank, P. Gerlaugb, and KM. Bethke. 1950. Prcli- kngth of fasting on intake, in vitro digestibility and chemical composi- minary observations upon factors intluencing cellulose digestion by tion of forage samples collected by esophageal tistulated sheep. J. Anim. rumen microorganisms. J. Nutr. 409-24. Sci. 45:885-890. Cogswell, C., and L.D. Kamstm. 1976. The stage of maturity and its effect Siam,P.L., G.R. LoveIl, and D.F. Hervey. 1971. Seasonal trendsin h&age upon the chemical composition of four native range species. J. Range and nutrient production of important sandhill grasses. J. Range Manage. Manage. 29&o-463. 24:55-59. EBB, W.C., C. Lascano, R. Teetar, aud F.N. Owem. 1982. Solute and Steel, R.G., and J.H. Tonic. 1980. Principles and procedums of statistics. particulate flow markers. p. 37-56. In: F.N. Owens (Ed.) Protein McGraw-HiU Book Co., New York. requirements for cattle: Symposium. Oklahoma State Univ. Press, Svejcu, T., and M. Vavm. 198!X Seasonal forage production and quality Stillwater. on four native and improved plant communities incastern Oregon. Tech. GBl, J.L., and IID, Hafs. 1971. Analysis of repeated measurements of Bull. 149. Agr. Exp. Sta., Oregon State University, Corvallis. animals. J. Anim. Sci. 3333X-336. Taylor, C.A., M.M. Kotbma~~~,L.B. Merrill, and D.P. ElIedge.1980. Diet Goerlug, ILK., and P J. Van Soat. 1978. Forage fiber analyses: Appara- selection by beef cattle under high-intensity low-frequency, short- tus, reagents, procedures and some applications. Agr. Handb. 379, ARS. duration, and Merrill grazing systems. J. Range Manage. 33:428434. USDA, Washington, DC. TBley,J.M.A.,audR.A.Terry. 1%3.A two-stagctechniqucfortheinvitro Goatach, A.L., and ML. GaIyan. 191)3.Ruthenium phenanthroline, dys- digestion of forage crops. J. Brit. Gmssl. Sot. 18:104-l 11. prosium and ytterbium as particulate markers in beef steers fed an Udeu, P., P.E. Colu~d, and PJ. Van Soat. 1980. Investigation of chro- all-alfalfa hay diet. Nutr. Reo. Int. 27:171-178. mium, c&urn and cobalt as markers in digesta. Rate of passage studies. Gronuu, W.L.; aud VJ. WBlium. 1973. Rate of passage of dig&a in J. Sci. Food Agr. 313625632. sheep. 3. Differential rates of passage of water and dry matter from the Watberburn, M.W. 1%7. Phenol-hypochlorite reaction for the determi- reticula-rumen and caecum and proximal colon. Brit. J. Nutr. N):231- nation of ammonia. Anal. Chem. 39971-975. 240. WI&e, L.M. 19113.tional changes in yield, digestibility, and crude Harris, L.E. 1970. Nutrition research techniques for domestic and wild protein of vegetative and floral tillers of two grasses. J. Range Manage. animals. Logan, Utah. 36402-404.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 479 Association of relative food availabilities and locations by cattle

D.W. BAILEY, L.R. RITTENHOUSE, R.H. HART, D.M. SWIFT, AND R.W. RICHARDS

AbStlUCt Four yearUng steers were trained and observed in a parallel-arm largest to smallest reward (Hulse and O’Leary 1982, Roberts and maze. Tbe purpose was to determine if cattle bad tbc ability to Van Veldhuizen 1985). This is evidence that they can remember associate locations witb relative food availabilities. Tbc study con- where they have foraged and the amount of food found there. sisted of 3 phases. In phase 1, ail 5 arms contabred 0.4 kg of gabt. The objectives of this study were to determine if cattle could In phase 2, the amount of grahr hr each arm was systematicilly distinguish and remember the amount of feed consumed in differ- varied from 0.1 to 0.8 kg. In phase 3, placement of grain was ent arms of a parallel maze and use that informtion for selecting reversed. Steers performed efficiently in all 3 phases of tbe study. arms during the next trial on the following day. The overaIl-mean number of correct cboicee in tbe first 5 entrancea Materials and Methods was 4.69 as compared to 3.73 by chance. Arms selected for choices 2,3, and 4 during tbe Iast 5 MaIs of pbase 2 were dIffercut(pherbivores appear to match the time 2 spent foraging in a plant community with the forage resources of in the community (Senft et al. 1987). Although this matching pattern has been described, the underlying mechanisms or behaviors that 3 explain this response pattern are not completely understood. In order to predict nutrient removal in different habitats, hypotheti- cal mechanisms or spatial-decision rules that herbivores might L 1 utilii should be developed and tested. One possible mechanism is that herbivores return to nutrient- rich, productive plant communities more often than to less- productive plant communities. This mechanism requires that her- Fig. 1. Five-arm parallel maze. bivores have the ability to remember where they have foraged and the resource level that was found there. Bailey et al. (1987) found that cattle have an accurate spatial memory in radial- and parallel- Preliminary training consisted of guiding each steer into the arm mazes. Other animal species, such as rats and pigeons, have decision area, assisting it down each arm, and allowing it to con- also performed accurately and efftciently in radial-arm mazes sume the 0.4 kg of grain mix placed in an opaque feeder at the end (Olton 1978, Roberts and Van Velduizen 1985). Rats and pigeons of each arm. This was continued once per day for 5 days to can learn to distinguish among quantities of food associated with 4 familiarize the steers to the maze. different arms in a radial maze and order their arm choices from After preliminary training, a trial consisted of guiding a steer into the decision area and allowing it to freely choose arms until all Authors arc former uatc research associate and rofesror, Range Science Department, Colorado LTdtate Univeristy, Fort Colliaa 805P 3; range scientist, USDA- grain was consumed. If the animal had not consumed all the grain ARS, Hi Plains Grasslanda Research Station, Chcyc~e, Wyoming 82001; asso- within 40 min, the steer was assisted (herded) into an arm still cute pro $cssor, Range Science Department, Colorado State University rofewor, Department of Psychology, Colorado State University. Research was fun*% cd in part containing grain. If the observer only started the animal moving by the Colorado state University Agricultural Experiment Station, Project 150771. and avoided influencing its choice, a push was recorded, and the Manuscript accepted 9 March 1989.

480 JOURNAL OF RANGE MANAGEMENT 42(6). November 1989 subsequent choice was termed non-assisted. A correctchoice was required very little assistance from the observers. On 2 occasions, a defined as entry into a previously unentered arm without assistance steer was herded into an arm still containing grain after a 40-min- followed by consumption of grain. A mistake or repeat was defined time period had been exceeded (assist). On 5 trials, a steer was as entry into an arm where grain had been consumed previously. started moving (push) after the time period was exceeded. Assist- Each steer was placed in the maze and observed daily until the end ance from the observer was required only for the last choice. of the study, and the order of placement was randomized. During Of greater importance was the order in which the arms were the interval between observation periods, steers were kept in a chosen. Arm 3 (nearest to the entrance gate) was the first arm nearby pasture. To minimize the chance of steers using grain odor chosen in 93% of the trials. In phase 2, all steers quickly learned to for a cue in arm selection, additional grain (0.4 kg) was placed at choose arms containing 0.8, 0.6 or 0.4 kg of grain in the first 3 the end of each arm. This grain was placed in an open container choices. During the last 5 trials of phases 2 and 3,85% of the first 3 behind the feeder and beyond the reach of the steer. choices were of arms that contained 0.8, 0.6, and 0.4 kg of grain The study consisted of 3 phases that were distinguished by the (Table 2). These arms would be expected to be selected for the first amount and location of grain. For the first 5 days (phase I), 0.4 kg Table 2. !Jequenceln wbkb vuloue amouote of grain (kg) were conearned of grain was placed at the end of each arm. During the next 7 days lntbelastStrleleofpbaxa2and3. (phase 2), steers within each identical-twin set were randomly assigned to either the ascending or the descending treatment. For the ascending treatment, 0.1,0.2,0.4,0.6, and 0.8 kg of grain were sequence Percent of Sequence’ frequency total trials placed in arms 1, 2 ,3, 4, and 5 (Fig. 1), respectively. For the descending treatment, 0.8,0.6,0.4,0.2, and 0.1 kg of grain were .4.8 .6 .2 .l 11 27.5 placed in arms 1,2,3,4, and 5, respectively. At the end of phase 2, .4.6.8 .2 .I 11 27.5 .6.8 .1 .2.2 animals were switched to the other treatment for 10 days (phase 3). 9 22.5 .6.8 .4 .l .2 2 5.0 Sign tests (Conover 1980) were used to compare arms chosen on .4.8 .6 .l .2 1 2.5 choices 1 to 5 between identical twins for the last 5 trials of phases 2 .4 .8 .2.6 .l 1 2.5 and 3. Signs had to be reversed in phase 3 in order to pool responses .4 .6 .2 .l .8 1 2.5 from phases 2 and 3, because of the crossover type of experimental .4 .2 .8 .6 .l 1 2.5 design. A difference in the order of arm choices between identical .4.2.4 .l .6.8 .8.6 .l.2 1 twins could indicate the ability of cattle to use spatial memory to 1 E .4.2 .I .6 .8 1 2:5 distinguish among locations with various levels of reward. Total 40 100.0 Sign tests were also used to compare arms selected by a steer for ‘Incorrect choices wcrc not included in this analysis. There were no incorrect choices choices 1 to 5 for the last 5 days of phases 2 and 3. Results from in the first 3 entrances. both steers in a treatment group (ascendingdescending or descending- ascending) were pooled and analyzed together. This analysis com- 3 choices on 10% of the trials by chance. Few errors made in phases pares arm selection of a steer (not a comparison between twins) 2 and 3 involved reentering arms that contained 0.1 or 0.2 kg of during phase 2 and 3. A difference in order of choices during phase grain. Arms which had contained 0.8, 0.6, and 0.4 kg of grain 2 and 3 would indicate that steers can form (and later modify) accounted for 32,26, and 32% of incorrect choices, respectively. associations between locations and levels of reward. Twins selected different (KO.05) arms for choices 2,3, and 4 in For both statistical analyses described above, the primary the last 5 trials of phases 2 and 3. There was no difference between emphasis should be given to choices 2 and 3. Gate location may twins (JPO.05) in arms selected for choices 1 and 5. have biased the first arm chosen. Choices 4 and 5 were occasionally Arms selected for choices 2,3, and 4 of phase 2 were different incorrect which could also confound the results. (KO.05) from those chosen in phase 3 for steers in both the Spearman’s rho rank correlation coefficient (Conover 1980) was ascendingdescending (steers 10A and 15B) and descending- used to compare the sequence of arm selection. The arm numbers ascending (steers 10B and 15A) treatments. Arm selection on (Fig. 1) were used as ranks. The rank correlation coefficient was choice 5 was different (X0.05) for the descending-ascending used as an indicator of the relationship between 2 sequences of arm treatment but not for the ascendingdecending treatment (JPO.05). choices. There was no difference (130.05) in arms selected for choice 1 for either treatment. Results and Discussion Rank correlation analysis of the entire choice sequence generally The steers performed very efficiently in all 3 phases of the study agreed with the separate analysis of each choice. The correlation, (Table 1). The overall-mean number of correct choices in the first 5 however, added little to interpretation of choice sequences and was entrances was 4.69. The number of correct choices in the first 5 removed from this presentation. entrances that would be expected by chance was 3.73 (calculated Steers may not have distinguished among all quantities of grain, from equations from Beatty and Shavalia 1980). The steers but they definitely avoided arms that contained 0.1 or 0.2 kg of grain until the fourth and fifth choices. Steers may have divided Table 1. Mean performance of etetn in a S-arm parallel maze wltb equal arms into those containing higher or lower quantities of grain. (pbeee l), vukd (pbeee 2) end revereed (pbeee 3) rewards. Steers quickly learned the quantity of grain (high or low) associated with each arm. It took somewhat longer for the steers to change their choice sequence when the placement of grain was Correct choices Number in first Total incorrect reversed (phase 3). The preference for higher quantities of grain did PlUtSe of trials 5 entrancesa choices not overcome steers’s bias for choosing the arm directly in front of the gate in choice 1. 1 20 4.70 0.35 2 28 4.54 0.71 Efficient performance of the steers in&mates that they have an 3 40 4.80 0.33 accurate spatial memory (Bailey et al. 1987). Animals apparently overall 88 4.69 0.45 use working memory to remember which arms were visited (Olton The number of arms that would be entered correctly by chance alone is 3.73 (calcu- 1978). Specific stimuli and responses that are required for a given lated by methods presented by Batty and Shavaiia 1980). trial are stored in working memory and can be forgotten at the end

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 481 of a trial. The location of each arm and the fact that each arm LiteratureCited contains a certain amount of food at the beginning of a trial are Bailey, D.W. 1988. Characteristics of spatial memory and foraging behav- stored in long-term or reference memory (Honig 1978). Rats and ior in cattle. Ph.D. Dim., Colorado State Univ., Fort Collins. pigeons apparently can store in reference memory which arms of a R&y, D.W., L.R. Rittenhouae, RX Hart, and R.W. Rkbuda. 1987. maze contain food and the amount of food consumed in each arm Spatial memory of heifers in radial- and parallel-arm mazes. Proc., West. (Hulse and O’Leary 1982, Roberts and Van Veldhuizen 1985). Sec. Amer. Sot. Anim. Sci. 38:7-10. Results from this study indicate that cattle can develop reference Rutty, W.M., and D.A. Shavalia. 19gO. Spatial memory in rats: time memory that also includes information on the relative amount of course of working memory and effects of anesthetics. Behav. Neural Biol. 28~454462. food contained in each arm. Conover, WJ. 19M). Practical nonparametric statistics, 2nd Ed. John As in the study by Hulse and O’Leary (1982), animals reentered Wiky and Sons, New York, New York. arms that contained larger amounts of food more often than arms Cook, C.W. 1966. Factors affecting utilization of mountain slopes by that contained less food. We speculate that the concentration of cattle. J. Range Manage. 19:200-u)4. mistakes in arms containing more grain may be a result of steers Gillen, R.L., W.C. Krueger, and R.F. Miller. 19(14.Cattle distribution on avoiding arms with less grain until they are certain that all grain in mountain rangelandin northeasternOregon. J. Range Manage. 37549453. Honig, W.K. 197g. Studies of working memory in the pigeon. p. 21 l-248. other arms was consumed. If this is true, reference memory may be In: S.H. Hulse, H. Fowler, and W.K. Honig (l&is.). Cognitive processes as important as working memory in choosing where to forage. in animal behavior. Erlbaum, Hillsdale, New Jersey. Animals could avoid areas that contain fewer resources until other Hulse, S.H., and D.K. O’Leary. 19112.Serial pattern learning: Teaching an areas are determined to be depleted. alphabet to rats. J. Exp. Psychol. Animal. Behav. Processes 8:260-273. Both reference memory and working memory should be consi- Miller, R.F., and W.C. Krueger. 1976. Cattle use on summer foothill dered when studying animal distribution and grazing behavior. rangelands in northeastern Oregon. J. Range Manage. 29367-371. Mueggler, W.F. 1965. Cattle distribution on steep slopes. J. Range. Man- Bailey (1988) found that cattle have the ability to remember where age. 18:255-257. they have foraged for periods up to 8 hours, and this indicates an Olton, D.S. 1978. Characteristics of spatial memory. p. 341-373. In: S.H. accurate working memory. Results from the present study show Hulse, H. Fowler and W.K. Honig(Eds.). Cognitive Processes in Animal that cattle can also develop a reference memory of locations that Behavior. Erlbaum, Hillsdale, New Jersey. involve tags for different food amounts and utilize that informa- Roberts, W.A., and N. van Veldhtdzen. 1985. Spatial memory in pigeons tion 24 hours later. Further studies are needed to determine if a on the radial maze. J. Exper. Psychol. Anim. Behav. Processes 11:241-260. memory-based (cognitive) mechanism is appropriate for predict- Senft, R.L., M.B. Coughenour, D.W.Bailey, L.R. Rlttenhouse, O.E. Sala, ing cattle grazing patterns. and D.M. Swift. 1987. Large herbivore foraging and ecological hierar- chies. Bioscience 37:789799. Senft, R.L., L.R. Rittenbouq and R.G. Woodmamee. 1983. The use of regression models to predict spatial patterns of cattle behavior. J. Range Manage. 36553-557.

Notice to Authors A FORM TO BE USED WHEN SUBMITTING MANUSCRIPTS is published on the last two pages of the Journal. A copy of this form must accompany all manuscripts submitted after July 1, 1989.

482 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 Herbivore effects on seeded alfalfa at four pinyon-juniper sites in central Utah

STEVENS. ROSENSTOCKAND RICHARDSTEVENS

Effects of nbbitr @pus cawomicus, Sylvihgus spp.), mule baugh 1983). It is highly palatable and nutritious to wild and deer (Odoeoikus hemion@, 8nd livestock on seeded 8if8if8 domestic herbivores (Dietz et al. 1962, White and Wight 1984). (Iitledicago sativa) were ttuditd on 4 sites in ctntnl Ut8h. Sites Rangeland alfalfa varieties have been successfully established on were domhuted by pinyon pint (Pinas t&i+nd Ut8h juniper pinyon-juniper sites receiving >25 cm annual precipitation, and (Jun@trus ost tosptma) 8nd were doubled-cbrined ind seeded are compatible with other species commonly used in mixtures with a mixture of gr8sses, forbs, 8nd shrubs between 1959 8nd (Stevens 1983). 1962. A +w8y txciosun WMbuilt on t8cb site in 1962 which Herbivore use can be an important influence on the survival and included the foiiowing tratmtnts: (1) control (r8bbits, deer, 8nd productivity of rangeland alfalfa seedings (Ries 1982). Published iiveatock excluded), (2) nbbit access, (3) deer 8ccess, urd (4) r8bbit information on the effects of grazing on alfalfa in the pinyon- plus deer 8ccess. The fifth treatment (outtidt the txciosurt) ~8s juniper type is limited and largely descriptive. Phillips (1979) com- rccttsibit to r8bbits, deer, md livestock. Aifaif8 density 8nd pro- pared alfalfa production in plots protected from grazing by deer duction were estimated 8t l- to S-ytu inttrvab between 1963 and and livestock with plots accessible to these herbivores. Rumbaugh 1986. Aif8if8 growth form WMmeasured in 1986. !jtrnd densities and Pedersen (1979) evaluated the effects of grazing by rabbits, de&red from 0.5 to 8.5 phtnts/mz to 0.5 to 2.5 phnts/m~ during deer, and livestock on alfalfa survival. Rabbit damage to seeded the 23-year sampiing period. Reproduction by seed WMnot tvi- alfalfa has frequently been observed (Tausch 1973, Lavin and dent. Aif8tf8 production fiuctu8ted greatly (4 kg/ha to 4,104 Johnsen 1977, Phillips 1977, Johnsen and Gomm 1981). kg/ha) with precipit8tion md dtcrttttd with incrtmtd herbivore Growth forms of alfalfa grazed by livestock have been studied access. Trtrtmtnt effects vuied. Rabbits hrd 8 nt@vt tffttt on only under tame pasture conditions (Heinrichs 1954, Kehr et al. rlWf8 density 8t 2 sites, but no effect on 8if8lf8 production. Deer 1963, Daday 1968, Gdara 1985). Results of these studies indicated use h8d inconaisttnt effects on 8lf8lf8 density, but reduced 8if8if8 that alfalfa persistence under grazing is closely related to growth production at 2 sites. The addition of livestock use reduced 8if8if8 form. density at 1 site, 8nd 8if8lf8 production 8t 3 sites. Guxxhrgtreat- This project was undertaken to quantify wild and domestic mtnts h8d 8 marked tffttt on 8if8if8 growth form. Dttrtttts in herbivore effects on the density, production, and growth form of height 8nd incrt8sea in b8tui cover were associated with intrttttd seeded alfalfa in a semiarid rangeland. Data were collected herbivore access. Results of this study indierrtt th8t 8if8K8c811 be between 1963 and 1986 from exclosures and adjacent unexclosed 8n importrnt md persistent component of seeding mixtures used areas at 4 chained and seeded pinyon-juniper sites in central Utah. on semiuid pinyon-juniper surges. Methods Key Words: Mtdicago sativa, rlfdfq rtngt tttding, herbivore Study Sites effects, txciosures The study was conducted at 4 pinyon pine (Pinus eduli+Utah Pinyon-juniper woodlands occupy approximately 7.1 million ha juniper (Juniperus osteosperma) big game range modification pro- of the Great Basin of the Western United States (Tueller et al. jects implemented between 1959 and 1962 in Sanpete County, 1979). The pinyon-juniper type is important habitat for wild ungu- central Utah. They are 242 to 400 ha in size, at elevations of 1,755 to lates and rangeland for domestic livestock, but forage values have 2,143 m, with long-term average annual precipitation of 29 to 46 suffered from extensive depletion of understory vegetation. This cm. Soils are limestone-derived cobbly loams in the Fontreen depletion has been attributed to a number of factors, including Series (Soil Conservation Service 1981). Prior to tree removal, all 4 overgrazing, tire suppression, and dominance by woody species sites were dominated by mature pinyon-juniper stands with (Arnold et al. 1964, Tausch et al. 1981). depleted understory vegetation. All 4 sites were double-chained in A great deal of effort has been directed toward increasing forage October or November. Between chainings a seed mixture of native production on pinyon-juniper lands. Most pinyon-juniper modifi- and introduced grasses, forbs, and shrubs was applied by fixed- cation projects have involved removal of the tree canopy by wing aircraft (Table 1). Alfalfa was the most heavily seeded forb, at mechanical means (chaining, cabling, or dozing), fiie, or herbi- rates of 0.9 to 2.2 kg/ha. A mixture of 3 cultivars (‘Ladak’, cides. Seeding is often necessary where desirable forage speciesare ‘Nomad’, and ‘Rambler’) was used. Post-treatment vegetation on absent or too sparse to respond to treatment. all sites was dominated by seeded perennial grasses. Alfalfa (Medicago sativu) is the most commonly seeded forb on Post-treatment grazing by domestic livestock varied among the pinyon-juniper modification projects (Plummer et al. 1968, Rum- 4 sites. All were rested for 2 to 12 years to enhance establishment of seeded species and recovery of native vegetation (Table 2). Follow- Authors arc ipde ndent consultant and fortner gradyate student, Department of Fishery and Wild lip”c Biology, Colorado State Umventty, Fort Collins 80523; and ing rest, use was limited to spring (May and June) at stocking rates research biologist, Utah Divtsion of Wildlife Resources, Great Basin Experimental of 1 to 5 ha per animal unit month (AUM). However, trespass Range, Ephrakm 84627. The U.S. Department of Agriculture, Forest Service, lntermountain Research grazing occasionally occurred. Most livestock use has been by Station facilitated this study through cooperative a cattle (Table 2). Utah Division of Wildlife Resources (UDWR) a fi~!$.%%;%:~ State University and UDWR. Funds were also provided through Federal Aid in A Cway exclosure was constructed on each site in 1962 (Fig. 1). Wildlife Restoration Project W82R Studyl. A. Perry Phtmmer, Stephen B. Monsen, The 4 grazing treatments inside the exclosure were: (1) control and Donald R. Christensen were responsible for initiation of thii study and early data collection. Three anonymous reviewers provided valuable comments on the manuscript. (rabbits, deer, and livestock excluded), (2) rabbit access, (3) deer Manuscript accepted 16 March 1989.

JOURNAL OF RANGE MANAGEMENT 42(6). November 1999 483 Table 1. Seed mixturea and rata (kg/b@ applied on 4 pinyonjunlper cb8bdng study rite8 in ccnlral utab. 91.4 m _I_) ,...... ;.----B--B. Location’ it iI Species EM MF MYF SH Rabbits, Fairway wheatgrass (Agropyron cristatum) 3.5 3.5 3.5 3.5 Deer and Deer il Intermediate wheatgrass 0.9 1.8 0.9 0.9 t Livestock Access ;I (Agropyron intermedium) iI Western wheatgrass (Agropyron smithii) 0.4 - 0.4 0.4 Excluded Pubescent wheatgrass 0.4 1.8 0.4 0.4 E iI (Agropyron trichophorum) ;I 0 ...... W...... , Smooth brome (ilromus inermis) 1.8 0.9 1.8 1.8 . I Russian wildrye 1.8 0.9 1.8 1.8 I (Psathyrosrachys junceus) s I Alfalfa (h4edicago sativa) 2.2 0.9 2.2 1.8 Rabbit Yellow sweetclover 0.9 0.4 0.9 0.9 Rabbit and (Mellotus ofjcinalis) Access Deer i Small bumet (Sangusorba minor) 0.4 - 0.4 0.4 Big sagebrush (Artemisia tridentata) 0.4 0.4 0.4 0.4 Access I Rubber rabbitbrush (Chrysothamnus - 0.4 - - I nauseosus) I .-----e-w I ‘EM = E. Mafield, MF = hlanti Face, MYF : May&Id Face, SH = S. Hollow.

Table 2. Grazing blstory 1963 to 19116of 4 phyon-juniper chaining study 2.4m wire net fence 3Ocm mesh site.8in central Utah. ---- 6lcm wire net fence 30cm mesh Post seeding rest No. Years Grazed Site period (years) cattle Sheep ...... -.SSa.SSBicm chicken wire fence E. Mafield 7 13 Manti Face 12 5 ; Fig. 1. pour-way exclosure design used at Ipinyon-juniper chaining study Mayfield Face 2 18’ - sites in central Utah. S. Hollow 5 14 2 actions, means within each sampling year were compared with Qcludcs both cattle and sheep use. Tukey’s test at the 0.05 significance level. Type I error probabilities for within-subject effects (year and year X treatment interactions) access, and (4) rabbit plus deer access. The fifth treatment (outside were based on a Multivariate Analysis of Variance Wilk’s Lambda the exclosure) was accessible to rabbits, deer, and livestock. F-approximation @Tour and Miniard 1983). Data Collection and Analysis In June 1986, a random sample of 50 alfalfa plants at full-bloom At each site, a set of 5 permanent 30.5- by 0.3-m belt transects stage was selected from each grazing treatment at 3 sites (E. May- was randomly located within each of the 5 grazing treatments. field, Manti Face, and S. Hollow). Height of each plant was Each transect was divided into 10 subplots of 3. l- by 0.3-m. Sam- measured to the nearest centimeter. Basal cover of each plant was pling occurred at l- to S-year intervals between 1963 and 1986, measured to the nearest 100 cm* within a 40- by 50-cm cover frame usually in early July. The number of alfalfa plants (seedling and divided into lo- by lo-cm squares. Grazing treatment effects were mature) in each subplot was tallied, along with an ocular estimate tested with Analysis of Variance in a completely random design. of above-ground alfalfa production. Clipped alfalfa samples of Where the ANOVA indicated significant m.05) treatment known weight were used to improve accuracy of biomass estimation. effects, means were separated with Tukey’s test at the 0.05 level of Data from each site were analyzed separately due to unequal significance. sample sizes (sampling years) and nonconcurrent sampling. Yearly subplot data withineach transect were pooled. Because exclosures Results were not replicated on each site (a common problem in grazing Density studies at the time this project was initiated), the 5 transects within Alfalfa density changed considerably during the 23-year sam- each treatment were used as independent sampling units for statis- pling period (Figs. 2, 3). Reproduction by seed was not evident. tical analysis. Interpretation of these data assumed that site differ- Density at the time of initial sampling (1 to 4 years post-seeding) ences were small relative to treatment effects, due to careful loca- ranged from 0.5 to 7.5 plants/m*. There were considerable differ- tion of the exclosures. Hawkins (1986) and Guthery (1987) noted ences between treatments at 3 of the 4 sites. Density at the Manti that valuable inferences can still be drawn from experimental Face site was much higher than at the other sites, despite a lower designs lacking true replication. seeding rate (Table 1). The effects of grazing treatments on alfalfa density and produc- Downward density trends across most treatments were observed tion were tested with Repeated Measures Analysis of Variance over much of the sampling period at 3 sites (Figs. 2,3). This was (Winer 1971). Control, rabbit, deer, and rabbit plus deer treat- reflected in a significant year effect found at all sites in the ANOVA ments were analyzed in a 2 X 2 factorial design. Rabbit plus deer for rabbit and deer treatments and at E. Mayfield and Mayfield effects were tested against rabbit plus deer plus livestock effects in a Face in the ANOVA for livestock effects. However, alfalfa density l-way design. One or more years of data from each site were increased between 1963 and 1967 in most treatments at all sites omitted from the latter analysis due to missing observations. All except Manti Face. The 1967 measurements were the highest main (grazing treatment) effects were tested as single degree of recorded during the sampling period for most treatments at all freedom contrasts. If the ANOVA indicated significant m.05) sites. Density at E. Mayfield and Mafield Face declined after the treatment effects and nonsignificant @ro. 10) confounding inter- 1967 peak, and then stabilized. Alfalfa density at Manti Face has declined steadily since 1964. Density at the S. Hollow site has

484 JOURNAL OF RANGE MANAGEMENT 42(6). November 1989 - Rabbits. Dser and Livestock Excluded Rabbits. Deer end Livestock Excluded ------Rabbit Accese Rabbit Access --- Deer Access Deer Access -- Rabbit and Deer Access Rabbit and Deer Access - Rabbit, Oeer and Livestock Access Rabbit. Deer and Livestock Access 8 E. Mayfield 7 Mayfield Face 1 6 5 t

63 67 72 77 79 63 67 72 75 77 79 82 85 86 82 85 86 S. Hollow r\, Manti Face

63 67 72 77 79 82 85 86 64 66 67 69 71 72 75 77 7g 82 *5 86 Year Year Fig. 3. Density of seeded arfarfo 1963 to 1986 at Moyfzld Face and S. Hollow pinyon-juniper chainingstudy sites in central Utah. Fig. 2. Density of seeded arfarfa 1963 to 1986 at E. Mayfild and Monti Face pinyon-juniper chaining study sites in central Utah. fluctuated, but not declined appreciably. Mean alfalfa density for all sites and treatments in the final sampling (1986) was 0.5 to 2.5 plants/ m2. Effects of rabbit and deer use on alfalfa density varied among sites. Significant rabbit effects were found at E. Mayheld and Mayfield Face, where alfalfa densities under rabbit use were significantly lower than in the control in 1 of 9 sampling years and 2 of 8 sampling years, respectively. Deer effects were significant at 2 sites. Alfalfa densities were significantly lower than the control in 3 of 8 sampling years at Mayfield Face, but significantly higher in 3 of 8 sampling years at S. Hollow. Year X rabbit interactions were significant at all sites except Manti Face, while significant year X deer interactions were found at all sites. Alfalfa density under grazing by rabbits, deer, and livestock was lower than under the other treatments for much of the sampling period at E. Mafield (Fig. 2). Density in the rabbit plus deer treatment was significantly higher in 5 of 8 sampling years. Signifi- cant year X livestock interactions were also found at this site. 200

Production 0 Large year to year fluctuations in alfalfa production were 63 67 72 75 77 79 82 85 86 observed across all sites and treatments, and coincided with changes in precipitation (Figs. 4-7). The highest production Year occurred in 1985 and 1986, the wettest years during the sampling Cbbblts. Deer and Livestock Excluded period. The ANOVA for rabbit and deer treatments indicated ______Rabbit Access significant year effects at all sites, while the ANOVA for livestock ___ Deer Access effects indicated significant year effects at E. Maytield and May- Rabbit and Deer Access field Face. Rabbit. Dear and Livestock Access Mean alfalfa production (all years combined) generally decreased with increased herbivore access (Fig. 8). Mean production was Fig. 4. Precipitation (October to May, cm)andalfaljhproduction (kg/ha) 1963 to 1986 under 5 grazing treatments at E. Moyfwldpinyon-juniper chaining study site in central Utah.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 485 2000 ,...w.-i-“” 1000 Manti Face : 700 f Mayfield Face i’ 700 c I I’ 300 t .! 600 .i’ 500 250 400 200 300 150 200 100 100 50 0 n 64 67 71 75 79 85 66 69 72 77 82 86 ” 63 67 72 77 79 82 85 86 Year Year

Rabbits. Oeer and Liveetock Excluded Rabbite. Deer and Liveetock Excluded

““““““” ““““_” Rabbit Access Rabbit Aceera

______Deer Access Deer Accerr

-- -.- Rabbit and Oeer Acceee Rabbit and Deer Access hbbit. Deer and Livestock Access Rabbit, Deer end Livertock Access

Fig. 5. Recipitation (October to May, cm)anda~alfaproduction (kg/ha) Fig. 6. Precipitation (October to May, cm)andalfaljhproduction (kgJha) 1964 to 1986 under 5 grazing treatments at Manti Facepinyon-juniper 1963 to 1986 under Sgrazing treatmentsat Mayfwld Facepinyon-juniper chaining study site in central Utah. chaining study site in central Utah. highest in the control treatment at all sites except Manti Face, rabbit plus deer plus livestock treatment than in the deer and rabbit where the rabbit treatment had the highest mean production. plus deer treatments. Mean basal cover per plant was significantly Negative effects of rabbit use approached significance 0.057) different for all treatments (Table 3). only at Mayfield Face. Mean production in the deer treatment was less than in control Teble 3. Hcigbt end beeel cover per plent (meen, SE) of eeededelfelfe et and rabbit treatments at all sites (Fig. 8). Mean production under full bloom efter 23 yeue unda 5 greeing treetmentaet 3 pinyon-juniper deer use was significantly lower in 4 of 8 sample years at E. ebeiniq study eltee in centrel web. Mayfield, and 4 of 9 years at S. Hollow. Comparisons at Manti Face and Mayfield Face were precluded by significant rabbit X Height (cm) Basal Cover (cmz) deer interactions. Treatment Mean’ SE Mean’ SE Significant year X rabbit interactions were found at all sites except E. Mayfield. Year X deer interactions were significant at all Rabbits, deer, and livestock 78.91a 1.03 368.67a 13.81 sites except Manti face. excluded Rabbit access 68.77b 1.39 776.73b 55.61 Annual alfalfa production at all sites was often lowest in the Deer access 41.97c 1.10 1376.00~ 50.37 rabbit plus deer plus livestock treatment (Figs. 4-7). Mean produc- Rabbit and deer access 39.83e 1.01 1541.33d 46.45 tion (all years combined) in this treatment was also lower than Rabbit, deer, and livestock 29.9Sd 0.87 1135.33e 44.88 under the other treatments at all sites (Fig. 8). The addition of BCA%ZR livestock to rabbit plus deer use significantly reduced alfalfa pro- Weans followed by the same letter arc not significantlydiffcrent(Tukcy’r test 735 df, duction at 3 sites. Production in the rabbit plus deer treatment was pZO.05). significantly greater in 7 of 8 sampling years at E. Mayfield and 3 of 7 years at Mayfield Face and S. Hollow. Significant year X live- Discussion stock interactions were found at E. Mayfield and Mayfield Face. Deneity Growth Fona In the initial sampling, alfalfa density at all 4 sites showed no Consistent differences in growth form were observed after 23 consistent pattern among treatments, and could have reflected years under the 5 grazing treatments. Plant height at full-bloom unequal establishment of plants, site effects, or differences in her- ranged from 7 to 115 cm, and decreased with increased herbivore bivore use. The higher overall density at Manti Face probably access (Fig. 9). Mean plant heights were significantly different for resulted from better site preparation, for example, greater seedbed all treatments except deer and rabbit plus deer (Table 3). disturbance and non-frozen soil during chaining. Establishment Basal cover of individual alfalfa plants at full-bloom ranged could have also been enhanced by more favorable post-seeding from 100 to 2,900 cm*, and generally increased with increased growing conditions. Density at E. Mayfield and S. Hollow could herbivore access (Fig. 10). However, lateral spread was less in the have been reduced by greater herbivore use prior to construction of

488 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 30 1 Rabbits, deer and livestock excluded

30 L Rabbit Access I 20 - 10 - l-l

63 67 72 77 79 82 85 86 Year

- Rabbits, Qeer snd Livestock Excluded ---- Rabbit Access --- Deer Access -- Rabbit and Qeer Access - Rabbit, Rer and Livestock Access

Fig. 7. Precipitation (October to May, cm)andalfolfaproduction (kg/ha) 1963 to 1986 under 5 grazing treatments at S. Hollow pinyorrjuniper chaining study site in central Utah. 10 20 30 40 50 60 70 60 90 100 Height of Individual Plants 1100 (Class Midpoint, cm) 1000 Rabbits. Deer and Livestock Excludel Rabbit Access Fig 9. Height dtktribution of individualalfalfaphntsatfull-bloom in 1986 900 Deer Access after 23 years under 5 grazing treatments at 3 pinyon+ruper chaining lbbbit snd Dser Access study sites in central Utah (n=l5Oplants/treatment, SOfrom each site). Rabbit. Deer snd Livestock 800 n Access the exclosures. 700 Concurrent stand losses observed for all treatments at 3 of the 4 sites suggest that herbivore use can be a secondary factor in alfalfa 600 survival. Results of other studies indicate high initial establishment 1 followed by rapid decline is typical on semiarid sites, even when protected from grazing. Holechek et al. (1982) found alfalfa mortality rates of 84 to 86% over 5 years on a southern Montana mined-land reclamation seeding. Rumbaugh and Pedersen (1979) attributed similar declines on central Utah big sagebrush (Artem& sia tridentata) and pinyon-juniper sites to drought stress and competition with other seeded species. Intraspecific competition may also become important at high densities. Rumbaugh (1982) observed greater post-seeding mortality in alfalfa stands with higher initial density (19 plants/m*) than in those with lower initial E. Hayfield Hanti ttayfield S. Hollow density (10 plants/m*). This may explain the consistent declines Face Face observed at Manti Face. Alfalfa mortality in rangeland seedings has also been attributed Locat ion to damage by pocket gophers (Bleak 1969, Townsend 1982, McGinnies and Townsend 1983). Gopher activity was not mea- Fig. 1. Meanproduction (hg/ha)of seeded alfalfa (mean, SE) 1963 to 1986 sured in this study, so the impact could not be assessed. under 5 grasing treatments at 4 pinyon-juniper chaining study sites in Similarities in alfalfa density across all sites and treatments 24 to central Utah. 27 years after seeding suggest that 2.5 plants/m* may be a maxi-

JOURNAL OF RANGE MANAGEMENT 42(0), November 1989 487 been previously described. However, deer use of alfalfa is well Rabbits, deer and livestock documented (Kufeld et al. 1973). The results of this study are inconclusive, showing both positive and negative effects. This excluded could be due to site differences as well as different levels of deer use. However, without replication of exclosures, further interpretation was not possible. Comparisons between rabbit plus deer and rabbit plus deer plus livestock treatments are most useful, representing realistic management alternatives for rangeland seedings. Results suggest that the addition of intermittent early-season, short-duration livestock use may not increase alfalfa mortality above that induced by wild herbivores and environmental factors. Variability in grazing treatment effects among sites may reflect different levels of herbivore use and/ or site effects. This could not be tested, however, as herbivore use was not measured and exclo- Deer Access I sums were not replicated. In general, results of this study concur with those of Rumbaugh and Pedersen (1979), who found greater alfalfa survival under protection from grazing by rabbits, deer, and livestock. The observed year X grazing treatment interactions suggest that the effects of herbivore use on alfalfa density vary with precipita- Rabbit and Deer Access tion. Brownlee (1973) observed decreased alfalfa survival on dryland sites in Australia with increased grazing frequency. This effect was exacerbated by moisture stress. Rumbaugh and Pedersen 20 - (1979) also indicated that grazing-induced alfalfa mortality was 10 - ll,f-Hl , . . ., 1 compounded by drought. These year X grazing treatment interactions could also reflect year to year variation in herbivore use.

Rabbit, Deer and Livestock Access Production Fluctuations in alfalfa production with changes in precipitation are typical on rangeland seedings. Similar variation has been reported for dryland sites in Saskatchewan (Campbell 1961) and pinyon-juniper sites in Utah (Phillips 1979). Alfalfa growth can 2.5 5 7.5 1012.51517.52022.52527.5 show a strong positive response to increased moisture availability (Plummer et al. 1968). Rabbit effects on alfalfa production approached significance Basal Cover of Individual Plants only at Mayfield Face, where effects on stand density were also (Class Midpoint, cm * x100) significant. This further suggests that the effects of rabbits alone were minimal, except at peak population levels. Deer use had a greater negative impact than rabbit use on production. The signifi- Fig. 10. Basal cover distribution of individual alfayaplants atfull-bloom cant rabbit X deer interactions found at 2 sites indicate that the in 1986 after 23 years under 5 grazing treatments at 3 pinyon-juniper effects of these herbivores on alfalfa production may not operate chaining study sites in central Utah (wl50 plantsltreatment. 50 from each site). independently. The addition of livestock to native herbivore (rabbit and deer) mum persistence density for these sites. Rumbaugh (1982) observed use had a marked effect on alfalfa production. Significant declines stand declines from 10 to 19 plants ms to 1.5 to 1.7 plants/ mr over a were observed at 3 sites, all of which received frequent livestock 2-year period on a central Utah dryland seeding. Leckenby and use. Livestock effects were not significant at Manti Face, which Toweill(l979) found a density of 2 plants/m2 after 4 years on a received the longest post-seeding rest (12 years) and least frequent seeded south-central Oregon Western juniper (Juniperus occiden- use. These results suggest that the frequency of livestock use is an rulis) site. Kilcher and Heinrichs (1969) proposed that density of important influence on alfalfa production. seeded alfalfa in semiarid regions reaches a balance with prevailing Because livestock effects on production were not always signifi- climatic regimes, the upper limit being determined by moisture cant, grazing was not necessarily deterimental. However, signifi- stress conditions. Successful alfalfa reseeding has been reported in cant production decreases were observed in ungrazed periods fol- semiarid pasture seedings (Rumba@ 1982). The lack of repro- lowing years with significant livestock effects, indicating that duction by seed observed in this study suggests that initial stand grazing-induced declines in alfalfa production can carry over from establishment is critical on semiarid pinyon-juniper sites. How- year to year. ever, reseeding could have occurred in unsampled years between There is little published information concerning grazing effects 1963 and 1967. on production of rangeland seeded alfalfa. Results of this study Severe rabbit damage has often been reported in rangeland concur with those of Brownlee (1973), who observed decreased alfalfa seedings (Tausch 1973, Lavin and Johnsen 1977, Phillips production on Australian dryland sites with increased grazing 1977, Johnsen and Gomm 1981). In this study, rabbits had occa- frequency. Phillips (1979) also noted that alfalfa production on sional significant effects on alfalfa density at E. Mayfield and pinyon-juniper sites in Utah decreased from overuse by rabbits, Mayfield Face. Observers reported unusually high jackrabbit pop- deer, and cattle. ulations at both sites for several years immediately after seeding. The presence of year X grazing treatment interactions suggests These results suggest rabbit use may not affect seeded alfalfa that herbivore effects on production are closely tied to precipita- density when populations are below peak levels. The effects of mule deer on survival of rangeland alfalfa have not 488 JOURNAL OF RANGE MANqGEMENT 42(6), November 1989 tion. These interactions could also reflect year to year variation in Dadry, H. 1968. Heritability and genotypic and environmental correla- herbivore use. The lack of significant year X livestock interactions tions of creeping root and persistency in Medicago saliva L. Aust. J. Agr. at 2 sites indicates livestock effects may be independent of precipi- Res. 19~27-34. Chemical composition tation with infrequent grazing (Manti Face) or on sites with 146 Die& D.R., R.H. Udall, end L.E. Yuger. 1962. and digestibility of selected forage species, Cache La Poudre Range., cm annual rainfall (S. Hollow). Colorado Colo. Game and Fish Dep. Tech. Pub. 14, Denver. Growth Form Gdara, A.O. 1985. The effect of grazing and various intensities of clipping on the regrowth and survival of several alfalfa cultivars. Ph.D. Thesis, Differences among the grazing treatments indicate that wild and Colorado State Univ., Fort Collins. domestic herbivore use can have a significant effect on alfalfa Guthary, F.S. 1987. Guidelines for preparing and reviewing manuscripts growth form. The observed decrease in height and increase in basal based on field experiments with unreplicated treatments. Wildl. Sot. cover with increased herbivore access could reflect selection or Bull. 15:306. plasticity at the individual plant level. Gf the 3 cultivars seeded, 2 Hewkine, C.P. 19%. Pseudo-understanding of pseudoreplication: a cau- are root proliferating (Nomad and Rambler), and the third tionary note. Ecol. Sot. Amer. Bull. 67:1&l-185. Heinriche, D.H. 1954. Developing creeping-rooted alfalfa for pasture. Can. (Ladak) is rhizomatous. However, a single strain can display a J. Agr. Res. 34269-280. combination of growth patterns (Kehr et al. 1963). Hole&&, J.L., EJ. DePuit, J. Coenenberg, and R. Valdez. 1982. Long- Clipping studies by Carlson et al. (1964) and Gdara (1985) term plant establishment on mined lands in southeastern Montana. J. demonstrated that root proliferation was stimulated by defolia- Range Manage. 35522-525. tion. However, frequent, intense clipping decreased lateral shoot Johnsen, T.N., Jr., and F.B. Gomm. 1981. Forage plantings in six Arizona spread. A similar pattern was observed in thii study. Basal cover pinyon-juniper subtypes. J. Range Manage. 3413-16. Per- per plant increased with increased wild herbivore access, but K&r, W.R., E.C. Conud, M.A. Akunder, end F.C. Owen. 1963. formance of alfalfas under five management systems. Nebr. Agr. Exp. decreased under the added pressure of livestock use. Sta. Bull. 211, Lincoln. Prostrate growth form has frquently been associated with Rilchu, M.R., end D.H. Heinrkhe. 1969. Influence of row spacing on yield alfalfa survival under grazing (Heinrichs 1954, Kehr et al. 1963, and creeping root development of Rambler alfalfa in a semiarid region. Daday 1968). Two theories have been proposed to explain the Can. J. Plant Sci. 49307-311. higher survival of creeping alfalfa. Washko (1966) suggested that Kuleld, R.C., O.C. Wellmo, end C. Feddemr. 1973. Foods of the Rocky Mountain mule deer. USDA Forest Serv. Res. Pap. RM-Ill, Fort creepers with submerged crowns were capable of faster regrowth Collins. following defoliation and resistant to trampling damage during LeTour, S.A., end P.W. Mlniard. 1983. The misuse of repeated measures grazing. However, the adaptive value of rapid regrowth on semi- analysis in marketing research. J. Marketing Res. 20~45-57. arid sites is questionable. Berdahl et al. (1986) concluded that high Levin, F., and T.N. Jobnaen, Jr. 1977. Species adapted for planting A& regrowth potential was not conducive to long-term alfalfa persist- ona pinyon-jumper woodland. J. Range Manage. 30~410415. ence under grazing. Gdara (1985) proposed that the persistence of Leckenby, D.A., end D.E. ToweiU. 1979. Seeded plant densities in the first, second, and fourth growing seasons after treatment of six juniper vegeta- creeping plants reflected their ability to evade complete defolii- tion subtypes on a southcentral Oregon mule deer winter~range.Oregon tion. Tall tap rooted plants were more likely to have all stems DCD. Fish and Wildl. Fiil Ret%Proi. W-70-R3. grazed to ground level than creepers with greater spread and a M~Cinnlea, W J., and C.E. Town&I. i983. Yield of 3 range grasses grown higher number of stems. Results of this study appear to support the alone and in mixtures with legumes. J. Range Manage. 36399401. latter hypothesis. Increased basal cover per plant was observed in Phillips, T.A. 1977. An analysis of pinyon-juniper chaining projects in the the rabbit, deer, and rabbit plus dear treatments, which were not Intermountain Region. USDA Forest Serv. Range Improvement Notes (Sept.): l-20. subjected to trampling by livestock. Phillips, T.A. 1979. North cedars pinyon-juniper studies. USDA Forest Management Implications Serv. Range Improvement Notes (Nov.):l-13. Plummer, A.P., D.R. Chrieteneen, end S.B. Monsen. 1968. Restoring big Results of this study show that with proper management, alfalfa game range in Utah. Utah Div. Fish and Game, Salt Lake City. can be an important and persistent component of seeding mixtures Rles, R.E. 1982. Environmental factors and alfalfa persistence in dryland used on semiarid pinyon-juniper ranges. The lack of reproduction pastures and rangeland, p. 4-6. In: Alfalfa for dryland grazing. USDA- by seed indicates that initial stand establishment is of critical ARS Agr. Info. Bull, 4444, Washington, DC. Rumba& M.D. 1982. Reseeding by eight alfalfa populations in a semi- importance. Alfalfa production is closely tied to precipitation, and arid pasture. J. Range Manage. 35384-86. generally decreases with increased grazing pressure. Long-term Rumba@, M.D. 1983. Legumes-their use in wildland plantings, p. management of alfalfa seedings should include close monitoring of 115-122. In: S.B. Monsen and N. Shaw (compilers), Managing inter- stand condition and livestock use. Use of cultivars with a prostrate mountain rangelands-improvement of range and wildlife habitats. growth form may enhance stand longevity. USDA ForestServ. GTR iNT-157, Ogden. - Rumbauch. M.D.. end M.W. Pedemen. 1979. Survival of alfalfa in five Literature Cited semia& range seedings. J. Range. Manage. 3248-51. Soil Coneervetion Service. 1981. Soil survey of Sanpete Valley Area, Utah. Arnold, J.F., D.A. Jemeeon, end E.H. Reid. 1964. The pinyon-juniper type USDA Soil Conserv. Serv., Washington, D.C. of Arizona: effects of grazing, fire, and tree control. USDA Prod. Res. Stevens, R. 1983. Species adapted for seeding mountain brush, big, black, Rep. 8. and low sagbrush, and pinyon-juniper communities, p. 78-82. In: S.B. Berdabl, J.D., A.C. Wilton, R.J. Lorenz, and A.B. Frenk. 19M. Alfalfa Monsen and N. Shaw (compilers), Managing intermountain rangelands- survival and vigor in rangeland grazed by sheep. J. Range Manage. improvement of range and wildlife habitats. USDA Forest Serv. GTR 3959-62. INT-157, Ogden. Bleak, A.T. 1969. Growth and yield of legumes in mixtures with grasses on Tauech, RJ. 1973. Plant succession and mule deer utilixation on pinyon- a mountain range. J. Range Manage. 21:259-261. jumper chaiings in Nevada. MS Thesis. Univ. Nevada, Reno. Brow&e, H. 1973. Effects of four grazing management systems on the Tausch, RJ., N.E. Weet, end A.A. Nabi. 1981. Tree age and dominance production and persistence of dryland luceme in central western New patterns in Great Basin pinyon-juniper woodlands. J. Range Manage. South Wales. Aust. J. Exp. Agr. and Anim. Hush. 13~259-262. 34259-264. Campbell, J.B. 1961. Continuous versus repeated-seasonal grazing of Townsend, C.E. 1982. Diseases, insects, and other pests of rangeland grass-alfalfa mixtures at Swift Current, Saskatchewan. J. Range Man- alfalfa, p. 13-14. In: Alfalfa for dryland grazing. USDA-ARS Agr. Info. age. 1472-77. Bull. 444. Culeon, GE., V.C. Sprague, end J.B. Washko. 1964. Effects of tcmpera- Tueller, P.T., C.D. Beerron,RJ. Tauecb, N.E. Weet, end K.H. Ru. 1979. ture, daylength, and defoliation on the creeping-rooted habitat of alfalfa. Pinyon-juniper woodlands of Great Basin: distribution, flora, vegetal Crop Sci. 14:2&I-286. cover. USDA Forest Serv. Res. Pap. INT-229, Ogden.

JOURNAL OF RANGE MANAGEMENT 42(0), November 1989 489 Washko, J.B. 1966. Environmental influences on the creeping-rooted char- Wirier, B.J. 1971. Statistical principles in experimental design. McGraw acters. Proc. Alfalfa Improv. Conf. 20:35-38. Hill, New York. White, L.M., aod J.R. Wigbt. 1984. Forage yield and quality of dryland grasses and legumes. J. Range Manage. 37~233-236. Emergence and root growth of three pregerminated cool- season grasses under salt and water stress

D.M. MUELLER AND R.A. BOWMAN

Red~gult~~~~oilsunda~miuidcoaditioarnitbout Cultural practices which encourage rapid growth at condition! irrigation ia difficult. High salt concentratiom both delay and suboptimal for germination should increase seedling emergence decrease germination and emergence, which iocreuea the time 8 and reduce moisture requirements during emergence. Hauser soil must rem& moist for germhtion 8nd emergence to t8ke (1983) showed that grass seed that was germinated before sowing place. Delayed germination can also affect a plant’s capability to (pregerminated seed) emerged sooner and produced more plants withtand summer drooght became of limited root development. than conventionally sown seed (untreated seed). Many plants are Cult~~rrrlpracticea that encourage rapid growth lt conditiona sut~ more sensitive to salinity during germination than in later stages of optimal for germination rhould increase eeedling emergence and growth (U.S. Salinity Laboratory Staff 1954, Ross and Hegarty reduce moisture requirement8 for emergence. We determined from 1979). Two greenhouse studies were initiated under different levels greenhouse studies tbe effecta of different leveb of eoil salinity and of soil salinity (measured as electrical conductivity, EC) and soil soil rater on emergence and on root and Soot growth of 3 prc water to compare emergence and root and shoot growth of plants germhtedc- granes: ‘Nordan’crested wheatgrass (Ape- grown from pregerminated and untreated cool-season grass seed: pyron dcscrtorrmr) (L.) Gaertn.), ‘FUntlock’ western wbeatgrana ‘Nordan’crested wheatgrass (Agropyron desetronrm) (L.) Gaertn.), (Pascopyruns m&W (Rydb.) A. Love), and ‘Vinall’ Russian nil- ‘Flintlock* western wheatgrass (P~~copyrum smithii (Rydb.) A. drye (Psotlryr&zc&ys junccrr (Fischer) Nevski). Seed pregermi- Love), and ‘Vinall’ Russian wildrye (Psufhyrosruchys juncea nated prior to sowing resulted in more rapid emergence than (Fischer) Nevski). untreated seed for all species at alI levels of soil salinity and soil Materiala and Methods water. Salinity and water mtrewdelayed and/or reduced emergence more in the untreated than pregermlnated seed of Russian wildrye The studies used seed which was pregerminated in a growth and western wheatgrass. Regerminating seed before planting also chamber and had radicles less than or equal to 3 mm in length. resulted in greater root biomass for all speck and greater root One-hundred seed per standard germination dish (32.5 X 33.8 X lengths for the 2 wheatgrass species than did untreated seed. 7&m) were germinated on Kimpaci and blotter paper soaked with 100 ml of distilled water. Crested wheatgrass and Russian Key Wordsz germination, saline soil, seeding, seed treatment wildrye were germinated in an 8-hr light period at 30“ C alternating Saltgrass meadows occupy an estimated 500,000 ha in Colorado with a 16hr dark period at 20” C. Western wheatgrass was germi- and Wyoming (Osbom 1974). These meadows are frequently more nated in the dark at temperatures that alternated between l5O C for moist than upland sites and have good production potential if 16 hr and 30° C for 8 hr [Association of Official Seed Analysts low-value saltgrass is replaced by more palatable species (Ludwig (AOSA) 19811. and McGinnies 1978). However, reclaiming salt-affected soils study I under semiarid conditions without irrigation is difficult. Most The effects of 4 levels of salinity (0,4,8, and I6 dS/m) on the grass seed are small and must be seeded at shallow depths. Seed- emergence of untreated, untreated + gel, and pregerminated seed lings are subjected to stress from extreme temperatures, strong were evaluated in Study I. The study design consisted of a random- winds, and excessive evaporation (Cook et al. 1974). Salt accumu- ized block with 4 replications. Approximately 2 ml of SGP* gel was lation near the surface of saline soils can create an even harsher applied to each pregerminated and untreated + gel seed before environment for germinating seed. High soluble salt concentra- sowing. The gel prevents damage to the radicle during sowing of tions decrease both germination rate and final germination percen- pregerminated seed (Seamy and Roth 1981). Fifty seed per treat- tage (Uhvits 1946, Mill&ton et al. 1951, Dewey 1962). The longer ment were planted I .25-cm deep in 60 X 40 X 7.5cm plastic lined it takes for a species to germinate, the longer the soil must remain flats in an Ascalon sandy loam soil, previously screened through moist, thereby increasing the amount of water required for germi- 1.25-cm openings. The Ascalon sandy loam soil is a fine-loamy, nation and emergence. Delayed germination also delays root mixed, mesic Aridic Argiustoll. growth, which results in increased stress from summer drought. Mention ofanytradcname bforinformation onlyrnd does not imply endorsement. Authon arc range a&&t, USDA-ARS High Plains Grasslands Research Station, %GP Absorbent Polymer h a General Milla Chemicala’ trade name for P “super” 8408 Iiildrcth Road, Cheyenne, Wyomin& 8u)o9; roil scicntiat, USDA-ARS Crops P”a” r compoacd of a natural polymer and e synthetic polymer nude of acrylamids Research Laboratory, Colorado State Umversity, Fort Collins 80523. rm sodium (or potaenium) acrylate. SGP ia manufactured under a USDA patent The authora would like to acknowled~ Dr. Gary Richardson, USDA-ARS statirrti- license and i available from the Hcnkle Corp., 4620 W. 77th Street, Minneapolis, cian, Ft. Collins, Colorado. for his amwarxc in stati&al analysis. Minn. 55435. Mamucript accepted 28 February 1989.

490 JOURNAL OF RANGE MANAGEMENT 42(6), November 1089 T&b 1. hit&l rwrpacr(we~- 17 dayn ah phntblg (%) of crested wheatgrassYwestera wheatWas& and Rti we Pre- -amI ttDtHaWsHddIrdlWIIlttQk.W!hQC).

Seed EC (dS/ m) Species g=rW=e treatment 0 4 8 16 Meam crutcd wlleatgra21 Initial MP= 2.0 2.5 2.3 4.0 2.7x Untreated 5.0 4.8 6.3 8.8 6.5 Means 3.k 3.6a 4.38 6.4b #WIv Seventeen presmn 89.5 83.5 62.5 g 71.4y dV Untreated 77.0 75.3 36.8 55.8x after Means 83.3b 79.41, 49.4a 42:2a Pltuitino dya western wheatgrass Initial preeerm 3.1 2.7 ;7 5.3 3.7x Untreated a.3 7.2 9.8 11.5 9.2y Mains 5.7ab 4.90 6.8b 8.4c

P=germ 89.5 87.5 7li.; 70.0 81.4~ Untreated 85.4 79.7 55.7 47.0 66.9x Mall8 87.41, 83.6b 67.10 58.5a

dp Russian wildrye P=gcnn 2.3a 2.3a ha 3.3a 2.7 Untreated 5.88 6.01~ 7.8b 1osc 7.5 MtallS 4.1 4.1 5.4 6.9

II7 N seventeen 79.Ob 68.5b 69.Ob 49.On 66.48 days 80.3 79.oc 59.2b 21.7a 60.0 after 79.6 73.7 64.1 35.4

‘MUM with tllc umc lcttcr UC not Iipifialntly diicrwt (pIo.Os). The kttcn e tluougll c arc UIcd for compuhna within rowl and the let* x &.g&”thro comnuisotu within columns. Where there is a unnifian interaction, rignificant_ dNercnca between main dkcta are not thown. Wkrc there u no Iigl&ldiicrwccI hctwccn just maincffcct~ xic dabown.

Salinity treatments were prepared using a solution of NaCl and water contents for the moderately dry and dry treatments were CaCls that gave the desired EC from a saturated extract (U.S. obtained by bringing the cylinders to CC and then allowing them to Salinity Laboratory Staff 1954). The saline solution was added to dry 7 and 14 days, respectively, before planting. After planting no the soil before planting. Calcium chloride was used to keep the additional water was added to these treatments. Soil water was sodium adsorption ratio (SAR) below 3 (U.S. Salinity Laboratory determined gravimetricalIy on planting day (12 July 1986) and 3,7, Staff 1954). Electrical conductivity of the saturated extract was 11, and 18 days after at depths of O-2.5, 2.5-5.0, 5.0-7.5, and determined at the end of the study on soil samples randomly taken 7.5-18.75 cm. Soil water contents of 15.05,10.13,7.87, and 7.06% from the surface (O-2.5 cm) of each treatment. Treatment EC’s at were determined for matric potentials of -0.03, -0.1, -0.3 and -1.5 the end of the study period were 2.3, 3.8, 8.7, and 17.4 dS/m. MPa, respectively (Klute 1986). Soil water content at the surface Osmotic potentials calculated from treatment EC’s obtained at the (O-2.5 cm) during planting was 11% for the moderately dry and 9% end of the study were approximately -0.27, -0.28, -0.63, and -1.27 for the dry treatment, which resulted in matric potentials of MPa, respectively, at container capacity (CC) (U.S. Salinity approximately -0.09 and -0.17 MPa, respectively. After 17 days, Laboratory Staff 1954, Rawlins and Campbell 1986). Container soil water content at the surface for the moderately dry and dry capacity (Cassel and Nielson 1986) was maintained gravimetrically treatments was less than 1% (Fig. 1). at 19% throughout the 17day study period (13 June 1985 - 9 July Each cylinder contained 16 seeds from a single seed treatment. 1985). Plastic linings prevented salts from leaching out the bottom Both pregerminated and untreated seed were planted 1.25 cm deep of the Ilats. Greenhouse temperatures averaged 18” C at night and in holes that had been punched in the soil. A syringe was used to 33O C during the day. Emergence measurements were taken daily inject gel (2 ml/seed) into each hole containing a pregerminated the fast week and every other day the following 10 days. seed. The seeds were then covered with soil. Artificial lighting (569 uE/ s/ mr) was set to come on at 0600 and study II go off at 2000 hr each day. Temperatures ranged from 38O C in the The effects of 3 soil water treatments (wet, moderately dry, and day to 190 C at night. dry), and 2 soil salinity levels (0 and 8 dS/ m) on emergence and Emergence was measured daily the first week and every other root and shoot weights and lengths of untreated and pre- day the following 17 days. The cylinders were split open at the end germinated seed were evaluated in Study II. Polyvinyl chloride of the study period (24 days), the soil was washed away, and the cylinders measuring 20 X 60 cm were sealed at the bottom with root and shoot weights and leqths were measured. redwood disks and silicon. Each cylinder contained the same For comparison with emergence from pregerminated seed, screened Ascalon sandy loam soil used in Study I. The study emergence from untreated seed was expressed as percentage of consisted of a randomized complete block with 3 replications. pure-live-seed (PLS). PLS determination was by AOSA (1981) The wet treatment was maintained near CC by weighing water into each cylinder daily throughout the 24day study period. Soil methods. The term ‘initial emergence’is used throughout the discussion as

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 491 multiple range test was used for mean separation. There were no significant differences between untreated and untreated + gel treatments in Study I or between salinity treatments in Study II; therefore, these treatments were combined in the final statistical analysis after their variances were found to be homogeneous. Unless otherwise stated, all tests of sign%ance were at P90.05.

Study I High salinity (8 and 16 dS/m) delayed initial emergence for untreated Russian wildrye seed. The 16 dS/m salinity level also delayed initial emergence for both pregerminated and untreated seed of crested and western wheatgrass (Table 1). Emergence for both seed treatments and all 3 species continued to be delayed by salinity after initial emergence had taken place (Fig. 2). MY3 KTER P!JNllNG Total emergence for pregerminated crested and western wheat- + EPlN-c-!uan 0 -25-5.6cm . Dwln-5.6-7.5m X -7.546&m grass seed was greater than total emergence for the respective untreated seed (Table 1). A significant seed treatment by salinity Fig. 1. Relationship betweenpercent soil water at 4 depths andnumber of interaction (K=O. 10) was exhibited for western wheatgrass. Total days without supplemental wuter in both the moderately dry and dry treatments. emergence for untreated but not for pregerminated western wheatgrass seed was lowered by salinity levels of 8 and 16 dS/ m (Table the day emergence was first observed within a treatment and a 1). A seed treatment by salinity interaction also occurred for Rus- species. The use of initial emergence makes it possible to determine sian wildrye. There was a more rapid and a more pronounced whether treatment effects on rate of emergence are due to a change decrease in total emergence with increasing salinity for the in emergence time. untreated than for the pregerminated seed (Table 1).

Study I and II Study II Standard analysis of variance techniques were used in both Salinity slowed and reduced emergence in both studies. How- studies to compare initial and total emergence. Duncan’s new ever, salinity affects were not significant in Study II because of

Table 2. Initial emergence (days) and emergence 24 day8 after photing (46) of crested wbuttgur, wemtern wbclltpur, and Rusdan wildrye prc- pmhuted and untreated wed at 3 aoil w&r levels.

Water levels SCCd Speck3 Emergence treatment FC ModDry Dry Means

Crested wheatgrass Initial Pregerm 2.5 2.8 --J 2.8 2.7x Untreated 4.3 4.0 5.5 4.6y Means 3.4s 3.4a 4.la

Twentyfour p=YF~ 59.4 34.4 -42.2 45.3x days Untreated 57.5 48.5 47.0 51.0x after Means 58.4b 41.4a 44.6a planting days Western wheatgrass Initial Pregerm 2.8a 2Sa 3.Oa 2.8 Untreated ::o” 7.k4.8 17.0l.Ob 8.4 Means

Twentyfour 79.7 56.3 ‘“46.9 61.0~ days UnGcated 74.3 17.8 33.9x atIer Means 77.Ob 37.Oa 2% planting days Russian wildrye Initial Pl-egel-m 2Sa 3.3a 2.9 untreated 6Sa ll.Ob 8.0 MutrB 4.5 7.1

Twentyfour PmgtIlll 42.2b 48.4b -43.8a 44.8 days Untreated 49.3c 28.8b 1?.8a 32.0 after MaIl!J 45.8 38.6 30.8 planting lMeanswith the same letter (UC not rigniticantly diifennt (EO.05). The ktters a through c ar6 1164 for compari6ohs within TOW6 aad the kt’” x throu comparisons within columns. Wh6r6 ther6 is 6 s@ticant interaction, significant dierace between main 6ff6ct.6are not shown. When thcr6 IS no s igl&X2% sigmfkamt diiffcrmca b6tw6en juIlt m6in effectr are shown.

492 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 100

40 _____ EC-8 fR~036) ‘# 2c=16 (rrLo.sr: ‘/ ‘. / / 4. ,;----- . ‘* 2D / *.* /$ .-_::IT4 0 100 r WESTERN WHEATGRASS WEslERN wHENGRAss lReATm - -mum - -m ____ -lEG --- ______kL ______

5 / / ii? Idlo / 2c-6 (R~O.61) izi 4 , ____ CC-16 (l?2-0.66 l- iPO ,‘I ,,/:,y s

LO //,I,” / /,‘Y” / 4_------uooDlw(R%.1 0

100

RUSSIAN WILDWE RUSSIAN WLDRYE -lREAlm - -TREATED - URlnaTED ---- UNTREATED---

20 ---_-___ EC-6 (l&6.62)

UOO ORf (R-.9

20 ,--..-- - - EC-16 (Ra _____ DRY (RL0.21)

0 0 4 6 12 16 20 24

DAYS AFTER EMERGENCE

Fig. 2. Relationship between time (days)andpercent emergence of crested wheatgrass, western wheatgrass, and Russian wildryefiompregerminored and untreated seed at 4 salinity (EC) levels and 3 soil water fi) levels. Regression curves are discontinued at total emergence.

JOURNAL OF RANGE MANAGEMENT 42(6). November lQ89 493 Table 3. Root ad shoot weight (mg) usd leqth (cm) ofcrated whutgma, waternwheatpa, and Buakn wildwe gmwofrom plstcmrhuted and tmtreedaeed8t3d~kvek.

Plant Seed water kvelr Species Measurement treatment FC ModDry Dry Means w Crated wheatgrass Root weight preserm 25.7 20.3 26.8y Untreated 8.7 11.8 13.9x Means 17.2s 16.la

Shoot weight PmgeMl 77.0 54.3 -29.5 77.Oy Untreated 49.0 30.0 35.8 38.3x Means 63.Ob 42.2a 32.7a cm Root length Prrgci=llI 34.0 40.7 36.8 37.2~ untreated 28.2 %a 30.8 27.3x MC.anS 3l.la 33&i :m Shoot kngth PregerllI 23.5 17.7 17.5 19.6x untreated 21.8 13.2 17.3 17.4x Means 22.7b 15Sa 17.4a

Western wheatgrass Root Weight Pregerm 13.3 19.7 -9.7 14.2~ Untreated 13.2 7.8 6.5 9.2x Means 13.3a 13.8a 8.18

Shoot weight 51.0 45.0 44.0x 41.5 32.0 E:! 32.7x 46.3a 38Sa 30.4a

Root length mm 28.2 29.6 31.0 29.6~ untreatfid 21.3 22.2 23.0 22.2x Means 24.8a 25.9a 27.Oa cm Sheet length 19.3 17.0 15.7 17.3x 17.0 13.8 14.3 15.1x 18.2a 15A 15.0s mg Russian wildrye Root weight PregerIll 26.5 23.5 6.8 18.9y Untreated 12.3 5.2 9.2 8.9x Means 19.4b 14.4b 8.Otl

Shoot weight Pregeml 37.2b 46.7b -13.2a 32.3 untreated 2SSb 14.8a 16.2a 18.8 Means 31.4 30.8 14.7 srn Root length pwv~ 25.0 33.5 20.8 26.4x Untreated 23.8 18.5 26.3 22.9x Means 24.4a 26.Oa 23.6a

Sheet length

lMans with the same lcttcrarc not si ’ lcantly diierent (PSO.05). The letters a through c arc used for comparisof~~ within rows and the letten x comparisons within columns. Where rf.m ts a sgnificant interaction. @ifkant differences between mam ctTcctaarc not shown. Where there u1no I signdiiant diierences between just main c&cts an shown. variability in evaporation between cylinders, so those treatments ment regression lines for both species show that water stress were combined in Table 2. delayed emergence from untreated seed more than pregerminated Western wheatgrass and Russian wildrye failed to emerge from seed after initial emergence (Fig. 2). Water stress slowed emergence untreated seed in several of the moderately dry and/or dry treat- from untreated more than pregerminated Russian wildrye and ment cylinders, which increased the variability between cylinders western wheatgrass seed by delaying initial emergence from the within these treatments. The increased variability obscured signifi- untreated seed (Table 2). Decreasing soil water content had no cant differences between western wheatgrass seed treatment corre- effect on crested wheatgrass initial emergence (Table 2) or rate of lation coefficients in the moderately dry and dry treatments, and emergence after initial emergence (Fig. 2). between Russian wildrye seed treatment correlation coefficients in Total emergence of western wheatgrass and Russian wildrye was the dry treatment. However, a visual observation of the dry treat- lower for untreated than pregerminated seed for both the moder-

494 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 ately dry and dry soil water treatments (Table 2). Crested wheat- conditions. The greater root biomass in all 3 species and greater grass grown from pregerminated seed gave the same total emer- root length in crested and western wheatgrass should aid in e&al+ gence as crested wheatgrass grown from untreated seed for all soil lishment and survival of plants from pregerminated seed over those water treatments (Table 2). from untreated seed during critical early stages of seedling devel- Root weights of crested wheatgrass and Russian wildrye plants opment. However, to insure successful field planting of pre- grown from pregerminated seed were 93,54, and 112% greater, germinated seed, cultural practices need to be developed that pre- respectively, than root weights of plants grown from untreated vent and/or minimiie desiccation of the radicle. The ability of seed (Table 3). Pregerminating crested and western wheatgrass seedlings from pregerminated seed to rapidly establish roots in seed before planting resulted in 36% and 33% greater root lengths deeper moist soil under high osmotic stress reduces the number of than those obtained from untreated seed. Russian wildrye roots days the soil surface must remain wet for seedling establishment. were much more fragile than the roots of the other species. Loss Therefore, practices could include planting in wet soil and/or during separation of roots from the soil is believed to account for planting during predicted periods of extended rainfall (3 to 4days). the lack of differences in root lengths between Russianwildrye seed The use of irrigation may also become more economical with treatments. pregerminated seed because less water would be needed for initial Shoot weights of Russian wildrye were lower for plants grown establishment. from untreated rather than pregerminated seed at the moderately dry soil water treatment (Table 3); however no differences between JAter8ture cited seed treatments were found at the dry soil water treatment. Anochtion of OflIe& Seed Analyh. 19gl. Rules for testing seeds. J. Seed Although not statistically significant, root weights of Russian wild- Tech. 6(2):1-12X rye and root and shoot weights of western wheatgrass grown from Cook, C.W.,R.M. Hyde, md P.L. Sk. 1974. Rcvegctation guidelines for untreated seed were lower at the moderately dry and dry soil water surface mined areas. Colotado State Univ., Range Sci. Dep. Ski. Series treatments than root weights of the same species grown from 16. pre-germinated seed (Table 3). Cunel D.K., and DR. Nielson. 19)(. Field capacity and available water capacity. In: A. Klutc (cd.), Methods of Soil Analysis. Agron. Amer. Seed treatments did not affect shoot length in any of the species. Sot. Agron. Mad&son,Wii. 9901-924. Soil water treatments, however, did affect shoot lengths. Shoot Dewey, DR. 1962. Germination of crested wheatgrass in salinized soil. lengths of crested wheatgrass and Russian wildrye were reduced by Agron. J. 541353-355. moderately dry and dry treatments (Table 3). Western wheatgrass Haoser, V.L. 19g3. Grass establishment by bandoleers, and germinated shoot length was also reduced by lack of soil water (pI=O. 10). seeds. Trans. Amer. Sot. Agr. Engin. 26:74-78.80. Khtte, A. 19g6. Water retention: Laboratory methods. In: A. Klutc (cd.), Methods of Soil Analysis. Agron. Amer. Sot. Agron., Madison, Wii. Discussion and Conclusions 9635-660. Russian wildrye and western wheatgrass seed take longer to Ludwig, JR. 1976. Rcvcgctation of a saltgrass meadow in northcastcrn germinate than crested wheatgrass seed. Untreated seed began to Colorado. Ph.D. Thesis,Range Sci. Dep., Colorado State Univ., Fort germinate in about 4 days for Russian wildrye and 5 days for Collins. Lndwig, J.R., and W.J. McCinnlas. 197g. Revegctation trials on a saltgrass western wheatgrass. Surface soil water in the dry treatment meadow. J. Range Manage. 31:308-311. dropped from 9% to approximately 5.4% in 4 days and from 9% to McChmla, W.J. 1960. Effects of moisture stress and temperature on 4.2% in 5 days (Fig. 1). Matric potentials went from approximately germination of six range grasses. Agron. J. 52:159-162. -0.17 MPa to less than -1.5 MPa before seed began to germinate. Mflffngton, AJ., G.H. Bunill, 1. lB. Mardr. 1951.Salt tolerance, gcrmi- Thus, a slow germination rate allowed the soil to dry at planting nation and growth tests under controlled salinity conditions. J. Agr. depth prior to and during the onset of germination, which progres- West. Aust. 28:19&210. Gabora, L.W. 1974. Soil-plant rcfationships in a saltgrass meadow. MS. sively delayed and/ or prevented emergence (Fig. 2) Growth lead- Thesis, Range Sci. Dep. Colorado State Univ., Fort Collins, Colo. ing to emergence from pregerminated seed begins almost imme- Ross, HA., and T.W. Heguty. 1979. Sensitivity of seed germination and diately after planting. Therefore, plants grown from pregerminated seedling radicle growth to moisture stress in some vegetable crop species. seed were able to take advantage of the more favorabk surface soil Ann. Rot. 43241-243. water conditions that existed at planting (9% for dry treatment) Rawfins, S.L., and G.S. Campbell. 19g6. Water potential: Thermocouple and send roots down into moist soil (Fig. 1 and 2). psychromctry. In: A. Klutc (cd.), Methods of Soil Analysis. Amer. Sot. Agron., Madison, Wis. Ludwig (1976) has shown Russian wildrye to be more sensitive Ubvfts, R. 1946. Effect of osmotic pressure on water absorption and to NaCl-induced osmotic stress during germination than other germination of alfalfa seeds. Amer. J. Rot. 33278-285. grass species, including crested wheatgrass. McGmnies (1960) Saarcy S.W., and L.O. Rotlt. 19gl. Precision metering and planting of attributed Russian wildrye’s poor field establishment to water pmgcrminated Beads for seedling production. ASAE Publication No. stress during germination. In both Studies I and II, Russian wild- IO-gl. ASAE, St. Joseph, Mich.49085. rye emerged more rapidly from pregerminated than untreated U.S. SaRnKy Laborstory St&. 1954. Determination of the properties of saline and alkali soils. p. 7-33. Zn: L.A. Richards (cd.) Diioscs and seed. Emergence of Russian wildrye and western wheatgrass from improvement of saline and alkali soils. Agr. Handb. L7SDA;U.S. Gov. pregerminated seed was less affected by NaCl, water stress, and Print. off., Washington, D.C. 1954. Plant rcsnonse and crol) selection drying after planting than emergence from untreated seed. There- for saline and alkali ioils; p. 55-68.In: L.A. Richards (cd.) Diagrtoscsand fore, planting pregerminated seed is a promising technique for improvcmcnt of saline and alkali soils. Agr. Handb. USDA, U.S. Gov. establishing Russian wildrye and western wheatgrass under saline Print. off., Washington, D.C.

JOURNAL OF RANGE MANAGEMENT 42(0), November 1989 Morphological and physiological variation among ecotypes of sweetvetch (Hedysarum boreale Nutt.)

DOUGLAS A. JOHNSON, TIMOTHY M.J. FORD, MELVIN D. RUMBAUGH, AND BLAND Z. RICHARDSON

A)wtriCt TbL rtudy considered seedling establishment chprrrcteristics, palatable early spring forage for big-game and domestic livestock, nitrogen fixation crprbility, nutritive v8lue, ind clustering reia- and no toxicity problems have been reported for sweetvetch tionsbips rmong 11 put8tivc ecotypea of sweetvetcb (Hedyawum (Plummer et al. 1968). Improper grazing has apparently eliiinated bore& Nutt.). A tot81 of 44 morpbohgic8i urd physiological sweetvetch from much of its original distribution (Plummer et al. vui8blm were evaluated in greenhouse rad field experiments. 1968). &cause sweetvetch is widely distributed throughout the Sweetvetcb root systems b8d hrge nodules tb8t were c8pobie of foothill and lower mountain elevations of the Intermountain West fixing nitrogen, 8 potentirlly useful rttribute in the recknation of and apparently exhibits considerable ecotypic variation for many nitrogen-limited environments such 8s mine spoiis in tbe western characteristics, it may be useful in the revegetation of a wide United Strtea. Sweetvetcb provided for8ge during tbe spring md diversity of sites (Redente and Reeves 198 1). The use of sweetvetch summer, but little forge we@8v8ihble during the f8ll and winter. for revegetation has been limited, primarily because seed com- An ecotype collected near Orem, Utah, exhibited superior seedling monly is harvested from wildland stands and, consequently, seed est8blisbment cburcteristics under maic conditions while 80 eco- production is erratic and seed is expensive. As a legume, sweet- type from nerr Ducbesne, Uhb, est8blisbed well under xeric con- vetch may also improve the nitrogen status of rangeland soils ditions. An ecotype from Hobble Creek, Utrb, showed superior through biological nitrogen fixation. rhizome development, a potenthlly useful cb8racteristic for st8bil- The primary objective of this study was to evaluate 11 diverse ixing bigbly erodible areas. Altbougb cluster uulysis procedures ecotypes of sweetvetch for their seedling establishment characteris- indicated differences among the ecotypes, this clustering w8s not tics, nitrogen fixation capability, and nutritive value. This study 8iwrys cleuiy reMed to cb8racteristicaof the collection site. Suffi- also evaluated relationships among the 11 ecotypes by cluster cient genetic diversity was present 8mong ecotypes to unsure8d8p analysis. Wion to 8 wide IMY of sites md to f8ciiitpte improvement tbrougb breeding md selection. Materials and Methods Key Words: Ut8b sweetvetcb, nortbem sweetvetcb, legume, for- !3eedsourceJl age, nitrogen fiution, seedling eshblisbment, seed production Seeds of 11 sweetvetch populations (putati!~ ecotypes) were obtained from a broad range of collection sites (Table 1). This Introduced forbs are becoming more widely used in rangeland seedings in the Intermountain West (Wasser 1982, Shaw and Table1. F.cotypeMcntifiutk~n nwnber, coUecUon lo&ion, and collection Monsen 1983). ‘Appar’lewis flax (Linum Zewesii L.), ‘Delar’small lite cllu8cterMJcafor 11 ecotypeaof mel?tvetcb. bumet (Sanguisorba minor Stop.), yarrow (Achilles milkfolium L.), several penstemons (Penstemon spp.), globemallows (Sphaer- alcea spp.), cicer millcvetch (AsrragaZ~ cicer L.), alfalfa (Medicago Site characteristica sariva L. and M. falcara L.), biennial sweet clovers (Melilotus kotypc Annual officinalis Lam. and M. alba Medikus), and sainfoin (Onobrychis identiti- ptipita- tion vicizyolia Stop.) are some of the more commonly used forbs in the cation- Collection Elevation Soil number location texture (cm) semiarid Intermountain West. Herbaceous legumes native to the (ml Intermountain West have not been widely used in seeding arid and Benmore, UT 1,739 Clay loam 33 semiarid rangelands because of questionable forage value, limited : Arco,ID 1,525 Silty loam 28 3 Duchesne, UT 1,525 Loam 33 seed availability, and livestock toxicity problems (Rumba@ 4 Ballard Mine, ID 2,074 Miiedump 48 1983). Orem. UT 1,586 Cobbly loam 38 Sweetvetch (Hedysarum boreale Nutt.) is a perennial legume : Salt Lake City, UT 1,586 Upland stony 46 species that is native to the Intermountain West and has received 7 Hobble Creek, UT 1,556 Stony loam 42 some attention as a possible species for use in revegetating Inter- Moab, UT 1,373 Sandy 19 mountain rangelands. Sweetvetcb was first described in 1818 ! Garden City, UT 1,830 Mountain 57 Loam ,(Redente 1980). The name “sweetvetch” is synonymous with Utah 10 Musselshell Co., MT 1,586 Loam sweetvetch and northern sweetvetch. The most recent and com- 11 Rio Blanc0 Cc., CO 1,739 Loam z plete description of the H. boreale complex was by Northstrom and Welsh (1970), as revised by Northstrom (1976). Sweetvetch original seed was used in both the greenhouse and field experi- growth in the early spring provides a considerable amount of ments. Seed from the Hobble Creek Utah site, (Ecotype 7) and the Authors arc plant physiolo ’ t, USDA Agricultural Research Service, Foray and Ballard Mine, Idaho site (Ecotype 4) came from plants established Range Research, Utah State 9p-9mvemity, Logan 84322-6Mo;former research asmutant, from seed from Orem, Utah, site (Ecotype 5) in 1963 and about De . Range Saence, Utah State University, Lolpn 843224230; rexarc ’ ’ USp: A Agricultural Research Service, Forage and Range Rcscarch,!J%<~~ 1%5, respectively. University, Logan 843224300, and research forester (retired), USDA Intermountain Forest and Range Experiment Station, Lo n, Utah 84321. Tii Ford currently is Greenhouse Experiment (Seedling Estrblisbment Chuycteristics) agronomist, Desert Land and Livestock, P. $ . Box250, Woodruff, Utah 84086. Seeds were removed from the loments, mechanicaily scarified, Cooperative investigations of the USDA Agricultural Research Service, the USDA Forest Service, and the Utah Agricultural Experiment Station. Journal Paper 3717. and inoculated with approximately 0.1 g of Rhizobium (Hedysa- Manuscript accepted 13 February 1989.

496 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 rum Spec. 2 inoculum, Nitragin Co.‘, Milwaukee, Wi8.j per 250 arranged in a randomized complete block design. For the seed seeds. They were sown in containers (length 20 cm, volume 195 cc) weight measurements, seed was bulked over replications within filled with a mixture of I/ 3 sand, I /3 peat, and 1/ 3 loam soil on 26 ecotypes and 4 aliquots of 25 seeds of each ecotype were weighed. March 1984. The plants were propagated in a greenhouse main- The second experiment quantified root and shoot biomass, root tained at approximately 30/ 15” C day/ night temperature regime length, nitrogen furation activity, and forage quality. Fifteen repli- with no supplemental lighting. Plants were kept watered at near cations of single plant plots were arranged in a randomized com- field capacity for 35 days before being placed under a greenhouse plete block design. Some plant mortality occurred, but an average line-source sprinkler system on 30 April 1984. The greenhouse of 12 plants were harvested for each ecotype. Soil at the site is a line-source sprinkler system has a sprinkler nozzle that traverses uniform Nibley silty clay loam (fine, mixed, mesic, Aquic Argius- along a central line above a groundbed and applies an amount of toll) with 0 to 3% slopes. Precipitation at the site occurs primarily water that decreases as distance from the central line increases as winter snow and spring rain and averages 44.2 cm annually (Johnson et al. 1982). Ten plants from each ecotype were placed at (United States Department of Agriculture 1974). During the 2 equal intervals along the irrigation gradient in 3 replications using years of the experiment, the annual precipitation was 77.2 cm and a randomized complete block design. Total water applied during 48.0 cm, respectively. the approximate 38-day growth period along the gradient ranged Plants of each ecotype were harvested during 20-23 June and from 9.80 cm (123 ml) to less than 0.04 cm (0.5 ml). Surviving again during 4-6 October 1984. Ten adjoining replications were plants were combined into 3 groups according to the total amount excavated the first harvest to minimize damage to plants that were of water applied. The 3 groups represented abundant water (plants to be excavated at the second harvest. A tractor-mounted tree lifter 1 through 3), adequate water (plants 4 through 6), and limited was used to cut the root systems at a depth of approximately 35 cm. water (plants 7 through 9); mean total water application in the Plants were then carefully removed from the soil by hand to groups was 8.7,4.2, and 1.2 cm, respectively. Virtually all plants at recover the root systems. This resulted in an excavated soil block the tenth position on the gradient died and were not included in the that was approximately 55 cm long, 45 cm wide, and 35 cm deep for analysis. each plant. Nodulated root sections were placed in 120 cc syringes During 6 to 8 June 1984, leaves were removed from the seedlings and nitrogen fixation was estimated for each accession using the under the line-source sprinkler and leaf area measured with a previously described acetylene reduction technique. Roots and Licor’ area meter. Stems, petioles, and leaves were then oven-dried shoots were weighed fresh, oven dried for 72 h at 65O C, and at 75’ C for 48 hand weighed. Nitrogen fixation was estimated on a reweighed for dry weight. whole plant and specific nodule activity basis by the acetylene Four plants of each ecotype were randomly selected from the reduction procedures described by Johnson and Rumba@ excavated plants for forage quality analysis. Shoots of these plants (1981). Each root system was extracted from the soil without were ground in a Wiley mill using a 1 mm screen and placed in washing and placed in a 6O-cc plastic syringe. The syringe was filled air-tight Mason jars until analysis for percent nitrogen by the with 54 cc of ambient air and 6 cc of washed commercial acetylene, macro-Kjeldahl technique (Harris 1970). The nitrogen content was creating an atmosphere of 10% acetylene and 90% air. Gas samples multiplied by 6.25 to estimate crude protein percent. In vitro dry were incubated 1 h at ambient temperatures and shielded from matter digestibility (IVDMD) for these plants was estimated using direct sunlight. After incubation, 11 cc gas aliquots were injected rumen fluid from a Holstein heifer fed an alfalfa hay diet. A 48-h from the syringes into 10 ml stoppered Vacutainers’. Ethylene fermentation period with the strained rumen fluid was followed by content of the gas sample was determined using a gas chromato- a 48-h incubation in hydrochloric acid-pepsin solution. graph quipped with dual hydrogen-flame ionization detectors Differences among ecotypes detected by analysis of variance maintained at 1500 C. procedures were tested for significance using the LSD procedure. The root systems were then washed and root lengths measured Data from the greenhouse and field experiments were combined with a Comair* root scanner. Nodules were removed from the roots (44 attributes) and taxonomic distances and clustering among the and counted. Roots and nodules were then oven-dried at 75O C for ecotypes computed by the unweighted pair-group method using 48 h and weighed. Total acetylene reduction activity was expressed arithmetic averages of standardized data (Rohlf 1987). as moles of ethylene per hour for the whole root system, whereas Results and Discussion specific nodule activity was expressed as moles of ethylene per gram of nodule dry weight. Data were analyzed by analysis of Greenb~ Experiment In the Intermountain West adequate soil moisture is usually variance, correlation, and multivariate procedures. Least signifi- available for imbibition and germination, but seedling establish- cant difference (LSD) values were computed to evaluate differen- ment and growth can be severely impaired by dry conditions ces among means. during early summer. The greenhouse experiment simulated Field Experiments (Mature Plaot Characteristics and Seed Pro- growth conditions during thii early summer period. Little mortal- duction) ity occurred under the abundant and adequate water level treat- For the field portion of the study, seeds from each sweetvetch ments. Under drought stress in the limited water application ecotype were inoculated with rhiiobia, germinated, and grown as treatment level, mortality was 30%. in the greenhouse experiment for approximately 30 days prior to Sweetvetch shoot mass decreased as less water was applied, transplanting. Plants of NC-83-l alfalfa (Kehr et al. 1975) served as whereas root mass was greatest at the adequate level of water a standard for comparison. Plants were transplanted in June, 1983, application (Table 2). Significant differences were observed among at a site 2 km south of Logan, Utah. Plants were placed 1 m apart in the ecotypes for both shoot and root mass. Ecotype 5 had the one-plant plots with additional plants providing one-row borders greatest shoot and root mass at the abundant water level, while to separate the seed production experiment and the destructive Fcotype 3 had the greatest root mass and the second greatest shoot harvest experiment. The first experiment measured seed produc- mass at the limited water level (data not shown). Some ecotypes, tion resulting from open-pollination and consisted of 8 replications e.g. Ecotype 1, maintained relatively constant shoot, root, and total biomasses as less irrigation water was applied, but the mass of Wention of a trademark, roprietary roduct, or vendor doea not constitute guaran- tee or warranty of the pr J uct by the s .S. Department of Agriculture and Utah State other ecotypes, e.g., Ecotype 11, declined markedly when less University. and does not imply approval to the exclusion of other product8 or vendors water was applied (data not shown). that may aho be suitable. The proportion of root mass to total plant mass increased as less

JOURbJAL OF RANGE MANAGEMENT 42(6), November 1999 497 Table 2. Shoot man, root mass, total root length, green leaf area, nodule root development rapidly (Blaisdell 1949, Platt and Weis 1985). man, number of nodolen, plant acetylene reduction aetivfty, and qed5c The seedling root needs to quickly explore a large soil volume and nodule 8ctidty of 11 aweetvetd~ ecotypa 8t 3 water kveh pnda the penetrate the soil profile as deeply as possible to aid in water ycenbouwlinc-lourcerprtnLkrr~m.~v~nprrrcntrthcmcrrn uptake as the soil profile dries. Consequently, total seedling root of 9 obaervation8 for tbe 8bund8at and limited water level8 aad 12 length is probably an important screening characteristic for suc- ob~~8tlo1~ for the adequte w&r kwel for each ecotype. cessful establishment in semiarid environments. Green leaf area could potentially be an important screening Signifii of F tests’ characteristic and aid establishment success of sweetvetch. During Differcnaa water high soil moisture periods, production of photosynthetic leaf area among 1CVCls would be important to optimize carbon gain during favorable Range of =otype by growing conditions. During drying periods, a plant that can main- Trait Water level Mean ccotype means means ccotypcs A tain green leaf area and photosynthetic activity may be able to Shoot mass (mg plant’) l * grow for a longer period and take advantage of additional precipi- Abundant 50 31 - 78 l tation. As a result, greater green leaf area at both higher and lower Adequate 42 31 - 98 l moisture levels probably confers a competitive advantage. Limited 21 9 - 31 * Nitrogen fixation is a unique biological process that may be Root mass (mi plant-‘) l * beneficial to plants growing in nitrogen-limited environments. Abundant 33 - 76 + Adequate z 37 - 86 * This benefit may be magnified during seedling establishment when Limited 31 11 - 49 l rooting density is minimal. However, the energy required for nitrogen fixation may mean that it may be one of the first plant Root length (m plant”) l * Abundant 1.36 0.80 - 1.77 l functions curtailed during drought. Consequently, it is important Adquate 1.63 1.10 - 2.09 l to determine if sweetvetch ecotypes vary in their ability to fur Limited 0.87 0.31 - 1.38 l nitrogen under water stress. Green kaf area (cm2 plant-‘) l * The number of nodules remained relatively constant for the first Abundant 4.3 2.1 - 7.0 l 8 plants along the water gradient, indicating that this trait was Adequate 3.0 2.1 - 3.8 l quite tolerant to decreasing water availability. Ecotypes differed Limited 1.0 0.4 - 1.8 NS significantly within each of the water levels, and their relative Nodule mass (mg plant-‘) .* rankings were consistent over the 3 water levels (Table 2). Nodule Abundant 7.2 2.5 - 11.5 l mass was reduced by 18% from the abundant to adequate water Adquate 5.9 3.2 - 9.4 * level and by an additional 51% from the adequate to the limited Limited 2.9 1.2 - 5.3 l water level. Ecotypes 2,4, and 5 had the most nodules and nodule Nodules (number plant”) NS mass at the abundant level of water application, and Ecotype 3 had Abundant 9.3 3.4 - 15.2 * Adquate 8.3 4.3 - 13.6 * the most nodules and nodule mass at the limited level of water Limited 4.7 2.0 - 9.4 * application (data not shown). Because nodule mass decreased with declining water, sweetvetch apparently allocates fewer resources to Plant acetylene reduction activity (am01 C& plant’s’) ** Abundant 0.381 0.178 - 0.769 I nodules as water becomes limiting. Adquatc 0.244 0.183 - 0.356 l Total plant and specific nodule acetylene reduction activities Limited 0.039 0.008 - 0.089 NS (Table 2) decreased with the decline in water between the abundant Specificnodule activity (nmols CJHI 8-l g-l nodule dry weight) NS and adequate water levels and were reduced sharply between the Abundant 67.7 12.8 - 110.0 NS adequate and limited water levels. Similar to nodule mass, acety- Adequate 54.0 27.6 - 83.5 NS lene reduction activity on both a whole plant and a specific nodule Limited 9.1 1.3 - 21.6 NS basis appeared more sensitive to drought than the other plant INS, ***Not aignifkantend signifkanta1 the 0.05and 0.01 kvels, rcspeaivcly. characteristics. Ecotypes 4 and 5 had the greatest acetylene reduction activity per plant at the abundant water level, but activities water was applied. Total biomass remained relatively constant for were not significantly different at the limited water level (data not the abundant and adequate water levels. Average root/ shoot ratios shown). Specific nodule activity did not differ among ecotypes. were 1. 1, 1.4, and 1.5 at the abundant, adequate, and limited water The interaction of ecotypes and water levels was significant levels, respectively. Total root length is an estimate of the soil (KO.01) for activity on a whole plant basis but not significant volume that the roots occupy, whereas root mass indicates below- (p>o.OS) on a nodule dry weight basis. ground carbon and nutrient investment. Similarly, green leaf area Field Experiments indicates the seedling photosynthetic area, whereas shoot mass Seed weights of the original collections and of the open- indicates the aboveground shoot carbon and nutrient investment. pollination increases varied significantly (X0.05) among ecotypes As less water was applied, the green leaf area, and thus transpira- (Table 3). Probably because growth conditions were more favor- tional area, decreased (Table 2). Total root length increased from able at the experimental location than at the sites of origin, weights the abundant to the adequate water level, but declined markedly at for 25 seeds from the open-pollination increases of the 11 ecotypes the limited water level. Under moderate water stress sweetvetch were heavier (mean of 0.234 g) than seeds from the original collec- i roots apparently increased the soil volume explored as soil mois- tions (mean of 0.195 g). Although the rankings of ecotypes accord- ture was depleted. Total root length and green leaf area differed ing to seed weight changed between seed generations, the 5 eco- :significantly among ecotypes. Ecotype 5 tended to have the grea- types in the original collection that produced the heaviest seed also test root length and the most leaf area at the abundant and ade- produced the heaviest seed in the open-pollination increases. quate water levels, while Ecotype 3 exhibited the greatest root Spearman’s rank correlation (r, = 0.73, X0.05) indicated an asso- length and most abundant green leaf area at the limited level (data ciation between the 2 types of s&s. not shown). In the greenhouse experiment, Ecotype 5 exhibited the best To ensure successful seedling establishment, a seed should ger- seedling growth at the abundant water level (data not shown) and minate quickly and initiate radical cell elongation and subsequent also had one of the highest seed weights (Table 3). Conversely,

498 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 seed weight(&J 25 seeds_‘) Ecotype ~otype Total identification original Test Seed production ident& Forage CNdC crude number collection envirottmettt (g plant-‘) cation weight protein protein IVDMD number (g plant-‘) % (g plad) (%) 0.166 fg 0.226 de 4.3 bc : 0.199 d 0.211 ef 3.6 bc Harvested 2Q-23 June 1984 3 0.137 h 0.213 d 3.2 bc Swectvetch: 4 0.243 a 0.271 ab 2.8 bc 1 54e 16.Oa 8.6 d 54.8 ab 5 0.223 b 0.278 a 9.0 a 2 97 cdc 15.9a 15.4 cd 48.8 bc 6 0.228 ab 0.260 abc 5.8 ab 3 117cd 16.6s 19.4 bc 50.0 bc 7 0.203 c 0.272 ab 3.0 bc 4 126bC 17.3a 21.8 bc 51.7 b 8 0.229 ab 0.255 bc 0.1 c 5 168b 15.6a 26.2 b 48.4 bc 9 0.188 de 0.239 cd 3.3 bc 6 12obc 15.2a 18.2 c 52.4 b 10 0.149 gh 0.149 g 0.2 c 122bc 17.Oa 20.7 bc 52.9 b 11 0.176 ef 0.199 f 1.6bc 64dc 16.2a 10.4 d 49.9 bc Mean 0.195 0.234 3.4 9 41 c 16.2a 6.6 c 52.4 b 10 69de 12.9a 8.9 d 49.1 bc Wana in the ~me column that an followed by a differentletter arc Ggnifiitly 11 64de 16.Oa 10.2 d 43.9 c dierent by LSD at the 0.0s level. Mean 95 15.9 15.1 50.4 Ecotype 3, which showed the best seedling growth at the limited Alfalfa: 279a 17.6a 49.1 a 61.8 a water level (data not shown), had one of the lowest seed weights (Table 3). These results indicate the heavier seeds may be impor- Harvested 4-6 October 1984 tant in screening for environments where precipitation is equival- SlWXtWtCh: ent to the abundant and adequate water treatment levels, but may 1 62 b 8.0 bc 5.0 b 35.7 2 82b 10.2 a 8.3 b 40.2 not be a desirable characteristic in screening for drought resistance. 3 83 b 7.8 bc 6.5 b 38.6 Similarly, Wright and Brauen (1971) found that Lehman lovegrass 4 87 b 7.8 bc 6.8 b 28.8 (hkagrostis lehmanniana Ncxs) lines with the heaviest seeds were 5 68b 9.4 ab 6.3 b 36.1 generally the most drought susceptible. 6 %b 8.8 abc 8.4 b 30.2 Seed weight is positively correlated with germination rate and 7 115b 5.4 d 7.2 b 37.6 seedling vigor in several plant species (Erickson 1946, Kneebone 8 85 b 8.1 bc 6.9 b 47.6 9 45 b 8.0 bc 3.6 b 28.8 and Cremer 1955, Beveridge and Wilsie 1959, Fransen and Cooper 10 135b 7.4 c 10.0 b 49.3 1976, Asay and Johnson 1980) and has been used to screen for 11 57 b 10.3 a 5.9 b 41.3 seedling vigor in several forage grasses (Tossell 1960, Asay and Mean 83 8.3 6.8 37.7 Johnson 1980). Seed weights of the 11 sweetvetch populations Alfalfa: 1,167 a 10.7 a 124.8 a 41.2 from the original collections were indicative of their seed weights when the populations were grown in a uniform environment. ‘Mcatw in the lame columnwithin a harvestthat arc followedby a diiermt letterare ~&ifiintly differentby LSD at the 0.05 level. However, because of the uniform environment, the open-pollination seed weights probably more accurately reflect the magnitude of the sweetvetch ecotypes at the first harvest but did at the second genetic differences among the accessions. harvest (Table 4). Alfalfa contained a higher percentage of protein Although some plants flowered, no viable seed was produced and greater total crude protein than any sweetvetch ecotype at both during the initial establishment year. Plants produced seed in their harvests. Sweetvetch ecotypes differed significantly (KO.05) in second growing season in the field nursery, and ecotypes differed total crude protein at the fast harvest but not at the second harvest. significantly (kYO.05) in seed yields (Table 3). Ecotype 5, originally Alfalfa tended to have a higher in vitro dry matter digestibility from a site receiving an intermediite level of precipitation, pro- (IVDMD) than any of the sweetvetch ecotypes at the fast harvest, duced the most seed in the test environment, while Ecotype 8 from but Ecotypes 8 and 10 were slightly more digestible than alfalfa at the most arid site produced the least seed. the second harvest. Ecotypes differed signiticantly (KO.05) in In the second field experiment plants were destructively har- IVDMD at the first but not the second harvest. vested during their second growing season to estimate forage avail- At the second harvest sweetvetch was much less valuable as a fall ability in early summer and fall. At the first harvest all sweetvetch forage than alfalfa (Table 4). At this harvest most of the sweetvetch ecotypes were in full bloom, whereas alfalfa was at one-tenth plants had senesced and had very few leaves. In addition, sweet- bloom. This difference in maturity limits the ability to directly vetch growth at this harvest was prostrate and would be difficult to compare alfalfa and sweetvetch. Sweetvetch provided more forage graze, particularly if covered by snow. Ecotype 10, one of the at the first than at the second harvest (Table 4); however, alfalfa inferior accessions at the first harvest, ranked at or near the top for produced significantly (KO.01) more forage than any sweetvetch forage weight, total crude protein, and IVDMD at the second ecotype at both harvests and more than 850% more forage than the harvest. Thii difference may have reflected phenological differen- most productive sweetvetch ecotype at the second harvest. Ecotype ces because Ecotype 10 originated from the northern-most location 7 produced the most forage annually (237 g), and Ecotype 9 pro- and appeared to begin growing later in the spring and grew later duced the least (86 g). Significant differences (X0.05) in forage into the summer than the other ecotypes. yields occurred among ecotypes at the first harvest but not at the Nodule mass varied significantly among the 11 ecotypes. Total second harvest. acetylene reduction per plant was highest for Ecotypes 3,4, and 5 at Percent crude protein did not vary significantly among the the first harvest (Table 5); Ecotypes 3 and 5 also had high specific

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 499 Table 5. Nodule man d acetylene reduction activity of field growo ranking of the ecotypes changed with harvests. Ecotype 10 had the plant8 of 11 ametvetcb ecotyper at 2 buvab. Ewb value npruenb the lowest nodule mass at the fit harvest, but exhibited the highest meanofl2meawements8ttheflr8tb8rvestand11 manuementa at the nodule mass at the second harvest. Nitrogen fixation characteris- second huvest~. tics of sweetvetch ecotypes could not be compared with those of alfalfa because none of the alfalfa plants in the field experiment Ecotype Acetylene reduction activity (nmol C&L) were nodulated. Other characteristics that may be important in breeding and identifica- SpCCifiC tion Nodule mass Total (g-l nodule dry selection programs include regrowth capability and rhizome number 041 @ant-? (plant-’ 8-l) weight 8-l) development. Regrowth may be important in fall forage produc- Harvested 20-23 June 1984 tion, and rhizome development would be important in erosion 0.33 be 10.5 bc 24.2 bc control and resistance to grazing. Seven of the 11 sweetvetch 0.30 bc 6.1 c 13.9 c ecotypes exhibited regrowth, whereas only Ecotypes 7,9, and 10 0.61 ab 25.0 a 43.6 a showed any rhizome development (data not shown). 0.98 a 22.5 ab 18.6 bc Field forage weight from Harvest 1 and seed production, proba- 1.01 a 26.9 a 31.1 ab bly the 2 most important economic variables examined in this 0.61 ab 10.6 be. 19.7 bc 0.59 ab 8.3 c 15.6 bc study, were significantly and positively correlated (r = 0.66). Total 0.33 bc 10.8 bc 16.7 bc crude protein and nodule mass from the field for Harvest 1 were O.lOc 4.4 c 18.9 bc significantly correlated with both forage and seed production from 0.05 c 2.8 c 7.8 c Harvest 1. Significant correlations were also found for the first 0.33 bc 8.6 c 21.4 bc field harvest between nodule mass and forage weight (r = 0.89) and 0.48 12.5 21.1 between nodule mass for the first field harvest and seed production (r = 0.62). In addition, total acetylene reduction at the first field Harvested 4-6 October 1984 harvest was significantly correlated (r = 0.73) with forage weight at 1 0.34 bc O.Oa O.Oa the first field harvest. These 3 correlations suggest that biological 2 0.11 c O.Oa 0.9a 3 0.51 b O.Oa O.Oa nitrogen fixation may have been important in the production of 4 0.14 c O.la 0.2il both forage and seed. 5 0.26 bc O.la O.la 6 0.38 bc O.la O.la Relationships among Ecotypa I 0.28 be O.Oa Oh The results of the cluster analysis for the 11 sweetvetch ecotypes 0.33 bc O.Za 0.4a are shown in Figure 1. The cophenetic correlation coefficient of ; O.la 0.78 indicated satisfactory agreement between the dendrogram and 10 0.231.41 abe % 0.3a 11 0.31 be 0:Oa 0.4a the resemblance matrix. Average taxonomic distance values for Mean 0.39 0.1 0.2 ecotype pairs ranged from 0.94 for Ecotypes 1 and 6 to 2.04 for Ecotypes 5 and 10 (Fig. 1). Ecotype 10, which represented the ‘Means in the samecolumn that arc followed by a diierent letter arc signiticantly different by LSD at the 0.05 level. northernmost collection, was distinctly separated from the other 10 sweetvetch ecotypes with a taxonomic distance value of 1.62. nodule activity. Acetylene reduction activity was extremely low in Ecotypes 4 and 7, which originated from seed from Ecotype 5, were plants from the second harvest, and no significant differences were closely clustered near Ecotype 5 baaed on the 44 morphological detected among the ecotypes. Nodule activity in sweetvetch appar- and physiological attributes evaluated. This clustering of related ently coincides with periods of most active growth. Nodule mass collections suggests that this analysis procedure is a useful tool for was generally lower for the second than the first harvest, and identifying relationships among the sweetvetch collections. Although

Origin -ID Benmore, UT l- Salt Lake City, UT 6 Duchesne, UT 3 Moab, UT 8 Garden City, UT 9 Rio Blanc0 Co., CO Arco, ID ! Ballard Mine, ID 4 Orem, UT Hobble Creek, UT ;

Musselshell Co., MT IV a 1 I 1 I I 0.8 1.0 1.2 1.4 1.6

Taxonomic Distance

Fig. 1. Clwter relationships of II sweetvetch ecotypes based on average taxonomic distance values computedfrom 44 morphological andphysiological attributes measured in greenhouse andjXd experiments.

!Xo JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 Ecotypes 8 and 9 originated from sites that differed greatly in their Jobmon, D.A., and M.D. Rumba@. 1981. NcduIation and acetylene precipitation (19 and 57 cm, respectively), their taxonomic dis- reduction by certain rangeland legume sptcies under held conditions. J. tance values were nearly equal (1.07 and 1.05, respectively), sug- Range Manage. 34:178-181. gesting that these ecotypes were genetically similar. Consequently, Johnson, D.A., M.D. Rumba@, L.S. Willardron, K.H. Auy, D.N. Rlneltart, and MR. Auruteb. 1982. A greenhouse line-source sprinkler even though ecotypes differed in their taxonomic distance values, system for evalusting plant response to a water application gradient. these differences were not always clearly related to specific charac- Crop Sci. 22469472. teristics of the collection site. Kehr,;W.R., D.K. Bama, E.L. Sorensen, W.H. Skrdla, C.H. Hanson, D.A. Mlkr. T.E. Themnsoa. LT. Carlaon. L.J. EBlnc. R.L. Taylor, Conclusions M.D. Rumbough, E.T. B&km, D.E. B&II, and M.K. M&r. i975: Registration of alfalfa germplasm pools NC-83-1 and NC-83-2. Crop Sweetvetch produced substantial amounts of forage during early Sci. 15604605. spring, but was apparently of little value for forage during the Kneebone, W.R., and C.L. Cremer. 1955. The relationship of seed size to late-fall and winter because of its prostrate fall growth and limited seedling vigor in some native grass species. Agron. J. 473472477. forage production during this period. However, it may be possible Northstrom, T.E. 1976. Typification and nomenclature of Hedysarum to select sweetvetch cultivars with an ability to regrow during the bowale Nutt. Great Etasin Natur. 36:45&464X Nortbstrom, T.E., and S.L. Welsh 1970. Revision of the Hedysarum fall. Acetylene reduction activities indicated that sweetvetch is boreale complex. Great Basin Natur. 30~109-130. capable of fixing nitrogen and would probably do well in nitrogen- PI&, W.J., sad I.M. WeIs. 1985. An experimental study of competition limited environments. Considerable morphological and physiolog- among fugitive prairie plants. Ecology 66:708-720. ical variability was observed in the 11 sweetvetch ecotypes studied. Plummer, A.P., D.A. Cbdstemen, ad S.B. Monsen. 1968. Restoring Ecotype 5 excelled under the high water levels of the greenhouse big-game range in Utah. Utah Div. Fish Game Pub. 68-3. line-source sprinkler system. Ecotype 3 was superior under the low Redente, E.F. 1980. The autecology of Ifedysarum boreale. Ph.D. Diss. Colorado State Univ., Fort Collins, Colorado. Diss. Abstr. No. 8028763. water levels and may be best for planting on arid, nitrogen-limited Redente, E.F., and F.B. Reeves. 1981. Interactions between vesicular- sites. Ecotypes 7 exhibited the greatest rhizome development and arbuscular mycorhixae and Rhizobium and their effect on sweetvetch may be valuable for stabilizing erodible sites. Ecotype 5, which growth. Soil Sci. 132:410415. exhibited superior seed yield, is a candidate for commerical seed Roblf, F.J. 1987. NTSYS-PC numerical taxonomy and multivariate analy- production. Sufficient genetic diversity was present among the sis system. Ver. 1.30. Exeter Publ., Setauket, New York. sweetvetch ecotypes to assure adaptation to a wide array of sites Rumbaugh, M.D. 1983. Legumes-their use in wildland plantings. p. 115-122. In: S.B. Monsen and N. Shaw (compilers). Managing Inter- and to facilitate improvement through breeding and selection. mountain Rangelands-Improvement of Range and Wildlife Habitats. USDA Forest Serv., Gen. Tech. Rep. INT-157. Literature Cited Sbaw, N., and S.B. Monsea. 1983. Nonleguminous forbs for rangeland sites. D. 123-131. In: S.B. Monsen and N. Shaw (comnilers). Manaxina hay, K.H., and D.A. Johnson. 1980. Screening for improvedstand estab- Intcrmountain Rangelands-Improvement of Range and Wildlife HibL lishment in Russian wild ryegrass. Can. J. Plant Sci. 60:1171-l 177. tats. USDA Forest Serv., Gen. Tech. Rep. INT-157. Bcverldp, J.L., and C.P. Wlldc. 1959. Influence of depth of planting, seed sixe. and variety on emergence and seedling vigor in alfalfa. Agron. J. Tosaell, W.E. 1960. Early seedling vigor and seed weight in relation to 51:731-734. breeding in smooth bromegrass, Bromus inermis Leyss. Can. J. Plant Blaladell, J.P. 1949. Competition between sagebrush seedlings and reseeded Sci. 40~268-280. grasses. Ecology 30512-519. United States Department of Agriculture. 1974. Soil survey of Cache Valley area, Utah. 192 p. Erickson, L.C. 1946. The effect of alfalfa seed sixe and depth of seeding upon subsequent procurement of stand. J. Amer. Sot. Agron. 38964-973. Wasser, C.H. 1982. Ecology and culture of selected species useful in revegetating disturbed lands in the West. USDI Fish Wildl. Serv. FWS/OBS- Fransen, SC., and C.S. Cooper. 1976. Seed weight effects upon emergence, 82156. leaf development, and growth of the sainfoin seedling. Crop Sci. Wright, L.N., and S.E. Brauen. 1971. Artificial selection for seedling 16434437. drought tolerance and association of plant characteristics of Lehmann Harris, L.E. 1970. Nutrition research techniques for domestic and wild lovegrass. Crop Sci. 11:324-326. animals: Volume 1. Utah State Univ., Logan, Utah.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 501 The effects of intra- row spacings and cutting heights on the yield of Leucaena leucocephala in Adana, Turkey

'i / \

AbStT8Ct This reaeucb was conducted in Adana, Turkey, between 1985 are received in December, January, and February (winter); 168 and 1987. Tbe study investigated tbe effects of the htra-row spat- mm in March, April, and May (spring); 29 mm in June, July, and inp and cutting heights on yields of 2 cultivars of Leucaenu hu- Au8ust @mmer); and 122 mm in September, October, and cwmph&, I(8 and Peru. November (fall), respectively. Tbe trial was a split-split plot ddgn in randomized blocka witb 4 Two L. kucocephala cultivar~, K8 and Peru, were used. The replications. Tbe main plots were tbe cultlvars, K8 and Peru; former, originating from Mexico, is-8 giant Salvadorian tree type. subplots were i&a-row qm- 25, SO,and 75 cm; and sub-sub The latter, ori8inating from Argentina, is a heavily branching plots were cuttiog beigbte, 20,4@,and 60 cm. Row spa&g was lm. shrubby type. Peru bad bigber leaf yields than KS.Increased i&a-row spacing Plots were arranged in a split-split plots design in random&d decreased both thin and thick dried stem yielda and foliage yield blocks with 4 replications. Size of the sub-sub plots were 3 X 10 = 30 per ba. Tbe effect6 of cutting bdgbts varied witb the years. Tbe mr. Main plots were cultivars; sub plots were intra-row spacings bigbest dried foliage yield was obtained from tbe plots cut at 40 cm (25 cm, 50 cm, and 75 cm) representing 3 planting densities as lntbe5rstyearand6Ol!miatbesecondyear. 40,000, 20,000 and l3,333 plants, ha”; and sub-sub plots were cutting heights (20 cm, 40 cm, and 60 cm). Key Worth intra-row spadng, forage yield, Le- kuco- Seeds of Lmcaena cultivars were treated with hot water for CrpAolrr about 3 minutes and soaked in the water overnight as described by Leucaena leucocephala is a multipurpose leguminous plant Skerman (1977). Soil was removed from under Leucaena trees 8rown in the tropical and subtropical re8ions of the world (Brew- which were previously planted with appropriate inoculum on 6 baker and Hutton 1979). This plant has been used as feed and July 1985. Thii soil was placed into the rows in which the water- forage for livestock, fuel, soil erosion COntrOl, ad biological fCItil- soaked seeds were hand sown, 3 seeds per planting position intra ixation materials, quality celhrlose for paper mills, and lumber. row spacin8. Seedlings were thinned at the 3 leaf stage, and only 1 L. leucocephala is naturalized between the latitudes of 30° North was left at each planting position. and 30” South (Skerman 1977). It has 3 distinct forms: Hawaiian, During the next 2 years the plants were cut when they reached an Peruvian, and Salvadorian types. These represent shrubby, shrubby- average height of 1.5 m Four cuts in 1986 (at 14 May, 27 June, 4 tree, and tree forms of the species (Skerman 1977). Aug., and 10 Sept.) and 3 in 1987 (at 23 June, 10 Aug., and 23 A preliminary research on the performance of L leucocephala in Sept.) were employed, respectively. Different cutting frequencies our region was conducted and published in Turkish with an Eng- were due to the climatic variation; the minimum temperature for lish summary (Tunq and Tiikel1986). In this paper, it was reported initiating growth of Leucaena in spring (15.50 C), as reported by that about 4.3-7.15 tons ha-’ green foliage and 470-780 kg ha-’ Skerman (1977) and Brewbaker and Hutton (1979), was met in crude protein yields were obtained when planted at 1 m spacing March 1985 and early May 1987. and cut at 60 cm height. These and other observations 8ave us some Plant materials cut from each plot were separated into 3 compo- indications that the L leucocephah may be grown further up to the nents: thick stems > 5 mm diameter; thin edible stems < mm 37” latitude than its commonly naturalized border and may also be diameter; and leaves. This separation is customarily used for Len- used as a summer forage crop for the small animal-raisin8 farms cuena (Gueverra et al. 1978). which usually experience gmn forage shortages during the peak of Results and Discussion summer months in the region. The main purpose of this current study was to further determine Tbkk Stem Yield optimum intra-row spaa and cutting heights for L kucocephala There was no statistically significant difference between the 2 in Adana, Turkey. years in the yield of thick stems in spite of the different cutting frequencies. Neither was there any si8nificant difference in the Material and Methods thick stem yield (Table 1) between the 2 cultivars. However, increasin8 intra-row spacing significantly decreased The study was conducted in the research plots of the Field Crops Department, Agricultural Faculty, Gukurova University, located the thick stem yield (Table 1). This finding was in support of the report by Gueverra and et al. (1978). at 37O 21’N and 3S’ 10’ E, throughout the years between 1985 and 1987. TMn stem Yield The soils were young alluvial deposits of the river Seyhan having Mean thin stem yield did not show any significant effects for the high lime contents, sandy-loam type textures, and classified as years and the cultivars (Table I). However, it&a-row spacing and Glass I land capability (Gxbek et al. 1974). cutting heights affected the thin stem yield. Giving more intra-row A typical coastal mediterranean climate prevails in the region spacing caused s@nifiit yidd decreases. The h@hest yield (1.04 with 8 total yearly precipitation of 642 mm. of this total, 323 mm tons ha-‘) was obtained from 25 cm in&a-row spacing but the lowest yield (0.78 tons ha-‘) was from 75 cm spacing. This was more Authon are associate professor, Deportment of Field Cropn, Cdlcge of Agricul- tun, ~ukurova Univcraity, A&M, Turkey; and march asshnk Dcputment of likely due to the fact that the larger intra-row spaced plants were Fiibzl Crop, Cdegc of Apiculttm. uknrova University, A&M, Turkey. branching more but producing less overall yield than the narrower Mmwcnpt accepted 28 February F989. spaced plants, which were abk to grow a greater number of plants

502 JOWRNAL OF RANGE MANAGEMENT 42(8), November 1999 Table 1. Dried yield of Leumem kucoqh& (tom ha-‘).

1986 * Ye8rs 1987 Mean Thick Forage Yield Thick Forage Yield Thick Forage Yield Treatments stems Thin stems Foliage stems Thin stems foliage StClUS Thii stems Foliage cultivars K8 2.46 0.73 4.19 2.49 0.89 3.70 2.48 0.81 3.95 b PCN 2.32 0.93 4.57 2.43 1.02 4.26 2.38 0.98 4.42 a In&a-Row Spacings (cm) 25 2.89 1.08 5.20 2.46 1.07 4.40 2.84 a 1.04a 4.80 a 50 2.32 0.79 4.17 2.29 0.94 3.80 2.30 b 0.86 b 3.99 b 75 1.97 a);70 3.78 2.31 0.87 3.76 2.14 b 0.78 b 3.78 b Cutting Heights (cm) 20 2.72 a 0.88 a 4.26 b 1.97 b 0.86 b 3.19c 2.35 0.87 3.73 4.36 : 2.402.06 bc 0.730.87 ab 4.304.59 ab 2.572.84 a 1.020.99 aa 4.144.63 ba 2.482.45 0.930.87 4.46 Mesns in each column with the same letters an not statistically different from each others at the 0.05 level of signifii~~ determined by the “LSD” teat. per square meter, producing fewer branches but more thin stem the plants cut at 4O-cm heights in the first year. However, dried yield. foliage yield was increased as the cutting heights increased, and the Effects of cutting heights on the thin stem yield revealed an highest yield (4.63 tons ha-‘) was reached at the MI-cm cutting inverse situation for the 2 consecutive years (Table 1). height in the second year. This was an interesting trend which might be explained by the combining effects of the cutting heights Dried FoKap Yield and the cutting frequencies applied in 2 different years. Cultivar Peru with a mean 4.42 tons ha-‘.dried foliage yield was significantly different and a higher yielding cultivar than K8 with a Literature Cited mean yield of 3.95 tons ha-‘. This difference was more likely due to Brewbaker, J.L., and E.M. Hutton 1979. Leucuena Versatile tropical tree the growth characteristics of the cultivars tested: the former a legume. p. 207-259 In: New Agricultural Crops, Edited by Gary, A. branching and the latter a gianttree type. Ritchie A.A.A.S. Selected Symposia Series. Westview Press, Boulder, Dried foliage yield was also significantly affected by the intra- Colorado. row spacings (Table 1). The highest yield (4.80 tons ha-‘) was Gueverra, A.B., A.S. Wbhey, and J.R. Thompum. 1978. Infiuencc of obtained from the 25-cm spacing. This result indicates that the intra-row spacing and cutting regimes on the growth and yield of Leu- caena. Agron. J. 70: 1033-37. dried foliage yield can not be increased by increasing intra-row Ozbek, IL, U. Dw, and S. Kapur. 1974. A complete soil survey of the spacing more than 25 cm. As mentioned initially, more densely camp area of Cukurova Univesity. Cukurova Univ. Col. of Agr. Pub. 73 sown L leucocephala plants produced more thick and thin stem Sci. Res. and Invest. 8 (In Turkish with English summary). yields. Therefore, dried foliage yields were increased as the thick Skerman, PJ. 1977. Tropical forage legumes. FAO Plant Production and and thin stem yields increased. These findings support the claims of Protection Series 2. Sumberg (1985) and Tune and Tiikel(1986), who concluded that Sumbarg, J.E. 1985. Note on estimating the forage yield of two tropical there was highly significant and close correlation between the stem browse species. Trop. Agr. (Trinidad) 26 (1):15-16. Tune, S.T., and T. Tukell986. A research on the effects of different cutting weights and the dried foliage yield for the first and second cuts of times and frequencies on yield and some agricultural characteristics of L.eucaena leucocephala. Leucaena kucocephala (Lam) de Wit. J. Agr. Coll., Cukurova Uniter- The highest dried foliage yield (4.59 tons ha-‘) was obtained from sity. l(I):+79 (In Turkish with English Abstract). _

JOURNAL OF RANGE MANAGEMENT 42(6), November 1980 Gambel oak root carbohydrate response to spring, summer, and fall prescribed burn&g

MICHAELG. HARRINGTON

Control of Gambel oak (Quercus gambd# Nutt.) for increased forage production and con& regeneration is difficult beawe of seasonal TNC variations between Gambel oak as a ponderosa pine its vigorous sprouting ability. Nonstructural root carbohydrate understory species and as a dominant overstory range species, and concentr8tion~ generally 8 good indicntor of sprouting potential, 3) to examine TNC changes in Gambel oak after single and were measured in undentory Gambel oak in a dense ponderosa repeated prescribed burns applied in different seasons. pint (PInus pondkrosu Laws.) stand follotig prescribed fire. Carbohydrates in roots of l- to Zyear-old sprouts after a single fire Study Site and Methods treatment were similar to tboae in unburned, mature oaks. Two The study area is on the San Juan National Forest in southwest- prescribed burna, 2 years apart during the summer carbohydrate em Colorado, 40 km northeast of Durango, at an elevation of depression, caused these root reserves to remrin low into fall 2,300 m. Aspect is south with less than 5% slope and precipitation dormancy and probably contributed to an observed oak reduction. at the study site averages 44 cm annually, with dry periods in the This summer carbohydrate depression, also observed in open- late spring and heavy midsummer thundershowers. Soil is gener- grown Gambel oak, can be recognized by rapid stem growth and ally a clay loam with about 30% clay and classified as a fine, mixed new leaf production. Mollic Eutroboralf. Overstory is an uneven-aged ponderosa pine Key Words: Quercusganbe& aproutingpotential,Pinu~pondcrosu stand averaging 740 trees/ ha and 28 m*/ ha basal area. Gambel oak Gambel oak grows on about 3.8 million ha in the West as a forms a dense understory, mostly less than 1.5 m tall. prominent shrub in several vegetation zones (Harper et al. 1985). It Four, 3-ha units were established, one for each of the bum ranges primarily from the higher pinyon-juniper (Ph.-Juniperus) treatments (spring, summer, or fall bums, or unburned controls). Homogeneous plant and site conditions existed between units woodlands into the ponderosa pine zone, and occupies much of the treeless grassland in Utah, Colorado, Arizona, and New Mexico before burning. Season of burning was related more directly to oak between elevations of 1,800 and 2,700 m. growth stages than to specific calendar dates. The spring unit was burned shortly after leaf-out in mid-June, when oak leaves were Gambel oak is a deep-rooted, highly competitive sprouter l/2 to 3/4 full size. Summer bums took place in mid-August, a few known for its suppression of herbaceous species in open range- weeks after the start of the rainy season, when stem elongation and lands (Jefferies 1%5, Marquiss 1973) and suspected of restricting new leaf development became obvious. Fall burning was con- forage production and conifer regeneration in forests as well ducted in late October or early November when oaks had become (DeVelice et al. 1986). Since forested areas of the Southwest have dormant and most leaves had fallen. important multiple-use range values as well as timber values, oak management is desired here in addition to open rangelands. Gam- The fall unit was burned first (1977), and the spring and summer be1 oak control has been largely unsuccessful because of a massive units were burned the following year (1978). Gambel oak roots system of underground buds (Tiedemann et al. 1987) and lack of were sampled and analyzed for TNC concentration beginning the proper timing of control practices (Engle et al. 1983). year after the spring and summer bums (1979). Roots and rhi- Most effective control of sprouting hardwoods generally follows zomes, which were not differentiated, were collected from 5 plants top-killing during the growing season (Hough 1968, Berg and which represented the general oak condition throughout each of Plumb 1972). Root carbohydrates, which are the energy source for the 3 bum and 1 control treatment. Sampled plants were sprouts in resprouting, are utilized in the spring for leaf development and burned units and mature oaks in the controls. Roots were collected later for flowering or additional plant growth. If a plant is top- at intervals of approximately 2 weeks from mid-May, before oaks killed shortly after new leaf development, additional carbohy- had broken dormancy, until leaf fall in mid-October. At each drates are required for resprouting (Cook 1966). This depleted collection period, a sample of branches, buds, and leaves were energy source theoretically leads to a reduction in both sprouting taken so oak phenology could be related to carbohydrate concentra- and general growth (Garrison 1972). Under frequent and intense tion. defoliation, reduction in the total carbohydrate concentration will The fall unit was burned again in October, 1979, and the spring normally follow, causing restricted plant development and even and summer units were rebumed in June and August 1980, respec- death. Using this strategy, repeated prescribed burning during low tively, when oak phenology was the same as in the 1978 bums. Oak carbohydrate periods has effectively controlled understory hard- root sampling was resumed for all treatments just before the 1980 woods in the Southeast (Hodgkins 1958, Grano 1970) and in the spring bum. Samples for this period were collected monthly until Lake States (Buckman 1964). mid-October. The burning and sampling schedule is shown in Because the sprouting potential of shrubs such as Gambel oak Table 1. depends primarily on root energy reserves, knowledge of seasonal During the initial bums, fire intensities and fuel consumption carbohydrate variation is important in timing of control applica- were very similar for all 3 seasons. In the repeat bums, fire intensi- tion. Objectives of this study were: 1) to examine seasonal changes ties and fuel consumption were greatest on the spring plot, and in total nonstructural carbohydrate (TNC) concentrations in roots lowest on the fall plot. Even though the fire missed some oaks, all of Gambel oak in a ponderosa pine understory, 2) to compare oaks sampled for TNC concentrations had been top-killed by the fire treatments. Author is with the USDA, Forest Service, Intermountain Research Station, Inter- mountain Fire Sciencea Laboratory, P.O. Box 8089, Mioula, Montana 59807. On each sample date, the selected oak plants were excavated to a Manuscript accepted 28 February 1989. depth of 30 cm. Only roots smaller than 1 cm in diameter were

804 JOURNAL OF RANGE MANAGEMENT 42(0), November 1989 Table 1. Schedule of burning treatments and root sampling for TNC uulya~.

YCM Treatment 1977 1978 1979 1980 Spring Bum(June) Sample(May-Oct.) Sample(June-Ckt.) Burn(June) summer Bum(Aug.) Sample(May-Oct.) Sample(June-Oct.) Bura(Aug.) Fall Bum(Ckt.) Samplc(May-Oct.) Sample(June-Oct.) Bum(Oct.) Control Sample(May-Oct.) Sample(June-Oct.) collected, cut into 5 to 7 cm sections, placed into plastic bags, and Results and Discussion set on ice. As soon as possible, the roots were dried in a forced-air oven at 50° C for a minimum of 48 hours. Samples were then TNC Response to a Sin& Burn ground in preparation for laboratory analysis. During the first sampling season, root TNC concentrations of Concentrations of TNC were determined using procedures des- oak sprouts from the 3 burning season treatments were very similar cribed by Trlica et al. (1977). Total nonstructural carbohydrates and, therefore, were combined and contrasted to those of control include reducing and nonreducing sugars, starch, dextrins, and treatment mature oaks. Figure 1 showsa lower, but insignificantly fructosens, which are a readily available source of energy. Hemicel- different, carbohydrate level in the burned oaks throughout the lulose and cellulose are not included. Values reported are milli- growing season. In addition, seasonal TNC average f SE of 88.3 f grams of nonstructural carbohydrates per gram of dry material. 7.5 mg/g in the oaks in the bum plots was not significantly lower T-tests were used to determine if carbohydrates differed between than the 94.5 f 6.8 mg/g of the unburned oaks. Small sample size burned and unburned oaks at each sample date following the first probably contributed to lack of significance. Gambel oak presum- bum. In addition, seasonal TNC averages of treatments were ably has abundant energy reserves to draw on because of its mas- compared using a t-test based on a seasonal variance for the mean sive underground biomass, much of which serves as storage organs of each treatment. This seasonal variance was computed by sum- (Tiedemann et al. 1987). Therefore, a single top-killing treatment ming variances of the means at each sample date and dividing by was insufficient to cause marked reductions in TNC. Engle and number of sample dates. Independence of observations at different Bonham ( 1980) reported no TNC reduction after single or repeated sample dates were assumed. An analysis of variance was conducted herbicide application on Gambel oak. to determine if differences, after 2 bums, existed among the 4 bum Comparhon 01 TNC Levels in Other Gambol Oak Sitea treatments and among the 5 sample dates. Because interaction was Seasonal trends in Gambel oak root TNC have been reported in present between date and bum treatment, a one factor analysis of 2 other situations-in untreated, mature oak plants (Marquiss variance, followed by Duncan’s multiple range test, was used to 1969) and in 2-year old sprouts resulting from a roller chopping compare carbohydrate differences between treatments at each treatment (Engle 1978). Both of these studies were conducted in sample date. Seasonal TNC averages of treatments were compared oakdominated communities, about 32 km southeast of the current by estimating a separate seasonal variance for each bum treatment mean and pooling such information as in a one factor analysis of variance. A 5% level of significance was used unless otherwise specified...... OPEN-GROTR. MATURR OUR -OPRR-GROWRRPROUTR

r - UNBURNED OAKS

55.' ' ' I, I I I II I JUNE ’ 1 AUC ’ aEPT ‘GCT 5/1R B/26/14 6/30 T/l2 T/25 8/S B/24 S/S S/22 lo/12 6~‘w JULY” DATE DATE Fig. 2. Comparison of root TNC cycles of Gambel oak in 3 growth stages; Fig.1. Seasonal TNCcyclesin bumedandunbumedGatnbeloak. Pheno- open-grown mature oaks (Marquiss 1969). open-grown sprouts (&de logical stages are in&&d. 1978) and utirstory mature oaks and sprouts (current study).

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 606 study site at approximately the same elevation. A major difference between these study sites and the present study site is that the latter is dominated by ponderosa pine with oak as a somewhat sup- pressed understory. Figure 2 compares Gambel oak TNC fluctuations from the 3 sites. Because TNC trends of burned (sprouts) and unburned (mature plants) were similar in the current study, those data were combined for each sample date. Absolute TNC values from the 3 studies differed; Marquiss’ (1969) values were consistently highest, and Engle’s (1978) data had both higher and lower points that those from the current study, but the data were quite similar (Fig. 2). Because oaks sampled in the current study were shaded by pine overstory and were presumably photosynthesizing suboptimally, less carbohydrate production and storage could be expected compared to open-grown oaks (Kramer and Koxlowski 1979). Other possible reasons for absolute TNC differences between studies could be dissimilar sites, climates, clone ages, and growth rates. These factors were shown to be influential in clonal TNC differences in Populus tremuloides (Schier and Johnston 1971) and Quercus havardii (Boo and Pettit 1975). Perhaps more important than absolute TNC values are compar- 6/19 7/24 8/14 9/6 10/9 isons of the timing of depletion and accumulation cycles, which DATE generally indicate periods of low and high sprouting potential, respectively (Berg and Plumb 1972). Oak root carbohydrates mea- Fig. 3. Comparison of root TNC cycles of Gambel oak burned twice in sured in the 3 studies had simiir trends, especially those of Mar- spring, summer, fall, or unburned. For each date, different letters indi- quiss (1969) and the current study. However, low points and sub- cate significant d~ferences (PCO.05). sequent peaks sometimes appeared at different times of the grow- mature oaks, open-grown oak sprouts, and understory oaks, this ing season. This could be partially explained by Engle’s (1978) period appears when leaves are approximately 1/ 2 full size. In all 3 infrequent sampling and also by differences in climate from year- situations, midsummer TNC low point would be aproximated by to-year which affects plant phenology and resulting TNC trends. l/ 2 full size of new summer leaves. It should be noted that because For example, Sweeney (1975) reported that late snowmelt delayed vigorous sprouting follows initial mechanical or fire treatments Gambel oak leaf-out. This, in turn, will delay the first depletion (Marquiss 1971, Harrington 1985), maintenance programs are and accumulation cycle. Also, the second cycle is apparently asso- required and should be repeated at expected TNC low points. ciated with the rainy season, which varies yearly. Because root carbohydrate storage cycles are not necessarily TNC Response to Two Burns related to specific dates, storage trends are better compared using Two years after initial fire treatments, oaks were burned again oak phenology. In all 3 studies, early TNC depletion was observed during the 3 respective seasons. Unburned oaks had a significantly or assumed in May during bud break and leaf expansion (Fig. 2). higher mean carbohydrate concentration (76.1 f 6.2 mg/g) over Marquiss (1969) reported that this depression occurred in mature, the entire growing season than summer-burned oaks (63.8 f 5.0 open-grown oaks until full-leaf size. However, 2-year-old, open- mg/ g) but not spring- or fall-burned oaks (66.7 f 5.0 and 66.1 f grown sprouts (Engle 1978), as well a both sprouts and mature 4.5 mg/g, respectively). oaks in an understory, showed TNC accumulation before full-leaf Figure 3 shows varying root carbohydrate storage trends. The stage. Thii implies that carbohydrate storage exceeds export in treatment by date interaction, (P = 0.08), indicated that TNC roots and rhizomes before leaves are fully mature. In all 3 studies, trends varied in different treatments depending on date. Accumu- midsummer reduction apparently was the result of a rapid lation/depletion cycles similar to those previously illustrated in regrowth period. Rapid stem elongation and appearance of new Figures 1 and 2 are denoted by the control trace (Fig. 3). Because leaves were reported by Engle (1978) and observed in the current sampling began late (mid-June) and was infrequent, initial TNC study. In both cases, root carbohydrates had been declining for reduction had passed and exact dates of peaks and low points may about 1 month before observed stem elongation. In the current have been missed. The fall unit was burned the previous October, study, new leaves were about 1/ 2 full size at the August 24 TNC and TNC depletion continued until late July, as supplemental minimum. Photosynthesis and translocation from both old and reserves were used for growth of new sprouts from underground new leaves evidently led to carbohydrate gain in late summer. All buds. Spring burning occurred immediately after the first sample leaves were full size by the early fall peak in both Engle’s (1978) and date resulting in a 2-month period of root carbohydrate utilization the current studies. Root carbohydrates declined by the last sample as resprouting occurred. Following depletion, root TNC accumu- date as oak leaves were falling indicating either unobserved lation in fall- and spring-treated oaks for the last 2 to 3 months as growth, perhaps in the roots, or translocation to stems and buds. carbohydrates were produced in excess of growth requirements Marquiss (1969) did not report detailed oak phenology and proba- and approached levels of the controls (Fig. 3). Berg and Plumb bly missed the last TNC reduction by terminating root collection (1972) also reported that early and dormant season defoliation early in fall. allowed for greater regrowth and TNC accumulation than active These root carbohydrate/ phenological comparisons imply that season defoliation. treatments for oak control should be applied at specific growth Oak root carbohydrate cycles in the summer unit were compar- stages, depending on age and position in the stand. Generally, able to those in the controls until the bum following the mid- reduced vigor and, therefore, best control of sprouting species may August sampk. As resprouting commenced, carbohydrates decreased be expected from repeated mechanical or fire treatments applied and remained low because plants failed to grow photosynthetic when carbohydrate reserves are lowest (Berg and Plumb 1972). For maturity before dormancy. Because restoration of spent energy

505 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 reserves was not apparent, summer-burned oaks entered dor- DeVeUce, R.L., W.H. Motr, and F. Ronee, Jr. 19g6+classifiition of mancy with TNC leveis significantly below those of other treat- forest habitat types of northern New Mexico and southern Colorado. ments (Fig. 3). With additional utilization of carbohydrates for USDA Forest Serv. Rocky Mountain Forest and Range Exp. Sta. Gcn. respiration during winter and for spring leaf-out, a 9-monthdeple- Tech. Rep. RM-131. Fort Collins, Colo. Bngle, D.M., 1978. Root carbohydrate reserves and hcrbicidal control of tion could be expected before any accumulation could ouutr. Gambcl oak sprouts. Ph.D. t&is. Colorado State Univ., Fort Collins. Reduced oak growth may be expected from this sequence and was En&. D.M., and C.D. Bonbm. 19BO.Nonstructural carbohvdrates in observed as oak density, canopy cover, and frequency declined r

We’ve waited for it. Now it’s here:

The G/OSSWJ/ of terms used5 in range management.

Available for $5.00 (US) per copy from the Society for Range Management, 1839 York Street, Denver, CO 80208. Telephone: (303)355-7070.

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 507 Predicting peak standing crop on annual range using weather variables

MELVIN R. GEORGE, WILLIAM A. WILLIAMS, NEIL K. MCDOUGALD, W. JAMES CLAWSON, AND ALFRED H. MURPHY

AbShCt Wide yearly fluctuations in peak standing crop on California predictability of functions relating weather patterns and peak annuai-type range arc largely cxpiained by temperature and pre- standing crop by including degree-days, dry periods, evaporation, cipitation patterns. The objective of this study is to improve the season start dates, and lengths, as well as precipitation distribu- predictability of functions relating weather patterns and peak tion. This paper presents the results of multiple regression analysis standing crop by including degree-days, dry periods, evaporation, of these variables. season start dates, and lengths and precipitation as independent variables. Beak standing crop was regressed on these independent Materials and Methods variables for the University of California Hopiand Field Station The study was conducted using peak standing crop, daily precip- (I-IFS) and San Joaquin Experimental Range (SJER). Fall and itation, and temperature data from HFS and SJER. Hopland is winter precipitation, winter degree-days, and longest winter dry located in Mendocino County, California. Approximately 64 km period were related to peak standing crop at HFS (Rz=O.61). from the coast at 39 N Latin the Coastal Range at elevations of 200 Spring precipitation, growing season degree-drys, winter evrpora- to 400 m, San Joaquin Experimental Range is located at 37 N Lat tion, md winter and spring &rt dates were related to peak stand- in the lower central foothills of the Sierra Nevada at elevations of Ing crop at SJER (R*=.72). The relationship of peak standing crop 200 to 500 m approximately 40 km north of Fresno, California. to accumulated precipitation on 20 November using 33 years of The independent variables which are listed in Table I were data (&0.34) was weaker than previously reported for the first 16 analyzed by stepwise multiple regression (BMDP2R, Dixon 1983) years (r2~0.49). This study suggests that timely prediction of peak and all possible subsets regression (BMDP9R, Dixon 1983) to standing crop may be possible at HFS but more difficult at SJER. determine those that were closely associated with peak standing Key Words: heat units, evaporation, modeling, precipitation, for- crop. As part of the regression procedure, residual plots were age yield, degree-days examined to assure fulfiiment of statistical assumptions, and mul- ticollinearity was checked to confirm compliance with the variance Wide yearly fluctuations in peak standing crop on California inflation factor criterion (Williims et al. 1979). annual-type range are largely explained by temperature and precipitation patterns (Talbot et al. 1939; Bentley and Talbot 1948, 1951; Talbot and Biswell1942; Heady 1956,1958; Jonesetal. 1963; Table 1. lndepcndent varhbia analyzed in multiple ngresaion umlysa 01 Hooper and Heady 1970; Murphy 1970; Duncan and Woodman- peak standing crop from Univerdty 01Califorah Hopiand Field Station see 1975; Pitt and Heady 1978; George et al. 1985,1988). (I-IFS)and San Joquin Experimental Range (SJER). Murphy (1970) obtained a correlation coefficient (r) of 0.70 QKO.01) between observed forage production and the amount of Independent Variables rain received by November 20 at the University of California Hopland Field Station (HFS). Pitt and Heady (1978) showed fall Di Fail degree-days D2 Winter degree-days and winter minimum temperatures, fall and spring periods of little D3 Spring degree-days or no precipitation, and spring precipitation to be important D4 Growing season degree-days weather variables associated with peak standing crop at HFS Pi Fall precipitation (mm) (R*=O.90). P2 Winter precipitation (mm) Duncan and Woodmansee (1975) found a weak relationship P3 Spring precipitation (mm) between forage yield and September, October, November or P4 Growing season precipitation (mm) Ni Fall longest dry period (days) December precipitation (K0.28) at San Joaquin Experimental N2 Winter longest dry period (days) Range (SJER). They found forage yield more closely related to N3 Spring longest dry period (days) precipitation in April (R-0.41), November and April (R=0.53), or N4 Growing season longest dry period November, January, and April (R=O.62), concluding that precipi- Gi Dry days in month following germination tation must be well distributed throughout the growing season to Fi Fail date ensure abundant forage yield. F2 Winter date F3 Spring date Continued collection of weather and peak standing crop data at F4 Summer date HSF and SJER have provided additional data for regression anal- Ll Fail length (days) ysis purposes. The objective of this study was to improve upon the Winter length (days) :; Spring length (days) L4 Growing season length (days) Senior,second, and fourthauthora arc extensionspkalist, professor,a$ cxtcn- sion specialist,Agronomy and Range Sci. Dept. Umv. of Cahfornia, Davis 95616. Third author is a farm advisor, Madera County, Univ. of California Coo r Ext., SJER ONLY K Berkeley. 94720. Fiith author iasuperintendent (retired), Univ.of California opland El Fall evaporation (mm) 95449. -- The authors wish to acknowkdgc the many research assistants and others who E2 Winter evaporation (mm) faithfully collected weather and standing crop data at these research stations for many Spring evaporation (mm) Pm. z Growing season evaporation (mm) Manuscript accepted.23 March 1989.

508 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 Daily temperature and precipitation data were available for 3 1 hirtum All.) have minimum germination temperatures near a daily growing seasons (fall 19%spring 1985) at HFS and 48 growing average of 5O C. (Young et al. 1973,1975a, 197b). seasons SJER (fall 1936spring 1984). A third analysis of the SJER Using the methods of George et al. (1988). the timing and length data was conducted that included evaporation data (Table 1) from of the fall, winter, and spring seasons that comprise the annual- nearby Friant Dam (11 km south of SJER) with 30 growing type range growing season were defined as follows: seasons (fall 1954-spring 1984). In addition, Murphy’s (1970) 1. Fall is the period between germination and the onset of cold analysis of HFS precipitation and yield data (fall 1952-spring weather. Germination is defined to begin the day after 25 mm of 1968) was repeated using HFS data from fall 1952 through spring precipitation occurred in a l-week period. 1985 and SJER data from fall 1936spring 1984. 2. The first day of winter was defined as the first cold day Accumulated degree-day values were determined using the sine (degree-days <3) in the first 7day period that averaged less than function method described by Logan and Boyland (1983). Nega- 3degmedays per day. tive values were equated to zero. The base temperature used in this 3. The first day of spring begins on the first warm day.(degee study was 5OC. Temperatures at or near 5’ C have been used as the days > 3) in the first M-day period that averages more than 3 base temperature in degree-day calculations and in reports of degree-days per day. minimum temperatures for growth for cool-season plants such as 4. The dry season (summer) was defined to begin 2 weeks fol- ryegrass (L&urn perenne L.) (Chang 1968), alfalfa (Medicago lowing the last rainfall total of 25 mm in 1 week. sativa L.), timothy (Phleumpratensis L.) and red clover (l’btifolium The precipitation criteria for estimating the fall germination date prutense L.) (Bootsma 1984), Durum wheat (liiiticum durum Desf. are widely accepted, having been first proposed by Bentley and var. Produru) (Sherwood et al. 1978), cool season grasses and Talbot (1951). The criteria for the start of the winter and spring legumes (Fitzpatrick and Nix 1970), and annual range (Bentley and seasons were empirically derived by analyzing accumulated degree- Talbot 1951). Representative species of the California annual day curves for over 100 growing seasons at HFS, SJER, University range such as filaree @odium botrys (Gw.) Bertol.), soft chess of California’s Sierra Foothill Range Field Station (SFRFS) and brome (Bromus mol& L.), ripgut brome (II. rigidus Roth), foxtail 30 other weather stations maintained in the University of Califor- fescue ( Vulpia megalura (Nutt.) Rydb.), and rose clover (Trifolium nia’s Integrated Pest Management data base. No fall season was

T&h 2. Pmk stmdiq crop and aeeamuiated effective preeipttationtotals lor 1952-53 through 1984-85 growing semen at Hopland Field Station.

Peak standing Precipitation(mm) Crop Nov. Nov. Nov. Nov. Nov. Nov. Nov. Dec. Jan. Feb. Mar. Apr. May Total Year (kg/ ha) 1 5 10 15 20 25 30 31 31 28 31 30 31 Season

52-53 0 0 0 68 68 68 498 770 778 880 958 1000 1018 53-54 z 43 43 73 158 170 228 248 560 693 835 903 903 930 54-55 1008 18 18 73 158 165 165 165 338 428 480 493 590 590 590 55-56 1344 10 13 13 38 78 120 120 605 983 1183 1205 1243 1260 1263 56-57 1456 78 95 273 483 573 638 730 733 57-58 3136 2z 2ii 2z 273 2; 2z 2:: 423 648 1135 1325 1475 1490 1510 58-59 1904 0 10 35 35 35 78 380 583 625 638 638 638 59-60 5; 53 :i 53 53 53 88 233 465 608 645 700 700 60-61 2464 20 z 23 80 90 130 155 463 560 708 745 765 768 61-62 1456 10 173 203 :E 353 625 748 763 778 778 62-63 3248 233 2::: 4: 2:: 2z 393 503 625 975 1005 1008 63-64 3920 93 120 158 200 235 E ii 320 483 490 578 598 610 w-65 2912 75 88 195 225 225 270 675 888 920 945 1075 1075 1075 65-66 2352 0 0 145 190 ii 240 363 600 685 728 755 760 765 2912 ii 198 225 238 433 675 685 875 993 1003 1028 iE;: 2464 5: 5: 63 ii 93 93 105 263 510 628 743 758 775 778 rt .57+ .61+ .63*+ .63** .70++ .61** .63+* 24 .19 .09 .20 .25 24 .21 68-69 2912 65 65 108 118 130 138 483 845 1095 1138 1198 1198 1200 69-70 75 75 ii 98 98 98 415 930 1030 1095 1103 1108 1123 70-71 :zi 108 108 123 130 130 145 2z 538 675 685 843 878 905 905 71-72 2352 10 25 58 58 58 270 345 428 523 535 535 72-73 3808 130 1:: 230 260 260 2: 400 728 948 1E 1048 1050 1050 73-74 3472 115 115 :: 355 425 445 498 680 948 1053 1375 1435 1438 1438 74-75 3360 38 38 45 45 50 73 73 210 270 555 815 858 858 860 75-76 2912 103 103 115 130 140 140 140 185 198 265 310 385 385 385 76-77 0 0 0 63 63 63 63 135 203 278 280 335 335 77-78 :zi 110 110 110 113 113 190 190 3;; 818 1023 1153 1293 1298 1298 78-79 2688 0 0 45 265 473 570 628 655 655 79-80 3136 2i 2i 2: 2i 2i 488 665 938 990 1048 1063 1073 80-81 2912 :: :: 38 45 53 228 463 538 625 648 660 660 El-82 3136 153 153 160 2.z 3ii 460 488 715 905 1020 1208 1355 1355 1355 82-83 2912 143 143 153 153 223 243 333 565 818 1095 1523 1675 1690 1693 83-84 3360 43 43 123 230 325 373 373 703 718 838 915 958 963 973 84-85 2912 70 100 138 238 293 320 368 438 455 505 633 635 635 635 rtt .49** .51** .55+* .51** .58** .53+” .51** .34’ .27 .28 34 .36+ 34 34 ~Correlatjon codli+cnts for 195243 through 196647 gro+ng seasons (Murphy 1970). t!!;T;mn codicmnts for 195243 through 198445 grown.9 seasons. l*=p<0:01

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 509 considered to occur in years where cold weather began before the germination criteria occur. The criterion for the start of the dry RC!SUlt# season (summer) was determined by reviewing dry matter produc- Murphy (1970) obtained a correlation coefficient of 0.70(+0.49) tion curves for 99 growing seasons simulated by the Annual Grass- between forage yield and the amount of rain accumulated by 20 land Ecosystem Model (Pendleton et al. 1983) using daily weather November based on data for the growing seasons of 1952-53 data from HFS, SJER, and SFRFS. through 1%7-68. When Murphy’s analysis was repeated using Daily precipitation, degreedays, evaporation, and days without rainfall data from the 1952-53 growing season through the rainfall were summed within each season to derive the independent 1984-85 growing season (Table 2), the correlation coefficient variables (Table 1). decreased to 0.58 (rW.34). When SJER data were subjected to Herbage yield at peak standing crop was estimated annually in Murphy’s analysis, precipitation accumulated by 31 January late spring in caged exclosures by clipping twenty and forty 0.09-m2 through 3 1 May and total precipitation were more closely related quadrats at I-IFS and SJER, respectively. The monitored sites were to peak standing crop (rTo.47, pCO.01) than fall precipitation graxed by sheep at HFS and cattle at SJER. (Table 3).

Tabk 3. Peak stmdiag crop ld accumukted efWive predphatton totab for 1936-37 tbrou@~1983-M flowing aeuona at San Joaqdn Expuimattal U.

Fcakstalxhg Precipitation(nun) Crop Nov. Nov. Nov. Nov. Nov. Nov. Nov. Dec. Jan. Feb. Mar. Apr. May Season Year fkal ha) 1 5 10 15 20 25 30 31 31 28 31 30 31 Total

36-37 la97 53 55 35 55 55 55 55 la3 264 434 544 559 559 . 559 37-38 2981 1 1 1 a 13 14 14 172 308 489 749 al2 816 3a-39 1626 41 43 49 50 50 92 163 225 297 305 E 39-40 2168 53 53 it 53 53 ii 312 443 491 521 :: 522 4041 2304 ii 42 42 42 42 42 42 2ii 339 522 587 700 700 700 4142 27 28 28 28 31 31 259 336 389 436 513 537 537 42-43 Et 0 11 11 m 47 47 z 101 209 387 416 416 416 4344 la27 5 5 5 13 m 20 74 135 it 279 324 338 339 44-4s 2905 2: 54 a3 150 150 150 150 222 230 336 444 452 462 463 4546 2033 45 45 48 50 52 68 78 170 192 324 326 2033 39 39 53 107 137 137 231 243 iii 316 E iii z 1670 48 : 61 61 65 65 65 a9 91 121 fit 331 344 361 48-49 1558 6 11 11 11 11 68 114 162 264 264 301 301 49-50 3183 0 0 :: :: 38 38 a7 216 281 397 50-51 2794 62 62 62 a7 2E 224 248 331 405 468 ii :z :: 51-52. 2869 ia la la 46 72 72 210 356 3% 539 5% 597 :: 52-53 2191 a a 4 65 65 65 222 274 277 309 351 366 377 53-54 11 1’: 11 64 72 72 106 174 235 335 366 383 54-55 z!i 0 0 0 34 60 ii 123 258 305 315 z: 405 405 5536 2206 2 2 41 63 :; 394 508 532 532 591 649 649 56-57 2199 45 4: 45 45 it 45 45 70 130 la5 285 353 357 57-58 2395 40 60 60 76 76 76 76 I50 261 423 z 765 776 777 58-59 a79 9 11 15 15 28 107 219 2m 243 246 2191 91 9; ;: ii 100 157 315 377 z 378 zli 1515 23 38 ?i 91 ii ;: 122 143 1% E 263 279 2% 304 61-62 2208 3 3 3 27 a2 128 191 435 494 499 500 500 62-63 3422 40 4i 40 40 ii 43 73 157 256 360 479 492 495 63-64 2793 z 112 145 161 161 172 m3 203 289 310 311 64-65 4226 56 dli iti 126 126 126 131 262 336 348 ii: 488 488 488 65-66 10 10 52 95 145 146 221 248 282 315 317 z 3 3 :‘: 21 55 67 2m 322 344 z ii: 681 694 z 2574 13 39 39 ii 111 150 191 257 268 288 290 6849 3051 : :: :: 102 102 106 106 202 438 632 695 768 768 793 69-70 3389 la la 49 49 49 49 49 120 272 30a 392 3% 3% 422 70-71 3054 2 4 12 16 16 21 98 220 249 258 293 316 378 378 71-72 1975 5 5 5 44 44 52 173 la3 221 221 270 72-73 3395 31 39 122 1c 146 146 20a 297 470 566 E it 582 73-74 3191 ti 64 66 97 134 136 138 256 385 452 569 644 645 647 74-75 2932 79 a0 a0 a0 106 106 174 208 313 4m 466 468 468 75-76 2862 :: 76 90 100 104 239 280 308 309 314 76-77 1167 59 59 iz ii ii ii 88 110 139 151 174 175 212 228 77-78 5067 2 16 16 16 37 37 la3 507 659 777 777 777 78-79 3758 :i 47 47 102 102 143 E 382 485 492 4% 507 4433 ii 39 :9’ 39 62 64 69 113 288 411 474 498 ~~ 2855 14 14 14 14 14 54 162 194 2% 328 z :z al-a2 4380 :;: ii 29 68 a4 a4 98 159 292 361 525 613 613 82-83 4066 88 I10 122 162 165 193 459 5% al2 a89 t: 83-84 2043 26 ifi 26 72 72 132 162 E 294 331 378 391 402 rt .02 .07 .I6 .I7 .m .2l .21 .23 .47+* .47++ .53++ .57+* .56*+ .56*+ Wmrdation cocffiioicnts for 193647 through 1982-83gro~ seasons. .::g_g

510 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 Tabk 4. Peak standing crop, winter degreedays @2), hII( and winter Table 5. PcllL standing crop, f&U(Pl), winter (P2) and spring (P3) predpi- (P2) precipiwioll rDd longul dry perhI (N2) for 1954-55 to 1984-M tion (mm) for 1936-37 throagl~1983-M at San Joaquin E&per. Range. growing wasons at Hopknd Fkld Station.

Pl P2 P3 Peak D2 Pl P2 N2 standing Year (mm) Crop 311 188 Yqar (W W (mm) (days) 36-37 1897 30 37-38 2981 58 126 484 5445 1008 284 95 426 17 38-39 1626 26 221 38 55-56 1344 166 0 1172 39-40 2168 36 178 264 56-57 1456 422 0 556 :: 4041 2304 3 0 696 57-58 3136 330 185 869 12 4142 2281 0 361 118 58-59 1904 362 0 572 24 4243 2718 179 159 59-60 232 2 473 12 4344 1827 : 259 0 60-61 z 137 15 268 19 4445 2905 113 295 8 61-62 1456 2% 717 15 4546 2033 0 164 84 62-63 3248 312 37; 232 27 46-47 2033 12 205 37 63-64 3920 274 156 328 15 47-48 1670 24 166 75 64-65 2912 275 143 707 8 48-49 1558 0 151 102 65-66 2352 135 503 11 49-50 3183 1 241 114 66-67 2912 E 158 14 SO-51 2794 287 159 0 67-68 152 64 z 19 51-52 2869 0 521 57 68-69 z 199 91 1021 6 52-53 2191 0 232 0 69-70 1904 153 55 375 20 53-54 2894 0 258 13 70-71 2128 2% 115 613 54-55 2443 29 245 11 71-72 2352 1 311 3 55-56 22% 0 531 116 72-73 3808 3;; 782 9 56-57 2199 0 86 107 73-74 3472 290 752 10 57-58 2395 36 379 322 74-75 3360 355 765 15 58-59 879 0 208 75-76 2912 281 191 23 59-60 2191 178 58 76-77 107 30 83 35 60-61 1515 3; 128 29 77-78 132 71 196 13 61-62 2208 0 478 19 78-79 2688 207 507 17 62-63 3422 70 379 79-80 3136 200 23; 406 22 63-64 2793 8: 113 80-81 2912 214 13 416 10 64-65 4226 39 253 1: 81-82 3136 404 420 810 15 65-66 2444 79 128 20 82-83 2912 162 85 490 11 66-67 3286 44 259 371 83-84 3360 229 18 67-68 2574 0 111 118 84-85 2912 95 10 68-69 3051 73 577 99 69-70 3389 0 223 125 Regression ccefflcient 2.2 2.76 -1.46 -56.7 70-71 3054 54 177 Standardized coefficient .33 .47 -.54 -.62 71-72 1975 0 187 ; R2 incrementtt .06 .18 .14 .22 72-73 3395 115 394 41 P 0.58 .002 .006 .OOl 73-74 3191 57 353 191 74-75 2932 92 294 18 R2 = 0.61, p

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 511 Table 6. Puk rtrnding crop, wtnter (F2) and rpring (F3) stmting date, dates tend to increase peak standing crop. The evidence shows that annual porrial: seasondegreedays (Dl), spring (P3) precipitation, and fall weather has a smaller influence on peak standing crop at SJER winter evapor8tion (E2) 195445 to l!M3-84 growing seuom at San than it does at HFS. The reason seems to be that the start of the fall Jorquh Experiment Range. season, which is dependent on fall rains, is earlier and more dependable at HFS than at SJER (George et al. 1988). Dependability of Peak precipitation, as Duncan and Woodmansee (1975) indicated, is Standina particularly important. At HFS spring rains are reasonably Crop P3 E2 dependable, while fluctuation of fall and winter precipitation con- (43/W Year n(day+)ED4 (mm) tribute greatly to the between-year variation in peak standing crop. 54-55 2443 80 1% 201 11 132 Precipitation is generally more variable at SJER than at HFS; 55-56 22% 72 191 74s 116 160 however, because most production occurs in the spring, between- s6-57 2199 9s 160 413 107 62 year variation in spring variables have a greater inthtence on SJER 57-58 2395 93 1% 1459 322 101 peak standing crop. 58-59 a79 73 I77 517 a2 Mild winters of adequate precipitation appear to be positively 59-60 2191 9s I153 si 99 187 associated with increased peak standing crop at HFS and SJER for 60-61 1SlS 7s 181 337 29 147 61-62 2208 76 19s 728 19 154 the following reasons. The relationship between winter evapora- 62-63 3422 90 IS2 914 379 so tion and peak standing crop is positive at SJER, suggesting that 63-64 2793 77 197 625 high evaporation is indicative of high solar intensity resulting in 64-65 4226 71 170 869 *ii !Z increased primary productivity. Winter precipitation is negatively 65-66 2444 171 563 20 106 associated with peak standing crop at HFS, suggesting that exces- 3286 66-67 ii 174 1286 371 97 sive precipitation and the associated cold, cloudy weather condi- 67-68 2574 a5 156 557 118 79 68-69 3051 82 197 1SlO 99 149 tion suppress production (Bentley and Talbot 1951). 69-70 3389 92 156 607 12s 78 Pitt and Heady (1978) discussed the utility of multiple regression 70-71 3054 99 1% 438 7 151 as a predictive tool for stocking rate decisions. Stocking rate 71-72 197s 73 419 9s decisions are generally made early in the growing season, requiring 72-73 339s 79 iii 1051 4: 185 a predictive tool based on fall and winter weather variables. Pitt 73-74 3191 77 178 914 191 127 and Heady’s (1978) regression model for June standing crop is a 74-75 2932 9s 222 18 228 75-76 2862 77 191 ;!z 0 171 strong relationship (Rr4l.90) but requires spring weather varia- 76-77 1167 a3 164 177 0 55 bles. Murphy’s regression model based on precipitation to 20 77-78 SO67 99 157 1738 433 76 November is simple, but the strength of the relationship has 78-79 3758 71 la4 a22 147 declined as additional years of data were used to derive the model. 79-80 4433 153 1181 2: This study proposes a regression model for HFS and similar envir- 80-81 2855 ii 162 770 147 ;: onments based on fall and winter weather parameters. A suitable 81-82 4380 84 166 1352 321 76 82-83 4066 68 Is4 1939 431 92 scheme is to use Murphy’s model: 83-84 2043 79 148 832 97 53 Y = 1958 + 4.1 X (1) Regression where X is accumulated precipitation (mm) to 20 November for Coefticient 22.6 -50.7 .s2 4.9 29.8 Standardiztd early prediction. Prediction based on Murphy’s relationship could Co&cient 24 -1.09 .16 .67 1.36 be adjusted at the end of winter (usually mid to late February) R2 incrementtt .04 .I8 .02 .27 .27 based on the relationship reported in this study: P .071 .OOl 208 .OO .Oo Y = 3640 + 2.8 Pl - 1.5 P2 + 2.2 D2 - 57 N2 (RW.61) (2) R2 = 0.72,p

512 JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 the same sites and improved monitoring of drought within the Hooper, J.F., and H.F. Heady. 1978. An economic analysis of optimum growing season should strengthen these relationships and identify rates of grazing in the Caliiomia aMd tytx-- massland._ J. Range Man- additional useful relationships. age. 23:%7-31 i. Jones, MS., C.M. McKell, and S.S. Wtoaos. 1963. Effect of soil tempera- Literature Cited turn and nitrogen fertilization on the growth of soft chess (Bromur mollis) at two ievations. Agron. J. 55:4&. Bentley,J.R., and M.W. Talhot. 1948. Annual-plant vegetation of the Logan, S.H., and P.B. Boyland. 1983. Calculating heat units via a sine California foothills as related to range management. Ecology 29:72-79. function. J. Amer. Sot. Hart. Sci. BR977-980. &Bentley,J.R., and M.W. Talbot. 1951. Efficient use of annual plants on Murphy, A.H. 1970. Predicted forage yield based on fall precipitation in cattle ranges in the California foothills. USDA Circ. 870. California annual grasslands. J. Range Manage. 23~363-365. Bootsma, A. 1984. Forage crop maturity zonation in the Atlantic Region Pendbton, D.F., J.W. Menke, W.A. Willbma, and R.G. Woodmansee. using growing dcgnwdays. Can. J. Plant Sci. 64~329-338. 1983. Annual grassland ecosysmm model. Hilgardia 51:144. Chang, Jen-Hu. 1968. Climate and agriculture, an ecological survey. Pitt, M.D., nnd H.F. Hudy. 1978. Responses of annual vegetation to Aldine Publishing Co., Chicago. temperature and rainfall patterns in California. Ecology. 59336-350. Dixon, W.J. (ed.). 1983. BMDP Statistical Software. Univ. of Calif. Press. Sherwood, B.I., R.D. Jackwn, and R.J. Regioato. 1978. Extending the Berkeley, Calif. “degreeday” concept to plant phenological development to include Duncan, D.A., and R.G. Woodmamee. 1975. Forecasting forage yield water stress effects. Ecology 59431433. from precipitation in California’s annual rangeland. J. Range Manage.. Talhot, M.W., H.H. Bbwell, and A.L. Hormay. 1939. Fluctuations in the 28327329. annual vegetation of California. Ecology 20~394402. Fitapatrick, E.A., and H.A. Nls. 1970. The climatic factor in Australian Talhot, M.W., and H.H. Bbwell. 1942. The San Joaquin Experimental grassland ecology. p. 8-11. In: Moore, MR. 1970. Australian Grass- Range: The forage crop and its management. Calif. Agr. Exp. Sta. Bull, lands. Australian Nat. Univ. Press, Canberra. 663. George, MR., W.J. Clawson, J. Menke, and J. Butolome. 19115.Annual Williams, W.A., C.O. Quabet, and S. Geng. 1979. Ridge regression for range forage productivity. Rangelands 7:17-19. extracting soybean yield factors. Crop Science 19869-873. George, M.R., K. Obon, and J.W. Menke. 1988. Range weather: a compar- Young, J.A., R.A. Evrg and B.L. Kay. 1973. Temperature requirements ison at three California range research stations. Calif. Agr. 4230-32. for seed germination in an annual-type. rangeland community. Agron. J. George, Melvin R., Chubs A. Raguse, W. James Claweon, Charlea B. 65656-659. Wibon, Robert L. Willoughby, Neil K. McDoupld, Don A. Duncan Young, J.A., R.A. Evans, and B.L. Kay. 19758. Dispersal and germination and Alfred I-I. Murphy. 1988. Correlation of degrecdays with annual dynamics of broadleaf filaree, Erodium botrys (Cav.) Bertol. Agron. J. herbage yields and livestock gains. J. Range Manage. 41:193-197. 6754-57. Heady, H.F. 19%. Evaluation and measurement of the California annual Young, J.A., R.A. Evans, and B.L. Kay. 197%. Germination of Italian type. J. Range Manage. 925-27. ryegrass seeds. Agron. J. 67~386389. Heady, H.F. 1958. Vegetational changes in the California annual type. Ecology 39402415.

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

FIELD CROP ABSTRACTS (annual field crops)

for coverage of the world literature on agricultural research

For specimen copies of these computer-produced monthly Commonwealth Bureau of journals and for lists of Pastures and Field Crops annotated bibliographies and Hurley, Maidenhead, other publications write to: Berks SL6 5LR, UK

JOURNAL OF RANGE MANAGEMENT 42(6), November 1969 513 Technical Notes

Direct effect of parasitism by Dinarmus acutus Thomson on seed predation by Acanthoscelides perforatus (Horn) in Canada milk-vetch

A. BOE, B. MCDANIEL, AND K. ROBBINS

AbOhCt Canada milk-vet& (Ash@us canudenh L.) ir a widespread Materials and Methods North American legume considered to be good forage in some In July 1987,4 inflorescences with mature pods were collected regions but potentially dangerour to livestock when it containa from each of 10 genotypes of Canada milk-vetch in a 250-space- high levels of 3-nitropropionlc acid. Larvae of the seed predator plant nursery establsihed on a Vienna loam, nearly level [fine- Acunthosceikks pqforutus (Horn) (Coleoptera:Brucbidw) oc- loamy, mixed Udic Haploborolls] at Brookings, S. Dak., in May curred in 77% of the mature pods from 10 genotypes of the legume 1985. The seed source for the nursery was a vigorous population growing in 8 nursery at Brook&, S. Dak., in autumn 1987. from a native prairie approximately 10 km northeast of Brookings, Dhuvnus u&us Thomson (Hymenoptera: Pteromalidae) par&- S. Dak. Inflorescences were placed in covered glass jars and main- ized 48% of the A. perforatus larvae and reduced numben of seeds tained at room temperature. Adults of the hymenopterous parasi- consumed by A. pcrjorotwhrvae by 23%. This study identifiedD. toid D. acutus began to emerge in large numbers from the pods in acutus aa a puaaitoid of A. pcrforroiurand indicated pursitoidn August 1987, but only a few adults of A. perforotus emerged during nuy play an import8nt role in recruitment of n8tive legumes. fall 1987. In January 1988,30 random pods from each inflores- Key Words: Astragalus canadensis L., Dinarmus acutus, Acan- cence of each genotype were carefully split along their septa, and thosc~ pafoatus, parasitoid, seed predation the 2 locules of each pod were examined under a dissecting micro- Canada milk-vetch (Astragalus canadensis L.) is the New scope at 10 to 30 magnifications. World’s most widely dispersed astragalus, ranging from the Gulf Frequencies of pod infestations by A. perforatus and parasitism Coast in eastern Texas to the Appalachians and lower St. Law- of A. perforatus by D. acutus were determined for 20 pods per rence River, west to the southern Rocky Mountains and the Great inflorescence (80 pods per genotype). Unparasitixed bruchid- infested locules contained a larva, pupa, adult or exit hole, and and Columbia Basins (Bameby 1964). It is perennial, rhixomatous, and usually quite robust. This legume is considered good forage in empty pupal case of A. perforatus. Parasitized bruchid-infested some regions (Hermann 1966), but can be toxic to livestock when it locules contained a larva, pupa, adult, or exit hole and exuviae of contains high levels of 3-nitroproprionic acid (James et al. 1980). D. acutus, and the mummitied remains of the parasitized bruchid Bruchid beetles (Coleoptera:Bruchidae) are important seed larva. The effect of parasitism by D. acutus on seed predation by A. predators of many North American legumes (Johnson 1970, Green pe@rotus was determined by counting intact seeds and testas of and Pahnbald 1975, Boe and Wynia 1985). The bruchid Acan- bruchiddestroyed seeds in unparasitixed and parasitized bruchid- thoscelidesperforatuss (Horn) has been reared from seeds of Can- ada milk-vetch (Johnson 1970), but its effects on seed production infested locules of a sample of 10 pods per infIorescence for 6 genotypes (40 pods per genotype). Analysis of variance was con- are unknown. ducted on inflorescence means for percentage and number of intact Lists of parasitoids of North American Bruchidae have been seeds in locules infested by parasitized and unparasitized larvae of compiled (Cushman 1911, Johnson 1970, Center and Johnson 1975), but little is known about effects of parasitoids on population A. perforatus. dynamics of North American legumes attacked by bruchid beetles. Results and Discussion Canada milk-vetch has forage and soil conservation potential, but evaluation for these purposes is dependent on adequate seed Seventy-seven percent of the pods of the 10 genotypes had at supplies of adapted ecotypes. Studies are needed to determine least 1 locule infested by A. perforotus and the overall mean genetic, physical environment, and biotic influences on seed pro- percentage seed loss to bruchid predation was 44% (Table 1). These duction. Objectives of this study were to quantify the effect of data showed A. perforatuscan seriously reduce viable seed produc- predation by A. perforatus on seed production of C. milk-vetch tion in Canada milk-vetch. Green and Pahnbald (1975) found seed and determine parasitism rate of A. perforatus by the small wasp predation by AcanthosceIidesfratercul~ (Horn) exceeded 60$&in Dinarmus acutus Thomson (Hymenoptera: Pteromalidae) in a natural populations of Astragalus cibarius Sheld. and Astragalus field nursery at Brookings, South Dakota. utahensis (Tom) T. and G., and postulated seed predation influenced population dynamics of these legumes. Authors are associate profewor, professor, and agricultural research technician, ‘vely, Plant Science Dep., South Dakota State Univ., Brookinga 57007. Thirty-six percent of the pods contained insects in both locules, “#iF authors extend their gratitude to Dr. C.D. Johmon, Northern Arizona Univ., but multiple occupancies in locules were infrequent (only 3 locules for providing identification of Acanthoscelide~ptyforatus;to Dr. E.E. Grirrcll, SF- tematic Entomology Laboratory, USDA, U.S. National Muacum, Washin on, D.C., contained 2 normal A. perforatus larvae). Since seed predation for providing idcntifmtion of Df-us acutw; and to Dr. R. Kibckhe? cr. Dr. G. levels were high, cannibalism (Nelson and Johnson 1983) and/ or Larson, and R. Bortncm for critically reviewing the mamucript. Contribution from the S. Dak. State Univ. Agric. Exp. %a., Brookings, SD 57007. competition (Boe et al. 1988) for food resources may have influ- Journal Paper No. 2329. Manuscnpt accepted 27 April 1989.

514 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 Table 1. Pod infestationand meedpred8tlon by the bruchid batk ha- bruchid beetles, and parasitic Hymenoptera is critical to under- thc&&.rpe+wu in a mmery of cmula adIk-vetckat Brook@, standing the ecology and population dynamics of native rangeland South Dakota and the infhncc of purultim by Dburmua mutm on legumes. seedpredationbytbebeetk. Seed predation levels in this space-plant nursery were comparable to those described for native stands of several other legumes. Seed predation in infested” Legume seed predation by bruchid beetles may exceed 50% (Green locules IllfCStd Overall ~ccd* Parasitized3 and Palmbald 1975, Boe and Wynia 1985) and occasionally reach Pods predation beetles Unparasitized Parasitized nearly 100% (Rogers and Garrison 1975) in natural plant commun- m ities, but considerable variation occurs among years and 10cati0n~ 77 f 144 44f 16 48*24 85 f 7 62 f 17 (Baskin and Baskin 1977). Seed predation in 37 populations of American licorice (Glycyrrhiza kpidora Pursh) from North and ‘Muns I’ *ksantly diienllt at the 0.05 level. ~McaM0 ?+ 24 mflorcsccnccs;10 pods per inflomceacc. South Dakota ranged from 7 to 7 1% with an overall mean of 41 f ‘Meam of 40 intlorcscewr; 20 pods per infloreacencc. 2% (Boe and Wynia 1985). The impact of bruchid beetles on ‘Standarddeviation. long-term seed production and the efficacy of hymenopterous cnced larval survival. parasitoids for reducing seed losses in cultivated nurseries and seed Fortycight percent of A. perforarus larvae were parasitized by increase fields of native legumes can only be determined after more D. acutus. D. acutus is found throughout the continental U.S. It extensive evaluation. has been reared from Bruchidius ater (Marsham) and &uchus brachiazis Fahr. (Krombein et al. 1979) and recorded from seed lots Literature Cited of legumes containing A. fraferculus and A. griseolus (Fall) (John- Buneby, R.C. 1964. Atlas of North American Asrragulus. Memoirs of the son 1970), but A. perforatus is a new host record. New York Botanical Garden Vol. 13. The New York Botanical Garden. Seeds of Canada milk-vetch are small (approximately 1.5 Bask&~, J.M., and C.C. Bukin. 1977. Predation of Cassia marilandica mg/seed), and individual larvae of A. perforurus consumed the seeds by Sennius ubbreviutw (Co1eoptera:Bruchidae). Bull. Torr. Bot. Club. 104~614. cotyledons, plumules, radicles, and parts of testas of 2 to 12 seeds. Boe, A., and R. Wynh. 1985. Seed predation. seedling emergence, and Unparasitized A. perforatus larvae destroyed 85% of the seeds in rhkomchamct&sticsofAmcrkanlicoricc. J. RangcManagc.38:~. locules they infested compared to 62% when parasitized by D. Boe, A.B., B. MCDmiel, and K. Robbim. 1988. Patterns of American act&us (Table 1). Uninfested locules contained 6.9 f 0.2 intact licorice seed predation by Acanrhoscelides ounrohu (Horn) (Coleoptcra:- seeds, and locules infested by parasitized beetles contained signifi- Bmchidae) in South Dakota. J. Range Manage. 41342-345. cantly (X0.05) more intact seeds (2.6 f 0.2) than locules (1.1 f Center, T.D., and C.D. Johnson. 1975. Host plants and parasites of some Arizona seed-feeding insects. Ann. Entomol. Sot. Amer. 69: 195-201. 0.2) infested by unparasitized beetles. D. acufus had a direct influ- Cheeke,P.R.,and L.R. Sbull. 1985. Natural toxicants in feeds and poison- ence on the present generation seed crop of Canada milk-vetch by ous plants. Avi Pub. Co., Inc. Westport, Corm. reducing the number of seeds consumed by larvae of A. perforatus. Cusbman, R.A. 1911. Notes on the host plants and parasites of some North Since normal larvae of the cabbage white butterfly (Pieris rapae American Bruchiie. J. Fcon. Entomol. 4489-509. (L.)) consumed more food than those parasitized by the solitary Green, T.W., and I.G. PaImbald. 1975. Effects of insect seed predators on braconid parasitoid Apanteks rubecula Marsh (Rahman 1970), we Astragalus cibarius and Astragalus utahensb (L-quminosae). Ecology 56: 1435-1440. speculate that parasitized bruchid larvae eat less than unparasit- Hemma, FJ. 1966. Notes on western range forbs: Cmciferae through ized larvae. Compositae. USDA Agr. Handbk. 293. In large-seeded legumes in which 1 or more bruchid larvae feed James, L.F., W J. Hartley, M.C. Willims, and K.R. Van Kampea. 1980. and develop within a single seed (Johnson 1981), germinability Field and experimental studies in cattle and sheep poisoned by nitro- and/ or vigor are usually reduced due to extensive damage to the bearing Asrragalus or their toxins. Amer. J. Vet. Rcs. 41:377-382. embryo (Boe et al. 1988). Although parasitism benefits large- Johnson, C.D. 1970. Biosystematics of the Arizona, California, and Oregon species of the &d beetle genus Acanthoscelides Schilsky seeded legumes by reducing the number of beetles that survive to IColeontera:Bmchidae). Univ. Calif. Pub. Entomol. 59~1-116. reproduce, it will not increase the number of viable propagules Jo~~~II;C.D. 1981. R&ions of Acunthoscelides with their plant hosts. from the present seed crop unless bruchid larvae are destroyed In: V. L.abeyrie(cd.), The ecology of bruchids attacking kgumcs (pulses). before they inflict severe seed damage. Dr. W. Junk Pub., The Hague. The importance of parasitic Hymenoptera as biological controls Krombein, K.V., P.D. Hurd, D.R. Smith, B.D. Burks, et. d. 1979. Catalog of insect pests of cultivated crops is well-known, but their influence of Hymenoptera in America north of Mexico. Washington: Smithsonian Institution Press. Vol. l:Symphyta and Apocrita (Parasitica). xvi + on recruitment and population dynamics of native legumes of l-1198. North American rangelands and prairies has not been investigated. Nelson, D.M., md C.D. Jobmoo. 1983. Selenium in seeds of Asrrugulur Many native legumes are important forage and soil conservation (Lcguminosae) and its effects on host preferences of bruchid beetles. J. plants, while others are undesirable because of their toxicity to Kans. Entomol. Sot. 56:267-272. livestock (Cheeke and Shull1985). This study identifibd D. acutus R&mm, M. 1970. Effect of parasitism on food consumption of Pieris rapae krvae. J. Econ. Entomol. 63:82&821. as a parasitoid of A. perforatus and revealed its direct influence on Rogers, C.E., snd J.C. Gaminon. 1975. Seed destruction in indigobush the reproductive capacity of Canada milk-vetch. These findings Amorpha by a seed beetle. J. Range Manage. 28~241442. suggest that elucidation of the interactions among native legumes,

JOURNAL OF RANGE MANAGEMENT 42(6). November 1989 515 Book Reviews

Viable Populations for Conservation. 1987. Michael E. “Introduction” (Chapter 1) and ‘Where do we go from here?” Soule (Ed.) Cambridge University Press, 32 East 57th St., (Chapter IO), plus Shaffer’s (Chapter 5), Maguire’s, et al. (Chapter New York, N.Y. 10022. $44.50 (cloth), $16.95 (paper). 8) and Salwasser’s et al. (Chapter 9) will provide a good perspective but you11 miss considerable modelling.--David K. Mann, Salt Viable Populations for Conservation is a collection of eight Lake City, Utah. papers bookended with an introduction and a look into the future by the editor, Michael E. Soule. The.question addressed is what are Mammals of Arizona. By Donald F. Hoffmeister. 1986. the minimum criteria for long-term persistence and adaptation of a species or population in a given place? To answer the question the University of Arizona Press, 1615 E. Speedway, Tucson, book focuses on the viable population, or the population that AZ 85719. 602 p. $49.95 cloth. “maintains its vigor and its potential for evolutionary adaption”. This weighty volume is the result of more than 35 years of System viability is not considered and the time frame is long term. research by the author on mammals in Arizona. During those 35 Although many numbers are proposed for minimum viable popu- years, Hoffmeister and colleagues and students working under his lations, we are cautioned that no single number has universal direction conducted field studies at nearly 1,000 localities in Ariz- validity. ona and examined more than 42,008 specimens of mammals from Goodman begins the minimum viable population discussion by the state. Readers of thii book thus will sense the warmth and examining the classical birth-and-death process model under aroma of many campfues. demographic and environmental variation, concluding that envi- Mammals of Arizona evidently was written primarily for stu- ronmental variation is more critical to species extinction than dents of mammalogy and wildlife biologists, but other kinds of demographic variation. biologists and avid naturalists will enjoy reading the life histories of Bolovsky continues to examine the classical birth-anddeath Arizona mammals. The Introduction contains a summary of fossil model and focuses on mammals. He concludes that the model mammals known from Arizona, a history of collecting and collec- catches the important aspects of the extinction process but over- tors of mammals in the state, and an explanation of information simplifies the dynamics and underestimates extinction probabilities. presented in the accounts of species. Chapter 1 summarizes hunt- Next up Ewens et al. build on the previous models, but introduce ing and trapping regulations in Arizona and includes brief genetics into the extinction process and examine what would accounts of game species, furbearers, predator problems, and rare happen in face of a catastrophe. These authors see a great need for and endangered species in the state. Range and wildlife officials in theory that brings together the genetic and demographic views of Arizona should read this chapter and consider the author’s com- the extinction process. They also point out that the catastrophe ments and suggestions. Chapter 2, a review of the four diseases model limits the applicability of the initial population size. Then (rabies, plague, tularemia, and spotted fever) in Arizona that are they give us a kemal of wisdom: It is quite difficult to mathemati- transmitted to man either directly or indirectly by mammals, is cally analyze any model that reflects biological reality at all accu- unique; we know of no other book on the mammals of a state that rately. Nonmathematicians don’t appreciate the liits of mathe- reviews this topic of concern to persons who work with wildlife. matical theory. Chapter 3 is a disappointment. It begins with a much too brief, In a fine paper by Shaffer he brings the ideas of the previous intuitive review of ways in which man’s activities in Arizona have three papers together and introduces another actor, chance. Popu- affected distributions and numbers of native mammals. Next are lations may become extinct just due to chance. He believes that in extremely superficial accounts of the topography, climate, soils, single, undivided, isolated populations “natural catastrophes set and vegetation in the state. There are no maps, and only one the limit to persistence”. reference to the published literature is provided for readers who Lande and Barrowclough focus on the genetic view of the extinc- wish additional information. Given Hoffmeister’s contention (p. tion process. “Genetic variation is necessary for evolution to occur xiii) that “The diversity of topography in Arizona. . .ia dlccted in and hence for adaptation to changes in the environment.” The the diversity of mammals...,” we expected more than he provided. effective size of the population will have to be measured or the The chapter ends with a somewhat informative account of how genetic variation monitored directly in problem populations. mammals have adapted to the adverse conditions that exist in the Although mentioned in some of the previous papers, Gilpin arid, hot deserts of Arizona. Chapter 4, on biogeography compen- introduces the concept of metapopulation. The idea is that natural sates for some of the weakeneases in chapter 3 and will be of interest populations occur in patches because of the patchiness of habitats. to field biologists and resource managers of all kinds. It includes These population patches make up the metapopulation. The adequate descriptions, detailed maps, and excellent photographs important aspect of the metapopulation in the extinction process is of the vegetative communities in Arizona. These are followed by a the flow of individuals between patches. detailed analysis of the geographic relationships of these communi- The case study on Sumatran rhino by Maguire et al. uses deci- ties to the mammalian fauna of the state and by a series of profusely sion analysis to choose between management alternatives for the illustrated faunal analyses. The latter, unfortunately, are not ade- endangered species. Decision analysis looks to be a good tool but quately explained or interpreted, and this will limit their usefulness. even the best tools can be made inoperable under politics. Chapter 5, accounts of species and subspecies of mammals in Salwasser, Schonewald-Cox and Baker conclude the eight core Arizona (pp. 37-564), consists of technical information on the 138 papers with one on interagency cooperation. They urge agencies species of mammals that are native to Arizona and 7 additional consider networks of land to accommodate minimum viable species of mammals that have been introduced to the state. A total populations. of 211 subspecies are recognized. Accounts include keys, detailed Who in range science or management should read Viable Popu- descriptions, tables of measurements and illustrations to facilitate lations for Conservation? If you want to know the current extinc- identification, well-prepared and useful distribution maps, infor- tion process theory and practice, then you11 want to study the mation on natural history and habitats (often with excellent pho- whole book several times. If the minimum viable population con- tographs of habitats and/ or the mammals themselves), and lists of cept is an area of interest but not your main thrust, then Soule’s

516 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 specimens examined and additional records from Arizona. It is resource which will no doubt remain under siege for the foreseeable especially noteworthy that many accounts contain original data future.-David L.. Scamecchio. Washington State University, not previously published and apparently resulting from analyses Pulhnan. conducted expressly for this book. The manuscript was completed in October, 1983, and the accounts do not reflect information published after that time. To Feed the Earth: Agro-Ecologyfor Sustainable DeveC The accounts of species are followed by a series of appendices opment. By Michael Dover and Lee M. Talbot. 1987. that review harvest data for game species, give common and scien- World Resources Institute, 1735 New York Avenue, NW, tific names of plants mentioned in text, and summarize type locali- Washington, D.C. 20006 (88 p.) $10. ties of mammals in Arizona. The bibliography lists nearly 600 This book is basicallya policy orientedstudy, though ecological references related directly or indirectly to mammals in Arizona. concepts and implications for agriculturearc well set forth. That Hoffmeister? long-awaited Mummuls ojArizonu is destined to there is a need for a change in agri-policy, particularly in the be the classic technical treatise on the mammalian fauna of that developing nations is evidenced in the statisticsgiven in the opcn- state for decades to come. It contains few obvious errors although ing chapter. For example, the fact that the world is awash with specialists will find details with which to quarral. The price and grain has not changed the fact that famine is more widespreadin technical nature of the book will discourage many potential pur- 1985 than in 1975. The United Nations Food and Agriculture chasers, but all persons who are interested in the fauna and flora of Organizations projects that by the end of the century 64 nations Arizona should be aware of Hoffmeister’s Mammals of Arizom will not be able to meet their food needs, and the resource base is and should at least have access to a library copy.-Jerry R. Choate shrinking. It is pointed out that agriculture is the mainstay of and Robert A. Nicholson, Fort Hays State University, Hays, economic development; so it follows that if agriculture can be Kansas. conducted on a sustainable basis so will economic development. They propose that the discipline of ecology provides the basis for a Rangelands: A Resource Under Siege. Proceedings of the systems approach to agricultural development that can help pro- Second International Rangeland Congress. 1986. Edited vide the solutions to the dual problem of productivity and by P.J. Joss, P.W. Lynch and O.B. Williams. Cambridge sustainability. University Press, Cambridge, CB2 IRP, UK. 634 p. Chapter Two addresses some of the problems of agriculture in developing nations. Aside from problems associated with growing $79.50. conditions in tropical climates, there are basic social and economic In this volume are the collected papers of the Second Intema- problems. Land tenure and government policies are identified as tional Bangeland Congress, held May 13-18, 1984, at Adelaide, important constraints to the development of a land stewardship South Australia. Also included are a list of participants along with ethic. transcripts of the business meeting, opening ceremony, opening Various ecological concepts that have come to prominence in the remarks and closing ceremony. The papers are divided into thir- last couple of decades are discussed in Chapter Three. Foremost teen sections addressing the areas of ecosystem dynamics, grazing among these concepts is that of diversity. The authors are quick to industries, mining and rangelands, primary producers, grazing point out that the concept that diversity leads to stability is over- management, technological improvement, UNESCO Man and the simplified if not dead wrong. However, diversity as it applies in Biosphere program, animal production, plant ecophysiology, agro-ecosystems is a major concept that ~118 through much of the developing challenges, range monitoring and administration, con- rest of the book. The point is made that the basic resource, soil, servation and wildlife, and the role of fire. Some sections are must be maintained in order to have stable agriculture at the field preceded by short introductory papers or statements by the conve- level as well as at higher levels. non. Only invited papers and lunchtime addresses are printed in Chapter Four is devoted to the analysis of multi-crop agriculture full. Contributed papers are published as unrefereed synopses of or polyculture. It is suggested that increased crop diversity ulti- approximately one page. The Proceedings contain no photographs mately leads to stability, ecologically as well as economically. and few figures and tables. The volume is well edited and well Many benefits of growing various crops in combination are cited, printed. Some of the work contained in the Proceedings is summar- such as increased capture of available sunlight by various crop ized from research papers published in other journals and books. canopies, less weed competition and fewer insect and disease prob- Other papers, especially some of the contributed ones, contain lems. Available nutrients are also used more efficiently by using preliminary results of research efforts. deep-rooted perennials that “pump” nutrients from deeper soil Certainly, a major value of the volume is that it allows a world- profiles to the surface in the form of crop litter used as mulch. wide survey of state-of-the-art rangeland research. But while the There are many claims made to the benefits of polyculture, but a papers in the volume show how rapidly range research and range strong body of scientific research seems to be lacking. management have advanced in recent years, they also expose some The final chapten of the book discuss policy issues raised by an lingering weaknesses of range science. For example, among the ecological approach to agriculture and what sort of steps need to be papers on livestock grazing there is a babe1 of terminology, vari- taken to move toward an acceptance of presented concepts. The ables, and concepts which, although accentuated by the multina- authors point out that the orientation of agricultural education is tional character of the contributors, show little uniformity even probably the most important constraint to the broader develop within national borders. Some differences are semantic, but others ment and application of ecologically sustainable agriculture. They show fundamental conceptual work which needs to be done in are also critical of the bulk of present agricultural research stress- developing range science into a rigorous, integrative systems ing that other measures such as stability and environmental quality science. Future International Bangeland Congresses may be need to be considered in addition to yield per unit of input. Devel- appropriate forums to address the development, formalization, opment assistance agencies are also criticized for often being andstat&n%mtionof&gesciena conceptsandterminologythrou8h designers of “top down” farming systems with little effort put into discussions and through publications in future Proceedings. Inter- fully understanding specific sites. Although very critical of the national communication would be improved, and a systematized, agricultural “establishment*‘, the authors try to stay on balance by standardized science would liiely prove useful in managing a recognizing the benefits of modem farming practices in approp-

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 517 riate situtions. The book ends with the authors offering a set of Chapter 1, the author index, lists 100 individuals who have criteria for assessing proposed research projects. Suggestions are collaborated and coauthored the annotated bibliography. (Dr. also offered for the development of national agricultural policies Fiiser is not listed here, since his name is common to all.) Chapters and international programs. 2 and 3 provide a topic index and a keyword subject index to the Thii book should be of interest to anyone interested in ecology bibliographic listings of chapter four. or agriculture, and should be particularly interesting to those Chapter 4, which runs from page 3% to 661, gets down to the involved in research and instruction in these fields.-Daniel J. essentials and serves up the meat and potatoes. Apparently every Holden Idaho Falls, Idaho written and spoken communication by Dr. Fisser and collabora- tors on this subject has been listed and indexed in this section. It represents an astonishing array of number of references, as well as Wyoming Shrubland Ecology. 1987. Herbert G. Fisser. breadth of subject matter. Citations are grouped chronologically, University of Wyoming Agr. Exp. Sta. Science Monogr. 1959 to 1985, with annual listings organized according to avenue of 49.661 pages, $32.50 (cloth). presentation, such as journal publications, experiment station pub- This is a unique book, the likes of which I have never seen before. lications, proceedings/abstracts, oral presentations, etc. Thii type Dr. Fisser has compiled and indexed his life’s work in arid and of organization leads to unavoidable repetition, with reference to a semi-arid rangeland ecological research into a monumental tome single experimental result appearing in several years, or type of that should make any researcher envious. It reports hi collabora- information release. But it also results in a very complete tion with more than 100 other authors (including 60 graduate inventory. students) summarizing research supported primarily by funding Dr. Fisser and the Wyoming Agricultural Experiment Station from the Bureau of Land Management and the University of are to be congratulated for making the effort to document this Wyoming Experiment Station. research so completely and making it available to the profession in It is organized into four chapters; the first three are indexes to published form. It so frequently happens that many important the fourth, which is an annotated bibliography. A system of six threads of thought and action of long-term research programs are digit alphanumeric numbers (year, source, and item) is assigned to forgotten and lost forever in buried files. This book gives assurance each bibliographic entry, following a system developed by John that at least this part of Wyoming’s record will not quickly Vallentine. These numbers identify cita{ions throughout each of disappear!-Grant A. Harris. Washington State University. the four chapters.

New for 1989! Covers all aspects of range management, including current and past ‘research- Range Management: Principles and Practices

Jerry L. Holechek, Rex D. Pieper, and Carlton H. Herbel, all of New Mexico State Unwerstty

This basic text provides lull cC?vWageOf and sclentlflc reports as a foundation Prentice Hall range management hlstory, policy on fed. for concepts presented Book Distribution Center eral lands, plant ecdogy. grazng and . Presenk economc and envtronmen- wildilfe management, range nutntlon. prob- tal consequences of range manage- Rt. 59 at Brookhill Drive ment practices lems In developing countries and range West Niack, NY 10995-9901 improvement l Dwussesgrazlng methodsand sup olemental feedma of range lIvestock (201) 767-5937 Practical in approach and l Shows how to se? stocklig rates targeted to the needs of . Includes 135 ftguresand90 tables to Price $48/bard illustrate VWIOUS concepts today’s range managers NaIlI.2 . Usescurrent and past publications Address Post/Zip Code

518 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 Index, Volume 42

A Brakcnsitk, D.L., 139 Deer, 128,361 Acocie fomesiana, 407 Bransby, David I., 425 Defoliation, 458 Accumulation of nitrate by aMUd goldencye Brcjda, John J., 104 Demmtnt, Montague W., 71 and showy goldencye, 1% Britton, CM., 289,372 Dctling. Jamts K., 149 Adams, Don C., 243,393 Bromus sp., 183,243,248 Dicktrson, Jr., Roy L., 400 Aerial photography, 442 Brothtrson, J.D., 98 Ditt and forage quality of intermcdiatt whcat- Aeschynomen americo~, 128 Browsing rcsponsc, 354 grass managed undtr continuous and short- Agrofomstry, 2.191,203,378 Bryant, Fred C.. 20,361 duration grazing, 474 Agropyron sp., 46,93,109,113,122.153,183, Brush, 28.36,332,354.407 Direct effects of parasitism by Dinarmus ocu- 243,316,376,458,474,490 Buchloe dactyloides, 343 tus Thomson on seed predation by Acan- Airborne video, 442 Buffalograss, 343 thoscelides per/orotus (Horn) in Canada Akcson, Mark, 233 Burning, I 1,295,504 milk-vttch, 514 Aldon, Earl F., 56,284 Do your digits betray you or dots rounding C Alfalfa, 28,228,248,312.483 raise your reputation? Viewpoint, 420 California, 71,233,281,415 Alpine grasses, 304 Dodd, Jtrrold L., 149 American jointvetch, 128 Call, C.A., 51 Donart, Gary B., 158 Canada, 173,252,368,412 American jointvetch improves summer range Donnelly. Dennis, 134 for white-tailed deer, 128 Canon, Stcphtn K., 468 Dormaar, Johan F., 252 Cardtnas. Manuel, 118,228 Andropogon sp., 183,213,332 Douglasfir, 2 Animal p&fo&ance, 11, 153, 163, 239,243, Cattlt, 31, 134, 143, 153, 243, 248, 337, 343, Dowhowtr, Stcvt L.. 143,239,468 346,393,431) 474,480 248.346.393.41s. 421.425 Dozing, 295 Cattlt nutrition and grazing behavior during Animal p&forma& and-diet quality as influ- Drawt, D. Lynn, 361 short-duration grazing periods on crested cnccd by burning on tallgrass prairie, 11 Dyer, Mtlvin I., 149 Animal pcrformancc and plant production wheatgrass range, 153 Dytrs woad, 5 Cattlt prcfercnccs for a hybrid grass: chemical from continuously graxcd cool-season rc- E claiicd and native pastures, 248 and morphological, 22 Economic conscqutnccs of altcmativt stock- Animal unit equivalent, 343 Centeureo maculosa, 222 ing ratt adjustmtnt tactics: a simulation Ard. Jant, 75 Ceroroider Itmeto. 178,228 Chaining, 98 approach, 165 Argentina, 382 Chambers, Jeanne C., 304 Economic modtls, 165,421,425 Arizona, 75,284 Characterization and germination of chasmo- Effect of cattle grazing on the growth and Artemisie sp. 228 gamous and basal axillary cltitogamous miscrotoxin conttnt of Columbii milkvctch, Asay, K.H. 109,122 Tht, 368 Ash; juniper, 295 florcts of Ttxas winttrgrass, 51 Characterization of seed germination and sccd- Effect of ftrtiiization datt and litter removal Association of rtlativt food availabilities and on grassland forage production, 412 ling survival during initial wet-dry pcri- the Ef$f$idofscEppinp and shttp grazing on dytrs ods following planting, 299 Astrogolus sp., 366,368,514 Cincotta, Richard P., 434 Atraxint and ftrtilizcr effects on Sandhills Efftcts bf goat browsing on gambcl oak com- Clawson, W. Jamts, 508 subirrigatcd meadow, 104 munitics in northtm Utah, 354 Cohtn, Will E., 361 Atriplex, sp., 28, 158,228 Effects of intra-row spacings and cutting heights Colt, David N.. 84 on the yitld of Leucoena kucocephalo in B Colorado, 327,343,454,504 Adana, Turkey, 502 Bailty,D.W., 480 Combreto spp., 332 Effects of nitrogen fertilization on spotted Balph, David F., 376 Comparing tht economic value of foragt on knapwecd and competing vtgttation in wtst- Balph, Martha H., 376 public lands for wildlift and livtstock. 134 cm Montana, 222 Bartlett. E.T.. 454 Comparison of hydrometer settling times in Effects of organic amendments on soil biota Beef production from nativt and stedtd soil particle sixc analysis, 81 on a dcaradtd ranacland. 56 Northtm Grtat Plaii ranges, 243 Comparison of wtight estimate and rising- Effects oFshortdu&ion &axing on winter B&her, Joyce W., 172 platt meter methods to measure htrbage annuals in the Ttxas Rolling Plains, 372 Bcrdahl, J.D.. 312 mass of a mountain meadow, 71 Effects of stocking ratt on quantity and quality of available forage in a southern mixed Bhattacharya, Ashok N.. 28 Computers, 79,225,421,434 442 Conner, J.R.. 165 grass prairie, 468 - Big blucstcm, 213 Effects of 2.4-D and atraxint on dtgradcd Control of huisacht and honey mcsquitt with Bio-cconomic evaluation of stocking rate and Oklahoma grassland, 217 suppltmtntary ftcding of a beef herd, 346 a carp&d roller htrbicidt applicator, 407 Elk, 134,233 Biological control, 2,5, 11,404 Cool-season grasses. 4% Elymus, sp., 122,243 Biomass production, 104. 113, 118, 149, 158, Copper dtficitncy in tult tlk at Point Rtyts, Emergence and root growth of three prcgcr- 191,203,217,252,281,284,412 California, 233 minatcd cool-season grasses under salt and Biondini, Mario E., 113 Correlation of steer avtragt daily gain with wattr stress, 4% Blackbum, Wiibtrt H., 378 diet quality and foragt phtnology in an Enhanced wheel-point mcthod for assessing Blue grama, 113 improved annual grassland, 415 covtr. structure and hcttrogtntity in plant Blucbunch whcatgrass, 412 Cotter, Paul F., 400 communities, An, 79 Boc, A., 514 Cow-calf vs yearling substitution ratio for &he&o sp., 31 Boggs, Ktith W., 222 shortgrass steppe, A, 343 Equilibrium, 266 Bohn. Carolyn C., 8 1 Crcstcd whcatgrass, 46,93,109,113,118,153. Dogrostis sp., 289 Bonham, C.D., 284 316,376.459 Erlinacr. L.L.. 3% Book Rtvicws, 86,260.346,437,514 Crested whtatgrass growth and rcplaccmtnt Estal&hmtnt of stvtn high yielding grasses Booth D. Ttrrancc, 178 following ftrtiliition thinning, and ntigh- on tht Texas High Plains, 289 Esttll, Richard E.. 118 Botanical composition, Il. 16, 98, 104, 172, bor plant removal, 93 Estimating production and utilization of jo- 191,203,209,217,222,239,248,281,284, Cuts cattlt use to avoid stepping on trtsttd joba, 75 327,372 wheatgrass tussocks, 376 Evaluating rcvtgctation practices for sandy Bothriochloa sp., 289 Currit, Pat O., 22,393 cropland in tht Ntbraska sandhiffls,257 Bouteloue sp., 113, 199,243,289,343 D Evaluation of pinyon sapwood to phytomass Bovty, Radnty w., 407 Dahl, Bill E., 400 relationships ovtr difftrtnt sitt conditions, Bowman, R.A., 490 209 Bradley, Lisa C., 361 Davies, HI., 420 Brady, W.W., 284 Davis, J.N., 98 Dc Moracs, Elino A., 239 Ekphurbia auka L., 172 Harrington, Michael G., 504 Landry, Virginii N., 415 Experimental evaluation of the grazing op Hart, Richard H. 421.480 Lang, Kent J., 61 timiration hypothesis, 149 Ha&ii, Ibrahim M.,-163 Laycock, William A., 90 Hatipoght, Rilstur, 502 Leafy spurge, 172 F He&salrum boreale. 496 Leafy spurge and the species composition of a Factors influencing intenill erosion from semi- Heitschmidt, Rodney K., 143, 165, 239,337, mixed-grass prairie, 172 arid slopesin N&v Mexico, 66 431,468 Legumes, 28,228,248,315 415,483,496 FadlaIla. Babo. 163 Henry, S.H., 284 Leininger, W.C.. 2 Farah, Kassim; 5 Herbage yield response to the maturation of a Leucaena kucocephakz, 502 Fecal collection, 3% slash pine plantation, 191 Lewis, Clifford E., 191 Fertilization, 93,213,222,327,412 Hcrbel, Carlton H., 386 Litter removal, 412 Fescue, 252 i/ Herbicides, 217,332,400,407 Loomis, John, 134 Festuca sp., 252 Herbivore effects on seeded alfalfa at four Louisiana, 128 Finley, J.W., 474 pinyon-juniper sites in central Utah, 483 Lowry, Stephen R., 104 Forage quality, 11,28,46,109,118,122,128, Hetrick, G.W.T., 213 M 153.163,228.239,243,323,415,431,468, Hipp, Billy W.. 225 474 Hof, J.G., 454 Macrophytes, 323 Forage selection, 22,122,143,153,434,480 Hofmann. L., 199.248 Majak, Walter, 368 Forage yield, 46,104,118,252,289,327,412, Holechek, Jerry L., 118,228 Malechek, John C., 153,376 502 Honey mesquite, 407 Manning, Mary E., 309 Ford, Timothy M.J., 4% Huisache, 407 Marietta, K.L. 289 Forero. L., 343 Hunter, Thomas K., 378 Matches, A.G., 316 Forest gnuing, 2 Hussein, Abdullahi, 16 Matric potential, 386 Founving saltbush, 158,228 Hybrid grass, 22 Matric potential of clay loam soils on arid Frank, A.B., 199,312 rangelands in southern New Mexico, 386 Frasier, Gary W., 299 I Mayland, H.F., 109,122 Freckman, Diana W., 56 Idaho, 122. 134 McDaniel, B., 514 Frost, Wii E., 281 Improved harness for securing fecal collection McDougald, Neil K., 281,508 bags to grazing cattle, An, 3% McGinmes, W.J., 327 G Indian ricegrass, 187 McKean, J.R., 454 Galyean, Michael L., 228 Infiltration and runoff water quality response McPherson, Gury R., 372 Gambel oak. 354.504 to silvicultural and grazing treatments on a Measurements, 71,79,81,209,299,393,420 Gambel oak root carbohydrate response to longleaf pine forest, 378 Medicago sativa, 28,228,248,312,483 spring, summer, and fall prescribed bum- Infiltration and sediment production as atIected Melgom, A., 20 ing, 504 by soil surface conditions in a shrubland of Melilotus sp., 248 Gcbhardt, Karl, 81 Patagonia, Argentina, 382 Meyer, Robert E., 407 Genetic variability of Mg, Ca, and Kin crested Infiltration parameters for rangeland soils, Mexico, 20 wheatgrass, 109 139 Milkvetch, 366,368 Geographic information systems, 442 Influence of mycorrhiil fungi and fcrtilira- Mineral dynamics in beef cattle diets from a George, Melvin R., 508 tion on big bluestem seedling biomass, 213 southern mixed-grass prairie, 431 Georgia, 191 Inlluence of native shrubs on nutritional sta- Mings, Thomas S.. 323 Germination characteristics, 36, 41, 51, 178, tus of goats: nitrogen retention, 228 Miserotoxin. 366,368 183,187,289,299,394,490,4% Insect grazers, 149 Mitchell, J.E., 343 Gibbens, Robert P., 386 Intm-row spacing, 502 Montana, 222 Goats, 228,354 Investigations of maternal and embryo/fetal Moore, John C., 113 Gogan, J.P., 233 toxicity of Qhedra viridis and Ephedra Moruhological and physiological variation Goldeneye, 1% nevadeneis in sheep and cattle, 3 1 among ecotypes of sweetvetch (Hedysarum Graff$ Paul s., 225 Isatis tinctoria, 5,382 boreale Nutt.), 4% Graham, Donald R., 222 Israel, 346 Morris, James G., 415 Grass tetany, 109,316 Moser, Lowell E., 104,183 Grazing behavior, 143.337,343,361,376,434, J Mosley, Jeffrey, C., 400 Jensen, Mark E., 275 Motivation of Colorado ranchers with federal Gii%g effects 149 158 199 239 248,252, Jessup, David A., 233 grazing allotments, 454 284,337,37i, 458: 468: 474,‘483 ’ Johnson. Do&as A.. 4% Msatiri, Dionix N., 332 Grazing management, 421,425 Johnson; Mari A., 128 Mueller, D.M., 490 Grazing systems, 16,20,143.153,337,425,474 Jojoba, 75 Murphy, Alfred H., 508 Great Plains, 178.243 Jones, T.A., 187 Green, Douglas M, 323 Jordan, Gilbert L., 41 N Greene, L.W., 431 Juniperus sp.. 228,295 Nebraska, 104,183,257 Griffin, Graham, 79 Justification for graxing intensity experiments: Needed research on domestic and recreational Griggs. T.C., 316 economic analysis, 425 livestock in wilderness, Viewpoint, 84 Growing degree days, 199,508 Nelson, Billy D., 128 K Growth dynamics of fourwing saltbush as Nelson, M.L., 474 affected by different grazing management Kansas, 213 Nevada, 8 1,209,275,309 systems, 158 Keegan, Thomas W., 128 New Mexico, 56.66, 118, 158, 196,228,386, Growth models, 393 Keel&, Richard F., 31 404 Growth patterns of yearling steers determined Kidambi, S.P., 316 Newman, Reg F., 412 from daily live weights, 393 Kie, John G., 71 Nichols, James T., 104 Growth regulators’ effect on crested wheat- King, Milton A., 183 Nielson, DC., 187 gmss forage yield and quality, 46 Kirbv. Donald R.. 323 Nitrogen, fuation, 4% Growth regulators (plant), 46 Knapp, Bradford W., 243,393 N-S Dakota, 61,178,199,203,249,312,323 Gutman, M., 346 Knapweed, 222 Noy-Meir, Imanuel, 266,346 Kothmann, Merwyn M., 239 Nunez-Hernanda, Gregorio, 118,228 H Kuykendall, Charles B., 118 Nutrient analysis, 28,122,163,228,233,316, Haferkamp, Marshall R., 41 323,415,431 Hall, John W., 368 L Halvorson, Gary A., 61 Iaca, Emilio A., 71

JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 Nutrient composition of selected emergent Rawls, W.J., 139 Severson, Kieth E., 203 macrophytes in Northern Prairie wetlands, Reclamation. 490.4% Sharrow, S.H., 2 323 Redente, Edward F.. 113 Sheep, 2,5,28,3 1, 122 Nutrient utilization of acacia, haloxylon, and Reece, Patrick E., 104 Sheep grazing as a silvicultural tool to sup atriplex species by Najdi sheep, 28 Reeder. J.D., 327 press brush, 2 Rehabilitation, 56.61,98, 118, 183,248,257, Shewmaker, Glenn E., 122 0 327 Shmnery oak, 400 Oaks, 281,295.354,400 Relationship among grazing management, Silicon, 122 Observation on cattle liveweight changes and growing desret_days. and morphological Silicon in C-3 grasses: Effects on forage qual- fed indices in Sudan, 163 development for native grasses on the ity and sheep preference. 122 Observation on white-tailed deer and habitat Northern Great Plains, 199 Simmoruisia chine&s, 75 response to livestock graxing in south Texas, Remote Sensing, 442 Sites, Robert W., 404 361 Remote sensing technology for mngeland man- Skousen, J.G., 98 Observations on biomass of dynamics of a agement applications, 442 SMART: A Simple Model to Assess Range crested wheatgrass and native shortgrass Response of a semidesert grassland to 16 years Technology, 421 ecosystem in southern Wyoming, 113 of rest from grazing, 284 Smoliak, Sylvester, 252 Observations on vegetation responses to im- Response of established foray on reclaimed Society for Range Management, 90 proved graxing in Somalia, 16 mined land to fertilixer N and P, 327 Soil amendments, 56 Oklahoma, 11,217 Revegetation of a salt water blowout site, 61 Soil analysis methods, 8 1 Oldfather, !I.. 257 Rhodes, B.. 2 Soil biota. 56,213.252 Olson, Bret E., 93,458 Rice, C.K., 217 Soil climate, 275 Olson, Kenneth C., 153 Richards, James H., 93,458 Soil climate and plant community relation- Onobrychis vicifoli 3 16 Richards, R.W.. 480 ships on some rangelands of northeastern Opportunistic management for rangelands not Richardson. Bland Z.. 4% Nevada, 275 at equilibrium, 266 Riddle’s groundsel, 464 Soil contamination, 61 Oregon, 2 R&hers, R.K., 165 Soil erosion, 66,378,382 Orysopsis hymenoides, 187 Ries, R.E., 248 Soil infiltration. 139.382 Owensby. C.E.. 213 Riggs. Robert A., 354 Soil monetties. 252.382 P RiFtian systems, 309 Soil &ity, 4% Panicum sp.. 183,257,289 Rittenhouse. L.R.. 343.480 Soil water, 309,386 Parish. SM.. 474 Robbins, K.; 514 ’ ’ Soltero, Sergio. 20 Parker, Lawrence, W.. 56 Root-feeding insects of Senecio riddclliii in Somalia, 16 Pascopyron sp.. 118,199,492 eastern New Mexico and northwestern Some effects of a rotational graxing treatment Passage rate, 474 Texas, 404 on cattle graxing behavior, 337 Pearson, Henry A., 378 Rooting characteristics, 309,3 12 Some effects of a rotational graxing treatment Phenological development, 415 Rooting characteristics of four intermountain on cattle preference for plant communities, Phillips Jr., Sherman A., 404 meadow community types, 309 143 Pieper. Rex D., 332 Rose clover, 415 Sorg-Swanson, Cindy, 134 Pinchak, W.E., 431,468 Rosenau, R.C., 122 Southward, G. Morris, 158 Pines, 191,2tl9,281,504 Rosenstock, Steven S., 483 Spatial memory, 480 Pinus sp., 191,209,281,504 Rostagno, C&u hf., 382 Species dominance, 468 Pmyon-juniper chaining and seeding for big Roundy, Bruce A., 75 Spoonts, B.O., 51 game in central Utah, 98 Rouse, Gerald B., 153 Staigmiller, Robert B., 243 Plant competition, 93,172,203,209,222.281, Rumba@, Melvin D., 4% Standing crop biomass, 20 Russian wildrye, 498 Standing crop patterns under short duration grazing in northern Mexico, 20 P:zcontrol, 2,5,98,217,222,332,400,484, Ruyle, G.B., 75 467.564, Steinberger, Yosef, 56 Poa sandbergii, 4 12 S Stevens, Richard, 483 Pocket computer program for collecting for- Saiioin. 316 st@a sp., 51, 199,243 age selection frequency data in the field, A, Sampling, 3% Stocking rate, 165,343,346,468 434 Sandberg’s bluegrass, 4 12 Story, Jim M.. 222 Poisonous plants, 3 1, I%, 366.368,404 Saudi Arabia, 28 St&kc, J.F., 217 Prasetyo, L.H., 420 Savabii MR., 139 Stromberg, M.R., 284 Predicting peak standing crop on ammal range Scamecchii. D.L., 474 Stubbendieck, James L., 183,257 using weather variables, 508 Sediment, 382 Succession. 266.295 President’s Address: Is it time for a change?, 90 Seed germination characteristics of selected Succession of secondary shrubs on Ashe juniper Price, David L., 158 native plants of the lower Rio Grandc Val- communities after dozing and prescribed Production model, 458 ley, Texas, 36 burning, 295 Rosopis gkandulosa. 407 Seed predation, 514 Sudan, 163 Psathyrostachya juncea, 490 Seed treatments, 36,178 Summer growth, 504 Pseudoroegneria spicatum, 4 12 Seed viability of alpine species: variability Supplemental feeding, 346 Pseudotsuga mensiesii, 2 tithin and among years, 304 Survival and agronomic performance of 25 Public land use, 134,454 Seedbed ecology of winterfat: cations in dias- alfalfa cultivars and strains interseeded into pore bracts and their effect on germination rangeland, 3 12 Q and early plant growth, 178 Svejcar, Tony J., 11,309 Quality and botanical composition of cattle Seedbed effects on grass establishment on Swanson, Sherman R., 389 diets under rotational and continuous grax- abandoned Nebraska Sandhills cropland, Sweetvetch, 4% ing treatments, 239 183 Swift, D.M., 480 Querm sp., 281,295,354,480,504 Seedbed preparation, 183,257 T Quinton, Dee A., 368 Seeded wheatgrass yield and nutritive quality Tanzania. 332 on New Mexico big sagebrush range, 118 R Tausch, Robin J., 209 Selenium, 366 Raguse, Charles A., 415 Taylor, R.G., 454 Senecio riakkllii. 404 Ramets, 458 Tembo, Ackim, 228 Ranchers, 454 Self-compatibility in’Paloma’ Indian ricegrass, Rasmussen, G. Allen, 295 187 Seligman, N.G., 346 Senft. Richard L., 458

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 521 Temperature responses and calculated heat Uresk, Daniel W., 203 Weigel, Jeffrey R., 372 u~~forge~nationofscveralrangcgrasscs Umess, Phiip J., 354 Weighing equipment, 393 and shrubs, 41 Utah,5,31,93,98,153,187,354,366,376,483, West, Neil E., 5 Test of grazing compensation and optimiza- 4% Westoby, Mark, 266 tion of crested wheatgrass using a simula- Wetlands, 323 tion model, A, 458 V Wheeler, J.L., 420 Texas,36,51,143,165,239,289,295,316,337, Valdex, Raul, 118,228 White, Larry M., 46 361,372,400,407,431,468 van Ryswyk, A.L., 412 Whitford, Walter G., 56 Texas wintergrass, 51 Variability for Ca, Mg, K, Cu, Zn, and K/O Wikeem. Brian M., 412 l71inopyrum sp., 316 + Mg) ratio among 3 wheatgrasses and sain- Wilcox, Bradford P., 66 Thurow, Thomas L., 16 foin on the southern high plains, 3 16 Wildliie forage, 98,128,134,233,361,483 Tier dynamics, 93 Vegetationand soil responses to short-duration Wildrye, 122,243 Tolleson, D.R., 3% grazing on fescue grasslands, 252 Williims, M. Cobum, 196,366 Toxicological investigations on Toano, Vegetation mapping, 442 Williams. William A., 508 Wasatch, and stinking milkvetches, 366 Vegetational response to herbicide treatment Williamson, Samuel C., 149 Tree canopy effects on herbaceous production for brush control in Tanzania, 332 Willms, Walter D., 252 of annual rangeland during drought, 281 Versatile data logging program in BASIC for Wilson, G.W.T.. 213 Trent, James, 309 the Laptop computer, 225 Wilson, Scott D., 172 Trifgruh hirtum, 4 15 Viguierahp.; 1% Wilton. A.C.. 312 Truscott, Doreen R., 22 Vole&v. Jerrv D.. 393 Winckel, Joy; 7 1 Tueller, Paul T., 209,442 Vera, Robin,-36 Winter forb control for increased grass yield Tukel, Tuncay, 502 W on sandy rangeland, 408 Winterfat, 178,228 Turkey, 502 Walker, Brian, 266 Walker, John W., 143,239,337 Wood, James C., 378 U Wood, M. Karl, 66 Wallace, Joe D., 228 Understory-overstory relationships in pond* Wright, Henry A., 295 Wailer, Steven S., 104,183,257 rosa pine forests, Black Hills, South Dakota, Wyoming, 113 203 Washington, 474

522 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999 Journal of Range Management

Official bimonthly publication of Society for Range Management

Volume 42,1989 Table of Contents Volume 42,1989

Number 1, Jammry Number 2, March Sheep graxing as a silvicultural tool to suppress brush--SM. Mar- President’s Address-Is it time for a change?-William A. row, W. C. Leininger. and B. Rhodes ...... 2 Laycock ...... 90 Effects of clipping and sheep graxing on dyers woad-Neil E. West Crested wheatgrass growth and replacement following fertilization, and Kassim 0. Farah ...... 5 thinning, and neighbor plant removal-Eret L Olson and James Animal performance and diet quality as influenced by burning on H. Richards ...... 93 tallgrass prairie--Tony J. Svejcar ...... 11 Pinyon-juniper chaining and seeding for bii game in central Utah- Observations on vegetation responses to improved grazing in J.G. Skousen, J.N. Davis, and J.D. Brotherson ...... 98 Somalia--Thornus L lhtuow and Abdulkahi J. Hussein ...... 16 Atraxine and fertilixer effects on Sandhii subirrigated meadow- Standing crop patterns under short duration graxing in northern John J. Brefda, Lowell E. Moser, Steven S. Wailer. Stephen R. Mexico-Sergio Soltero. Fred C. Bryant, and A. Melgosa . . . . 20 Lowry, Patrick E. Reece, and James T. Nichols ...... 104 Cattle preferences for a hybrid grass: chemical and morphological Genetic variability of Mg, Ca. and K in crested wheatgrass-H.F. relationships-Doreen R. lYuscott and Pat 0. Currie ...... 22 Maykmd and K. H. Asay ...... 109 Nutrient utilization of acacia, haloxylon, and atriplex species by Observations on biomass dynamics of a crested wheatgrassand native Najdi sheep-Ashok N. Bhattacharya ...... 28 shortgrass ecosystem in southern Wyoming-Kdward F. Redente. Investigation of maternal and embryo/fetal toxicity of K&ru?a vir- Mario E. Biondini. and John C. Moore ...... I 13 idis and Ephedra nevadensis in sheep and cattle-Kichard F. Seeded wheatgrass yield and nutritive quality on New Mexico bii Keeler ...... 31 sagebrush range-Jerry L Ho&&k, Richard E. Estell, Char&s Seed germination characteristics of selected native plants of the lower B KuykenaUl, Raul Yak&z,Manual G&ems, and Gregorio Nunez- Rio Grande Valley, Texas-Robin S. Vora ...... 36 Hemandez ...... 18 Temperaturc responses and calculated heat units for germination of Silicon in C-3 grasses: effects on forage quality and sheep pm- several range grasses and shrubs-Gilbert L Jordan and Marshall feren~Glenn K. Shewmaker, H.F. Mayland, R.C. Rosemu, R. Haferkamp ...... 41 and K.H. Asay ...... 22 Growth regulators’ effect on crested wheatgmss forage yield and American jointvetch improves summer range for white-tailed deer- quality-Larry hf. White ...... 46 Thomas W. Keegan. Mark K. Johnson, and Billy D. Nelson . . . 128 Characterization and germination of chasmogamous and basal axil- Comparing the economic value of forage on public lands for wildliie lary cleistogamous florets of Texas wintergrass-C.A. Call, and and livestock-John Loomis, Dennis Donnelly, and Cindy Sorg- B. 0. Spoonts ...... 51 Swanson ...... 134 Effects of organic amendments on soil biota on a degraded rangeland- Infiltration parameters for rangeland soils- W.J. Rawls, D.L Bra- Walter G. Whitford, Earl F. Aldon. Dkt~ W. Freckman, Yosef ken&k, and M. R. Savabi ...... 139 Steinberger, and Lawrence W. Parker ...... 56 Some effects of a rotational graxing treatment on cattle preference for Revegetation of a salt water blowout site-Gary A. Hakorson and plant communities-John W. Walker, Rodney K. Heitschmidt, Kent J. Lang ...... a 61 and Steve L Dowhower ...... 143 Factors irdluencing interrill erosion from semiarid slopes in New Experimental evaluation of the graxing optimization hypothesis- Mexico-Bra#ord P. Wilcox and M. Karl Wood ...... 66 Samuel C. Williamson, James K. Detling, Jerrold L Dodd, and Comparison of weight estimate and rising-plate meter methods to Melvin I. Dyer ...... I...... 149 measure herbage mass of a mountain meadow-Dnilio A. Luca, Cattle nutrition and graxing behavior during short-duration-grazing Montague W. Demment, Joy Winckel, and John G. Kie ...... 71 periods on crested wheatgrass range--Kenneth C. Olson, Gerald Estimating production and utilization of jojoba-Bruce A. Roundy, B. Rouse, and John C. Malechek ...... 153 G. B. Ruyle, and Jane Ard ...... 75 Growth dynamics of fourwing saltbush as affected by different grax- An enhanced wheel-point method for assessing cover, structure and ing management systems-David L Rice, Gary B. Donart. and heterogeneity in plant communities-Graham F. Grt#j% ...... 79 G. Morris Southward ...... 158 Comparison of hydrometer settling times in soil particle size Observation on cattle liveweight changes and fecal indices in Sudan- analysis-Carolyn C. Bohn and Karl Gebhardt ...... 81 Ibrahim M. Hashim and i%bo Fadkdla ...... 163 Viewpoint: Needed research on domestic and recreational livestock in Economic consequences of alternative stocking rate adjustment tac- wilderness-David N. Cole ...... tics: a simulation approach-R.K. Riechers, J.R. Conner, and Rook Reviews ...... : R. K. Heitschmidt ...... 165 Leafy spurge and the species composition of a mixed-grass prairie- Joyce W. &kher and Scott D. Wilson ...... I... 172 Reviewers for 1988 ...... 176

524 JOURNAL OF RANGE MANAGEMENT 42(6), November 1969 Number 3, May Number 4, July Seedbedecology of winterfat: cations in diaspon bracts and their Opportunistic management for rangelands not at equilibrium-Mark effect on germination and early plant growth-D. Terrunce Westoby, Brian Walker, and hand Nay-Meir ...... 266 Booth 178 Soil climate and plant community relationships on some rangelands Seedbed effects on grass establishment on abandoned Nebraska of northeastern Nevada-Mark E. Jensen ...... 275 Sandhills cropland-Milton A. King, Steven S. Walhr, Lowell E. Tree canopy effects on herbaceous production of annual rangeland Moser, and James L Stubbendieck ...... 183 during drought- William E. Frost and Neil K. McDougaid . . . 281 Self-compatibility in ‘Paloma’ Indian ricegrass- T.A. Jones and D. C. Response of a semidesert grassland to 16 years of rest from grazing- Nielson ...... 187 W. W. Brady, M. R. Stromberg, E. F. Aldon, C. D. Bonham. and Herbage yield response to the maturation of a slash pine plantation- S.H. Henry ...... * ...... 284 Clt@-ordE Lewis ...... *...... 191 Establishment of seven high yielding grasses on the Texas High Accumulation of nitrate by annualgoldeneye and showy goldcneye- Plains-K.L Ma&to and C.M. Britton ...... 289 MCobum WiBiams ...... 1% Succession of secondary shrubs on Ashe juniper communities after Relationship among grazing management, growing degree-days, and dozing and prescribed burning-G. Al&n Rasmussen and Henry morphological development for native grasses on the Northern A. Wright ...... 295 Great Plains-A. B. Frank and L Hofmann ...... 199 Characterization of seed germination and seedling survival during the Undcrstory-overstory relationships in pondcrosa pine forests, Black initial wet-dry periods following planting-Gary W. Frazier . . . 299 Hills, South Dakota-Daniel W. Uresk and Kieth E.. Sever- Seed viability of alpine species: variability within and among years- son ...... 203 Jeanne C. Chambers ...... 304 Evaluation of pinyon sapwood to phytomass relationships over dif- Rooting characteristics of four intermountain meadow community ferent site conditions-Robin J. Tmueh and Paul T Tuelhr . . . 209 types-Mary E. Manning, Sherman R. Swanson, Tony Svejcar, Influence of mycorrhizal fungi and fertilization on big bluestem seed- and James Dent ...... 309 ling biomass-B.A.D. Hetrick, G. W. T. Wilson, and C.E. Survival and agronomic performance of 25 alfalfa cultivars and Owensby ...... 213 strains interseeded into rangeland-J. D. Berdahl, A.C. Wihon, Effects of 2,4-D and atrazine on degraded Oklahoma grasslands- and A.B. Frank ...... 312 C.K. Riceand J.F. Stritzke ...... 217 Variability for Ca, Mg, K, Cu, Zn, and K/(Ca + Mg) ratio among 3 Effects of nitrogen fertilization on spotted knapweed and competing wheatgrasses and sainfoin on the southern high plains-S.P. vegetation in western Montana-Jim M. Story, Keirh W. Boggs, Ktdambi, A.G. Matches, and T.C. Griggs ...... 316 and Donald R Graham ...... 222 Nutrient composition of selected emergent macrophytes in Northern A versatile data logging program in BASIC for the Laptop computer- Prairie wetlands-Donald R. Kirby, Douglas M. Green, and Paul S. Graff and Billy W. Hipp ...... I...... 225 Thomas S. Mings ...... 323 Influence of native shrubs on nutritional status of goats: nitrogen Response of established forages on reclaimed mined land to fertilizer retention-Gregorio Nunez-Hemandez, Jerry L Hole&k, Joe NandP-J.D. Ree&rand W.J. McGinnies ...... 327 D. Walkzce, Michael L Galyean, Ackim Tembo. Raul Vakiez, and Vegetational response to herbicide treatment for brush control in Manuel Cardernat ...... I..... 228 Tanzania-Dioniz N. Msajirt and Rex D. mper ...... 332 Copper deficiency in tule elk at Point Reyes, California-Peter J. P. Some efftis of a rotational grazing treatment on cattle grazing Gogan, David A. Jessup, and Mark Akeson ...... 233 behavior-John W. Walker and Rodnev K. Heitschmidt . . . . . 337 Quality and botanical composition of cattle diets under rotational and A cowsalf vs yearling substitution ratio for shortgrass steppe-L continuous grazing treatments-John W. Walker, Rodney K. Forero, LR. Rittenhouse. and J.E. Mitchell ...... 343 Heitschmidt, Elino A. De Moraes. Merwyn M. Kothmann, and Bio-economic evaluation of stocking rate and supplementary feeding Steve L Dowhower ...... 239 of a beef herd-N.G. Seligman. I. Noy-Meir. and M. Beef production from native and seeded Northern Great Plains Gutman ...... 346 ranges-Don C. Adams, Robert B. Staigmiller, and Bradford W. Book Reviews ...... * ...... 349 KMPP ...... 243 Animal performance and plant production from continuously grazed cool-season reclaimed and native pastures-L Hofmonn and R.ERies ...... 24 Vegetation and soil responses to short-duration grazing on fescue grasslands-Johan F. Dormaar, Sylvester Smoliak, and Walter D. Willms ...... 252 Evaluating revegetation practices for sandy cropland in the Nebraska sandhills-S. Oldfather, J. Stubbendieck. and S.S. Walter . . . . 257 Book Reviews ...... 260

JOURNAL OF RANGE MANAGEMENT 42(6), November 1989 525 Number5, !Septembes Number 6, Novemk Effects of goat browsing on gambel oak communities in northern Remote sensing technology for rangeland management applica- Utah-Robert A. Riggs ond Phitip J. Umess ...... 354 tions-Paul T. lueller ...... 442 Observations on white-tailed deer and habitat response to livestock Motivation of Colorado ranchers with federal graxing allotments- grazing in south Texas- Will E. Cohen, D. Lynn Drawe, Fred C. E. T. Bartktt, R. G. Taylor, J. R McKeaG and J. G. Ho/ ...... 454 &yattt,andLisaC.&adky ...... 361 A test of graxing compensation and optimization of crested wheat- Toxicological investigations on Toano, Wasatch, and stinking milk- grass using a simulation model-Bret El OLpon,Richard L Senft, vetches-M. Cobum William . . . , . . . . . , ...... 366 and James H. Richards ...... 458 The effect of cattle grazing on the growth and miserotoxin content of Effects of stocking rate on quantity and quality of available forage in a Columbia milkvetch-Dee A. Quinton, Waiter Majak, and John southern mixed grass prairie-Rodney K. Heitschmidt, Steven L. W. Hall ...... 368 Dowhower. William E. Pinchak, and Stephen K. Canon ...... 468 Effects of shortduration grazing on winter annuals in the Texas Diet and forage quality of intermediite wheatgrass managed under Rolling Plains-Jeffrey R. Weigei, Guy R. McPherson, and Carl- continous and short-duration grazing-M.L Nelson, J. W. Fin- ton M. Britton ...... *.... 372 ley. D.L Scarnecchia, and S.M. Parish ...... 474 Cues cattle use to avoid stepping on crested wheatgrass tussocks- Association of relative food availabilities and locations by cattle- David F. Balph, Martha H. &a&h, and John C. Makchek . . . . 376 D. W. Bailey L R. Rittenhouse, R. H. Hart, D. M. Swvt, and R W. Infiltration and runoff water quality response to silvicultural and Richard ...... 480 grazing treatments on a longleaf pine fotcst-James C. Wood, Herbivore effects on seeded alfalfa at four pinyon-juniper sites in Wilbert H. Blackbum, Henrv A. Pearson. and Thomas K. central Utah-Steven S. Rosenstock and Richard Stevens . . . . 483 Hunter ...... -...... 378 Emergence and root growth of three pregerminated cool-season Infiitmtion and sediment production as affected by soil surface condi- grasses under salt and water stress--AM. Mueller and RA. tions in a shrubland of Patagonia, Argentina-C&sar M, Bowman ...... 490 Rostagno ...... 382 Morphological and physiological variation among ecotypesof sweet- Matric potential of clay loam soils on arid rangelands in southern vetch(Hedysatwm boreakNutt.)-Douglas A. Johnson, Timothy New Mexico-Carlton H. Herbel and Robert P. Gibbens . . . . . 3% M.J. Ford, Melvin D. Rumbaugh. and Bland 2. Richard- Growth patterns of yearling steers determined from daily live son ...... 496 weights-Pat 0. Cur& Jerry D. Voksky, Don C. A&, and The effects of intra-row spacings and cutting heights on the yield of Brdford W. Knapp ...... 393 Leucaena leucocephah in Adana, Turkey-?lmcay ticket and An improved harness for securing fecal collection bags to grazing Riistii Hatipoglu ...... 502 cattle-D. R Tolleson and L L. Erlinger ...... 3% Gambel oak root carbohydrate response to spring. summer, and fall Winter forb control for increased grass yield on sandy rangeland- prescribed burning-Michael G. Harrtngton ...... 504 Bill E. Dahl. Jeflrev C. Moslev. Paul F. Cotter, and Rov L predicting peak standing crop on annual range using weather Dickerson, Jr.-*. .-...... I,...... :. . . 400 variables--Melvin R George, William A. Williams, Neil K. Root-feeding insects of Senecio riddellii in castem New Mexico and McDougald. W. James Cknvson, and Alfred H. Murphy . . . . . SO8 northwestern Texas-Robert W. Sites and Sherman A. Direct effect of parasitism by Dinarmus acutus Thomson on seed Phi&s, Jr...... 404 predation by Acanthoscelidesperforatus (Horn) in Canada milk- Control of huisache and honey mesquite with a carpeted roller herbi- vetch-A. Bee, B. McDaniel, and K Robbins ...... 514 cide applicator-Rodney W. Bovey and Robert El Meyer . . . . . 407 Book Reviews ...... 516 Effect of fertilization date and litter removal on grassland forage production-&ian M. Wikeem. Reg F. Newman, and A.L van Ryswyk ...... 412 Correlation of steer average daiiy gain with diet quality and forage phenology inan improved annual grassland-Charks A. Raguse, James G. Morris, and Virginia N. Landry ...... 415 Viewpoint: Do your digits betray you or does roundiig raise your reputation?-J. L. Wheekr, LH. Prasetyo, and H.I. Davies . . . 420 SMART: a Simple Model to Assess Range Technology-Richard H. Hart ...... 421 Justification for graxing intensity experiments: economic analysis- David I. Bransby ...... 425 Mineral dynamics in beef cattle diets from a southern mixed-grass nrairie- W.E. Pinchak, L. W. Greene, and R. K. Heit- &tmidt ...... 431 A pocket computer program for collecting forage selection frequency data in the fiild-Richard P. Cincotta ...... 434 Book Reviews ...... 437

526 JOURNAL OF RANGE MANAGEMENT 42(6), November 1999