TABLE of CONTENTS: Vol

TABLE OF CONTENTS: Vol. 41, No. 6, November 1988

ARTICLES Invited Synthesis Paper 450 Stability of African pastoral ecosystems: Alternate paradigms and implications for development by James E. Ellis and David M. Swift

Plant Ecoloav 460 Plant Responsesto pine management and deferred-rotation yaxing in north Fior- ida by Clifford E. Lewis, George W. Tanner, and W. Stephen Terry 466 Species diversity and diversity profiles: Concept, measurement, and applh?ationto timber and range management by Clifford E. Lewis, Benee F. Swindel, and George W. Tariner 470 Control of threadleaf rubber rabbitbrush with herbicides by Steven G. Whisenant 472 Defoliation of Thurber needlegrass: herbap and root responses by David Ganskopp 477 Season of cutting affects biomass production by coppicing browse species of the Brazilian caatinga by Linda H. Hardesty, Thadis W. Box, and John C. Malechek 481 Effects of dormant-season herbage removal on Flint Hills rangeland by Lisa M. Auen and Clenton E. Owensby 483 Stocking rate effects on intensive-early stocked FM Hills bluestem range by Clenton E. Owensby, Robert Cochran, and Ed F. Smith 488 Determination of root mass ratios in alfalfa-grass mixtures using near infrared reflectance spectroscopy by M.D. Rumbaugh, D.H. Clark, and B.M. Pendery

Reclamation 491 Germination of green and gray rubber rabbitbrush and their establishment on ~081 mined iand by J.T. Romo and L.E. Eddleman

Plant-Animal Interaction 4% Floristic changes induced by flooding on graaed and ungrazed lowland grasslands in Argenthu by Enrique J. Chaneton, Jose M. Facelli, and Roland0 J.C. Leon 500 Grazing effects of the bulk density in a Natraquoli of the Flooding Pampa of Argentina by Miguel A. Toboada and Ralil S. Lavado 503 Stability of graxed patches on rough fescue gassiands by Walter D. Willms, John F. Dormaar, and G. Bruce Schaalje 509 Effects of phenology, site, and rumenfill on tail larkspur consumption by c8ttle by James A. Pfister, Gary D. Manners, Michael H. Ralphs, Zhao Xiao Hong, and Mark A. Lane 515 Habitat selection and activity patterns of female mule deer in the Front Range, Colorado by Roland C. Kufeld, David C. Bowden, and Donald J. Schrupp

TECHNICAL NOTES 522 Separating leaves from browse for use in nutritional studies with herbivores by J.F. Gallagher, T.G. Barnes, and L.W. Vamer Published bimonthly-January, March, BOOK REVIEWS May, July, September, November 524 Photosynthesis Volume I. Energy Convenion y Plants and Bacteria. Photosyntbe- Copyright 1999 by the Society for Range aia Volume II. Development, Carbon Metabolism and Plant Productivity Edited by Management Govindjce; A Handbook of Mexican Roadside Flora by Charles T. Mason Jr. and Patricia B. Mason; by lNDlVlDUALSUBSCRlPTlONisbymembershipin Ecology: Individuals, Populationa and Communities the Society for Range Management. Michael Begon, John L. Harper, and Colin R. Townsend; The Rise of Multiple-use LIBRARY or other INSTITUTIONAL SUBSCRIP- Management in tbc Intermountain West: a History of Region 4 of the Forest TIONS on a calendar year basis are $sS.gClfor the !Service by Thomas G. Alexander United States postpaid and 3SS.00 for other countries, postpaid. Payment from outside the United States should be remitted in US dollars by international money order or draft on a New York bank. BUSINESS CORRESPONDENCE, concerning sub- scriptions, advertising, reprints, back issues, and related matters, should be addressed to the Manag- ing Editor, 1939 York Street, Denver, Colorado 80206. ED~TOR~ALCORRESPONDENCE.~~~K%~~~~~~~~- scriptsor otheredhorial matters,should beaddressed to the Editor, 1939 York Street, Denver. Colorado 80206. INSTRUCTIONS FOR AUTHORS appear on the inside back cover of most issues. A Style Manual is also available from the Society for Range Manage- mentattheaboveaddress@$Z.Wforsinglecopies; $1.25 each for 2 or more. THE JOURNAL OF RANGE MANAGEMENT (ISSN W2249X) is published six times yearly for SSS.00 per year by the Society for Range Management, 1839 York Street, Denver, Colorado 90206. SECONDCLASS POSTAGE paid at Denver,Colorado. POSTMASTER: Return erttlra journal wltb address cbangtRETURN POSTAGE GUARANTEED-to Society for Range Management, 1939 York Street, Denver. Colorado 9CMtIS.

The Joumel of Ren9a Managemen tsarvesasa forum for the presentation and discussion of facts, ideas, and philosophiae pertaining to the study, management, and use of rangelands and their several resources. Accordingly, all material published herein is signed and reflects the individual views of the authors and is not necessarily an official position of the Society. Manuscripts from any source-nonmembers as well as members-are welcome and will be given every consideration by the editors. Submissions need not be of a technical nature, but should be germane to the broad field of range management. Editorial comment by an individual is also welcome and, subject to acceptance by the editor, will be published as a “Viewpoint.”

Maneglng EdItor ASSOCIATE EDITORS PETER V. JACKSON Ill WILL BLACKBURN RODNEY HEITSCHMIDT BRUCE ROUNDY 1939 York Street N.W. Watershed Res. Center Box 1858 325 Biological Sciences Denver, Colorado 80206 270 South Orchard Vernon, Texas 78384 East Building. Univ. Arizona Boise, Idaho 93705 Tucson, Arizona 95721 Editor N. THOMPSON HOBBS PATRICIA G. SMITH CARLTON BRITTON Colorado Div. of Wildllfe PAUL TUELLER Society for Range Management Range &Wildlife Mgmt 317 W. Prospect Range Wildlife h Forestry 1339 York Street Texas Tech University Fort Collins, Colorado 80528 UNR loo0 Valley Road Denver, Colorado 80208 Lubbock, Texas 79409 Rena, Navada 99912 (303) 3%7070 PETE W. JACOBY, JR. THOMAS A. HANLEY P.O. Box l&59 RICHARD S. WHITE Book Review Edltor Forestry Sciences Lab. Vernon, Texas 79394 KSU Extension Center GRANT A. HARRIS BOX20909 Colby, Kansas 67701 Forestry and Range Management Juneau, Alaska ggeO2 HOWARD MORTON STEVE WHISENANT Washington State Univarslty 2CrCxlE. Allen Road Tucson, Arizona 93719 Texas A&M Unlvesity Pullman, Washington 99194-9410 RICHARD H. HART USDA-ARS Dept. of Range Science 9409 Hildreth Rd. Collage Station, Texas 77843-2126 Cheyenne, Wyoming 92009

Invitational Synthesis Paper

TheEditorial Board of the Journal of Range Management invited James E. Ellis and David M. Swift to prepare this synthesis paper in recognition of their many, valuable contributions tu understanding of the ecology of grazing systems. Their work is characterized by a high level of imagination, by a steady commit- ment to thoroughness, and by distilled, clear thinking.

JIM ELLIS took undergraduate work in animal husbandry at the University of Missouri and also obtained his Master of Science degree there studying wildlife biology. In 1970, he received his Ph.D. in Zoology at the University of California at Davis, where he was a National Institute of Health trainee in systems ecology. Shortly thereafter, he held a National Science Foundation post- doctoral fellowship at the University of Bristol working on systems analysis of mammalian social systems. He joined the Natural Resource Ecology Laboratory of Colorado State University as a Research Ecologist in 1971. He is currently the Associate Director of the Laboratory. Jim has enjoyed immense success in developing research programs; during the last decade he directed or played a major role in 12 successful proposals, collectively exceeding three milliondollaninsuppott. He has publishedextensively, withmost Research Scientist. Dave has deep expertise in simulation of cco- of his work focussing on processes regulatinggrazing systems. Jim logical systems. He has worked on models investigating energy and has served as a consultant to the U.S. Senate, as well as the nitrogen balance in ruminants, control of diapause in grass- govcmmcnt of Saudi Arabia and the Norwegian Agency for Inter- hoppers, carrying capacity of elk winter range, effects of sulfur national Development. His must recent project uses a systems dioxide on grassland ecosystems, influences ofgrazing on primary approach to understand the controls on stability and persistence of production, and the nutritional benefits of herding. He frequently a pastoral ecosystem in East Africa. contributes to the applied as well as the theoretical ecological literature. Dave teaches courses in ruminant nutrition in the Range DAVESWIFTstudied forest botanyasanundergradutc at the Science and Wildlife Biology departments at Colorado State Uni- the New York State College of Forestry and received a master’s versity and has supervised the research of many masters and doc- degree in watershed resources at Colorado State University. In toral students. He is widely admired by those who have had the 1985 he earned his Ph.D. in Animal Science at Colorado State good fortune to study with him. His recent work has taken him to University. Since 1970 he has been a member of the staff of the East Africa, Pakistan, and the People’s Republic of China. Natural Resource Ecology Laboratory where he now serves as a

Stability of African pastoral ecosystems: Alternate paradigms and implications for development JAMESE.ELLlS ANDDAVID M.SWIFT

Abstrnd African pastoral ecosystems have been studied with the assumptions that these ecosystems are potentially stable (equilibrinl) sys- Our view of the world, OI our perception of any system, has a tems which become destabilized by overstucking and overgrazing. great deal of influence on how we go about dealing with that Development policy in these regions has focused on internal &era- system. Believing that illness is caused by spirits leads to very tions of system structure, with the goals of restoring equilibrium different sorts of treatments than the perception that viruses OT and increasing productivity. Nine yews of ecosystem-level research bacteria may be the cause. Likewise, our perception of how partic- in northern Kenya presents a view of pastoral ecosystems that are ular ecological systems operate determines the approaches that we non-equilibrinl but persistent, with system dynamics affected more advocate in attempting to modify or manipulate those ecosystems. by nbiotic than biotic controls. Development practices that fail tu For example, one school of thought held that unusually high recognize these dynamics may result in increased deprivation and concentrations of elephants were responsible for the wide-spread failure. Pastoral ecosystems may be better supported by develop destruction of dry woodlands in many areas of east and central ment policies that build on and facilitate the traditional pastoral Africa (Beuchnerand Dawkins 1961, Glover 1963, Laws 1970). An strategies rather than constrain them. alternative hypothesis proposed that elephants and trees were Key Words: pastoral ecosystems, equilibria1 ecosystems, nun- locked in a stable limit-cycle in which elephant populations equilibria1 ecosystems, pastoral development expanded at the expense oftree populations. Ultimately the loss of woodlands would cause reductions in elephant populations while woodlands would begin to recover, initiating a new cycle (Caugh- Icy 1976). The general perception of destructive elephant-tree interactions led to some large-scale programs of elephant “con-

460 JOURNAL OF RANGE MANAGEMENT 41(e), November ,988 trol” in an effort to preserve woodlands. However, later work pastoralists and their ecosystems in the literature. The work of suggested that climate change, varying water tables, or combina- early ethnographers tended to picture traditional pastoralists in a tions of herbivores, fire, and climate changes were responsible for rather romantic light, i.e., self-reliant nomads, moving freely, liv- the decline of woodlands. These results showed that elephant ing off the land, while defying governments and settled agricultu- control programs were, in some cases, a needless slaughter of a ralists (Evans-Pritchard 1940, Spencer 1973, Jacobs 1965). This scarce species (Western 1973, Sinclair and Norton-Griffiths 1979, perception seems also to have incorporated the idea, at least implic- Pellew 1983, Dublin 1987). itly, that pastoralists lived in some sort of balance with their In scientific parlance the elephant-tree interaction was perceived natural environment if not with their neighbors; pastoral systems in terms of a particular paradigm; that paradigm influenced the were not identified with environmental degradation. structure of scientific discourse, the types of analyses and models The idea that pastoral&s do not achieve a balance with their used, and the kinds of management solutions applied to related environment but routinely overstock and overgraze, is an old one problems. Alterations in management practices occurred when the (Stebbings 1935), but was stated most forcefully and coherently by dominant paradigm was questioned and another proposed. Such Brown (1971), the Chief Agriculturalist of colonial Kenya, and changes in scientific paradigms often promote the use of new more recently by Lamprey (1983). Brown’s viewpoint arose no models, different analytical approaches, and new solution regimes doubt from his own experiences, but also incorporated the concept (Kuhn 1970). of the irrationality of pastoral management practices first posed by We believe that the time is ripe to examine the paradigms which Herskovits (1926), whose “cattle complex”hypothesis spawned the govern our thinking about African pastoral ecosystems. Discus- view that accumulation of vast numbers of livestock was an irra- sions of these pastoral ecosystems usually revolve around prob- tional pastoral tradition, ingrained in the social system; a practice lems of low productivity, overstocking, overgrazing, drought, clearly incompatible with good environmental management. Brown range deterioration, dying livestock, starving people, and so on. In identified a different sort of irrationality; he asserted that pastoral- essence, the current paradigm focuses on pastoralism as a malad- ists were engaged in dairy operations under environmental condi- aptive and destructive system of exploitation. However, years of tions much more suitable for meat production and he suggested intervention into pastoral systems, by governments and develop- that these dairy operations tended to be very inefficient in terms of ment agencies, have failed to relieve the perceived problems; in milk produced per unit of forage available. He then presented a fact, development has sometimes rendered the target population numerical analysis of pastoral subsistence needs based on ineffi- and its ecosystem somewhat worse off than before the intervention cient milk production and used this to substantiate the case for (Helland 1980, Swift 1977, Sanford 1983, Swift and Maliki, 1984). obligatory overstocking, overgrazing, and environmental destruc- This lack of success has been so pervasive that some research tion by pastoralists. organizations and major funding agencies have essentially given up Lamprey, in contrast, pointed out that pastoral management is on pastoralists and arid lands in Africa and are investing their adaptive and rational from the perspective of human survival, but resources elsewhere (ILCA 1987). It is possible that some interven- does not incorporate environmental conservation as a manage- tion failures could be the result of technical incompetence on the ment objective (1983). Further, Lamprey’s incisive ecological per- part of development experts or intransigence on the part of pastor- spective brought some important assumptions into the open which alists. However, it seems unlikely that experts are always techni- are often held implicitly, but seldom stated. There is, in this view, cally incompetent or pastoralists always intransigent; the near the potential for density-dependent equilibria between herbivores universal failure of pastoral development suggests that something and vegetation in the regions of East Africa occupied by pastoral- more fundamental is amiss. Universal failure would be expected if ists and such equilibria are frequently achieved in natural ecosys- invalid paradigms underlie the development interventions. tems, not inhabited by man. Lamprey reasoned that overstocking Two sorts of inquiries are needed to assess the validity of the by pastoralists causes departures from these equilibria1 conditions dominant pastoral paradigm and the models and interventions and rangeland degradation. Clearly, continuing degradation of the which follow therefrom. First, it is necessary to identify the domi- environment would eventually lead to extinction of pastoralists, nant paradigm, the assumptions on which it is based, and their but this has not occurred over the thousands of years of pastoral implications. Secondly, the paradigm needs to be tested against occupation of Africa because, “Either they could move on from the extant observations and empirical data. Do the observed dynamics degraded lands into new territories, or they could adapt their fit the dominant paradigm? If not, the paradigm obviously needs to pastoral practices to increasingly marginal conditions (for in- be altered to provide a more realistic model of pastoral ecosystems stance, by herding camels instead of cattle) and remain where they from which better intervention and management procedures were” (Lamprey 1983, p. 664). would hopefully derive. Thus the pastoral paradigm materializes: pastoral systems are In this paper we propose that the dominant paradigm of pastoral potentially stable systems but departures from equilibria1 condi- systems is based on assumptions (1) that African pastoral ecosys- tions necessarily follow from the inefficiency of pastoral resource tems are potentially stable (equilibrial) systems; (2) that these exploitation and the resultant need to overstock the ecosystem. potentially stable systems are frequently destabilized by improper This environmentally destructive pattern has not led to extinction, use on the part of pastoralists; and (3) that alterations of system however, because new regions were exploited while the degraded structure (reducing livestock numbers, changing land-tenure pat- ones presumably recovered (this might now be considered a patch- terns, etc.) are needed to return these systems to an equilibria1 and dynamic equilibrium), or because alternate equilibria1 states were more productive state. We then provide evidence from our own possible (when a system was degraded by cattle it could still be work (1) that stable equilibria are not achievable in many pastoral utilized by camels or goats); i.e., these systems contain multiple ecosystems, although long-term persistence is; (2) that interven- stable points (Noy-Meir 1975). tions aimed at achieving stability in non-equilibria1 systems are The final element of the dominant paradigm proposes that in likely to be irrelevant at best or disruptive and destructive at worst; many cases development has exacerbated degradation in pastoral and (3) that successful interventions will be designed to accommo- ecosystems. Veterinary care and reductions in tribal raiding are date system dynamic variation rather than aimed at maintaining said to have released livestock populations from former density- equilibria1 conditions. dependent constraints, while curtailment of nomadism, losses of grazing lands to agriculture, security problems, and the settlement The Dominant Paradigm: Pastoral Degradation of some pastoralists have combined to reduce the area of rangeland of Equilibria1 Ecosystems available while herds are on the increase (Lusigi 1981, Lamprey There are at least 3 separate but interrelated perceptions of and Yussuf 1981, Lamprey 1983). Overstocking and overgrazing

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 451 have intensified while the reductions in rangeland area have immediate and direct impacts on pastoralists and it also may have a removed the possiblity of achieving patchdynamic equilibria. certain appeal to some local planners or officials. The assumption Resulting acceleration of degradation is hypothesized to be caus- is that some form of privatization will alleviate the imbalances ing widespread shifts from cattle to camels in northern Past Africa supposedly induced by communal grazing. This is the viewpoint (Stiles 1983); while vegetation removal by livestock in the Sahel is expressed in Hardin’s ‘tragedy of the commons’ concept (Hardin believed to have increased soil surface albedo, leading to reduc- 1968) as applied to pastoral systems. Group ranches, grazing tions in rainfall and rapid desertification (Charney et al. 1974, blocks, grazing cooperatives, etc., are all schemes aimed at attach- Sinclair and Fryxelll985). To summarize, the dominant paradigm ing groups of pastoralists to specific tracts of land which will of pastoral ecosystem dynamics is based on perceptions that: presumably induce better management practices (Helland 1980, l Pastoralism is basically an inefficient and therefore environ- Oxby 1982). Implicit in the strategy is the premise that the tracts of mentally destructive resource exploitation strategy. land have the capacity to support stable, balanced populations of l The ecosystems occupied by pastoralists have the capacity to livestock and people if only good management prevails, i.e., equi- support stable (equilibrial) populations of herbivores, but pastoral librial conditions are attainable (Fig. 2). The results of interventions strategies necessarily lead to overstocking and tend to move the system away from the potential equilibrium conditions. l Pastoralists have avoided large-scale extinctions by moving to new areas after degrading previously occupied environments or by Ecosrsr~n OESTAGILIZEO EGUILISRIUW EcOSlSTEMS changing strategies to accommodate the new but somewhat RESTORE0 degraded .environmental state. l Modernization has made things worse by reducing the potential for large-scale movement and by removing density-dependent ALTER i:-,j IAND constraints on livestock populations (Fig. 1). ELIMINATION OF TENURE 0”ERGRAZING SYSTEMS

ALTERATION t IN hWROVE0 TECHNICAL DEVELOPMENT PASTORAL EFFICIENCY INTERYENTIONS PSACTICES PRACYICES I i I INCREASED REDUCE ,r SECONDARY STOCKING PROGUCTION RATES

Fig. 2. Development practicesand expected results based on the quilibrial I/ MOVETO ‘“SW Prrdw* OVERGRAZING ""DEGRADED AREAS f and schemes based on these assumptions are not encouraging. I PASTORAL They include disruption of pastoral societies and ecosystems and o"EGS:OCKING RESPONSE ALTERGRAZING TO SELF- the retreat of development agencies from the arid and semi-arid TRADITIONAL PRIORITILS TO XNDUCEO PASTORAL f ACCOb!MODATE OLGRAOA- lands of Africa(Helland 1980, Swift and Maliki 1984, ILCA 1987). PR*cTIcES INEFFItxENT DEGRADED TION RESOURCE ECOSIGTEWS Some Theoretical Considerations EXPLOITAT1ON The idea that there is a balance of nature has always had great Fig. 1. The equilibrhl paradigm of pa&oral ecosystem dynrmics. appeal to scientists, naturalists, and others. When the concept is stated in more analytical terms, it generally is accepted to mean Implications of the EquilibriaI Paradigm that natural systems exist in some sort of homeostatic state, and that this state is maintained by interactions among the components A similar perception of pastoral systems has been termed the of the system. The concept thus implies that the dominant type of “mainstream view” because it has dominated attitudes toward system interaction is negative feed-back control of rate processes pastoral&s and shaped pastoral development policy (Sandford which maintain the system in a dynamic equilibria1 state or within 1983). Some of the major consequences which Sandford attributed some limited domain of attraction. The antithetical view of natural to the mainstream view include the assumption that whatever system dynamics emphasizes the role of external control mecha- pastoralists are doing is inappropriate and therefore range man- nisms, i.e., drivers, which arenot subject to feed-back control from agement programs aimed at reducing presumed degradation within the system. When system dynamics are dominated by exter- should be applied universally and rapidly. Likewise changes in nal forces, the opportunity for the development of feed-back con- land-tenure systems and existing institutions are assumed to be trol is much reduced and the persistence of the system depends desirable and are therefore undertaken without consideration of upon the development of other sorts of stabilizing mechanisms. what may be useful or valuable in the existing systems. These varying perceptions of natural system dynamics have given We think that the key assumption in the equilibria1 paradigm has rise to famous scientific debates such as the density-dependent an even more fundamental impact on development policy than the versus density-independent population control controversy (Nichol- mainstream view. The paradigm assumes that the ecosystems son 1958, Nicholson and Bailey 1935). occupied by pastoralists generally function as equilibria1 systems While these contrasting views of system dynamics are both now which are regulated by densitydependent feed-back controls; recognized as legitimate models of biological reality (Noy-Meir however, pastoralists override these feed-back controls to the det- 1979/80, Wiens 1984), the concept that systems operate in a riment of themselves and their ecosystems. If this assumption is homeostatic fashion, based on equilibria1 dynamics, has had by far accepted, it is logical to reason that internal alterations in system the more pervasive. role in ecological thought. The ideas, that structure can correct the imbalances and restore the system to communities are structured by competitive interactions and that equilibria1 conditions. The most obvious adjustments to make are evolution proceeds through a continuous process of adaptive “fine those involving the number of livestock per unit area. Hence two tuning”, are based on an equilibria1 view of system dynamics. In types of development procedures follow: reduction of stocking addition, the development of the concept of ecological stability and rates and alteration of land-tenure systems. Destocking is a very the analysis of the stability properties of ecological models pro- direct means of altering system structure, but it is hard to sell to ceeded from an assumption of equilibria1 conditions (Lotka 1925, pastoralists. Alteration of land-tenure has the advantage of less Volterra 1928, Rosenzwieg and MacArthur 1963). The equilibria1

452 JOURNAL OF RANGE MANAGEMENT 41(6), November 1968 assumption, its implications for the competitive structuring of communities, and its validity in mathematical analyses, have been questioned (Wiens 1977, 1984, Caswell 1978, Connell and Sousa 1983); however, it remains a powerful concept in both field and mathematical assessments of population, community, and ecosys- REDUCTION tem dynamics. Clearly then, we should not be terribly surprised to OF FEEOBACU find that equilibria1 assumptions dominate our perceptions of pastoral ecosystems. A major challenge to the equilibria1 assumption was posed by Wiens, who after several years of field research on avian communities in arid and semi-arid environments, concluded that equilibria1 conditions occurred only occasionally in these environments. He asserted that abiotic controls regulated bird populations in arid ecosystems and that since equilibria1 conditions (those required for STNOG COMPENSATIONS ASXOTIC S”SsYIT”TION competitive interactions to develop) were seldom attained, the role PERTURSATIONS of competition in shaping these communities was minimal (Wiens 1977). He extended these conclusions to propose that ecosystems exist along a gradient from equilibria1 conditions where biotic interactions structure communities to non-equilibria1 conditions DESYASILX J5 where abiotic controls determine system structure and dynamics p&l&g$Z %cg (Wiens 1984). This scheme has been further elaborated by DeAnge- lis and Waterhouse (1987), based on their review of model analyses Fig. 3. Facton promoting destabilization and pemistcncc of ecosystems. of ecological stability. They propose that the spectrum of potential (Modified from DeAngeL end Wtierhoose 1987). system behaviors centers on the concept of stable equilibria1 sys- instabilities in pastoral ecosystems. Why then is there such a poor tems, but that systems frequently fail to demonstrate equilibria1 record of success in the development of these ecosystems and why behavior due either to the disruptive effects of stochastic elements do degradation and famine continue in Africa? It is, in our view, (such as abiotic controls) or due to instabilities induced by strong because pastoral ecosystems are often strongly controlled by internal feed-backs. Strong feed-back or overconnectedness is external forces rather than, or in addition to, internal biotic fac- exemplified where herbivores or predators over-exploit their food tors. Therefore the nominally applied development procedures are resources, resulting in departures from equilibrium points as pro- often irrelevant to the systems of interest, or worse, they comprise posed for the case of overgrazing by pastoralists. When strong additional perturbations which act to further destabilize system feed-backs are accompanied by time-lags, stable-limit cycles may dynamics, rather than damping them. result. This is the dynamic suggested to explain the interdependent How does one determine whether a pastoral ecosystem is domi- interactions between elephants and trees (Caughley 1976). nated by internal biotic controls or by external perturbations? The question posed by DeAngelis and Waterhouse (1987) (and Theoretical considerations and model analyses of grazing systems many theoretical modelers before them), focuses on how natural suggest that internal control mechanisms will come to dominate systems manage to persist in the face of pervasive destabilizing system dynamics when plant growing conditions are relatively forces. Their resolution of the question is based on the operation of invariant over time, whereas major variations in growing condi- a number of different mechanisms incorporated into theoretical tions will prevent the development of strong biotic interactions and models to stabilize model performance. The crucial point is that feed-back controls on system dynamics (Noy-Meir 1975, Wiens different sorts of stabilizing mechanisms are requiied to ameliorate 1977, 1984). A number of other structural characteristics and the effects of different destabilizing forces. Where destabilization overall patterns of system dynamics may also be used to differen- results from strong stochastic forces, such as abiotic perturbations, tiate between equilibria1 and non-equilibria1 grazing systems persistence may be maintained by the introduction of compensa- (Table 1). tory mechanisms or by increasing the spatial scale of the model ecosystem. Where instability is caused by strong biotic interac- Non-Equilibria1 Dynamics in a Pastoral Ecosystem tions, stability may be regained by the introduction of disturbances Nine years of research among pastoral nomads in northern which influence either resource or consumer density, or by incor- Kenya has given us a very different view of the dynamical behavior porating mechanisms which reduce the strength of system interac- of pastoral ecosystems than that portrayed in the equilibrial para- tions, e.g., interference, inefficiency, etc. Thus, the specific mecha- digm. It has also caused us to question the appropriateness of nisms employed in attempting to stabilize model ecosystem per- development procedures which are based on the paradigm’s formance depend upon what one perceives as the critical destabil- assumptions. The results of our work in Ngisonyoka Turkana, one izing force. Although this analysis focuses on model ecosystems, it of 15 sub-tribal territories in Turkana District, reveal anything but provides some important insights into the different ways in which an equilibria1 ecosystem. Here in the arid northwest comer of real ecosystems may persist under the influence of potentially Kenya, pastoralists are locked in a constant battle against the destabilizing controls and interactions (Fig. 3). vagaries of nature and the depredations of neighboring tribesmen. The implications for the persistence of African pastoral ecosys- Droughts and raids are ongoing stresses; drought-induced live- tems are clear. If these are potentially equilibria1 systems, con- stock mortality frequently diminishes herds by 50 percent or more trolled mainly by strong biotic interactions, then destabilizations (McCabe 1985, in press, Ellis et al. 1987). Infant mortality rates are and degradation may, in fact, be caused by overstocking and high, human nutritional status is quite dynamic, and emigration strong biotic coupling and feed-back from livestock to plants. This from the pastoral sector is common (Brainard 1986, Galvin 1985, is a basic presumption of the dominant paradigm. Appropriate McCabe 1985). However, despite the dynamic nature of the ecbsys- stabilizing procedures under these conditions would indeed involve tern there is little evidence of degradation or of imminent system fine-tuning the interaction between plants and livestock through failure (Ellis et al. 1985, Coughenour et al. 1985, McCabe and Ellis such means as destocking and land privatization, provided these 1987). Instead, this ecosystem and its pastoral inhabitants are procedures actually reduced consumer density relative to vegeta- relatively stable in response to the major stresses on the system, tion resources. In other words, the currently applied development e.g., frequent and severe droughts (Ellis et al. 1987). In theoretical practices may be well-suited to correcting internally induced terms this is a non-equilibria1 but persistent ecosystem (Holling

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 453 Table 1. Cbaruteristics of cquilibrial and non-equilibrial grazing systems. 160 D : dry season W: wet season Equilibria1 Grazing Non-equilibria1 Grazing Systems Systems

Abiotic Abiotic conditions Stochastic/variable Patterns relatively constant conditions Plant growing Variable plant growing conditions relatively conditions invariant Plant- Tight coupling of Weak coupling of Herbivore interactions interactions Interactions Feedback control Abiotic control Herbivore control of Plant biomass plant biomass abiotically controlled Population Density dependence Density independence Patterns Populations track Carrying capacity too carrying capacity dynamic for close 60, population tracking Limit cycles Abiotically driven cycles Community/ Competitive struct- Competition not Ecosystem ing of communities expressed Character- istics Limited spatial extent Spatially extensive Externalities critical to 00 Self-controlled systems SEASON DWDWD DWDWD system dynamics 81 82 82 83 83 81 82 82 83 83 YEAR 82 83 84 82 83 84 1973, DeAngelis and Waterhouse 1987). Our results do not support the dominant paradigm, and they demonstrate that at least for NGISONYOKA TARAcH (TURKANA) some pastoral ecosystems, the assumptions of equilibria1 dynam- Fig. 4. Seasonal and annual NDVI dynamics. ics, and the intervention practices which follow, are inappropriate. Climatic Variation and Plant Production in Ngisonyoka, Turkana indicative of equilibria1 dynamics. A critical assumption about equilibria1 grazing systems is that The major perturbations on the Turkana ecosystem are droughts plant growing conditions are relatively invariant over time. Noy- of a year’s duration or longer. In central Turkana, rainfall has Meir (1975) studied the stability properties of potentially equilib- dropped 33% or more below the long-term averages 13 times in the rial model grazing systems. His assumptions about the characteris- last 50 years, i.e., once every 3-4 years, and at least 4 of these have tics leading to equilibrium included time invariant growth during been severe multi-year droughts (Ellis et al. 1987). NDVI values the growing season and no use of the forage resource outside of the recorded during a single-year drought show growing season values growing season. These are very restrictive assumptions indeed and that are only l-l .5 times those of the preceding dry season, indicat- would not be expected to be met in many grazing ecosystems. It is ing that plant biomass production was one-half or less of that probable, however, that such restrictive assumptions are not neces- achieved in a normal year (Fig. 4). This value may be even lower in sary for equilibrium dynamics to prevail. Equilibrium is probably the second year of a multi-year drought; production levels are possible or at least approachable in systems which demonstrate known to drop to 113 or l/4 of average values during droughts in intraseasonal and intra-annual variation in plant growth provided some arid areas (Paulsen and Ares 1961). The dramatic response of the level of interannual variation is not too great. If the timing and plant biomass to dry seasons and drought demonstrates the perva- magnitude of primary production is more or less predictable on an sive role of climate in the vegetation dynamics of this ecosystem. annual basis, a more or less constant carrying capacity develops The high degree of seasonality plus the great interannual variabil- and density dependent population processes permit the herbivore ity lead to pulsed and undependable plant growth, rather than the community to track gradual changes in the forage resource. constant or at least predictable growing conditions necessary for Pastoralists generally occupy arid or semi-arid environments the development of equilibria1 grazing interactions. where climatic variability causes distinct pulses of plant produc- Herbivory and Plant Production tion followed by long periods of plant dormancy, but in which the Another fundamental assumption about the operation of equi- pulses of production are not predictable in terms of time or magni- librial grazing systems is that herbivores play a major role in tude. Rainfall averages between 200 mm and 600 mm over most of controlling plant biomass through consumption and offtake (Noy- Ngisonyoka territory. The growing season ranges from 60 to 90 Meir 1975, McNaughton 1979). Using ecosystem-wide estimates of days (April-June) during normal years, leaving a 9-10 month dry forage production and livestock consumption, we calculated that season with little or no plant growth. Occasional good years occur total livestock offtake is on the order of lo-12% of forage produc- when a short rainy period interrupts the dry season in October or tion during a good year in Ngisonyoka, Turkana. Likewise, Ngiso- November, causing another 20-30 days of plant growth. Satellite nyoka stocking rates are less than one-fourth the theoretical max- derived normalized vegetation index values (NDVI’s) suggest that imum carrying capacity in a good year, allowing for 50% forage in the most productive regions of Turkana District, green plant consumption (Coughenour et al. 1985, Ellis et al. 1987). Given biomass increases by 2-2.5 times during the growing season, rela- these stocking levels and offtake rates it seems unlikely that live- tive to the long dry season (Fig. 4). Intra-annual growing condi- stock exert a major control on plant biomass. tions are highly variable even during the best of years and are not

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 Field and laboratory assessments of livestock grazing effects on 0) herbaceous and dwarf shrub vegetation have revealed no signifi- 2 4 - cant grazing effects on plant production (Bamberg 1986, Cough- enour et al. in prep). On the other hand, livestock do influence the morphology of important dwarf shrubs (Bamberg 1986, Mugambi in progress) and increase germination and establishment of important trees (Ellis et al. in prep, Reid in progress). Thus, while livestock may, in the long-run, alter the structure and composition 1 6 - of the plant community, they appear to have no role in regulating yearly plant production, only a minor role in regulating biomass levels and consequently little or no role in regulating the amount of 1 2 _ forage available. The strong force exerted by climate on forage production and the minimal influence of livestock on forage availability means that there is little opportunity for the develop ment of strong feedbacks from livestock to plants. The plant- herbivore interaction is therefore only loosely coupled in this respect, and probably operates as a density-independent relation- 4- ship most of the time. This form of interaction is indicative of non-equilibria1 systems, and is inconsistent with the characteristics I I I I of equilibria1 grazing systems (DeAngelis and Waterhouse 1987). z : Livestock Dynamles 5d) Despite the fact that Ngisonyoka livestock consume only a small % proportion of the total forage produced in a good year, livestock nutritional status and production rates closely track the seasonal dynamics of plant production. This is because nutritional condition and secondary production are limited by forage quality (pro- ,,: :; tein content and digestibility) during the long dry seasons even though forage quantity is usually adequate. For most livestock \ species, diet quality drops to maintenance levels by the middry \ -* \ r’ % season (Fig. S), and loss of condition and reduction of production \\ / ‘. / continue for several months, until the following rainy season. ‘\ #’ Though camels are able to do better than the other species, main- 35 \ \\ /* taining a diet of adequate nitrogen content, the digestibility of their \ 0’ \\ 0’ diet declines to the point that their nutritional state is compromised \ / \\ / also (Coppock et al. 1982, 1985, 1986a,b,c). \ 0’ In drought years forage quantity as well as quality becomes V limiting. While the quantity of forage removed by livestock is only 25 moderate, that removal plus consumption by termites and losses due to decomposition and weathering eventually deplete the forage supply. During droughts, livestock are on starvation rations; nutri- w MD LD W ED tional condition and production spiral downward. Only the length 1981 1982 of the drought determines how serious the effects on livestock condition and subsequently human nutrition, will be (Coppock et Season al. 1986b, c; Ellis et al. 1987). If droughts last two years or more, the Fig. 5. !%uonal concentrations for (a) crude protein,(b) IVDDM (in vitro decrements in livestock condition also begin to influence livestock digestible dry matter) for Iivestock diets in South Turkuu, Kenya, population size through effects on reproduction and mortality. during 1981-82. Specks coded ae goats (- - -), sheep (--), cattle We do not propose that there is no connection between livestock -a-), end cemels(-). Seasons are coded M W (wet, April-May), ED density and the degree of nutritional stress experienced during the iearly dry, June-July), MD (middry, Au~usbOctober), and LD (Iate- annual dry period or during more extended droughts. Any time a dry, Novemba-March). AU vahesare based on 100% dry matter. (mod- forage resource varies in quality, there is an opportunity for density ified ftom Coppock et al. l!NMb). dependence to operate through competition for small quantities of population mteractions in which livestock numbers are controlled the best available forage (Hobbs and Swift 1985). Thus, there is the by forage availability (Noy-Meir 1975). But in Turkana, drought potential for density dependent condition change during the dry perturbations appear to regulate livestock populations and this season. In the case of longer droughts, it could be argued that a control operates independent of livestock density under most cir- lower stocking rate prior to the drought results in a greater quan- cumstances. Single-year droughts undermine livestock condition tity of cured forage to carry the animals through the period of and may reduce reproductive rates but do not induce livestock stress. This is correct to a point but termites, microbes, and abiotic mortality at low to moderate stocking rates. However, even at low factors deplete this resource regardless of the density of grazing stocking rates, multi-year droughts cause drastic increases in live- animals. In the case of either seasonal dry periods or longer stock mortality and reductions in reproduction. A two-year droughts there is simply no forage available that will maintain drought in 1979-80 caused losses of SO-70% of the livestock popu- animals in positive nitrogen and energy balance (Coppock et al. lation in parts of Turkana District. But recovery from these losses 1986~). Once the forages cure, the animals have begun the process was, in many cases, relatively rapid; some herds had returned to of starvation; and the length of time that the stress must be endured predrought levels four years after the drought ended. This rapid is more critical to the outcome than the number of animals endur- rate of recovery was largely the result of two mechanisms: rapid ing the stress. A similar conclusion was reached by Wallmo et al. reproductive rates of small stock (goats and sheep) and immigra- (1977) regarding the carrying capacity of mule deer winter range. tion of cattle into the District after the drought ceased. Much of While density has a role, it is small compared to the role of this immigration involved the return of animals which had been environmental uncertainty. taken out of the region during the drought, while other recruitment Equilibria1 grazing systems exhibit strong density-dependent was likely the result of raids on neighboring tribes (McCabe 1985,

JOURNAL OF RANGE MANAGEMENT 41(e). November 1988 455 onstrating long-term persistence in a difficult environment. Over man’s lands” where the risk of intertribal raiding is especially high. the past decade, which included a serious drought in 1979-80 and a Alternately, access to other tribes’resources is sometimes obtained single-year drought in 1984, there has been no famine in Ngisony- by force, usually a difficult and bloody task (McCabe 1985, in oka territory and no overt evidence of environmental degradation press, Dyson-Hudson and McCabe 1983, McCabe and Ellis 1987, (Ellis et al. 1984, 1987). Short-term persistence was maintained Ellis et al. 1987). Thus, a major strategy for coping with severe despite large-scale livestock losses in a situation in which 80% of drought is to utilize fully the extensive scale of the ecosystem, or to pastoral food intake is derived directly from livestock products obtain resources which are actually outside that system. (Galvin 1985, Galvin and Waweru 1987). Pastoral persistence in this ecosystem is achieved through a Compensation for Reductions in Livestock Production series of stabilizing strategies which vary with the strength of the The other major strategy for dealing with severe stress is com- environmental stress. Dry season decrements in forage availability pensation for reduction in livestock products, accomplished either and livestock production are mild stresses; the Ngisonyoka com- by substituting other products which are available, or by reducing pensate for dry season diminution in livestock products by switch- human demand. Alternative food products may include foods ing foods and by reducing their own activity levels and energy gathered from the bush; purchased foods, usually grain, obtained requirements. Nevertheless, dry seasons regularly result in small by selling livestock; meat from livestock which have died from declines in human nutritional status and weight, averaging 34% of starvation or disease: and relief foods, ifthese are available (Galvin total body weight (Galvin 1985, Galvin and Waweru 1987). Like- 1985, Galvin and Waweru 1987, Ellis et al. 1987). wise, there are indications of seasonality in human reproductive Reduction in human demand entails people actually leaving the patterns with the majority of children born in the wet season, pastoral system. During the multi-year drought of 1979-80, many indicating that most conceptions occur when women reach their Turkana pastoralists left the system temporarily by moving to peak nutritional status in June and July (Leslie and Fry 1988). villages or famine camps within Turkana or to highland areas Single-year droughts are only slightly more stressful than nor- where they stayed with friends or relatives or became laborers on mal dry seasons for the Ngisonyoka; that is, the strategies farms, in abbatoirs, etc. (McCabe 1985). As much as 20% of the employed to cope with a normal dry season are usually adequate Turkana pastoral population may have emigrated from the pas- for dealing with those relatively harsh dry seasons which follow a toral sector in the early 1980’s although many had returned by 1984 poor rainy season. Multi-year droughts provide a much more (Ecosystems Ltd. 1984). Individuals undertaking such temporary formidable stress and require more drastic responses. These gener- emigrations are usually of lower social status and not essential for ally take one of two forms: (1) expansion of the spatial scale of the maintenance of pastoral herds, e.g., unmarried or widowed exploitation, or (2) compensation for losses of livestock products female relatives, young men who are not herders, etc. Their emigra- by wholesale food substitution or reduction in human demand. tion does not substantially inhibit the normal functions of the Expansion of the Spatial Scale of Exploitation system. This combination of stabilizing strategies-food substitution, Ngisonyoka territory covers about 9,000 kmr. In normal years demand reduction through emigration, and expansion of the spa- only a portion of this territory is utilized by the pastoralists and tial scale of exploitation-allow the Ngisonyoka to persist through their livestock (Ecosystems Ltd. 1984, McCabe 1985); however, periods of severe stress without famine, without degrading their during severe droughts the spatial scale of exploitation expands. ecosystem, and without permanently reducing their own popula- This is accomplished through a variety of tactics, each of which tion. Each of these strategies depends to some extent on utilizing entails increasing levels of risk. resources which are not exploited during non-stress periods and As environmental stress increases, pastoralists divide their live- which lie outside the spatial scale of routine exploitation. It can be stock herds and and accompanying human groups into smaller and concluded that this ecosystem cannot and does not support the smaller units. These small units tend also to move more frequently extant populations of humans and livestock during periods of than do the larger units which exist during wetter periods (McCabe severe environmental stress which occur about once per decade 1985, in press). Dispersion of the livestock and human population (1096-2% of the time). If external/peripheral resources were not is facilitated by a widely distributed water supply; few large por- available, the human and livestock populations would have to be tions of the ecosystem are unusable because of lack of dry-season maintained at considerably lower levels the remaining 8090% of water (Dyson-Hudson and McCabe 1983). Division into smaller, the time to avoid excessive livestock losses and human famine, more dispersed groups is referred to as “risk spreading” in the during droughts. The ecosystem is not balanced, does not operate literature (den Boer 1968); however, smaller groups of pastoral&s in an equilibria1 fashion, and cannot be treated as if it did. are actually more vulnerable to raids by bandits or other tribes, therefore incurring somewhat more risk. The real value of this dispersion is to spread grazing pressure more broadly and evenly Implications for Development over the region. The dominant paradigm assumes that pastoral ecosystems are The fact that the Ngisonyoka have unused space to move into potentially equilibria1 grazing systems and that destabilization of during periods of drought stress suggests that the system is stocked these systems is due to overstocking and overgrazing by pastoral- well below its average ecological carrying capacity. This is true and ists. Conventional development practices are based on these it is a critically important feature of the system; but not for the assumptions; the goals include restoration of equilibria1 conditions usual density-related reason that stocking too close to carrying and increases in productivity. Conventional development proce- capacity creates the risk of overgrazing and subsequent range dures involve the establishment of group ranches, grazing blocks, degradation. In this case, the low level of stocking buys, not or grazing associations where pastoralists are confined to particu- protection for the plant resource, but time for the pastoralist in the lar tracts of land. Resources are developed and technical interven- form of an ungrazed reserve. A single year drought is thus survi- tions are provided within those tracts to raise productivity and to vable. Large scale herd losses are avoided if the rains return before better regulate the interaction between animals and plants. In a second year of drought is encountered. theory these interventions have the potential to achieve the desired In some cases it is possible to move livestock completely out of effects if, in fact, the problem is properly diagnosed; that is if they the territory, into regions occupied by other tribes or into other are applied in ecosystems which have the potential to operate as subsections of Turkana, if the other pastoralists are agreeable. This equilibria1 systems, and which have truly been destabilized by tactic is especially useful if the other regions have been less affected overgrazing. by the impinging drought. If access to external grazing lands is not We have attempted to show that in Ngisonyoka Turkana and permitted, it may be necessary to move into boundary areas or “no most probably in many other arid or semi-arid pastoral ecosys-

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 457 terns, equilibria1 conditions are not attainable. Rather, ecosystem technical issues? The main message which we wish to convey here is dynamics are dominated by the stochastic perturbations of multi- that appropriate policies and technical interventions can be app- year droughts. Under these conditions large-scale destocking lied only if the fundamental dynamics of the target systems are would result in immediate deprivation for pastoralists even during clearly understood. No one is better qualified than range ecologists mild stress periods. Likewise, confining pastoral&s to grazing to analyze and describe the dynamics of pastoral ecosystems and blocks or ranches would reduce the spatial scale of exploitation through this activity to provide the basic understanding necessary and result in disaster during serious droughts. The obvious conclu- for enlightened development policy and intervention. Unless pas- sion is that conventional development procedures are destabilizing toral ecosystem dynamics are considered and used as guidelines for influences in ecosystems which are dominated by stochastic abiotic development policies, interventions are likely to be random activi- perturbations and which operate essentially as non-equilibria1 ties which comprise development by trial and error, a practice with ecosystems. unfortunate implications for the ecosystems and people on which Can non-equilibria1 pastoral systems be improved by develop these “development experiments” are performed. ment interventions? Or should pastoralists living under these conditions be left alone? In our view the latter is not an option. Pastoraliits are coming under the influence of external forces Acknowledgments regardless of their remoteness or of the relative success or failure of their traditional exploitation systems. Therefore we must explore The research upon which this paper is based was supported by grants the appropriateness of different development interventions for from the US National Science Foundation, Ecosystems Studies and different ecosystems and design interventions to fit the dynamics of Anthropology programs. Additional funding was provided by the Norwe- specific target systems. A cautious approach to pastoral develop- gian Aid Agency for International Development (NORAD). Our thanks to ment is to ask if intervention strategies can be formulated which Jesse Logan, John Wiens, and Mike Coughenour for comments on the will build upon the best aspects of traditional systems, rather than paper and to Sharren Sund for help with the manuscript. imposing wholesale alterations upon them. The Ngisonyoka and several other Turkana groups seem adept at resisting the affects of severe droughts and at maintaining a Literature Cited sustainable and persistent, albeit unstable and dynamic, exploitation system. Their main strategies for maintaining this system in Bamber8, LA. 1986. Effects of clipping and wateringfrequency on production of the African dwarf shrub Indigoferu spinoso. MS. Thesis, Colo- the face of perturbations are: rado State University, Fort Collins. - _ - l expanding the spatial scale of exploitation during stress Beuchner. H.K.. and H.C. Dawkins. 1961. Veaetation chance induced bv jMiOdS; elephants andfire in Murchison Falls Natioial Park, Ugaida. J. Wildi and Manage. 27:36-53.

l compensating for productivity losses by product substitution Brahard, J. 19%. Differential mortality in Turkana agriculturalists and and adjusting for loss by demand reduction. pastoralists. Amer. J. Physical Anthro. 70(4):525-536. Since these practices are key to pastoral success and persistence in Brown, L.H. 1971. The biology of pastoral man as a factor in conservation. non-equilibria1 systems, interventions which facilitate rather than Biol. Conserv. 3(2):93-100. constrain these strategies should be considered. In the first case, it Caawell, H. 1978. Predator mediated coexistence: a non-equilibrium model. Amer. Nat. 112:127-154. is clear that an extensive spatial scale is a prerequisite for a success- Caughley, G. 1976. The elephant problem-an alternative hypothesis. E. ful pastoral system where droughts are frequent. Reductions in Afr. Wild]. J. 14265-283. scale or confining pastoral&s to ranches is an invitation to disas- Chamey, J.G., P.H. Stone, end W.J. Quirk. 1975. Drought in the Sahara: a ter. There is no inherent reason why the maintenance of spatial biopdysical feedback mechanism. science 187~434-455. scale could not be included as a development objective. However, Connell, J.H., end W.P. Soum. 1983. On the evidence needed to judge development schemes often emphasize the delivery of “technical ecological stability or persistence. Amer. Nat. 121:789-824. packages” and the improvement of resources within “tractable” Coppock, D.L., J.E. Ellie, J. Wicnpahl, J.T. McCabe, D.M. Swift, and areas and consider such issues as spatial scale beyond their K.A. Galvin. 1982. A review of livestock studies of the South Turkana purview. Ecosystem Project. pp. 168-172. In: Proceedings, Small Ruminant CRSP Workshop, Kenya. SR-CRSP, Nairobi. In regard to the second strategy, food substitution and demand Coppoek, D.L., J.T. McCabe, J.E. Ellis, K.A. Gelvin, end D.M. Swift. reduction can best be facilitated by maintaining an open economy 1985. Traditional tactics of resource exploitation and allocation among in which there is a free flow of goods both in and out. Developing nomads in an arid African environment. In: Proceedings of the Intema- strong interactions with the national economy through viable tional Rangeland Development Symposium, Society for Range Man- marketing systems seems the best way to assure this flow. agement, Salt Lake City, Utah. Improved economic flows could eventually reduce the necessary Coppock, D.L., J.E. Ellis, and D.M. Swift. 1986a. Livestock feeding spatial scale, if the commodity flows included supplemental live- ecology and resource utilization in a nomadic pastoral ecosystem. J. stock feeds. However, these flows depend upon welldeveloped Appl. Ecol. 23(2):573-583. market and credit systems at local, regional, and national levels. Coppock, D.L., D.M. Swift, end J.E. Ellis. 1986b. Seasonal nutritional characteristics of livestock diets in a nomadic pastoral ecosystem. J. These development procedures, maintaining the spatial scale of Appl. Ecol. 23(2):585-595. pastoral ecosystems and facilitating the flow of goods into and out Coppock, D.L., D.M. Swift, J.E. Ellis, and K. Galvin. 1986e. Seasonal of the ecosystem through market development, are policy-oriented, patterns of energy allocation and production for livestock in a nomadic rather than technical solutions. This is not to say that technical pastoral ecosystem. J. Agr. Sci., Camb. 107347365. procedures are not useful; it simply suggests that they must be Coughenour,M.B., J.E.EUis,D.M. Swift,D.L.Coppock, K. Galvin,J.T. imbedded in a progressive pastoral development policy in order to McCabe, and T.C. Hart. 1985. Energy extraction and use in a nomadic be effective. Policy implementation is usually a national function, pastoral ecosystem. Science 20~619-625. therefore, successful pastoral development probably can not be Cou8henour, M.B., D.L. Coppoek, J.E. Ellis, and M. Rowland. Herbace- divorced from on-going national development. This suggests that ous biomass and productivity in an arid pastoral ecosystem in Northwest Kenya (in prep.). we may need more emphasis and research on pastoral development Coqbenour, M.B., D.L. Coppock, M. Rowland, and J.E. Ellis. Dwarf policy and how best to use technical interventions to augment shrub (Indigofera spp.) ecology in South Turkana, Kenya. (in prep.) those policies, rather than emphasizing technical interventions DeAngel& D.L., and J.C. Waterhouse. 1987. Equilibrium and nonequili- alone, and ignoring policy-level concerns. brium concepts in ecological models. Ecol. Monogr. 57(1):1-21. Should range ecologists and managers be involved in policy denBoer, P.J. 1968. Spreading the risk and stabilization of animal analysis and application as well as providing expertise on more numbers. Acta Biotheoretica (L&den) 18:165-194.

458 JOURNAL OF RANGE MANAGEMENT 41(S). November 1988 .

DubIin, H. 1987. Decline of the Mara woodlands: the role of tire and McCabe, J.T. 191)s. Livestock management among the Turkana: a social elephants. Ph.D. Thesis, University of British Columbia. and ecological analysis of herding in an East African pastoral popula- Dyson-Hudson, R., and J.T. McC8be. 1983. Final Report to the Norwe- tion. Ph.D. Diss. State University of New York-Binghamton. gian Aid Agency on South Turkana Water Resources and Livestock McC8be, J.T. Drought and recovery: Livestock dynamics among the Ngi- Management. sonyoka Turkana of Kenya. Human Ecology (in press). Don8ldson, T.J. 1986. Pastoralism and drought: a case study of the Borana McCPbe, J.T., 8nd J.E. EMS. 1987. Beating the odds in arid Africa. Nat. of southern Ethiopia. M. Phil Thesis, University of Reading. Hist. 96(1):3241. Ecosystems, Ltd. 1984. Turkana District resource survey. Turkana Reha- Nicholson, AJ. 1958. The self-adjustment of populations to change. Cold bilitation Project, Draft Final Report, July 1984. Ministry of Energy and Spring Harbor Symp. Quant. Biol. 22153-173. Regional Development, Government of Kenya, Nairobi. Nicholson, A J., 8nd V.A. Bliky. 1935. The balance of animal populations. J. Animal Ecol. 2:132-178. Ellis, J.E., D.L. Coppock, J.T. McC8be, K. G8lvin, ld J. Wienp8hl. 1984. Norton-GrIfBths, M. 1979. The influence of grazing, browsing and tire on Aspects of energy consumption in a pastoral ecosystem: wood use by the the vegetation dynamics of the Serengeti. In: A.R.E. Sinclair and M. southern Turkana. In: C. Barnes, J. Ensminger and P. OXeefe, (eds.) Norton-Griffiths (eds.), Serengeti: Dynamics of and Ecosystem. Univer- Wood, Energy and Households, Perspectives on Rural Kenya. Energy, sity of Chicago Press, Chicago. Environment and Development in Africa (6). The Beijer Institute and the Noy-Meir, I. 1975. Stability of grazing systems: an application of predator- Scandinavian Institute of African Studies, Sweden. prey graphs. J. Ecol. 63:459481. E~s,J.E.,D.S.Schimel,M.B.Cou~cnour,T.C.~rt,J.G.Wyant,andS. Noy-Meir, I. 1979/80. Structure and function of desert ecosystems. Isr. J. Lewis. Enhancement of tree establishment by pastoral nomads in an arid Bot. 2&l-19. tropical ecosystem. (in prep.). Oxby, C. 1982. Group ranches in Africa. Overseas Development Institute ElIis, J.E., K. GaIvIn, J.T. McCabe, D.M. Swift. 1987. Pastoralism and Review. 2:2-13. drought in Turkana District, Kenya. A report to NORAD, Nairobi. P8ulsen, H.A., 8nd F.N. Ares. 1961. Trends in carrying capacity and 294PP. vegetation on arid southwestern range. J. Range Manage. 14:78-83. Evmns-Pritcbud, E.E. 1973. The Nuer. Oxford University Press, London. Peliew, R.A. 1983. The impacts of elephant, giraffe and fire upon the G8lvin, K.A. 1985. Food procurement, diet and nutrition of Turkana Acacia rortilis woodlands of the Serenneti. Afr. J. Ecol. 21:41-74. pastoralists in the ecological and social context. Ph.D. Diss. State Uni- Rosenzweig, M.L., 8nd R.H. M8eArthuL 1963. Graphical representation versity of New York, Binghamton. and stability condition of prey-predator interactions. Amer. Nat. G8lvin, K., 8nd S.K. W8weru. 1987. Variation in the energy and protein 97:209-223. content of milk consumed by nomadic pastoralists of northwest Kenya. &tdford, S. 1983. Management of Pastoral Development in the Third In: A.A.J. Hansen, H.T. Homlli and V.J. Quinn (eds.) Food and Nutti- World. John Wiley and Sons, New York. tion in Kenya. A Historical Review, UNICEF, Nairobi. Sinckir, A.R.E., 8nd J.M. Fryxell. 1985. The Sahael o Africa: ecology of a Glover, J. 1963. The elephant problem at Tsavo. E. Afr. Wildl. J. 1:30-39. disaster. Can. J. Zool. 63:987-994. Hardin, G. 1968. The tragedy of the commons. Science 162124348. Spencer, P. 1973. Nomads in Alliance. Symbiosis and Growth Among the HeRand, J. 1988. Five Essays on the Study of Pastoralists and the Devel- Rendille and Samburu of Kenya. Oxford University Press, London. opment of Pastoralism. Occasional paper No. 20. Sosialanthropologisk Stebbings, E.P. 1935. The encroaching Sahara. Geo. J. 86510. Institut, Universitetet i Bergen, Bergen, Norway. Stiles, D.N. 1983. Camel pastoralism and desertification in Northern Herskovits, MJ. 1926. The cattle complex in East Africa. American Kenya. Desert. Contr. 8:2-8. Anthropologist Vol. 28. Swift, J. 1977. Pastoral development in somalia: herding cooperatives as a Hobbs, N.T., 8nd D.M. Swift. 1985. Estimates of habitat carrying capacity strategy against desertification and famine. In: M.H. Glantz (ed.), Deser- incorporatingexplicit nutritionalconstraints. J. Wildl. Manage. 49:814-822. tification: Environmental Degradation in and Around Arid Lands, HoUing, C.S. 1973. Resilience and stability of ecological systems. Ann. Boulder Westview Press. Rev. Ecol. Syst. 4:1-23. Swift, J., 8nd A. MaBkI. 1984. A cooperative development experiment ILCA (Intem8tiotul Livestock Centre for Afriu). 1987. ILCA’s Strategy among nomadic herders in Niger. Overseas Development Institute, and Long-term Plan: A Summary. Addis Ababa, Ethiopia. paper I&. Jacobs, A.H. 1965. The traditional political organization of the pastoral VoIterra, V. 1928. Variations and fluctuations of the number of individuals Maasai. Unpublished Ph.D. Thesis, Oxford University. 427 pp. in animal species living together. Translated in R. N. Chapman 1931. Kuhn, T.S. 1970. The Structure of Scientific Revolutions, 2nd Ed. Univer- Animal Ecology. McGraw-Hill, New York. sity of Chicago Press. W8llmo, O.C., L.H. C8rpenter, W.L. Re8eIin, R.B. Gill, 8nd D.L. Baker. L8mphere, J. 1976. The Traditional History of the Jie. Oxford University 1977. Evaluation of deerhabitat ona nutritional basis. J. Range Manage. Press, London. 30:122-127. Lamprey, H.F. 1983. Pastoralism yesterday and today: the overgrazing Wiens, J.A. 1977. On competition and variable environments. Amer. Sci. problem. pp. 643666 In: F. Bourliere (ed.), Tropical Savannas: Ecosys- 65590-597. tems of the World, Vol. 13. Elsevier, Amsterdam. Wiens, J.A. 1984. On understanding a non-e.quilibrium world: myth and Lamprey, H.F., 8nd H. Yussuf. 1981. Pastoralism and desert encroachment reality in community patterns and processes. In: D.R. Strong, Jr., D. in northern Kenya. Ambio l&131-147. Simberloff, L.G. Abele, A.B. Thistle (eds.), Ecological Communities: Laws, R.M. 1970. Elephants as agents of habitat and landscape change in Conceptual Issues and the Evidence, Princeton University Press, Prin- East Africa. Oikos 21:1-15. ceton, New Jersey. Leslie, P.W., 8IId P.H. Fry. Extreme seasonality of births among South Western, D. 1973. Structure, dynamics and changes of the Amboseli Basin Turkana pastoralists. (submitted). ecosystem. Ph.D. Diss. Nairobi University. LusigI, W J. 1981. Combatting Desertitication and Rehabilitating Degraded Population Systems in Northern Kenya. Tech. Rep. A4, UNESCO- UNEP Integrated Project in Arid Lands, Nairobi.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 459 Plant responses to pine management and deferred-rotation grazing in north Florida

CLIFFORD E. LEWIS, GEORGE W. TANNER, AND W. STEPHEN TERRY

Abstract Responses of herbaceous and woody plants to combinations of 4 This study was designed to test 2 general hypotheses. First, pine management and 4 grazing management systems were tested understory species’ frequency of occurrence and woody species on a wet-flatwoods site in the pintwiregrass vegetation type of foliar cover are not altered by 4 timber management treatments. north Florida. Frequency of occurrence of herbaceous species and Secondly, implementation of 3 deferred-rotation grazing systems foiii cover of woody species were determined in natural stands of among the timber management treatments will not affect herbace- SO-year-old slash and iongleaf pine (Pinus elliottii Engelm. and P. ous species, woody species, and planted pine responses. We are pahstrisMill.) and compared to similar forest sites that were reporting the responses of herbaceous and woody vegetation to harvested and site prepared by double-chopping and not replanted this intensive type of multiple-use management. with slash pine, or replanted to 1,112 trees/ha in single- and Methods double-row configurations. In addition, these sites were ungrazed or grazed using 3 deferred-rotation systems. Prescribed burning in Study Area the natural stands increased occurrence of most herbs’and sthnu- The experimental area was located on 73 ha of the University of iated new species to occur, but had little effect on woody plant Florida’s Austin Cary Memorial Forest, Alachua County, Florida, composition. However, harvesting of pines and double-chopping on a typical flatwoods site of the pine-wiregrass vegetation type. resulted in the occurrence of many new herbaceous species and The area originally was occupied by a natural stand of 50-year-old increased occurrence of most initially present. Pineiand threeawn slash (pinus elliottii Engelm.) and longleaf pine (P. palustris Mill.) (histido strictu Michx.), the major herb, initially deereased in with a basal area of 20.7 m2 per ha. The area had been protected occurrence with intensive site disturbance. Six years after distur- from fire and other disturbances for many years until it was lightly bance, most herbaceous species were declining in occurrence. prescribed burned during the winter of 1975-76. Grazing or growth of replanted pines had little infiuence on occur- The climate of this area typically is subtropical and humid with rence of herbaceous species. Both burning and mechanical distur- long, hot summers and short, mild winters. The frost-free period bances initially reduced foiiar ground cover of most woody speck, averages 276 days. Rainfall averages about 140 cm per year, with however, few species were eliminated from the community. Most about half falling during June through September. woody species were recovering within 6 yr from treatment, but Soil on about 70% of the experimental area is a sandy, siliceous, succession was somewhat slower on mechanicaiiy treated areas. hyperthermic Ultic Haplaquod of the Pamona series, which is acid Survival and growth of planted pines were not affected by grazing, and contains an organic stain layer between 41 and 61 cm, with a nor did planting configuration affect pine growth. clay layer at about 109 cm. Basinger fine sand (a siliceous, hyperthermic Spodic Psammaquent) occurs at slightly lower elevations Key Words: pine-wiregrass vegetation, succession, mechanical and occupies about 15% of the area. Sparr sand (a loamy, siliceous, site disturbance, prescribed burning hyperthermic Grossarenic Paleudult) and Adamsville sand (a Management for simultaneous production of 2 or more pro- hyperthermic, uncoated Aquic Quartzipsamment) occurs at slightly ducts often creates conflicts and competition. Intensive site prepa- higher elevations on the remaining 15% of the area. ration and forest management practices in the South affect all Twentyeight l-ha test pastures were selected for timber and components of an ecosystem since both overstory and understory grazing treatments. The remaining area was divided into 3 holding vegetation are manipulated by harvesting, site preparing, and units for cattle when they were not grazing the test pastures. planting density. At the same time, use of livestock to harvest Twenty-two test pastures and the holding units were clearcut dur- native forages also can affect a number of resources. Research on ing the fall and winter of 1976-77, then were double-chopped with single-product management has yielded little information about a roller-drum chopper during August and September 1977. As is long-term effects of intensive timber or range management on each typically done in grazed forests of the South, the 6 pastures left in other or other forest resources. Several attempts have been made to natural stands were prescribed burned with fairly hot back-tires synthesize the fragmented literature into evaluations of multiple- during February 1978 and again in February 1981. A deer-proof use management (Lewis 1973,1975,1977; Pearson 1975,1980), but fence was built around the entire 73-ha study area in the fall of 1977 these attempts do not replace the need for actual research involving and 5 white-tailed deer (Odocoikus virginiunus Zimmerman) were multiple-use objectives. Past research has shown that intensity and placed on the area. Slash pine was planted during the winter of season of grazing affects planted pine survival along with produc- 1977-78. Interior pasture fences were built in the summer of 1978. tion and species composition of the forage resources (Halls et al. Cattle were placed on the area in March 1979 and grazed in test 1956, Hilmon et al. 1962, Pearson et al. 1971). Also, tree planting pastures from May 1979 until the study was discontinued in density, thinning practices, and length of timber rotations greatly December 1983. affect understory plants for livestock and wildlife use (Grelen et al. Treatments 1972, Lewis 1975, Pearson 1980, Wolters 1982). Deferred-rotation The study was a factorial experiment of 4 timber treatments and grazing is reputed to improve forage conditions, but has not been 4 grazing treatments in which 2 treatment combinations were not tested in pine-wiregrass vegetation. used (Table 1) in a randomized complete-block design with 2 Authors arc range scientist, USDA-Forest Service, associate professor, and range blocks. One block was somewhat drier than the other. Timber research biologist (formerly), Department of Wildlife and Range Sciences, University treatments were: (1) natural stands of 50-year-old pines; (2) clear- of Florida, GamaviUc 32611. Terry is currently land resources manager, Miccosukec Indian Reservation, Miami, Florida 33144. cut, chop, no planting of pines (open); (3) clearcut, chop, plant Manuscript accepted 2.4 May 1988. slash pine ( I,1 12 trees/ ha) at a spacing of 2.4 X 3.7 m (single rows);

460 JOURNAL OF RANGE MANAGEMENT41(6), November 1988 ing important trends, or being important cattle or wildlife foods. Table 1. Combinations of 4 timber conditions and 4 grazing conditions contained in the fmtorial experiment on the Auetin Gary Memorial Results and Discussion Forest. Natural Tree Stands Woody Plants Grazing treatments Cattle grazing had no significant impact on woody plant species Timber Graze (days) 0 45 30 30 in the natural tree stands (Table 3). It was apparent that species, treatments rest (months) 12 3 4 6 such as saw-palmetto [Serene repens (Bartr.) Small], gallberry [Z/es &bra (L.) A. Gray], and other shrubs that had dense stands Natural stand, 50 yr old pines X X X initially tended to maintain this condition regardless of burning or Site prep only- no pines planted X X X grazing. It should be noted that any significant differences between Site prep, single-row pines- 2.4 X 3.7 m x X X X Site prep, double-row pines- ungrazed and grazed treatments in 1976 and 1978 were the result of 11.2 X 2.4) 12.2 m X x. x X natural variation and not the result of grazing, since cattle were not placed in the test pastures until May 1979. Mature slash and longleaf pine crowns averaged 62% crown and (4) clearcut, chop, plant slash pine ( 1,112 trees/ ha) in a double- cover at the time of initial measurements in 1976. Hardwood tree row configuration at 1.2 X 2.4 m, with 12.2 m between double species contributed little cover. With burning, most of the hard- rows. Slash pine seedlings were from genetically selected seed that wood trees and shrubs were topkilled to ground-line but usually were nursery grown for about 1 yr. Grazing treatments compared resprouted. Conditions were similar in both grazed and ungrazed no-grazing with 3 deferred-rotation systems: (1) graze 45 days then treatments. Natural regeneration of pines under this mature can- rest the pasture for 3 months before grazing again; (2) graze 30 days opy, however, was minimal. with 4 months rest; and (3) graze 30 days with 6 months rest. The 14 Saw-palmetto and gallberry were the dominant shrubs prior to treatment combinations were randomly located in each block. and following treatment (Table 3). Burning in February of 1978 We used 2 herds of beef cows ina put-take system to simulate the and 1981 reduced saw-palmetto and gallberry foliar coverage in grazing treatments. The overall objective was to remove about 50% those years but, at the end of the study, coverage slightly exceeded of the available forage during each grazing period. About 5 anim- initial estimates in both ungrazed and grazed conditions. Huckle- als were placed in a pasture at the beginning of a period for a short berry (Guylwsuciu spp.) and runner oaks [Quercuspumilu Walt., time (2 or 3 days) then removed for several days. This would be Q. minima (Sarg.) Small] initially were the next most important repeated throughout a prescribed grazing period in an effort to shrubs, and they held similar positions at the end of the study. simulate the type of use that forage plants usually receive as cattle Practically all shrubby species reacted to fire with reductions move through an area and periodically graze the same location or occurring during the year of burning but recovery in the following plants. Under this system the cattle were used only to impose years. The only species that seemed intolerant of burning were grazing treatments on the vegetation and not as measures of animal dwarf pawpaw [Asiminu purvifloru (Michx.) Dunal] and St. responses to treatments. John’s wort (Hypericum spp.). Vines were a minor component of the understory woody cover- Measurements Three 30.5-m line transects were randomly placed in each test age throughout the study. Muscadine grape (Vitk rotundifoliu pasture at 45” angles to the tree row orientation. To determine Michx.) and greenbriers (Smilux spp.) were the most common and herbaceous species frequency of occurrence; each transect was are known to provide important wildlife food resources in the form divided into 100 30.5-cm segments. A species was tallied as present of fruit and browse (Harlow and Jones 1965). Vines responded to if any living part of the plant interdepted the transect. Results are fire in a manner similar to that of shrubs: they decreased the year of expressed as frequency of occurrence per 30.5-m transect. Crown burning with a gradual increase in cover between the tires. intercepts of all woody vegetation were measured to the nearest The responses of woody species in these natural stands to period- centimeter by species along the line transect. Plant measurements ic prescribed burning in the winter were similar to Langdon’s were made in 1976 before treatment and subsequently in the fall of (1981) report of plant persistence but reduction in size and crown 1978, 1980,1981, and 1983. coverage after 30 years of repeated burning. Reduction of woody plant cover due to livestock browsing apparently was not a signifi- Tree survival and growth was determined in each planted pasture by measuring 10 trees permanently located near the center of cant factor. Kalmbacher et al. (1984) reported diets of cattle on the each line transect. Total height, stem diameter, height to crown, south Florida range to be composed of 9 and 26% browse in summer and winter, respectively. diameter of crown, mortality, and damage by animal, disease, or insect were determined annually. Herbaceous Plants . The most common herbaceous plants were pineland threeawn Statlatieal Analyses Significance in analyses of variance was determined at the 0.05 (Aristidu strictu Michx.), grassleaf goldaster [Hetero- level. Single degree-of freedom contrasts tested for significance in thecu gruminifoliu (Michx.) Shinners], and bracken fern [Pteri- the timber treatments were natural tree stands vs. harvested/ dium uquillinum (L.) Kuhn.] (Table 2). Several species noticeably chopped, chopped/ not planted, vs. chopped/ planted, and single- increased following introduction of fire, eg., sensitive partridgepea row vs. double row configurations. Grazing treatments were tested (Cussiu nictituns L.) (the most common legume), eupatorium for: ungrazed vs. grazed, 45-day grazed vs. both 30day grazed (Euputorium spp.), and low-panicums (Dichunthelium spp.). Oth- treatments, and 30-4 vs. 30-6 grazing, ers that responded positively, probably more from increases in Statistical analyses of the data showed that most of the signifi- plant size rather than new plants, were pineland threeawn and cant differences in timber treatments were between the natural tree creeping bluestem (Schizuchyrium stoloniferum Nash). stands and the clearcut and site prepared stands. Most of the Grazing had little negative effect on plant frequencies of occur- significant differences due to grazing treatments were between ungrazed and grazed conditions. Therefore, the data were com- rence. The low-panicum grasses continued to increase with grazing bined for ungrazed and grazed treatments from both natural tree and following burning. Bracken fern was less frequent in grazed as pastures and mechanically disturbed pastures. Of the 142 herb and compared to ungrazed pastures, but this seemed to be the result of 68 woody plants encountered, species presented are those having an initially low frequency of occurrence rather than a response to significant differences, being prevalent in most treatments, show- grazing. In 1981 eupatoriums, legumes, and other forbs were signif-

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 461 Table 2. Frequency of occurrence of herbeceour plnnte end cover by woody vegetation in pestawe with ao overstory 01 SO-ycu-old MIMIIand longkaf pinea and an undemtory tbet was preecribed burned in Febrwry 1978 and 1981. Dote collected in September and October of year indkated!

Ungrazed pastures Grazed pastures2 Plant type and species 1976 1978 1980 1981 1983 1976 1978 1980 1981 1983 foliar cover (%) Shrubs Gaylussacia spp. 5.1 2.2 9.9 8.5 1.4 6.4 3.2 12.8 II.8 Ilex glabra (L.) Gray 10.2 12.8 11.6 17.0 X 19.8 12.2 2:.:* Lyoniu Iucti (Lam.) K. Koch 2.6 f % 6.6 ::: 4.5 0.6 1.6 1.0 1:4 Myrica cerifera L. 1.6 0.5 2.9 2.0 2.1 0.4 0:2 0.6 1.0 1.3 Quercus minima, pumiha Walt. 7.1 4.0 7.9 II.5 7.5 5.7 2.2 5.6 9.2 5.0 Rubus spp. 2.5’ 0.6. 0.8* 0.7* 3?z 0.17.8 3:; 2:: 31.80.1 Serenoa repens (Bartr.) Small 20.5 6.1 21.2 19.9 2!? Vaccinium myrsinites Lam. 3.9 ::: 1;:: 3.0 419 5:s 1.7 7:8 0:6 8.0 Other shrubs 6.9. 15.1* 16.0 1.9 0.6 3.1 1.3 2.9 Vines Smilax rotundifolia L. I.1 0.2 1.1 1.0 1.2 0.2 0.1 0.2 0.3 0.2 Vitis rotundtifolia Michx. 0.0 0.0 0.0 0.0 0.0 3.6 0.8 3.0 0.8 0.6 Other vines 0.0 0.0 0.2 0.1

Grasses Andropogon capiUipes Nash 0.3 0.0 0.8 0.5 1.0 1.6 1.1 I.6 I.1 I.6 Andropogon virginicus L. 1.3 2.0 3.8 3.8 3.8 3.2 5.8 1.9 Aristida stricta Michx. 17.2 19.7 26.8 2: 24.3 7.9 16.8 19.6 18.3 zo”:: Dichanthelium spp. 0.0 2.8 4.3 417 4.5 0.0 3.6 7.5 6.0 Panicum spp. 2.3 0.3 3.9 0.5 0.8 2.9 0.2 4.0 1.4 :3 Panicum verrucosum Muhl. 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.4 0:o Paspalum setaceum Michx. 0.0 1.0 0.5 0.0 0.0 0.0 0.5 0.4 0.2 0.0 Schizachyrium stoloniferum Nash 0.3 2.2 1.8 2.8 1.3 0.7 2.5 1.0 1.9 1.0 Other grasses 4.4 4.6 8.3 7.3 7.4 5.9 3.6 3.1 4.3 2.8 Grass-likes cvpenrs SPP. 0.3 0.3 0.2 0.0 0.1 0.5 0.0 0.0 0.0 Rhynchospora spp. 0.0 0.2 88 0.5 0.5 0.0 2.3 1.2 0.4 0.1 Scleria spp. 1.8 2.3 2:2 2.8 2.0 1.9 2.2 I.5 1.8 2.4 Other grass-likes 0.2 1.3 0.8 0.5 0.0 1.0 1.3 1.2 0.8 0.0 Forbs Aster spp. 0.0 I.5 0.7 1.3 0.9 2.8* 0.8 0.9 1.2 0.8 Diodia teres Walter 0.0 0.5 0.0 0.0 0.0 0.0 0.2 0.0 0.0 0.0 Eupatorium spp. 0.0 2.4 0.7 2.0’ 0.2 0.0 2.5 0.0 0.2 0.0 Heterotheca graminifolia (Michx.) Shiuners 9.2 5.3 10.5 10.0 19.8 3.1 5.2 6.5 6.8 7.7 Legumes (Fabaceae) 3.9 6.3 0.5 6.6’ 0.2 0.9 3.2 0.5 1.5 0.1 Rhexia spp. 0.2 0.6 0.5 0.0 1.4 0.9 0.1 Xyris spp. 0.0 0.0 0.0 0.0 0.2 0.0 :: :: 0.1 Other forbs 6.4 5.5 4.8’ 4.8 0.8 2.7 3:o 1:9 I.5 Ferns Pteridium aquilinum (L.) Kuhn 37.5* 47.0. 28.3’ 44.0’ 6.2 6.2 2.6 4.8 _- __ 26.0’ 6.2 Other ferns 0.0 0.3 0.8 0.8 0.0 1.1 2.4 2.1 3.6 2.8 ‘Values for ungrazed treatments arc averages from two pastureswhile values for grazed arc averages from 4 pastures with 3 line-transects in each. Means marked with an asterisk are signiiicantly larger than the corresponding mean for that year in the other grazed treatment. *Livestock grazing did not begin until 1979. Tanopy cowrage by 50-year-old longleaf and slash pines was estimated only in 1976 because treatments had no effect on pine canopy. icantly more frequent in ungrazed pastures but not in 1980 nor age (Table 3). Overstory harvest and site preparation almost elimi- 1983, the other years when cattle grazed the pastures. Since 1981 nated all woody cover, but the surviving individuals began a slow was extremely dry during the spring, these species may have been recovery. Water oak (mercus nigru L.) seemed to recover most grazed more heavily than in other years. quickly. Planted slash pine eventually provided the greatest coverage among tree species. Naturally regenerating longleaf pine occurred in both grazed and ungrazed pastures. By the end of the Harvested and Site Prepared Stands Woody Plants study, tree canopy was dominated by pines and oaks. Trees, other than pines, did not initially contribute much cover-

462 JOURNAL OF RANGE MANAGEMENT 41(6), November 1666 Table 3. Frequency of occurrence of berbaceoua plants cod cover by woody vegetation in putura that were ckucut end doubk-cbopped III19n; rvmgu of phted end unplanted peetum. Date coUected In September 8nd October of year indicated.1

Ungraxed pastures Grazed pastures2 Plant type and species 1976 1978 1980 1981 1983 1976 1978 1980 1981 1983

Shrubs GaylussacicrSQQ. 5.4 0.7+ 4.3 6.6. 2.0 4.4 0.4 2.8 2.3 1.0 Ilex ghsbra (L) Gray 10.6 0.9 9.0 10.6 19.0 12.7 0.9 10.4 9.6 16.4 Lvonia lucida (Lam.) K. Koch 0.6 0.4 1.2 1.7 1.8 1.7 1.1 5.1 3.3 tiyrica cergera I.,. ’ 0.7 co. 1 2.1 7.0 0.6 0.1 3.5 6.7 Q4ercu.s minima, pumiht Walt. 3.3 0.5 2.4 ;: 1.3 2.5 0.4 2.1 1.7 1.8 Rubus SQQ. 0.3 1.0. 5.2 413 4.6 0.1 0.4 5.3 5.8 3.7 Serenoa repens (Bark) Small 31.6 1.5 7.4 8.9 12.9 27.2 1.0 6.8 7.4 10.7 Vaccinium myrsinites Lam. ::: 0.32.1 6.0,1.9 1.8 2.1 1.9 0.2 1.5 1.2 1.6 Other shrubs 2.9 2.8 4.6 2.0 5.4 4.4 4.3 Vines Smilox rotundijolia L. 0.2 0.2 0.7 0.6 0.5 2.0 1.7 2.0 Vitis rotundifolio Michx. 2.9 0.1 0.4 0.1 0.2 0.6 0.6 0.9 Other vines 0.6 co. 1 2.4 2.3 3.3 4.1 4.7 8.7 Trees Pinus elliottii Engelm. 0.0 0.1 4.3’ 6.2 14.8 0.0 0.1 2.6 PinUS SQQ.' 65.4 - - 59.3 - - 4.7 13.4 Quercus nigra L. 7.2 0.4 3.4 2.6 8.0 4.5 0.2 1.5 2.2 4.9 Quercus spp. 1.6’ 0.0 0.7 0.8 0.1 0.1 0.5 1.3 1.1 Other trees 3.3 0.1 2.9 5.9 3.1 0.6 2.6 2.6 4.2

Grasses Andropogon capillipes Nash 0.4 5.3 43.2 46.1 44.3 0.3 3.2 10.8 34.4 Andropogon virgirticus L. 0.5 3.1 15.6 12.7 10.3 1.9 1.9 23.4 20.6+ g;. Aristida stricta Michx. 8.3 5.3 9.2 11.2 13.7 7.7 2.7 4.7 10:1 Dichanthelittm spp. 44.9 7.1 16.1 14.5 0.0 55.9. 25.4* 3% 37.9. Panicum spp. 84 14.0 19.0 3.3 7.1 1.0 23.0 20.0 14.8 13.5 Panicurn verrucosum Muhl. 0:o 8.3 1.0 0.3 0.0 0.0 2.8 2.1 0.0 Paspalum setaceum Michx. 0.1 2.6 2.1 2.7 0.8 0.0 i:: 4.4 2.5 0.3 Schizachyrium stolontj-erum Nash 1.2 0.6 1.8 4.4 4.7’ 0.3 0.7 1.1 2.0 1.3 Other grasses 2.1 5.0 6.4 5.4 6.7 3.7 4.3 9.7 7.7 8.4 Grass-Likes cvpenrs SPP. 0.4 11.4 0.2 0.8 0.0 0.3 6.9 2.7 1.4 0.3 Rhychospora SQQ. 0.0 9.2 4.8 1.1 0.2 0.0 8.7 7.2 3.1 3.6+ Scleria spp. 0.5 7.9 5.1 4.5 5.2 0.9 13.6 9.4 6.7 Other grass-likes 0.3 9.3 3.3 1.2 1.1 2.5 7.5 5.0 ::: 2.2 Forbs Aster spp. 2.0 1.4 3.6 1.7 1.4 4.5 1.7 0.9 1.5 Diodia teres Walter 0.3 17.6 0.4 0.1 0.0 1:.: 1.3 0.2 0.1 Eupatorium spp. 0.0 3.0 :2 1.8 1.1 0.0 2:6 3.3 2.9 2.1 Heterotheca graminifolia (Michx.) Shinners 3.4 4.2 6.8 6.0 14.1. 2.0 1.5 3.2 2.4 Legumes (Fabaceae) 0.5 3.4 1.1 1.1 0.1 2.4 4.7 2.3 z Rhexia spp. 0.4 1.7 3.3 2.8 0.3 0.1 1.2 5.3 :.; 0:5 Xyris spp. 0.3 3.6 2.5 4.2 1.7 0.5 2.6 5.2 3:2 1.2 Other forbs 1.1 11.6 11.3 7.1 4.8 2.6 3.7 17.9 12.5 9.6. Ferns Pteridium aquilinum (L.) Kuhn 21.4 27.0 17.8 22.9 23.4 17.2 25.4 17.8 19.9 18.6 Other ferns 1.4 1.4 1.1 1.7 2.6 1.7 1.7 2.4 2.0 1.1 1Values for ungmzed treatments arc rwwages from 6 pastums while values for grazed arc averagea from 16 pasturct with 3 tine-transects in each. Means marked with an asterisk a~ significantly larger than the corrcaponding mean for that year in the other grazed condition. ‘Lwcstock grazing did not begin until 1979. Tanopy coverage by SO-year-old longkaf and slash pines was estimated only in 1976.

Saw-palmetto, the dominant shrub prior to treatment, was provided greater coverage at the end of study than initially. Black- greatly reduced by double-chopping. However, surviving saw- berry (R&US spp.), fetterbush [Lyoniu lucidu (Lam.) K. Koch], palmetto plants increased in cover annually, but still provided less and southern waxmyrtle (Myrica certjka L.), species that gener- than half the original cover after 6 years. Runner oak and huckla ally increase with disturbance, had greater coverage at the end than berry were reduced by chopping and, like saw-palmetto, never at the beginning of study. Most of the minor shrubs remained regained their former degree of cover. Gallberry, the second most unchanged after the initial reduction from chopping. dominant shrub, was reduced by chopping but recovered rapidly. Before treatment, muscadine grape had greater foliar cover than This plant root-sprouts prolifically on chopped areas and, thereby, all other vines. Reductions in vine coverage with disturbance was

JOURNAL OF RANGE MANAGEMENT 41(S). November 1666 463 followed by some increases for most species. Greatest recovery was Table 4. Influence of grazing treatments on average survival, height, by other vines, which were primarily yellow jessamine [ Gelsemium diameter (DBH), height to first living limb (crown base), and crown sempervirens (L.) Aiton f.], a species poisonous to livestock. diameter of planted slash pine after 6 growing seasons. Grazing began in Grape, greenbrier, and yellow jessamine are known to be preferred May of the second growing aeason.’ foods for white-tailed deer (Harlow and Jones l%S). The responses of woody plants to mechanical disturbance by clearcutting and chopping were similar to those reported by Conde Grazing Crown Crown treatment2 Survival Height et al. (1983) on a similar site. Most species persisted and began to Diameter base diameter regain canopy cover over time following the initial reduction from +?& -m- -cm- -m_- -m-- chopping. In south Florida, similar trends were noted, especially O-12 91. 4.1 6.1 1.2 1.5 for saw-palmetto (Lewis 1970, Moore 1974, Tanner et al. 1988). 30-4 92’ 4.3 7.0 1.1 1.7 Grazing tended to reduce the cover of some shrubs such as 30-6 96’ 3.9 5.8 1.3 1.4 ground blueberry (Vaccinium myrsinites Lam.), runner oak, and 45-3 80b 4.0 6.3 1.2 1.5 saw-palmetto, but the only statistically significant reduction was ‘Means within a column for any a that are marked with differing superscri ts are on huckleberry in 1981. Although cattle will browse on shrubs significantly different at the 0.0 !e level. Unmarked means are not SL‘gnifLntly different. (Kalmbacher et al. 1984) and young hardwoods, they usually do *The deferred-rotation grazing treatments are identified by the number of days grazed not contribute much to cattle diets. Thill(1984) found very few followed by the number of months rrsted before being grazed again, e.g., 30-4 woody plant species to comprise,more than 5% of both cattle and indicates a 3day grazing period followed by 4-months rest. deer diets on Louisiana pine-hardwood sites. Herbaceous Plants grazed for 45 days and rested for 3 months following the 6th Pretreatment measurements indicated that pineland three- growing season (Table 4), the additional mortality of trees that awn, grassleaf goldaster, and bracken fern were the most common year was not caused by cattle. Rather, these tree deaths were caused herbaceous plants in the understory on these pastures (Table 3), by permanent stunting after 80% needle defoliation by an unknown similar to the previously discussed natural tree stands. Following insect during the second growing season and by suppression from a clearcutting and site preparation, a heavy influx of new herbaceous dense stand of over-topping brush, primarily gallberry. species was measured, particularly among the forbs and grass- Tree growth was not affected in any year by any deferred- likes. Many species, such as eupatorium persisted throughout the rotation grazing treatment (Table 4). Heights averaged 4.1 m at study, particularly where grazing kept a more open condition. plantation age 6 while diameters averaged 6.3 cm. These growth After an increase immediately following disturbance, a trend rates were similar to other plantations on flatwood sites of this toward decreasing occurrence was apparent in legumes, flatsedge region (Lewis et al.. 1985). Also, crown development was not (Cyperus spp.), maidenhair sedge (Eleocharis vivipara Link), affected by grazing treatments. Although cattle sometimes break sloughgrass (Scleria spp.), poor joe (Diodia teres Walter), and lower limbs by rubbing on stems, this did not change the height to eupatorium similar to old-field succession. Although grassleaf the lowest living branch. At the end of the 4th growing season, goldaster maintained its frequency of occurrence under grazing, it when pines were at a height to be particularly susceptible to this increased in ungrazed pastures to a level of being significantly form of injury, only 0.2% of the trees were noted to have limb greater than in grazed pastdres after 6 years. Occurrence of other breakage, 0.9% were leaning due to rubbing attributed to cattle, forbs decreased after the initial surge following disturbance and no and 0.2% had partial girdling of the stem from hoof action. During grazing, while occurrence in grazed pastures remained high and the 6th year, the only other year with cattle-related damage, 0.2% significantly greater in 1983. Rhynchospora (Rhynchospora spp.) were found to be leaning from rubbing. Although limb breakage responded in a similar manner. Bracken fern remained quite con- could influence crown spread, crown diameters were similar on stant following site disturbance and was not affected by grazing. grazed and ungrazed treatments in all years. Grasses, such as chalky bluestem (Andropogon capillipes Nash), Injury to planted pines can come from several sources: fire, broomsedge bluestem (Andropogon virginicus L.), and low- atimals, insects, and diseases. Insects were the primary form of panicums that are important foods for cattle, increased injury in this study: pine tipmoth (Rhyacionia spp.) were active in greatly following mechanical disturbance. Responses in both the early years infesting 14.1,2.3, and 0.2% of the trees during the ungrazed and grazed pastures were similar initially and little harm- 2nd, 3rd, and 4th seasons, respectively. Southern fusiform rust ful impact after grazing began in 1979 was observable on most [Cronartium quercum (Berk.) Miyabe ex Shirai f. sp. fusiforme species. Broomsedge and low-panicums continued to increase and (Burdsall and Snow)] was noted on 0.7% of the trees after the 4th attained significantly higher amounts with grazing. Occurrence of season and had increased to 5.1% during the 6th season. Pitch creeping bluestem, a very desirable cattle food, typically is stimu- canker (Fusarium moniliforme Sheld. var. subglutinans Wollenw. lated by site disturbance associated with chopping (Yarlett 1965). and Reink.) had infested 1.6% of the trees during the 6th season. In this study creeping bluestem occurrence also increased and was Overall, cattle were a minor source of injury to the pines and significantly greater in ungrazed than in grazed pastures by the end probably did not affect survival or growth considering the minor of the study. amount or type of injury (Lewis 198Oa, 1980b). Pineland threeawn is a relatively undesirable forage grass Overall, planting pines in double-rows with wide spaces between, because of its low nutrient content and poor digestibility (Lewis et as compared to the normal, single-row configuration, did not al. 1975). This plant suffers considerable mortality when chopped influence tree growth. Although height to the first living branch or disked but is seldom eliminated. Although pineland threeawn was significantly greater for pines planted in single rows after 3 frequency of occurrence on site prepared areas was approximately growing seasons, differences were not significant in other years. one-half of natural stands, slow, but steady increases occurred These responses to planting configuration follow those of other throughout the study. Under the open, site prepared conditions, studies reported by Lewis et al. (1985). increases in crown spread of surviving plants accounted for its increase in occurrence through time. Conclusions Planted Pine Responses to Grazing Site preparation for planting pines is intended to reduce early Grazing per se and deferred-rotation treatments had no sign& competition to assure good establishment and rapid early growth. cant influence on any of the parameters measured on planted pines Since neither burning, clearcutting and site preparing, not grazing through age 5. Although survival was significantly less in pastures had a major or lasting affect on species comp&tion and foliar cover of most woody plants, these management practices can be

464 JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 continued in the pine flatwoods without harmful effects on these Halls, L.K., O.M. HaIc, 8nd B.L. Soothwell. 19%. Grazing capacity of components of the environment. wiregrass-pine ranges of Georgia. Georgia Agr. Exp. Sta. Tech. Bull. Reintroduction of prescribed burning within natural stands N.S. 2,38 p. increased the occurrence of most grasses and forbs and allowed HuIow, R.F., 8nd F.K. Jones. 1965. The white-tailed deer in Florida. Fla. new species to occur. Periodic tire helped maintain the presence of Game and Fresh Water Fish Comm. Tech. Bull. No. 9.240 pp. most species and, therefore, is recognized as being desirable for Hilmon, J.B., C.E. Lewis, 8nd J.E. Bethune. 1962. Highlights of recent maintaining a vigorous, diverse understory for forage and wildlife results of range research in southern Florida. Proc. Sot. Amer. Forest. habitat. 196273-76. With mechanical disturbance, many new herbaceous species Kdmb8cher, R.S., K.R. Long, M.K. Johnson, 8nd F.G. M8rtin. 1984. Botanical composition of diets of cattle grazing south Florida range. J. appeared and most other species increased in occurrence, while Range Manage. 373334-340. pineland threeawn decreased. The general trend of most plants was Lmgdon, O.G. 1981. Some effects of prescribed fire on understory in to increase immediately following disturbance and then to slowly loblolly pine stands. P. 143-153. In: Prescribed Fire and Wildlife in decrease over time as ecological succession proceeded. However, Southern Forests Symp. Proc. Clemson Univ. some very desirable forage species persisted or increased, which Lewis, C.E. 1970. Responses to chopping and rock phosphate on south resulted in an improved herbaceous strata for livestock grazing. Florida ranaes. J. Range Manaxe. 23276-282. Lewis et al. (1984) reported an increase in forage yields following Lewis, C.E. 1608. Simulated cattle injury to planted slash pine: Combina- site preparation. tions of defoliation, browsing, and trampling. J. Range Manage. 33:340-345. Planted slash pine had little influence on occurrence of herbace- Lewis, C.E. 198Ob.Simulated cattle injury to planted slash pine: Girdling. ous species through plantation age 6 years. However, crown clo- J. Range Manage. 33:337-340. sure was incomplete in all stands, especially in the double-row Lewis, C.E. 1973. Integrating management of forest and range resources. P. configuration. In contrast, both tree and brush canopies tended to 69-78. In: Range Resources of the Southeastern United States. Amer. keep occurrence of herbaceous species low in the natural stands. Sot. Agron. Spec. Pub. 21. The biggest changes resulted from tree harvest and site preparation Lewis, C.E. 1975. Grazing considerations in managing young pines. P. which initially removed most woody plants. Their gradual return 160-170. In: Management of Young Pines Symp. Proc., 1974. USDA was accompanied by declines in some herbaceous species. These Forest Serv., Southern and Southeastern Forest Exp. Sta. and Southern Area State and Private Forestry. successional trends are explored more fully in a companion article Lewis, C.E. 1977. Principles and status of integrated management from the (Lewis et al. 1988). range viewpoint. P. 26-34 In: Mutual Opportunities for Forest, Range, Compared to ungrazed areas, the effect of these deferred- Wildlife Management Symp. Sch. Forest Resour. and Comet-v. Rep. 4. rotation grazing systems exerted little influence on most herbace- Lewis, C.E., R.S. Lowrey, W.G. Monwn, 8nd F.E. Knox. 1975. Seasonal ous species. Apparently, the rest periods following grazing allowed trends in nutrients and cattle digestibility of forage on pine-wiregrass most species to maintain a satisfactory level of vigor. Also, tram- range. J. Anim. Sci. 41:208-212. plingdisturbance of the soil and keeping the understory more open Lewis, C.E., B.F. Swindel, L.F. Conde, 8nd J.E. Smith.1984. Forage yields through removal of plant material by cattle may maintain more improved by site preparation in pine flatwoods of north Florida. South species than occurs with total protection from grazing. In ungrazed J. Appl. Forest. 8:181-185. treatments, the lack of defoliation by grazing or prescribed tire Lewis, C.E., B.F. SwIndel, and G.W. Tanner. 19%8.Species diversity: appeared to cause some species to stagnate and decline in vigor. Concept, measurement, and application to timber and grazing measurement. J. Range Manage. 41469472. This integrated pine-cattle management study shows that multi- Lewis, C.E., G.W. T8Mer, 8nd W.S. Terry. 1985. Double vs. single-row ple products can be simultaneously produced without major pine plantations forwood and forage production. South. J. Appl. Forest. impact on each other. This lends validity to conclusions drawn 955-61. from synthesis efforts to evaluate combining cattle, timber, and Moore, W.H. 1974. Some effects of chopping saw-palmetto-pineland wildlife management. It appears that neither site preparation, threeawn range in south Florida. J. Range-Manage. 27:101-104. burning, nor proper grazing has an extremely harmful nor long- Peuaon. H.A. 1975. Ranue and wildlife onnortunities. P. 19-27. In: Man- lasting effect on woody and herbaceous species or on planted slash agement of Young Pines Symp. Proc. USbA Forest Serv. Southern and pine. Regulating grazing intensity and reducing pine density or Southeastern Forest Exp. Sta. and Southern Area State and Private Forestry. crown coverage through planting configuration will maintain good Person, H.A. 1980. Livestock in multiple-use management of southern forage resources, wildlife habitat, and wood yields. forest range. P. 75-88 In: Southern Forest Range and Pasture Sympo- sium. Winrock Internat. Literature Cited Peuson, H.A., L.B. Wkhker, 8nd V.L. Duv8U. 1971. Slash pine regeneration under regulated grazing. J. Forest. 69744-746. Conde, L.F., B.F. Swindel, 8nd J.E. Smith. 1983. Plant species cover, TIMer, G.W., W.S. Terry, 8nd R.S. K8lmb8eher. 1988. Mechanical brush frequency, and biomass: Early responses to clearcutting, chopping, and control on south Florida ranges. J. Range Manage. 41:245-248. bedding in Pinus elliottii flatwoods. Forest Ecol. Manage. 6307317. Thill, R.E. 1984. Deer and cattle diets on Louisiana pine-hardwood sites. J. Grelen, H.E., L.B. WhBaker, and R.E. Lohrey. 1972. Herbage response to Wildl. Manage. 48:788-798. precommercial thinning in direct-seeded slash pine. J. Range Manage. Wolters, G.L. 1982. Longleaf and slash pine decreases herbage production 25435437. and alters herbage composition. J. Range Manage. 35:761-763. Yarlett, L.L. 1965. Control of saw palmetto and recovery of native grasses. J. Range Manage. 18:344-345.

JOURNAL OF RANGE MANAGEMENT 41 (S), November 1988 485 Species diversity and diversity profiles: Concept, measurement, and application to timber and range management

CLIFFORDE. LEWIS,BENEE F. SWINDEL,AND GEORGEW. TANNER

AbstnCt The concepta 8nd m of sevcnl diversity 8ssessments 8re pre- diversity indices and comparative diversity profiles. sented 8nd 8ppiied to 8 pnctierl situ8tion. Burning, mech8niul Analytical Methods methods of site prepurtion, rad c8ttie pizing ire common dis- turb8nces in forest8of the South. Their influence on pi8nt diversity Patil and Taillie (1982) showed that species diversity should be indices ue exrmined in 8 iongle8f-si8sb pine forat of north Flor- defined as average (community) species rarity and that familiar id8. Species richness, Shnnon’s index, md Simpson’s index indices each measure average rarity. Thus, if we denote showed incrwes in diversity shortly following burning md site Bi = abundance of i* species (i = 1,2. “’ , s) in a community prepu8tion 8nd 8 trend towed pre-tre8tment conditions 8fter 6 containing s species, ye8rs. Deferred-rot8tion gr8zing systems hrd no infkence. Corn- S pwtive diversity profiles showed simik trends but were more pi = ai/ F1 ai = proportional abundance, and infonnrtive by providing both qaulit8tive md qu8ntit8tive infor- R(p) = a measure of rarity based on proportional abundance, mrtion. These technique8 8re us&11 for newsing community reaponeee to m8n8gement prrctic* th8t is, they 8re effective then average species rarity for the community is simply methods for underst8nding the imp8cts of forest nunrrgement 8nd A=.$R(Pi). nnge management pnctices on pleat community structureand succewion. Now taking R(p) to be, respectively, Key Work species richness, Sh8nnon’s diversity index, Simp R&i) q (1 - Pi)/Pi* son’s diversity index, compurtive diversity profiies, phnt succes- R&pi) = -log pi, and sion, pine-wiregrmssvegetrtion Rd(pi)=l-Pi Literature on the mcasurcment of specks diversity is vohnni- and computing average rarity, we have nous. Many indices have been proposed, and there has been much CPi&(n)=A,=S-1 = species count, debate concerning their use and meaning. Analogous measurements appear in ecology, genetics, linguistics, information theory, Z pi R,I, (pi) = ha = - 8, pi log pi q Shannon’s index, and and economics. Fortunately, significant theoretical advances have recently appeared in the ecological literature (e.g. Grassle et al. Xpi&(pi)=A,=l-$lP: = Simpson’s index. 1979). Patil and Taillie (1982) proposed a compelling definition of Whittaker (1975) did diversity analyses via importance-value species diversity and showed unified, intuitive motivations for the curves obtained by ranking the proportional abundances so that most popular indices along with their relatedness to measures of p1 L pz 1 ... 2 pi I “’ 2 PO rarity. These authors and others (e.g. Solomon 1979, Taillie 1979) VI introduced concepts of intrinsic diversity orderings. When intrinsic and plotting these pi against rank order, i.e., plotting the points orderings exist between 2 communities, any diversity index will (1, Pl) agree with the indicated ordering, so these orderings are index free. (2. pz) Some forest and range management practices (especially clear- (39Pa) cut harvesting with intensive site-preparation and range improve- . . . ment through chopping or discing) conjure vivid images of ecosys- Comparative diversity profiles (Taillie 1979, Patil and Taillie tem destruction. Hence, many conservation and environmental 1982, Swindel et al. 1987) compare diversities in 2 communities groups have suggested a need for assessment of the effects of such based on plottings of left tail sums of ranked proportional abun- practices on species diversity. Therefore, we need better tools for dances from 2 communities against each other. Given 1 community assessing and presenting the impacts of management practices on described by data [l] and another described by community structure. Indeed, the National Forest Management Act of the United d>pl1...1pf>...IR’ PI States [Federal Register 44(181), 219.13(6)] requires that managed (the k* tail sums are then pl + ps + ... pk and p’ + pI + ... + p’ for practices maintain the diversity of forest ecosystems as demon- community [l] and [2], respectively), the comparative diversity strated by quantitative comparisons of the diversities of natural profile consists of the points and managed forests. The Act is based not only on aesthetic preferences, but on sound biological principles as well-for exam- (Pb r.47 ple, a diverse community represents a larger gene pool, and is less (Pl+ P%P; + I4 (Pl+ P2+ PS,ti + pi + Pa susceptible to devastation by catastrophic events such as pest .. . epidemics. In this paper we analyze recent data from integrated pine- If these points all plot below the diagonal of a unit square then grazing management research to determine trends in species occur- the second community is intrinsically more diverse (Patil and rence, cover, and certain treatment effects on such trends qsing Taillie 1982), and any diversity index will agree with this ordering. (Points all above the diagonal imply the converse, index-free, Authon are range scientist and nsearch forcster, USDA-Forest Service, and asso- diversity ordering.) If the profile crosses the diagonal, then the 2 ciate professor, Department of Wildlife and Range Sciences, University of Florida, communities are not intrinsically ordered, and different diversity GaincsviUe 3261 I. Manuscript accepted 24 May 1988. indices can then disagree on the diversity ordering between the 2 communities.

466 JOURNAL OF RANGE MANAGEMENT 41(6), November 1968 Effects of Pine Management and Grazing Management were combined. To illustrate the overall results, data from the on Species Diversity chopped-only treatment for frequency of occurrence of herbaceous species were used to compare diversity (Fig. 1). It is evident that Study Site herbaceous species diversity increased following site disturbance The experimental area was on 73 ha of the University of Flori- (for about 4 years) then declined. These same trends occurred da’s Austin Cary Memorial Forest, northeast of Gainesville, Flor- whether disturbed by prescribed fire; harvest and chopping; har- ida, which was occupied by a natural stand of 50-year-old slash vest, chopping, and normal planting; or harvest, chopping, and pine (pinus elliottii Engelm.) and longleaf pine (P. pulusrrti Mill.). double-row planting of pines. This trend is in agreement with The understory was dominated by common gallberry [Zlex glabra (L.) A. Gray)] and saw-palmetto [Serenoa repens (Bartr.) Small] with few herbaceous plants. Bracken fern [Pteridium aquilinum (L.) Kuhn.)] and pineland threeawn (Aristidu stricta Michx.) were the most common herbs. Various bluestem grasses (Andropogon and Schizachyrium spp.), panicum grasses (Panic-urn spp.), low- panicum grasses (Dichanthelium spp.), lopside indiangrass [Sorghas- trum secundum (Ell.) Nash], and a number of composites occurred sporadically over the area. The climate is subtropical and humid with long, hot summers and short, mild winters. The frost-free period averages 276 days. Annual rainfall is 140 cm with half falling during June through September. Soils on most of the area are siliceous, hyperthermic sands. The acidic sands contain an organic stain layer between 41 to 61 cm and IO - I a clay layer at about 109 cm. r”--- Twenty-eight l-ha test pastures were selected for treatments * /” (Lewis et al. 1988). The remaining area was divided into 3 holding 0’ 1 * I a.0 units for cattle when they were not grazing the test pastures. 76 76 60 61 63 Twenty-two test pastures and the holding units were clear-cut UNGRAZED GRAZED during the fall and winter of 1976-77, then were double

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 467 0.92 0.94 0.96 0.98 I \ ’

0.92

Fig. 3. Diversity profiles comparing post-treatment communities to the pretreatment community for occurrence of herbaceous species in the clearcut. chopped, and not-replanted treatment. The upper right-hand corner is shown at IOX magnification.

0.0 02 0.4 06 0:s 1.0

CUMULATIVE PROPORTIONAL ABUNDANCE: PM-TREATMENT

Fig. 4. Diversity profiles comparing post-treatment communities to thepretreatment communityfor crown cover of woody plants in the treatments that were clearcut and chopped. The upper right-hand comer is shown at 10X magnifcation.

-010 Oi 014 0;6 0.6 1.0

CUMULATIVE PROPORTIONAL ABUNDANCE: PRE-TREATMENT similar to those in Figures 1 and 2. ments are sometimes thought to produce only monocultures of Diversity changes were less pronounced following fire than fol- slash pine plantations and are less widely recognized as methods lowing substantial mechanical site disturbance. The former treat- for improving wildlife habitat quality and forage production. ment has long been recognized in the South as a tool for both However, these results and others (Lewis et al. 1984, Swindel et al. forage and wildlife habitat improvement (widely deemed to be 1987) indicate improvement in wildlife habitat and forage yields positively correlated with increased diversity). The latter treat- are common.

468 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1988 Effects on Diversity Profiles determining trends following disturbance. Our evaluations indi- Comparative diversity profiles for herbaceous vegetation (Fig. cate that Shannon’s index is especially useful in describing ecologi- 3) and woody vegetation (Fig. 4) used the pretreatment community (1976) as a baseline to which post-treatment communities were cal trends. This is probably because it adequately includes both species count and evenness. compared for 1978, 1980, and 1983. The most obvious feature is that all 3 comparative profiles plot below the diagonal. This shows Diversity profiles are even more useful for investigating ecological conditions because they not only illustrate intrinsic diversity that post-treatment plant communities were intrinsically more diverse than the plant communities prior to disturbance. orderings, they also illustrate other important information about community structure. For example, diversity profiles indicate the Temporal trends in diversity are also evident. For occurrence of relative contribution of abundant species and the relative contribu- herbaceous plants (Fig. 3), the major increase in diversity occurred tion of rarer species to diversity. At the same time they can show within a year after disturbance (by 1978) but continued to increase successional trends of a community over time as affected by over the next 2 years. However, by 6 years after disturbance there abundant and rare species. was a return of diversity toward the pretreatment condition (Figs. 1 All of these diversity assessments possess the qualities of being and 2). Trends in diversity of cover for woody species (Fig. 4) give unaffected by size (area) of the community or by the size of the additional information on the community response when the more plants. They are also indifferent to species’names while being only common species (left-tail) and rare species (right-tail) crossed over concerned with the length of species lists and equitablity of propor- in the relative ordering. In other words, the rare species contrib- tional abundance. These characteristics make them especially use- uted the most to diversity of the community in 1978 while there was ful as measures of diversity in meeting requirements of the an additional increase in the diversity of common species in 1980. National Forest Management Act. Information on other ecological properties of communities can Conservationists and wildlife biologists should be especially be derived from comparative diversity profiles as compared to interested in the use of diversity profiles since the right-tails of these diversity indices. Specifically, the left-tail points indicate the con- profiles present information on the rare species of the community. tributions of common species while the right-tail points indicate Rare species influence the perpetuation of some desirable com- the proportion of rare species. For example, by counting the tick munity components more than do abundant species and are fre- marks along the x-axis in Figure 3, it can be seen that only 2 quently suggested as being required in order to maintain biological herbaceous species contributed 80% of the abundance and that 19 diversity. Although these diversity assessments are indifferent to species comprised the remaining 2% of the community in 1976 plant names, graphical methods for examining individual species (base-line). Then by counting the number of points along the y-axis responses to treatment are available (Moore et al. 1982a, 1982b). in any year, the number of common species contributing to this Our empirical study of the influence of disturbance either by 80% is seen to be 6 in 1978 and 8 in both 1980 and 1983. It is burning or by harvesting and site preparation in pine-wiregrass impossible to determine the exact number of species in the rare vegetation revealed that diversity increased for 2 to 4 years then category (right-tail) because of .the convergence of points. How- began to return to pre-treatment levels. Scarce herbaceous species ever, if the number of species (richness) in the original data set is increased rapidly following disturbance but changes in the number known, it is easy to determine the number of species in the rare of abundant species were limited. By 6 years following disturbance group. For this data set it was 19,37,39, and 32 in 1976,1978,1980, all components were moving slowly toward pre-treatment condi- and 1983, respectively. These procedures applied to Figure 4 tions. These results show that common timber management and showed there were 3,8,9, and 7 common woody species and 2516, grazing management practices do not decrease species diversity. 21, and 22 rare woody species in 1976,1978,1980, and 1983. Since Rather, these disturbances can increase or help maintain diversity there was no major invasion of new woody species, as occurred in southern forests. This knowledge alleviates much of the concern with herbaceous species, it becomes clear that the increase in about environmental degradation. Diversity profiles are especially diversity was the result of transferring abundancy among the exist- useful for presenting these results. inn snecies. ‘n;e diversity profiles also reveal interesting information about Literature Cited extremely rare species. By drawing a vertical line from the 0.99 tick Crassle, J.F., G.P. Paul, W. Smith, and C. Taiiiic (ed.) 1979. Ecological mark on the horizontal axis, the contribution of extremely rare diversity in theory and practice. Int. Coop. Pubi. House, Fairland, species to percent occurrence by herbaceous species and percent Maryland. cover by woody species can be estimated. The point at which this Lewis, C.E., C.W. Tanner, and W.S. Terry. 1988. Plant response to pine vertical line intercepts the diversity profile indicates the relative management and deferred-rotation grazing in north Florida. J. Range contribution of extremely rare species. This is best illustrated by Manage. 41:463468. the woody plants (Fig. 4). Where the cumulative proportional Lewis, C.E., B.F. Swindei, L.F. Conde, and J.E. Smith. 1984. Forage yields contribution of extremely rare species was 1% in 1976, it increased improved by site preparation in pine flatwoods of north Florida. South. J. Appi. Forest. 8:181-185. to 5.2% in 1978. Thereafter, it decreased to 3.1% and 2.8% in 1980 Moore, W.H., B.F. Swindei, and W.S. Terry. 1982a. Vegetative response to and 1983, respectively. The number of species in these extremely clearcutting and chopping in a north Florida flatwoods forest. J. Range rare categories can be roughly estimated by counting the number of Manage. 353214218. points along each profile to the right of the vertical line. However, Moore, W.H., B.F. Swindei, and W.S. Terry. 1982b. Vegetative response this becomes an impossible task when the points fall close together, to prescribed tire in a north Florida flatwoods forest. J. Range Manage. as in Figure 3. Yet, knowing that only 4 species contributed to this 35:386-389. 1% in 1976 (base-line), it is apparent that many additional species P&ii, C.P., and C. Taiiiie. 1982. Diversity as a concept and its measure- occurred in each post-treatment year. However, the post-treatment ment. J. Amer. Statis. Assoc. 77548-561. cumulative proportional abundance for extremely rare herbaceous Solomon, D.L. 1979. A comparative approach to species diversity. P. 29-35 species was 6.0% in 1978,6.8% in 1980, and 5.6% in 1983. For both In: Grassie et al. (ed.), 1979. Swindei, B.F., L.F. Conde, 8nd J.E. Smith. 1984. Species diversity: con- plant types, these profiles indicate an ecological trend where plant cept, measurement, and response to clearcutting and site-preparation. diversity began to return toward pre-treatment conditions by 6 Forest Ecoi. Manage. 8: 1i-22. years after disturbance. Swindei, B.F., L.F. Conde, and J.E. Smith. 1987. Index-free diversity Discussion and Conclusions orderings: Concept, measurement, and observed response to clearcutting and site-preparation. Forest Ecol. Manage. 20~195-208. Diversity assessments are useful tools in ecological evaluations. T8iiiie, C. 1979. Species equitability: A comparative approach. P. 5 1-61. In: Diversity indices, such as species count, Shannon’s index, and Grassie et al. (ed), 1979. Simpson’s index have been used for comparing diversity and for Whittaker, R.H. 1975. Communities and ecosystems (2nd ed.). MacMillan, New York, N.Y.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 469 Control of threadleaf rubber rabbitbrush with herbicides

STEVEN C. WHISENANT

Abstract Foliar sprays of 2,4-D [(2&dicblorophenoxy)acetlc acid]), pic- tives of this study were to (1) compare threadleaf rubber rabbit- loram (4-amino-3,5,6-Mchloro-2-pyridineeuboxylic acid), dicamba brush control achieved withclopyralid (3,6dichloro-2-pyridinecarbo- (3,6dichlor&-methoxy ,benzoic &I), or clopyralid (3,&dichioro-2- xylic acid) versus control achieved with previously recommended pyridinecarboxylic add) were applied in 30 or 150 L of total spray herbicides and (2) to evaluate the influence of spray volume on solution ha-1 to threadleaf rubber rabbitbrush [Clrrysothomnur herbicide efficacy. nauseosus ssp. consimi#s (Greene) Hall & Clem] in Garfield Materials and Methods County, Utah. Additional herbicide treatments were applied in 150 L ha-1 in Sevier County, Utah. Herbicides were less effective when Study Sites applied in 30 L ha-1 than when lppiied in 150 L of total spray Experiments were established near Antimony (Garfield County) solution ha-*. Mortaiity was 74 to 87% following applications of and Salina (Sevier County), Utah. The Garfield County site was 4.4 kg a.e. (acid equivalent) 2,4-D ha-l. Dicamba applied at 3.3 kg dominated by mountain big sagebrush [Artemiria tridentata ssp. ha-1 resulted in 70 to 87% mortality, and picloram applied at 0.8 kg voseyana (Rydb.) Beetle] and threadleaf rubber rabbitbrush at ha-1 resulted in 56 to 79% mortality. The greatest mortalities (84 to 2,290 m elevation with a mean annual precipitation of 36 cm. 97%) occurred on areas treated with 2.2 kg clopyralid ha-l. Mortal- About 50% of the precipitation falls during the 90day growing ity of threadleaf rubber rabbitbrush increased an average of 28,17, season. Soil at this site was a Codley silt loam (tine, silty, carbo- 33, and 27% following applications of 2,4-D, dicamba, picloram, natic, frigid Ustic Torriorthent). and clopyralid respectively, by using 150 L spray volume. Greatest The Sevier County site was dominated by threadleaf rubber increases were at the lowest herbicide rates. Applying herbicides in rabbitbrush growing on silty loams (fine, silty, mixed, frigid Xerol- greater amounts of carrier (water) significantly increased both lit Haplargid). This site is at 1,919 m elevation with a mean annual mortality md canopy reduction of threadleaf rubber rabbitbrush precipitation of 37 cm. About 80% of the annual precipitation falls for at least 39 months. between November and April. The mean frost-free period is about Key Words: clopyralid, dicamba, picloram, 2,4-D, spray volume 100 days. At least 8 species of rabbitbrush (Chrysorhumnur spp.) occur in Herbicide Applications the Great Basin area of the western United States. They are most Herbicide treatments at both sites were applied to 3 replications abundant on open plains, valleys, foothills, and mountains to 3,300 of 3- by 30-m plots in randomized complete block designs. Herbi- m (Hitchcock et al. 1969). Rabbitbrush frequently increases fol- cides were applied in water with a COs-powered backpack sprayer lowing removal of big sagebrush (Artembia WidentQta Nutt.) with in 30 or 150 L total spray solution ha-’ at the Garfield County site fire, heavy grazing, or herbicides. Most herbicide applications for and at 150 L ha-l at the Seiver County site. Application rates for rabbitbrush contol have been directed at rubber rabbitbrush each herbicide were based on previous research, however, because [Chrysothamnus nauseosus (Pallas) B&t.] or green rabbitbrush clopyralid had not been reported previously, a wider range of rates [Chrysothumnus viscidtjlorus (Hook.) Nutt.]. (0.6 to 2.2 kg ha-‘) was used. A commercial surfactant (a mixture of Rubber rabbitbrush is most susceptible to conventional herbi- alkyl-polyoxyethylene glycols, free fatty acids, and isopropanol) cides when new leader growth reaches 6 to 9 cm (Hyder et al. 1958, was included at 0.5% (v/v). Mohan 1973, Cluff et al. 1983) and is less susceptible in dry years. Treatments at the Garfield County site were applied on 20 July Mohan (1973) suggested spraying rubber rabbitbrush with 2.2 kg 1983 and I6 June 1984. Treatments were the propylene glycol butyl 2,4-D [(2,4dichlorophenoxy) acetic acid] ester ha-’ only when new ether ester of 2,4-D at 2.2 and 4.4 kg 8.e. ha-‘; the dimethylamine leader growth exceeded IO cm and enough water is present in the salt of dicamba at 3.3 and 4.4 4 a.e. ha-‘; the potassium salt of upper 10 cm of soil for rapid growth of rubber rabbitbrush. picloram at 0.6 and 0.8 kg a.e. ha and; the monoethanolamine salt Cluff et al. (1983) evaluated the efficacy of picloram (4-amino- of clopyralid at 0.6, 1.1, and 2.2 kg a.e. ha-l. New leader growth of 3,5,6-trichloro-2-pyridinecarboxylic acid), dicamba (3,6dichloro- threadleafrubber rabbitbrush was 5 to 12 cm during both the 1983 2-methoxybenzoic acid), silvex [2-(2,4,5-trichloro-phenoxy)pro- and 1984 applications. Air temperature was 19 to 23O C and pionic acid], 2,4,5-T [(2,4,5&chlorophenoxy) acetic acid], and relative humidity was 39 to 47% during the 1983 applications. triclopyr {[(3,5,6-trichloro-2-pyridinyl)oxylacetic acid) on thread- Temperature and relative humidity were 18 to 24O C and 48 to 54% leaf rubber rabbitbrush [ Chrysothamnus nauseosus ssp. consimilis respectively, during the 1984 applications. Soil-water content was (Pallas) Britt.] in central Nevada. They reported threadleaf rubber not measured, but the associated herbaceous vegetation was rabbitbrush mortalities averaged 87% on areas sprayed with 2.2 kg actively growing during both the 1983 and 1984 applications. 2,4-D ha-‘. Mortality following applications of triclopyr (3.4 kg Treatments were applied at the Sevier County site on 14 June ha-‘), silvex (3.4 kg ha-‘), or dicamba (3.4 kg ha-? were similar to 1984. Dicamba (3.3 kg ha-‘), clopyralid (1.1 and 2.2 kg ha-‘), and 2,4-D. 2,4-D (3.3 and 4.4 kg ha-‘) were applied in 150 L of total spray Despite data indicating adequate control of rubber rabbitbrush solution ha-‘. Ambient temperature was 17 to 25’ C, relative with available herbicides, many resource managers report poor humidity was 28 to 40%, and threadleafrubber rabbitbrush leader results. Newer herbicides and/ or different application techniques growth was 4 to 8 cm when herbicides were applied. Associated may increase efficacy of rubber rabbitbrush control efforts. Objec- herbaceous vegetation was still green when herbicides were applied, but grass leaves were beginning to fold from water-stress. The author is associate professor, Dept. of Range Science, Texas A&M University, College Station, Texas 77843-2126. Evaluations This research was conducted while the author was on the Botany and Range Science Treatments were evaluated in Oct. 1985 and Oct. 1986. The 1986 Faculty at Brigham Young University. Manuscript accepted 29 June 1988. evaluations followed third and fourth post-treatment growing seasons of the 1984 and 1983 applications, respectively. No significant

470 JOURNALOF RANGE MANAGEMENT 41(6), November 1966 differences were found between Oct. 1985 and Get. 1986 evalua- Table 2. Percentage canopy reduction and mortdity of threadhfrubber tions of plots treated in 1983. Consequently, Get. 1986 evaluations rabbitbnub following berbkide applkdiom in Scvkr County, Utah. were used as the final evaluation on experiments initiated in 1983 and 1984. Percentage canopy reduction (CR) of each woody plant rooted Treatment within a 25-m long by 2-m wide belt transect in each plot was Rate Canopy Mortality visually estimated. Totally defoliated plants were examined for Herbicide Ikn ae ha? reduction I%) (46) presence of basal or stem sprouts. The proportion of completely None 4 0 defoliated plants with no sprouts to the total number of plants was 2,4-D 3.3 90 85 used as an estimate of mortality. 2,4-D 4.4 91 87 Dicamba 3.3 84 77 Statistical Analysis Picloram 0.6 50 37 Data were subjected to arc siq/x transformation prior to analy- Picloram 0.8 78 69 sis of variance. The herbicide treatment by spray volume interac- Clopyralid 1.1 81 70 tion was significant at the Garfield County site for both mortality Clopyralid 2.2 93 88 and CR. Neither the application date by spray volume nor the L.S.D. (0.05) 12 15 application date by herbicide treatment interaction was significant at the Garfield County site. Analysis of variance was pooled over (water) significantly increased both mortality and canopy reduc- application dates for the Garfield County site and presented by tion of threadleaf rubber rabbitbrush for at least 39 months. These treatment for each site. Least significant difference (LSD.) at the data suggest that improved efficacy is possible with higher spray 5% level was used to separate statistically different treatment volumes. Improved efficacy with higher carrier volumes is proba- means. bly due to improved coverage of herbicide on both the shrub canopies and leaf surfaces;possibly resulting in increased absorption. Results and Dhcussion Rubber rabbitbrush leaf surfaces are densely tomentose which Threadleaf rubber rabbitbrush mortality for the 150 L ha-’ minimizes contact of spray droplets with the leaf surface. For treatments was similar to those reported by Cluff et al. (1983) for example, turkey mullein [Eremocurpus serigerus (Hook.) Benth.] dicamba. Mortality following 4.4 kg 24-D ha-’ averaged 87% for plants were sprayed with 20 to 780 L ha-’ total spray volume and both sites when the higher spray volume was used and 74% with the the leaves examined with the cathodoluminescence detection mode lower spray volume. Shrub mortality exceeded 90% at the Garfield of a scanning electron microscope (Hess et al. 1974). The amount of herbicide reaching the cuticle surface increased with increasing County site on areas sprayed with 4.4 kg dicamba ha-’ or 1.1 to 2.2 kg clopyralid ha-’ at 150 L ha-’ spray volume (Table 1). No treat- spray volumes (Hess et al. 1974). Similar results may be expected ment produced over 88% mortality at the Sevier County site (Table on rubber rabbitbrush when low spray-volumes are used. Threadleaf rubber rabbitbrush mortality and CR were lower at 2). the Sevier County site than the Garfield County site for all herbi- Threadleaf rubber rabbitbrush mortality increased an average of cides tested in 150 L ha-’ total spray volume (Tables 1 and 2). 28, 17, 33, and 27% following 2,4-D, dicamba, picloram, and Differing responses to herbicides may have been related to the clopyralid applications respectively, by using the higher total spray winter precipitation regime of the Sevier County site compared to volumes. Those increases were always greatest at the lowest herbi- the Garfield County site, which receives more summer precipita- cide rates. Applying herbicides in greater amounts of carrier tion. By the time leader growth was sufficent for spraying at the Sevier County site, soil water contents appeared to have been low; Table 1. Canopy rduction ($) end mortdtty (96) of tImdeaf rubber resulting in less effective threadleaf rubber rabbitbrush control. rabbitbrusb following herbicide applications in Gufldd County, Utah. Picloram applied at 0.6 kg ha-’ killed 32 to 52% of the threadleaf rubber rabbitbrush (Tables 1 and 2) but appeared to provide inconsistent control of big sagebrush (data not shown). These Treatment observations and the data of Tueller and Evans (1969) and Whis- Rate Canopy Mortality enant (1987) suggest that low rates (0.3 to 0.6 kg ha-‘) of picloram Herbicide (kg ae ha? reduction (%1 @?A1 may be used to control rubber rabbitbrush without eliminating big -30 L ha-’ total spray volume- sagebrush. None 0 The Siever County site also contained founving saltbush [Atri- 2,4-D 2.2 5: 40 plex cunescens (Pursh) Nutt.] in at least 1 replication of each 2,4-D 4.4 80 74 treatment. All herbicides except clopyralid killed founving salt- Dicamba 3.3 84 70 bush. Fourwing saltbush was also resistant to clopyralid in a west Dicamba 4.4 89 81 Texas study (Jacoby et al. 1981). Clopyralid has been shown to be Picloram 0.6 39 32 highly selective between certain plant families in other studies Picloram 0.8 72 56 (Bovey and Meyer 198 1, O’Sullivan and Kossatz 1982, Whisenant Clopyralid 0.6 41 26 1987). Clopyralid also effectively controls big sagebrush without Clopyralid 1.1 86 81 Clopyralid 2.2 91 84 causing significant damage to such important shrub species as antelope bitterbrush [Purshia tridenrutu (Pursh.) DC.] or Saska- -150 L ha-’ total spray volume- toon serviceberry (Amelunchier ulnifoliu Nutt.) (Whisenant 1987). This potential for increasing the selectivity of rangeland herbicide None2,4-D 2.2 705 5: applications should be examined in other communities. 2,4-D 4.4 93 87 Dicamba 3.3 90 87 These data suggest that threadleaf rubber rabbitbrush can be Dicamba 4.4 97 95 effectively controlled (L87% mortality) with sprays of clopyralid at Picloram 0.6 61 52 2.2 kg ha-‘, dicamba at 4.4 kg ha-‘, or 24-D ester at 4.4 kg ha-’ Picloram 0.8 88 79 applied in spray volumes of 150 L ha-‘. Spray volumes of 30 L ha-’ Clopyralid 0.6 66 57 will decrease the efficacy of those herbicides. In a winter- Clopyralid 1.1 96 95 precipitation regime, soil-water may be largely depleted before new Clopyralid 2.2 99 97 leader growth reaches the recommended 10 cm length. Threadleaf LAD. (0.05) 6 7 rubber rabbitbrush mortality may be reduced under those conditions.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 471 Literature Cited Jecoby, P.W., C.H. Meadors, and M.A. Foster. 1981. Control of honey mesquite (Prosopis juliflora var. glandulosa) with 3,6dichloropicolinic Bovey, R.W., and R.E. Meyer. 1981. Effects of 2,4,5-T, triclopyr, and 36 acid. Weed Sci. 29~376-378. dichloropicolinic acid on crop seedlings. Weed Sci. 29:256-261. Mohan, J.M. 1973. 14 years of rabbitbrush control in central Oregon. J. Cluff, G.J., B.A. Roundy, R.A. Evms, and J.A. Young. 1983. Herbicidal Range Manage. 266845 1. control of greasewood (Sarcobatus vermiculorus) and salt rabbitbrush O’Sullivan. P.A.. and V.C. Kossatz. 1982. Selective control of Canada (Chrysorhamnus nauseosus ssp. consimilis). Weed Sci. 31:2X-219. thistle in rapeseed with 3,6dichloropicolinic acid. Can. J. Plant Sci. Hess, F.D., D.E. Bayer, and R.H. Falk. 1974. Herbicidal dispersal pat- 62989993. terns: I. As a function of leaf surface. Weed Sci. 23:394401. Tueller, P.T., and R.A. Evans. 1969. Control of green rabbitbrnsh and big Hitchcoek, C.L., A. Cronquist, M. Ownbey, and J.W. Thompson. 1969. sagebrush with 24-D and picloram. Weed Sci. 17:233-23X Vascular plants of the Pacific Northwest. Part 5: Compositae. Univ. Whisenant, S.G. 1987. Selective control of mountain big sagebrush (Arte- Washington Press, Seattle, Wash. misia rridentata ssp. vaseyana) with clopyralid. Weed Sci. 35120-123. Hyder, D.N., F.S. Sneva, D.O. Chileote, and W.R. Furtick. 1958. Chemi- cal control of rabbitbrush with emphasis upon simultaneous control of big sagebrush. Weeds 6:289-297. Defoliation of Thurber needlegrass: herbage and root responses

DAVID GANSKOPP

Abstract Thurber needlegrass (S@r tiwrbericuroPiper) is an important nomenclature is Pseudorogneria spicata (Pursh)A. Love by Bark- component of both forested and shrub-steppe communities of the worth et al. 19831, Idaho fescue (Festuca idahoensis Elmer), or Pacific Northwest and Great Basin regions, and little is known of needle-and-thread grass (Spa comata Trin. & Rupr.) (Dauben- its tolerance to defoliation. A study was conducted on the !Squaw mire 1970). Thurber needlegrass is valuable forage for livestock Butte Experimental Range to determine the response of container- and wildlife, and its cured foliage maintains a slightly higher crude ized Thurber needlegrass to single defoliations (2Scm stubble) protein content than other bunchgrasses sharing the same envi- throughout the growing season. Dates of treatment spanned veget- ronment (Hickman 1975). Only 3 reports have addressed the sensi- ative through quiescent stages of phenology. Response variablea tivity of this species to defoliation or grazing. Tueller (1962) found included: summer regrowth, number of reproductive stems, fall Thurber needlegrass decreased with grazing, and while monitoring growth, and subsequent spring herbrp production, change in recovery of grasses from a severe l-year drought, Ganskopp and basal area, and root mass. Vigor of Thurber needlegrass was Bedell (1981) found a significant reduction in height of Thurber reduced most by defoliation during the early-boot stage of devel- needlegrass plants with a history of heavy, summer use. In a opment. Impacts were successively leas severe from vegetative, community dominated by Thurber needlegrass, Eckert and Spencer I&-boot, and anthesis treatments, respectively. Cumulative her- (1987) detected significant reductions in basal areas of plants heav- bage production the year of treatment was reduced from 38 to 64% ily grazed during a single growing season and subsequently by defoliation at the early-boot stage. The same treatment reduced deferred for 4 growing seasons. Where more moisture was avail- subsequent spring growth by 46 to 51% and root mass the next able and Thurber needlegrass shared the environment with blue- spring by 34 to 45%. Treatment effects were somewhat reduced bunch wheatgrass, there were no reductions in basal area (Eckert when temperature and moisture regime-s allowed substantial and Spencer 1987). regrowth after defoliation. Defoliation during or after anthesis had Defoliation of other cool-season bunchgrasses is typically most little effect on plant response. Managers should be aware that a injurious during the middle portion of their rapid growth period single defoliation, particularly during the boot stage, can signifi- when substantial regrowth is prevented by inadequate moisture cantly reduce subsequent herbage production and root mass and (McIlvanie 1942, Stoddart 1946, Blaisdell and Pechanec 1949, possibly lower the competitive ability of Thurber needlegrass. Blaisdelll958, Cook et al. 1958). Given the sparsity of information available on responses of Thurber needlegrass to defoliation, a Stipa thurberha, Key Words: clipping, grazing tolerance, her- study was initiated in 1982 to evaluate herbage and root responses bage production, roots of this species to l-time defoliations at various phenological stages Thurber needlegrass (Slipa thurberiana Piper) is a component of during the spring-summer growing season. The objective was to several steppe, shrub-steppe., and forest communities from central identify at which phenological stage defoliation most adversely Washington to California and east to southwest Montana and affected regrowth and subsequent spring growth and root mass. northeast Wyoming (Hitchcock and Cronquist 1974, Franklin and Description of Study Area and Methods Dymess 1973). It may dominate the herbaceous layer (Culver 1964, Hironaka et al. 1983) or exist as a subordinant in communities The study was conducted at the Squaw Butte Experimental where the vegetation is characterized by bluebunch wheatgrass Range (119’ 43W, 43” 29’N) approximately 72-km west-southwest [Agropyron spicatum (Pursh)Scribn. & Smith; recently revised of Bums, Oregon. Mean annual precipitation is 27.6 cm with peak accumulations in November, December, January, and May (rang- Author is mnp scientist, USDA-AR& Eastern Oregon Agricultural Center, ing between 2.9 and 3.6 cm); and a minimal accumulation (0.8 cm) Sauaw Butte Station. Star Rt. 14.51 Hwv. 205. Bums. Oreson 97720. *khcauthor wishes to express gratitude to Con&e Fischer&d Kay Marietta for their in July (NOAA 1986). Herbaceous plant yield is most strongly assistance in data collection. correlated with September-June precipitation accumulations (Sneva Technical Paper 8330. Oregon Agricultural Experiment Station. Manuscript accepted 21 June 1988.

472 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 1982). Mean annual temperature is 7.6O C with recorded extremes tling occurred, additional soil was provided to create a continuous of -29 to 42O C (NOAA 1952-1986). Overstory dominants of major surface across the top of the bags and the surrounding soil. Sep- plant communities include western jumper (Juniperus occidentalis tember through June precipitation for the ensuing establishment, Hook.), low sagebrush (Artemisia arbuscula Nutt.), and 3 subspe- treatment, and harvest years (1983-1987) was 109, 119,77, 111, cies of big sagebrush (Artemisia tridentata Nutt.) including basin and 97% of the long-term average (25.2 cm, NOAA 1986). Emerg- big sagebrush (subsp. tridentata Nutt.), Wyoming big sagebrush ing weeds were removed when necessary. (subsp. wyomingensis Beetle), and mountain big sagebrush (subsp. One-half of the plants were scheduled for defoliation treatments vaseyana (Rydb.) Beetle). Understory dominants include blue- in 1985 and total plant harvest in 1986, with the remainder desig- bunch wheatgrass, Idaho fescue, Thurber needlegrass, Sandberg’s nated for a replication of the study in 1986 through spring of 1987. bluegrass (Poa sandbergii Vasey), and cheatgrass (Bromus tecto- This provided 2 and 3 growing seasons, respectively, for plants to rum L.). recover from the rigors of transplanting and acclimate to contain- In August 1982 after quiescence, 150 plants were excavated from ers. Plants scheduled for 1985 treatments were clipped to a 2.5cm sites supporting a Wyoming big sagebrush-Thurber needlegrass height in October 1984 to remove standing litter. Plants scheduled community. Descriptions of vegetation and soils of the community for replication of the study in 1986 were similarly clipped in are provided by Doescher et al. (1984) and Lentz and Simonson October 1985. (1986). An effort was made to preserve soil-root integrity of plants A treatment consisted of a single defoliation (2.5-cm stubble) on to a 15-cm depth. Bunches were transported to a common garden, 1 of 7 dates during the spring-summer growing season. Biweekly trimmed to approximately equal dimensions (7.5-cm diameter), treatments spanned the vegetative through quiescent stages of and replanted in polyethylene bags with a 2%m diameter and phenology. Treatment l(24 April) occurred when plants were in a 61ern depth. The 61-cmdimension was selected because it approx- vegetative stage of growth. Treatments 2 and 3 (8 May and 22 May) imated the depth of unweathered bedrock in the plants’ natural were applied during the early and late-boot stages of development, habitats (Lentz and Simonson 1986). respectively. As apical meristems began to elevate, bases of repro- Bags were constructed by cutting sections from a continuous ductive tillers became visibly swollen during the early-boot stage.. tube of polyethylene material, folding 1 end 3 times, and securing In the late-boot stage, uppermost awns were emerging from the the folds with heavy duty staples. This closure allowed water to boot, but remaining portions of reproductive stems were still con- drain from the bottom but excluded outside roots. Sifted alluvium cealed by protective sheathing. Treatment 4 occurred during (6.4-mm mesh) was used as a potting soil to facilitate later removal anthesis (5 June), and treatment 5 (19 June) occurred when seeds of roots. Particle composition of the soil was 770/Osand, 9% silt, and were tilling and just beginning to harden. Most seed had hardened, 1490clay. Filled bags were placed vertically in 61-cm deep trenches some had fallen from the plant, and portions of the foliage were and exterior voids filled with sifted soil. After planting, a l-time beginning to senesce when treatment 6 was applied (3 July). A irrigation of approximately 2 liters of water per plant was provided count of seed stalks was also conducted on this date to provide an to enhance soil:root contact and facilitate soil settling. When set- index of reproductive tiller development in all treatments. Treat-

T8blc 1. Biomrw (ml/cm* b88818re8)of pretnrtment herb8ge production, regrowth (h8rvested 31 July), f8U growth (huveeted IS Oct.), 8nd herb8ge 8nd roots the rukcquent epring @uvCaed 5 M8y) of Thurber needlegr88#pl8ntr defoli8ted on 1 of 7 d8te8 rp8noing the gringaummer growing 8e88ons of 1985 8d M&i. Phntr of the 17 July tre8tment grew u&hibited 8nd were viewed 88 controls. Me8ns in rows sh8ring 8 common letter 8re not signifk8ntly different (p>O.OS)..

Dates of defoliation 24 April 8 May 22 May 5 June 19 June 3 July 17 July Phenology Vegetative Anthesis Hard seed Boot Soft dough Seed shatter 1985 treatments Pretreatment herbage 4Oa 54a 116b 128bc 171c 227d 232d Regrowth (31 July) 79c 31b 46b 33b la

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 473 ment 7, applied on 17 July when all seed had shattered and all of 1986, however, mean daily temperatures were 7.2” C above foliage was brown, was viewed as a control. Material harvested in those of the same period in 1985 (NOAA 1986). Because soil the progressive application of treatments (pretreatment herbage moisture levels were generally greater during the 1986 growing production) was retained and dry weights obtained to provide an season than in 1985 (Fig. l), higher levels of regrowth were exhi- index of seasonal growth patterns. bited by the 1986 treatments. Immediately after application of a treatment, basal areas of Thurber needlegrass root and shoot biomass were most adver- treated plants were ascertained by measuring (to the nearest 1 mm) sely affected by defoliation during the early-boot stage of devel- major and minor axes of the crowns and later solving for the area opment during both years plants were treated. Plants defoliated of an ellipse. Individual portions were measured and the derived during the early-boot stage generally had lower regrowth, total areas totaled if a crown had fractured into obvious components. herbage, spring growth, and total roots than plants defoliated at After acquisition of crown dimensions, soil samples (5 to 4Ocm earlier or later phenological stages (Table 1). depth) were collected from 5 bags and soil water content measured As would be expected in an environment where deep soil mois- gravimetrically. Attempts to sample soil moisture at lower depths ture is a product of winter precipitation, soil water content was of the column were confounded by contamination from the drier gradually depleted as the growing season advanced, and potential and often unstable upper portions. for regrowth after defoliation progressively declined. In 1985, Response variables in this study included: regrowth, seed stalk regrowth exceeded pretreatment production only when plants density, fall growth, total herbage production during the year of were clipped in the vegetative stage (Table 1). By 19 June 1985 soil treatment (= sum of pretreatment herbage, regrowth, and fall moisture content was reduced to 4.2%, and plants defoliated on or growth), and the subsequent spring’s herbage production, root after that date produced only trace amounts of regrowth. Across mass, and change in basal area between treatment application and all treatments in 1985, regrowth constituted roughly 16% of total final harvest. herbage production. The greater moisture availability in 1986, Regrowth was harvested from all treatments on 31 July when however, allowed a greater accumulation of regrowth in even the most foliage had turned brown. Fall growth occurred in both years latest treatments. Regrowth exceeded pretreatment production for that treatments were applied, and was harvested in mid-October. the vegetative through anthesis treatments, and contributed suffi- Spring growth of all treatments was harvested to ground level on cient biomass to the 3 July treatment to exceed the control plants in 5 May (late-vegetative phenology), and live crown dimensions total herbage production. Averaged across all treatments, regrowth remeasured on all plants. Root bags were excavated, severed into 3 constituted 44% of total herbage production in 1986. increments (O-20, 20-40, 40-61 cm) and roots manually washed Among all treatments and years, fall growth constituted only 3 free of soil over a 5-mm mesh screen. Additional detritus was to 4% of total yearly standing crop (Table 1). No significant treat- removed by floating and agitating roots in a shallow pan of water. ment effects were detected, and no consistent trends were apparent Because plant crowns were extremely compact in nature, and dead across treatments. and foreign material could not be satisfactorily’removed, crowns Significant treatment effects were detected in total herbage pro- were severed from the uppermost root mass. duction in both years (Table 1). In 1985, defoliation at any time All harvested materials were oven dried at 60’ C for 48 hours from the vegetative through softdough stages caused significant and weighed. When first transplanted, all bunches were trimmed to reductions in total herbage. Most severe, compared to the controls, approximately equal dimensions. During the years allowed for was a 63% reduction in total herbage when plants were defoliated plant establishment, however, variations in growth and rodent during the early-boot stage. Due to the greater contribution of damage inconsistently altered basal areas. To compensate for vari- regrowth to total herbage in 1986, differences between treatments ations in plant size all weights were expressed on a mg/ cm* basal were less apparent, but defoliation during the early boot stage was area basis. Pretreatment herbage, regrowth, fall growth, and den- again responsible for the greatest reduction in total herbage (46% sity of reproductive stems are based on the initial measure of basal less compared to 3 July treatment). area; and spring growth and root weights based on the final mea- Despite a two-fold difference across treatments, no significant sure of basal area. Root weights from the 3 different increments effects could be detected in spring growth of plants treated in 1985 were also converted to percent of total root mass to test the (Table 1). In the 1986 treatments, however, spring growth of plants hypothesis that relative distribution of root mass was altered by defoliated during the early-boot stage was significantly less than treatments. for those treated during or after anthesis, and plants defoliated The study was arranged in a completely randomized design with during the vegetative stage produced significantly less spring 7 treatments and 8 replications. Years were analyzed separately. growth than those clipped at seed-shatter. The extra plants provided replacements if rodent damage or plant Responses of roots mirrored those of the above ground compo- mortality occurred prior to scheduled treatment dates. When sig- nents. Root mass of all treatments was nearly 3 times higher in 1987 nificant differences were detected among treatments, mean separa- compared to 1986. Across treatments and years, roughly 68% of tion tests were made using least significant difference (LSD) the total root mass was concentrated in the upper 20 cm of the soil, procedures. Significance was assumed at KO.05. and significant effects were detected in this portion of the roots during both years (Table 1). Middle and lower portions of the roots Results averaged 18 and 14% of the total, respectively, and treatment effects were not consistently apparent within these portions of the Phenological development of plants was similar between 1985 root mass. When root weights from the 3 sections were converted and 1986, but timing of herbage accumulation varied somewhat to relative values (percent), no significant differences were found between the 2 growing seasons. Herbage production for control among treatments, indicating the distribution of roots remained plants, clipped on 17 July, was nearly identical between the 2 years, relatively constant across treatments. however. Pretreatment herbage data of Table 1 can be visualized as Plants defoliated during the boot stage in 1985 exhibited sign& the growth curve of Thurber needlegrass. In 1985, growth was most cantly less root mass in the O-2O-cm profile than in subsequent rapid during the boot stage (9 May-22 May) when there was a treatments (Table 1). Although treatment effects were significant 214% increase in herbage. In 1986, a 257% increase occurred in the 20-4O-cm depth, the only clear separations of means were during the late-boot through anthesis stage (23 May-S June). between the first-2 and last-2 treatments. When total roots were During the early portion of the 1986 growing season (25 April-22 considered, the reduced mass in the early-boot stage treatment May), mean daily temperatures ((maximum + minimum)/ 2) aver- differed significantly from treatments during or after anthesis. aged 3.7” C degrees cooler than in 1985, which probably depressed Similar trends were exhibited with the root systems of plants herbage production. In the late boot through early anthesis portion excavated in 1987. Treatment effects, however, were significant

474 JOURNAL OF RANGE MANAGEMENT 41(6), November 1666 only in the O-2O-cm depth. Defoliation during the early-boot stage reduced root production in the O-2O-cm depth by 33 to 43% when 17 compared to treatments applied during or after anthesis. Summa- 16 tion of data from the top, middle, and lower depths, however, enhanced variation to such a degree that significant responses could not be detected in analysis of total root mass. Maximum densities of reproductive stems and mature seed were attained by the July treatments which allowed seed to mature prior to defoliation (Table 2). In both years densities of reproductive stems were reduced nearly 50% by the 24 April clipping when plants were described as being in a vegetative stage of growth. Possibly, apical meristems were just beginning to elevate, and nearly 50% were removed by clipping, or tillers failed to differen- tiate or were aborted. By 8 May in both years, reproductive stems were visibly swollen, and 80 to 98% of the meristems were harvested. On 22 May, awns were just extending from the uppermost sheaths, and clipping removed all apical meristems in 1985. In 1986, however, when early growth rates were slowed by cool Fig. 1. Mean gravimetric water content and waterpotential of soil (MO- temperatures, a few reproductive meristems were low enough to cm depth) in Thurber needlegrass containers on 7 dates whenplants were escape the 22 May defoliation. June treatments removed all apical defoliated in the 1985 and 1986 growing seasons (IFS). m&terns in both years, and no replacements developed. Basal areas of plants in all treatments declined during both years severely defoliated and the small amounts of photosynthetic tissue plants were defoliated. Declines averaged 45% for the 1985 treat- remaining can not support development of new shoots, mainte- ments and 16Yc for the 1986 (Table 2). Trends suggested earlier nance of the plant and regrowth of foliage initially depend on treatments were most affected, but no significant differences stored carbon (White 1973, Caldwell et al. 1981). Possibly, plants among treatments were detected in either year. clipped at the early-boot stage were forced to initiate regrowth with less stored carbon than other treatments. More recent efforts Discussion and Conclusiona (Richards and Caldwell 1985, Brown 1985) have found little corre- Responses of Thurber needlegrass to defoliation during the lation between carbohydrate levels and regrowth and suggest lim- growing season paralleled those reported for other cool-season itations of meristematic tissues or variations in carbon allocation bunchgrasses common to the Pacific Northwest and Great Basin patterns might be more important. regions (Hanson and Stoddart 1940, Mueggler 1970, 1972; Bran- A 50% reduction in herbage production the subsequent spring son 1956, Eckert and Spencer 1987). Clipping at any time prior to further suggested plants were most heavily stressed by defoliation seed shatter depressed total herbage accumulation for the current at the early boot stage. Treatment differences may have been year. alleviated if plants had been allowed to grow for a longer period, Due to a progressive depletion of soil moisture in late spring and but past efforts by others indicate more than 1 growing season may early summer, greatest opportunities for regrowth were exhibited be necessary for complete recovery of weakened grasses (Biswell by the earliest treatment. Intuitively, one would expect an orderly and Weaver 1933, Hanson and Stoddart 1940, Eckert and Spencer decline in the production of regrowth as plants were defoliated on 1987). successively later dates. Early-boot stage treatments, however, Perhaps the most important response of Thurber needlegrass to exhibited depressed regrowth in both years when its production defoliation during the early-boot stage was the reduction in root was aligned with earlier and subsequent treatments. Early-boot mass. While the present study detected no significant changes in stage defoliation removed approximately 88% of aboveground the relative distribution of roots in the soil column, significant material, and due to the compact nature of the crowns, probably an differences in mass intuitively imply differences in absorptive sur- equivalent or greater proportion of photosynthetic tissue. Patterns faces. Mass of roots in all treatments were nearly 3 times higher in of carbohydrate withdrawal and allocation vary among species 1987 compared to 1986. Under controlled conditions, Mohammad (Jameson 1963, White 1973, Caldwellet al. 1981, Brown 1985), but et al. (1982) noted reduced root production of plants exposed to data from other cool-season bunchgrasses suggest nonstructural increased water stress. One can only speculate, however, as to carbohydrate pools are at their lowest concentration at this stage of whether the disparity between years was a product of 3 years of phenology (McIlvanie 1942, Donart 1%9). When grasses are undisturbed establishment time, a function of less moisture stress

Table 2. Density (number/cm1 beul area) of reproductive etcme (sampled on 3 July) and pereed change lu basal uess (between treatment date sod 5 May the subeequent spring) of Thurber needkgraes detolkted to I 2.S-cm stubble on 1 of 7 dates spanning tbe spring-eummer growing seasons of 1985 end 19g6. Means in rows sbuing a common ktter are not slgntfkantly different (P70.05).

Dates of defoliation 24 April 8 May 22 May 5 June 19 June 3 July 17 July Phenology Vegetative Anthesis Hard seed Boot Soft dough Seed shatter Density of reproductive stems 1985 0.4b O.lab O.Oa O.Oa O.Oa O.& Ok 1986 OSbc O&b O.la O.Oa O.Ch l.Od 0.8cd % Change in basal atea 1985 -%a -67a -34a 42a -45a -59a -23a 1986 -21a -26a -19a -17 - 7a -1Oa -1Za

JOURNAL OF RANGE MANAGEMENT 41(g), November 1988 475 in 1986, or an interaction of the 2 factors. Culver, R.N. 1964. An ecological reconnaissance of the Arfemisia steppe Recent work has downplayed the importance of roots as a on the east central Owyhee uplands of Oregon. M.S. Thesis, Oregon source of carbohydrate reserves (Marshall and Sager 1965, Cald- State Univ., Corvallis. Daubenmire, R. 1970. Steppe vegetation of Washington. Washington Agr. well et al. 198 1). Nevertheless, their function as the primary means Exp. Sta., College of Agr. Washington State Univ. Bull 62. of water and nutrient absorption, makes them paramount to plant Doescher, P.S., R.F. Miller, and A.H. Wtawud. 1984. Soil chemical survival, particularly in arid environments. Grasses with weakened patterns under eastern Oregon plant communities dominated by big root systems suffer reductions in drought tolerance (Hanson and sagebrush. Soil Sci. Sot. Amer. J. 48~656-663. Stoddart 1940, Weaver and Albertson 1943, Crider 1955),cold and Donut, C.B. 1969. Carbohydrate reserves of 6 mountain range plants as heat tolerance (Biswell and Weaver 1933, Julander 1945), and related to growth. J. Range Manage. 22:4I l-415. competitive ability (Weaver 1930). Eckert, R.E. Jr., and J.S. Spencer. 1987. Growth and reproduction of Plants in this study existed in individual containers and were not grasses heavily grazed under rest-rotation management. J. Range Man- exposed to competitive influences of other vegetation. Where there age. &156-159. Franklin, J.F., and C.T. Dyrness. 1973. Natural vegetation of Oregon and is no competition, or where growing pastures receive uniform Washington. PNW Forest and Range Exp. Sta. USDA Forest Serv. grazing, defoliation may decrease moisture losses through transpi- Gen. Tech. Rep. PNW-8. ration, extend the retention of soil moisture, and thereby provide Canskopp, D.C., and T.E. Bedell. 1981. An assessment of vigor and an improved moisture regime for the remaining foliage or regrowth production of range grasses following drought. J. Range Manage. (Svejcar and Christiansen 1987). The greater longevity of green 34:137-141. foliage on regrowing plants in this study suggests this process may Hanson, W.R., and L.A. Stoddart. 1940. Effects of grazing upon bunch have occurred. Had this study been conducted under field condi- wheatgrass. J. Amer. Sot. Agron. 32~278-289. tions where neighboring grasses and shrubs could compete for Hickman, D.E. 1975. Seasonal trends in the nutritive content of important unused resources, the potential for regrowth may have been range forage species near Silver Lake, Oregon. PNW Forest and Range Exp. Sta. USDA Forest Serv. Res. Pap. PNW-187. reduced, and the negative aspects of defoliation even more HIronaka, M., M.A. Fosberg, and A.H. WInward. 1983. Sagebrush-grass enhanced (Weaver 1930, Mueggler 1970,1972). Anadditional bias habitat types of southern Idaho. College of Forest., Wildl., and Range to this study was that no concerted effort was made to separate live Sci. Forest., Wildl. and Range Exp. Sta., Univ. of Idaho, Moscow. Bull. and dead roots. Obvious dead and decaying roots were quite brittle 35. and were shattered and washed away in the separation of roots and Hitchcock, C.L., and A. Cronquist. 1974. Flora of the Pacific Northwest. soil. Some dead roots were certainly included in samples, however, Univ. Washington Press, Seattle. and presence of this material would lead to an understatement of Jameson, D.A. 1963. Responses of individual plants to harvesting. Bot. treatment effects on the live component. With these biases in mind, Rev. 29:532-594. extrapolation of these findings to field conditions should perhaps Julander, 0.1945. Drought resistance in range and pasture grasses. Plant Physiol. 20~573-599. be made with even a more conservative approach than is suggested Lentz, R.D., and G.H. Simonson. 19%. A detailed soils inventory and by these data. associated vegetation of Squaw Butte Range Experiment Station. Agr. Thurber needlegrass contributes substantially to the forage Exp. Sta. Oregon State Univ., Corvallis. Special Rep. 760. resource of the Pacific Northwest and Great Basin regions, and is Marshall, C., end C.R. Sager. 1965. The influence of defoliation on the obviously quite sensitive to defoliation during the growing season distribution of assimilates in L&urn multiflorum Lam. Ann. Bot. (Eckert and Spencer 1987). If we are to wisely manage this species, 29~365-372. future efforts should address its sensitivity to repeated defoliations, McIlvanIe, S. 1942. Carbohydrate and nitrogen trends in bluebunch the seasonality of its carbohydrate dynamics and root growth, and wheatgrass, Agropyron spicatum, with special reference to grazing influ- its competitive abilities and rate of recovery from severe or ences. Plant Physiol. 17:540-557. Mohammed, N., D.D. Dwyer, and F.E. Busby. 1982. Responses of crested repeated defoliations. wheatgrass and Russian wildrye to water stress and defoliation. J. Range Literature Cited Manage. 351227-230. _ Mueggler, W.F. 1970. Influence of competition on the response of Idaho Bukwortb, M.E., D.R. Dewey, and RJ. Atkins. 1983. New generic con- fescue. USDA Forest Serv. Res. Pap. INT-73. cepts in the lliriceue of the Intermountain region: Key and comments. Mueggler, W.F. 1972. Influence of cdmpetition on the response of blue- Great Basin Natural. 43561-572. bunch wheatgrass to clipping. J. Range Manage. 25:88-92. Biswcll, H.H., and J.E. Wc8ver. 1933. Effect of frequent clipping on the National Oceanic and Atmospheric Administration. 1952-1986. Climato- development of roots and tops of grasses in prairie sod. Ecology logical data annual summary, Oregon. 58-92:( 13). 14:368-390. Richards, J.H., and M.M. Caldwell. 1985. Soluble carbohydrates, concur- and Effects of herbage removal at Blaisdell, J.P., J.F. Pechanec. 1949. rent photosynthesis and efficiency in regrowth following defoliation: a various dates on vigor of bluebunch wheatgrass and arrowleaf balsam- field study with Agropyron species. J. Appl. Ecol. 22907-920. root. Ecology 30:298-305. Sneva, F.A. 1982. Relation of precipitation and temperature with yield of Seasonal development and yield of native plants on the Blaisdell, J.P. 1958. herbaceous plants in eastern Oregon. Int. J. Biometeor. 26:263-276. upper Snake River plains and their relation to certain climatic factors. Stoddrrt, L.A. 1946. Some physical and chemical responses of Agropyron USDA Tech. Bull. IWO. spicutum to herbage removal at various seasons. Utah Agr. Exp. Sta. Quantitative effects of clipping treatments on 5 range Branson, F.A. 19%. Bull. 324. grasses. J. Range Manage. 9~86-88. Svejcar, T., and S. Christiansen. 1987. The influence of grazing pressure on effect of severe defoliation on the subsequent Brown, R.F. 1985. The rooting dynamics of caucasian bluestem. J. Range Manage. 403224-227. growth and development of 5 rangeland pasture grasses of south-western Tueller, P.T. 1962. Plant succession on two Arremisia habitat types in Queensland. Aust. J. Ecol. l&335-343. southeastern Oregon. Ph.D. Thesis, Oregon State Univ., Corvallis. Caldwell, M.M., J.H. Richards, D.A. Johnson, R.S. Now& and R.S. Weaver, J.E. 1930. Underground plant development in its Elation to D,rurec. 1981. Coping with herbivory: photosynthetic capacity and grazing. Ecology 11543-557. resource allocation in 2 semiarid Agropyron bunchgrasses. Oecologia Weaver, J.E., and F.W. Albertson. 1943. Resurvey of grasses, forbs, and So:1624. underground plant parts at the end of the great drought. Ecol. Monogr. Responses of crested Cook, C., L. Stoddart, and F. Kinsinger. 1958. 13:64-l 17. wheatgrass to various clipping treatments. Ecol. Monogr. 28:237-272. White, L.M. 1973. Carbohydrate reserves of grasses: a review. J. Range Root-growth stoppage resulting from defoliation of Crider, FJ. 1955. Manage. 26:13-18. grass. USDA-SC!% Washington 4.C. Tech. Bull. No. 1102.

476 JOURNAL OF RANGE MANAGEMENT 41 (S), Novembw 1988 Season of cutting affects biomass production by coppicing browse species of the Brazilian caatinga

LINDA II. HARDESTY,THADIS W. BOX, AND JOHNC. MALECHEK

AbShCt This paper reports the effect of season of cutting on coppice affordable by subsistance farmers. Manual control, controlled biomass production by 5 tree species common in the semiarid burning and grazing are already practiced in the region, though the caatinga woodlands of northeast Brazil. Trees were cut early and full potential of these techniques is unreported. The ability of most late ln the wet and dry seasons and coppice biomass production caatinga species to sprout (coppice) has made control difficult. was monitored for 2 growing seasons after cutting. However, complete control of woody vegetation is usually not No mortality occurred as a result of cutting in any season. The desirable because the leaves of some woody plants can be a critical effect of season of cutting on subsequent coppice production was source of forage during the dry season (Pfiiter and Malechek most pronounced in tbe first year but differences persisted into the 1986). second year. Production by trees cut late ln the wet season lagged Coppice growth is a valuable forage resource. All of the desir- behind that of trees cut at any other time. This was true for all able browse species in the caatinga sprout prolifically, and coppice species except marmeliero (Croton hem&wwreus Muell. Arg.) dur- growth of many of the normally unpalatable species is palatable to ing both years. Pau branco (Auxmuno oncocdyx Taub.) produc- sheep and goats. In addition, coppice growth remains green for tion was maximized by cutting late in the dry season. Jurema preta several months after intact trees have shed their leaves. Foresters (Mtmosa acutistipula Bentb.) and catingueira (Caesa@inia pyra- have managed coppicing trees to achieve various objectives and mido#s Tul.) production was maximized by cutting early ln the dry manipulation of coppice growth may also be useful in managing season. The season of cutting does not affect marmeliero stem browse ranges. production. Except for the late wet season, no treatment signiS- Numerous factors have been suggested or demonstrated to cantly affected sabti Mimosa caesa&iniforto production. Stem influence coppicing including species, ecotype, site, season of biomass production Is affected more by season of cut tban is leaf treatment, type and frequency of injury, age, diameter and height biomass production. The different patterns of response among of the stump, origin and placement of the sprout, stem and root these species could be the basis of a selective cutting scheme to carbohydrate levels, and various growth regulators (Blake 1983). achieve objectives such as browse and wood production. Most of the work on the subject has been with temperate tree species and a few humid tropical species (Walters 1972, Walters Key Words: sprouting, woodland grazing, shrubland grazing and Wick 1973). The range of known coppicing responses is so Northeast Brazil is a densely populated agricultural region broad that assumptions about an unstudied species are risky, where many people live at or below subsistance level. Crop and especially in an unusual environment such as the caatinga. livestock production are mixed and livestock are an important The season of cutting is one factor that is easily manipulated. It is source of food and income. Small ruminants are the most common generally stated that coppicing is maximized by cutting during the livestock, accounting for 18 and 92% of the nation’s sheep and goat dormant season and minimized by cutting during active growth populations, respectively (USAID Unpub. Project Proposal 1980). (Smith 1986). Our paper reports on the effect of season of cutting Fluctuating forage supplies are a major constraint to livestock on coppice production of several caatinga species. The objectives production. The native vegetation, known as caatinga, is a com- were to determine: (I) if season of cutting affects subsequent plex mix of deciduous woody species with an annual berbaceous above-ground biomass production, (2) if cutting at particular times understory. During the 6-9 month dry season, forage disappears causes stump mortality, (3) if the ratio of leaf to stem biomass is rapidly as herbaceous plants and fallen tree leaves are either con- influenced by season of cutting, and (4) if seasonal variation in sumed or disintegrate. This annual period of nutritional stress cutting provides enhanced management opportunities for the results in low reproductive rates and substantial weight losses caatinga. (Huss 1976). Methods Traditionally, the caatinga is cut and burned for either cropland or pasture improvement. Several years after this treatment woody Study Sites plants reinvade, and the stand is abandoned for up to 20 years. This study was conducted at 2 sites within the state of Ceara. The Acceleration of this cycle is thought to account for widespread soil Sobral site was located at the National Goat Research Center of : erosion and undesirable changes in vegetation cover and composi- the national agricultural research organization (EMBRAPA), tion. More sustainable brush management techniques are needed near the town of Sobral at an elevation of 63 m. The Quixadl site to meet producers’ needs. was located at Fazenda Iracema, an experimental farm operated The literature provides numerous examples of brush manage- by the state agricultural research organization (EPACE), near ment techniques adapted to particular objectives (Hardesty 1984). Juatama at an elevation of 180 m. In Brazil we focused on techniques which would be acceptable and The interior of Ceara is semiarid with rainfall normally concentrated between January and April. This is variable, however, and Authors are assistant professor, Department of Forestry and Range Mana ment, extended droughts are characteristic. The average annual rainfall Washington State University, Pullman 991644410; dean, College of Natural PE”csour- ces, and head, Department of Ranp Science, Utah State University, Logan 84322. At for Sobral is 759 mm. During the study (1982-1984) the annual the time of this research, the scmor author was -ch associate, Department of rainfall was 650,447, and 986 mm, respectively. Average annual Range Science, Utah State University. This research was carried out as a part of the U.S. Agency for International rainfall at Quixada is 873 mm, but it was 710, 275, and 769 mm Development Title XII Small Ruminant Collaborative Research Su port Program during the study. The average annual temperature range for the under Grant No. AID/DSAN/XIIC4049, in collaboration with the g mpnsa Brasi- leira de Pesquisa Agropequaria, and Empress de Pcsquisa Agropcquaria do Ccati. region is 22-26“ C, with little daily or seasonal variation (Camargo The authors wish to thank Dan Hardesty and their Brazilian collaborators for 1965). Potential evaporation ranges from l,OOO-1,500 mm/yr logistical support and John Tebcrg and Dr. Roger Chapman for assistance with data (Camargo 1965, Eiten and Goodland 1979). analysis. Parts of this paper appear in the senior author’s Ph.D. dissertation at Utah State University. Both sites have shallow soils (C.5 m) underlain by a crystalline The authors acknowledge the approval of the College of Agriculture and Home shield. The dominant soils at the Sobral site are a sandy-gravelly Economics, Scientific Paper 7529. Manuscript accepted 29 June 1988. JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 477 lithosol, a non-calcic brown soil with a clay layer, and a eutrophic A randomized block design with 3 replications was used at both red-yellow podzol. All are moderately to well drained. Soils at the sites. Each treatment was applied to 1 plot in each of the 3 blocks. Quixadl site are lithosols and noncalcic brown soils (Ramos Within the plots, a minimum of 6 healthy trees of each species 1981). Both sites have gently rolling topography. studied were randomly selected from within the modal diameter Study sites were located in mature caatinga stands but the nature class for the species in that stand. Stumps were cut at a standard 30 of the vegetation at the 2 sites varied. The caatinga near Sobral is cm height and permanently marked with aluminum tags. dominated by pau branco (Auxemma oncocalyx Taub.), a species Double sampling (Cook and Stubbendieck 1986) was used to that does not occur in Quixada. Woody species that dominate the determine biomass production at the end of the first growing Quixadii site include jurema preta (Mimosa acutistipula Benth.), season after treatment (1983). Two coppicing stumps of each spe- catingueira (Caesalpinia pyramidalis Tul.), and sabi6 (Mimosa cies from each plot were randomly selected for harvest. The height caesalpinifolia Benth.). All of these species are common in Sobral, and diameter of the coppice clumps were measured, then all as are marmeliero (Croton hemiargyreus Muell. Arg.) and mofumbo regrowth was cut from the stumps. The harvested material was (Combretum leprosum Mart.). Moron5 (Bauhiniaforficata Link.) separated into leaf and stem fractions, oven dried at 65” C for 48 occurred more frequently in Sobral, and a variety of species which hours and weighed. Regression equations were developed relating together composed a small part of the flora at Quixada did not leaf and stem biomass production to height and diameter for each occur at all in Sobral. There is no comprehensive vegetation classi- species and treatment. The height and diameters of the coppice fication for this region making ecological interpretations of vegeta- clumps of all unharvested stumps were recorded and used as inde- tion largely speculative. Jurema preta dominates disturbed sites pendent variables to predict biomass production. At the end of the but once a canopy is established, it fails to reproduce, becomes 1984 growing season, coppice growth was cut from all previously decadent, and dies out of the maturing stand. The Quixada site had unharvested stumps, separated, dried, and weighed. considerably more live jurema preta, thus the stand may be some- Variables analyzed include stem, leaf, and total biomass and the what younger than the Sobral stand. Both stands are considered ratio of leaf to stem biomass. Homoscedasticity was verified with “old” (40-60 years) by longtime residents, and are dense stands Cochran’s test (Guenther 1964). Analysis of variance was per- (23,000 stems/ha in Quixada), 6-10 m in height, with closed formed using the SAS statistical package (Statistical Analysis canopies. Systems 1986) and means were compared with Fisher’s protected LSD test at a = .05 (Steel and Torrie 1980). species studied The species studied were sabia, catingueira, marmeliero, pau Results and Discussion branco, and jurema preta, all of which are common throughout the region. Sabil is a small tree, highly preferred for fenceposts and as 1983 Biomass Production long-, hot-burning firewood suitable for firing ceramics and other With the exception of evergreen jurema preta, cutting late in the industrial uses. Sabia’s foliage is extremely palatable to sheep and wet season does not allow enough time or suitable conditions for goats throughout the year. coppice growth. Responses to this treatment, therefore, are not Catingueira occurs in several plant associations and in more readily apparent until rains begin the following growing season. habitats than any of the other species studied. At both study sites, it For other treatments, however, some trends became apparent in was a dominant species in terms of size and frequency. This tree the first year. tends to become hollow with age and bums rapidly. Hence, its Based on comparisons of stem weight by the different species at value for wood or fuel is limited. Catingueira is one of the first trees the 2 sites, there are significant production differences among to produce leaves when seasonal rains begin, and the foliage is species in response to season of cut. Catingueira produced more eagerly sought by livestock. After a few days the foliage acquires a stem biomass when cut at either time in the dry season, than if cut strong, pungent smell and is ignored by livestock until leaves have in the early wet season (Tables 1 and 2). The season of cutting had dried and fallen at the end of the wet season. This characteristic little effect on marmeliero production. The stem weight of pau ensures that catingueira foliage will be available for consumption during the dry season. Table 1. Mcm coppiceproduction (g dry wt) of individualplants by aaaaon Marmeliero is a small tree, rarely exceeding 8 m in height or 10 of cutting in Sobml, 1983. cm in diameter. It has limited value for construction or firewood and is not normally browsed by livestock. If sufficient rain occurs Sample Stem Leaf Total Leaf/ stem during the dry season, marmeliero will produce leaves, and live------SiZC wt wt wt ratio stock will browse it at this time (Pfister and Malechek 1986). This n x x x x tree tends to form dense stands and is considered an invader, often dominating cleared pastures or disturbed areas. Treatment/species Catingueira Jurema preta was the only evergreen species studied. Although early dry 12 312a 286a 591 ab .89b avidly browsed year-round by all types of livestock, the compound mid-dry 30 391 a 341 a 731 a .91 b early wet 17 139c 271 a 410b 1.88 a leaves are very small and frequently beyond reach of browsers. The dense wood is valuable for fenceposts, firewood, and charcoal. Marmeliero Pau branco is used for construction, furniture, and fuelwood. early dry 8 215 a 124a 339 a .52 a The foliage is not normally browsed by sheep and goats, although middry 11 166a 138 a 304a .83 a cattle find it palatable (Braga 1960). early wet 12 104a 14Oa 245 a 2.02 a Pau Branco Treatments and Experimental Design early dry 20 569b 321b 890b .59 b Cutting treatment consisted of traditional land clearing by mid-dry 30 735 a 442a 1176a .67 b commercial woodcutters using hand tools. The date of the seasonal early wet 18 349c 275 b 624e .88 a treatments was determined by the rainfall pattern. Because trees are normally cut in the mid-dry season, this was considered a Sabia early dry 19 399a 241a 639 a .64C control treatment. The early dry-season plots were cut in July 1982, middry 31 297ab 212a 508 a .74 b mid dry-season plots in November 1982, early rainy-season plots in early wet 19 231 b 198a 429 a .89 a late January 1983, and late wet-season plots in May 1983. All For each species values within a column followed by the same letter arc not signifi- woody plants were cut, and the usable wood was removed. Plots cantly different (pI.05) by the LSD test (Steel and Tonic 1980). were protected from grazing throughout the study.

478 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 Table 2. mean coppice production (g dry wt) of individuel planta by eeaeon of cuttiag in Quiudi, 1983.

Stem wt Lcafwt Total wt Leaf/ stem ratio n 2 n j? n x n x Treatment/ species Catingueira early dry 18 249 a ------middry 28 204a - - - early wet 18 41 b -

Jurema Preta early dry ;: 3974 a 18 261 b E 4012 a 15 .08 b mid-dry 1837 b z 392 ab 229 b 20 .22 b early wet 22 1421 b 523 a 22 1944b 20 .4Ob late wet 17 17c 17 18c 17 6c 6 2.10 a Mamicliero early dry 19 137a -- mid-dry 23 ll34a - - -- early wet 19 214 a -- - Sabia esrly dry 18 326 b -- mid-dry 25 591 a ------early wet 19 351 ab ------

For each specks valueswithin a column followed by the same letter are not significant diiffmnt (X.05) by the LSD test (Steel and Tonic 1980). branco was significantly different for each season of cutting (Table In Sobral, the season of cutting did not affect leaf production of 1). Unlike most species, it produced most when cut in the mid-dry any species except pau branco. As with stem production, leaf season. There were differences in stem production by sabii on the 2 production was maximized by mid dry-season cutting. Dry season sites (Tables 1 and 2). In Sobral, maximum stem production leaf biomass of jurema is difficult to evaluate because the species is occurred after early dry-season cutting. Production following mid gradually exchanging leaves during this period. Leaf biomass from dry-season cutting was intermediate between production after late wet-season treatments was lower from other treatments. For early dry-season and early wet-season cutting. In Quixada, mid all species except marmeliero, the ratio of leaf biomass to stem dry-season cutting maximized sabia stem production. Jurema biomass decreased with increasing time after treatment, support- preta produced the most when cut in the early dry season (Table 2). ing the observation that leaf production dominates the early phase In 1983, leaf fall occurred simultaneously at both sites, preclud- of coppice growth. ing collection of leaf production data for most species in QuixadP. 1984 In Sobral, where leaf and stem data were available, the pattern of BiomassProduction Treatments were completed during the 1983 growing season, biomass production with season of cutting was similar for the stem allowing all the remaining stumps an equal amount of growing and leaf fractions. Although leaves are the primary forage compo- time during 1984. Data were not available for the Sobral site in nent, leaf material is more subject to loss or damage from storms or 1984, restricting further comparisons between the 2 sites. insects than stem material. Therefore, it should be noted that stem No mortality occurred as a result of cutting, and results from biomass was more readily measured and hence the most reliable 1984 were consistent with those from 1983. For all species except variable examined. Table 3. Mean coppice production (g dry wt) of individuel plents by season of cutting in Quhdi, 1984.

Stem wt Leafwt Total wt Leaf/ stem ratio n x n x n x n x Treatment/species Catingueira’ early dry 12 1604 11 620 11 2250 11 .58 mid-dry 18 796 18 279 18 1075 18 .33 early wet 10 898 11 466 10 1390 10 .66 late wet 12 406 12 225 12 631 12 .66 Jurema Preta early dry 12 5317a 12 524a 12 5841 a 12 .08 a middry 14 2082 b 14 191 a 14 2213 c 14 .lOa early wet 12 3189 b 12 506a 11 4004b 11 .13a late wet 11 767 c 9 167 a 9 1061 c 9 .39 a Marmeliero early dry 12 438 a 12 181 be 12 619a 12 .58 a mid-dry 16 613a 16 289 a 16 902a 16 .59 a early wet 15 581 a 15 263 ab 15 844a 15 .52 a late wet 12 313a 12 151 c 12 463a 12 .93 a Sabia early dry 13 1186a 13 591 a 13 1777 a 13 .52 a mid-dry 19 1584 a 19 754a 19 2338 a 19 .48 a early wet 13 1330 a 13 540a 13 1871 a 13 .41 a late wet 13 457 b 13 225 a 13 681 b 13 .53 a For each specka values within a column followed by the same letterarc not significantlydifferent (PC.03 by the LSD test (Steel and Torrie 1980). ‘This data not suitabk for analysisof variance-see text.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1999 479 marmeliero, stem biomass production by trees cut in the late stump and roots. How long the stump must depend on these wet-season continued to lag behind production by trees cut in any limited resources before new leaves are produced becomes a critical other season (Table 3). As in 1983, marmeliero stem production factor in its survival. was not affected by the season of cutting. Jurema preta production In summary, cutting trees in the caatinga during the dry season was maximized by cutting early in the dry season. Only late wet- allows time for the tree to coppice and the new shoots to produce season cutting affected subsequent sabi& production. An analysis and store carbohydrates’ before the next dry season. Cutting late in of variance was not done for catingueira because of large differen- the rainy season does not allow the stump time to replenish lost ces in the variance of the treatment groups. However, it appears carbohydrates, forcing it to survive the entire dry season on limited that early dry-season cutting might maximize stem production. resources. As a consequence, minimal reserves are probably avail- None of the species’ leaf biomass production was significantly able for early rainy-season growth, and this may affect future affected by the season in which the tree was cut, nor was the production. leaf/stem biomass ratio significantly different between treatments Literature Cited for any species in 1984. The season in which jurema preta, sabia, pau branco, and per- B&nger, R.P. 1979. Stump management incxtxses coppice yield of syca- haps catingueira are cut influences the amount of stem tissue more. South J. Appl. Forest. 3:101-103. regrowing from remaining stumps. This response has important BWre, TJ. 1972. Studies in the lignotubers of Eucalyprus obliqua L’Herit. management implications. Desirable species that respond to sea- III. The effects of seasonal and nutritional factors on dormant bud son of cut should be cut during the dry season to maximize future development. New Phytol. 71:329-334. BI8kc, TJ. 1983. Coppice systems for short-rotation intensive forestry: the production. Less desirable species should be cut late in the rainy influence of cultural, seasonal and plant factors. Aust. Forest. Res. season to minimize coppicing. This scheme favors production of 13:279-291. desirable species, perhaps giving them a competitive advantage in Br8y, R. 1960. Plantas do Nordeste especialmente do Ceari, 2nd ed. the regenerating stand. When browse production is a goal, dry Imprensa Gficial, Fortaleza, Brazil. season cutting reduces the impact of treatment on forage produc- Bucll, J.H. 1940. Effect of the season of cutting in sprouting of dogwood. J. tion in the year of treatment because fallen leaves can be utilized Forest. 3g549-650. before land clearing begins. C8nmrgo, M. 1965. An outline of the Brazilian soils. Proc. 9th Intemat. The lack of tree mortality suggests that seasonal cutting treat- Grassl. Cong. 9:9-16. ments alone are not sufficient to influence plant density. However, Cook, C.W., 8nd J. Stubbendieck (eds.). 1986. Range research: basic problems and techniques. Sot. Range Manage., Denver. a reduced biomass response may indicate increased vulnerability to Cremer, K.W. 1973. Ability of Eucalyptus regnans and associated ever- subsequent treatments such as browsing, burning, or slashing cop- green hardwoods to recover from cutting or complete defoliation in pice growth (Hardesty and Box 1988). different seasons. Aust. Forest. Res. 69-22. The fact that leaf production in the second year was not affected DeBeU, D.S., 8nd L.P. AUord. 1972. Sprouting characteristics and cutting by season of cut, and that the leaf/stem ratio tends to decline over practices evaluated for cottonwood. Tree Planter’s Notes. 23:1-3. time suggests that season of cutting cannot cause any long-term Eiten, C., 8nd R. Coodl8nd. 1979. Ecology and management of semi-arid shift in leaf production relative to stem production. Observations ecosystems in Brazil. P. 277-301. In: B.H. Walker, ed. Management of of coppice growth in the third and subsequent years demonstrate semi-arid ecosystems. Elsevier Scientific Publ. Co., Amsterdam. that the leaf/stem ratio continues to decline for several years Guenther, W.C. 1964. Analysis of variance. Prentice Hall, Englewood Cliffs, N.J. beyond the 2-year period reported here. H8rdeaty, L.H. 1984. The challenge of integrated brush management in the Browsing livestock consume large quantities of foliage and semi-arid tropics. Rangelands 6~249-253. smaller amounts of shoot tips, branches, flowers, and fruit (Pfister Hlrdesty, L.H., 8nd T.W. Box. 1988. Defoliation impacts on coppicing and Malechek 1986). By the end of the second growing season, browse species of the Brazilian caatinga. J. Range Manage. 4156-70. coppice growth had exceeded the reach of browsing animals, and Huss, D.L. 1976. Livestock production and range management in north- most browse consisted of fallen leaves. This growth form consti- east Brazil. Rangeman’s J. 3: 175-176. tutes a natural storage system for leaves which, once abcised, Knmer, PJ., 8nd T.T. KozIowski. 1960. Physiology of trees. McGraw- become an important forage source in the dry-season. This study Hill, New York. demonstrates that leaf production can only be increased by maxim- Wster, J.A., 8nd J.C. M8leckek. 1986. Dietary selection by goats and sheep in a deciduous woodland of northeast Brazil. J. Range Manage. izing total growth, a large portion of which will be stem. 39~24-29. Our study was designed to determine if season of cutting, a PrIngle, W.L., Elliott, C.R., 8nd J.L. Dobb. 1973. Aspen regrowth in factor known to affect coppicing of temperate trees, also affects pastures on the Peace River region. J. Range Manage. 26~260-262. coppicing in the caatinga. Although the study was not designed to R8mos, A.D. 1981. Nocoes de pedologia, Empresa de Pesquisa Agrope.- test hypotheses regarding the mechanisms involved, some cautious quaria do Ceari. Fortaleza, Brazil. speculation may be useful in interpreting our results. Smith, D.M. 1986. The practice of silviculture, 8th ed. John Wiley and Many temperate trees produce the greatest coppice growth when Sons Inc., New York. cut during the dormant season (Buelll940, Blake 1972, DeBelI and St8tistic8l Anrlysia Systems. 1986. SAS Version 5. SAS Institute, Gary, Alford 1972, Pringle et al. 1972, Steinbeck et al. 1972, Cremer NC. Steel, R&D., 8nd J.H. Torrie. 1980. Principles and procedures of statis- 1973, Belanger 1979). In the semiarid tropics, this period corres- tics. McGraw-Hill Book Co., New York. ponds to the dry season. In our study those species that were Steinbeck, K., R.G. McAlpine, 8nd J.T. M8y. 1972. Short-rotation culture affected by season of cut produced the greatest coppice growth of sycamore: a status report. J. Forest. 70~210-213. when cut at some point during the dry season. When a tree is W8lters, G.A. 1972. Coppicing to convert small cull trees to growing stock leafless, it still has a continuing energy demand for respiration. USDA Forest Serv. Res. Note PSW-272. Pac. S.W. Forest and Range This is true even if a major portion of its biomass is removed by Exp. Sta. Berkeley, Calif. cutting. Most nonstructural carbohydrates in trees are stored in the Wllters, G.A., 8nd H.L. Wick. 1973. Coppicing to convert cull Australian above-ground woody material (Kramer and Kozlowski 1960). toon, tropical ash, to acceptable trees. USDA Forest Serv. Re-s. Note PSW-283. Pac. S.W. Forest and Range Exp. Sta., Berkeley, Calif. When a tree is cut, the stump has only the reserves remaining in the

480 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1988 Effects of dormant-season herbage removal on Flint Hills rangeland

LISA M. AUEN AND CLENTON E. OWENSBY

Ab8trWt Intensive-early stocking in the Kansas Flint Hills has greatly standard plots, herbage was removed on 1 May by burning in 1984 increased livestock production efficiency. The potential grrziap of and by mowing and raking in 1985 because of an earlier spring. regrowth on intensive-early stocked Flint Hills pastures was stu- Hulbert (1969) showed no significant difference between plots died by monthly mowing to km height from October to April, denuded by burning and plots denuded by mowing and raking. The 1983-1985. Those treatments had no effect on total nonstructural 3 X 3-m plots were mowed to a 5cm stubble height with a sicklebar carbohydrates (TNC) in An&t~pogongerudiiVitmu~ rhizomes or mower, and the cut herbage was either raked off or returned evenly on herbage production the foliowlng seasons. Since there was no over the plots. reduction in herbage yield for any mowing date, cattle producera Herbage Production can apparently restock IES pastures after 1 October. Herbage production for all treatments was determined from Key Words: total nonstructural carbohydrates, Amfropogon ger- fixed rn2 subplots by hand clipping to 5 cm on 15 May, 1 June, 15 urd& winter removal, nexr-infrared reflectawe spectroscopy June, 1 July, and 15 July, 1984 and 1985. All herbage samples were dried for 72 hr at 55’ C and weighed to estimate dry-matter Kansas Flint Hills range is converted to saleable red meat more production. efficiently by intensivecarly (IES) stocking than by season-long stocking. IES results in higher animal gains per hectare without Total Nonstructural Carbohydrates (TNC) sacrificing individual animal performance. Thus, producers realize Rhizomes from 6 big bluestem plants were collected from each more profit compared to season-long systems. In addition, IES plot every 2 weeks from 1 October 1983 to 1 May 1984 and 1 cattle can be sold in July, when traditionally fewer cattle are sold October 1984 to 1 May 1985 and monthly from 1 June to 1 and prices are higher. September of both years. No plants were removed from the her- IES consists of stocking at double the season-long rates during bage production subplots. The rhizomes were placed in paper bags, the first half of the growing season, then removing the animals and dried for 72 hr at 55’ C, and stored. After completion of the study, allowing the forage to regrow the remainder of the season in order all rhizomes were cleaned with cold water washing, redried for 48 to replenish carbohydrate reserves (Launchbaugh and Owensby hr, roots were removed, and rhizomes ground with a Udy Cyclone 1978). Regrowth on IES pastures was responsible for higher her- Mill (l-mm mesh), and stored in a dark plastic vial. The rhizomes bage yields at the end of the growing season compared to season- were analyzed for TNC (mg l g’) using a Technicon InfraAlyzer long stocked pastures (Smith and Owensby 1978). Winter grazing 400 (I/A 400) near-infrared reflectance spectrophotometer. The of regrowth on IES pastures would be an alternative to feeding hay combination of wavelengths 1680,1778,1818,1940, and 2348 nm and provide producers greater flexibility in purchasing cattle. were used for analysis. These gave the best multiple regression The objective of this study was to determine the effects of winter equation between laboratory values and predicted values with a herbage removal on subsequent herbage production and on TNC 0.9871 multiple correlation coefficient (Wetzel 1983). reserves in big bluestem (Andropogon gerurdii Vitman) rhizomes, The I/ A 400 was calibrated using 40 rhizome samples from the a key dominant of the Flint Hills. comparison standard treatment and selected across all sampling dates to insure a range of TNC concentrations (Wetzel 1983). The Materials and Methods TNC content of these samples were predetermined using a dual- enzyme method (Khaleeluddin and Bradford 1986). TNC concen- Location trations have been shown to be highly correlated with pool sizes in The study area was on the Konza Prairie Research Natural Area another group of warm-season, tallgrass species, the Old World near Manhattan, Kansas (Hulbert 1985). The plots were located on bluestems (Bothriochloa sp.) (Coyne and Bradford 1987). a loamy upland range site with a Benfield-Florence complex soil (Benfield series: fine, mixed, mesic, Udic Argiustolls; Florence Soil Moisture series: clayey-skeletal, montmorillonitic, mesic, Udic Argiustolls). Soil moisture was determined gravimetrically on 1 May, 1 June, The site was burned by a wildfire in early spring of 1982. 15 July 1984, and 1 May, 15 May, 15 June, 15 July 1985. Two soil samples were taken from the O-15 cm layer of each plot. The mean Vegetation of those 2 samples was used in the analysis. No soil samples were The vegetation within the plots was typical of the Flint Hills taken from the herbage production subplots. range ecosystem. The dominant grasses were big bluestem, Indian- grass (Sorghasrrum nuruns (L.) Nash), and little bluestem (Andro- Experimental Design and Analysis pogon scoparius Michx.). The study was conducted using a randomized complete block design with 3 blocks. A varying 2-3% slope was the blocking Treatments factor. Data were statistically analyzed using SAS PROC GLM Treatments were mowing and removing herbage or mowing and (SAS 1982) for unbalanced data. Where significant differences leaving herbage on different plots on the first of each month from were found least squares mean separation was at P q 0.10. Other- October 1983 to April 1984and October 1984 to April 1985. These wise, probabilities of significance are reported. 2 treatments were applied on 7 dates over 2 years. Snow cover eliminated December and February treatments. For comparison Results and Discussion The authorsarc research assistant and professor of range science, Department of Agronomy, Kansas State University, Manhattan 66506. Herbage Production Contribution No. 88-262-J from the Kansas Agricultural Experiment Station. None of the winter mowing treatments reduced herbage yields in Manuscript accepted 1 I July 1988.

JOURNAL QF RANGE MANAGEMENT 41(6), November 1988 481 320 1983 1984 1985 _ 280 : 240 V !g 200 5 160

._=a 120 ; 80 a 40 z 0 JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND

d JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND Month Fig. 1. Total monthly precipitation (mm) and deviation from normal at Manhattan, Kansas from 1983 through 1985. the following season when compared to comparison standard ing winter mowing had more total soil moisture than plots on (p=O.OOOl), with the herbage yield averaging 3,145 kg/ha on 15 which herbage was removed (prO.04). There were no differences July 1984 compared to 2,386 kg/ ha on 15 July 1985. Precipitation among mowing treatments in total soil moisture in the O-15 cm during May and June was higher in 1984 than 1985 (Fig. 1). It is layer (m.16). likely that the greater available soil water during the 1984 growing season produced the higher herbage yields in 1984 versus 1985. Conclu!3ioll8 Removing all herbage by mowing to 5 cm on a given date during Total Nonstructural Carbohydrates (TNC) the winter should be more detrimental to subsequent herbage Two consecutive years of winter mowing did not significantly yields than selective herbage removal through the winter by graz- reduce TNC at any date for any treatment the second year ing. Since winter mowing to 5 cm on Flint Hills range did not (Hl.66). Owensby et al. (1974) showed that intensive herbage reduce herbage yields or TNC concentration in big bluestem rhi- removal by clipping through the growing season did not lower .zomes the following season, cattle producers apparently can use TNC reserves in big bluestem rhizomes until the second year IES pastures during the winter after sufficient regrowth has following the clipping. In addition, the time at which cattle are occurred. returned to the IES pastures for winter grazing would be crucial. Owensby et al. (1977) showed that TNC in big bluestem rhizomes Literature Cited was lower during the growing season under IES, but the removal of cattle on 15 July from the IES pastures gave adequate time for the Coyne, P.I., nnd J.A. Bradford. 1987. Nitrogen and carbohydrate parti- big bluestem to regrow and translocate carbohydrates to the rhi- tioning in Caucasian and WW-Spar Old World Bluestems. J. Range zomes before the onset of dormancy. They showed that under IES, Manage. 40:353-360. the translocation of carbohydrates occurred primarily during Sep Hulbart, L.C. 1969. Fire and litter effects in undisturbed bluestem prairie in tember, with little movement thereafter throughout the winter Kansas. Ecology 50:874-877. Hulbert, L.C. 1985. History and use of Konza Prairie Research Natural season. Therefore, restocking before 1 October would result in Area. The Prairie Scout 5:63-93. lower carbohydrate storage and lower herbage production in the Ktmleeluddin, K., and L. Bradford. 1986. Dual enzyme method for deter- following season. mination of total nonstructural carbohydrates. J. Assoc. off. Anal. TNC in big bluestem rhizomes was lower in 1984 than 1985 Chem. 69162-166. (BO.03). Precipitation during July and August of 1983 and 1984 Knapp, A.K. 1985. Effect of fire and drought on the ecophysiology of was lower than in July and August of 1985 (Fig. 1). Knapp (1985) Andropogon gerardii and Panicum virgatum in a tallgrass prairie. Ecol- reported that drought stress occurred in big bluestem from July to ogy 6643094320. mid-August of 1983, during which net photosynthesis declined to Lattnchbaugh, J.L., and C.E. Owensby. 1978. Kansas Rangelands: their near zero. Following precipitation in September, photosynthetic management based on a half century of tesearch. Kansas Agr. Exp. Sta. Bull. 622. rates in big bluestem returned to about half the season maximum Owensby, C.E., J.R. Raim, and J.D. McKcndrict. 1974. Effects of one (Knapp 1985). Lack of precipitation in July and August 1984 also year of intensive clipping on big bluestem. J. Range Manage. 27341443. could have caused some drought stress in big bluestem. Lower Owensby, C.E., E.F. Smith, and J.R. Rnbm 1977. Carbohydrate and photosynthetic rates caused by the mid-season drought likely pro- nitrogen reserve cycles for continuous, season-long and intensive-early duced less carbohydrate, resulting in less being translocated to stocked Flint Hills bluestem range. J. Range Manage. 30558-260. rhizomes. The combination of possible lower carbohydrate storage SAS Institute, Inc.. SAS User’s Guide’ Basics, 1982 Edition. Gary, NC SAS in 1983 and the mid-season drought in 1984 probably caused the Institute Inc., 1982. ’ lower TNC in 1984 compared to 1985. Smith, E.F., and C.E. Owenaby. 1978.Intensivezarly stocking and season- long stocking of Kansas Plint Hills range. J. Range Manage. 31:14-17. Soil Moisture Wetzel, D.L. 1983. Near-infrared reflectance analysis: sleeper among tpec- Plots on which herbage was returned to the soil surface follow- troscopic techniques. Anal. Chem. 55: 1165A.

482 JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 Stocking rate effects on intensive-early stocked Flint Hills bluestem range

CLENTON E. OWENSBY, ROBERT COCHRAN, AND ED F. SMITH

Stocking rate effects on intensive-early stocked Kansas Flint without affecting grassland productivity the following season. Hills range were studied from 1982 tbrougb 1987. Rates were 2X, Owensby et al. (1977) had indicated earlier that the lack of grazing 2.5X, and 3X normal season-long stocking rates for 200-225 kg from mid July until frost allowed for adequate storage of reserve steers. Study design was a randomized complete block witb 2 carbohydrates. Since Smith and Owensby (1978) also reported that replicates. Grass and forb standing crop (kg/ha) were estimated slightly more than 1,200 kg/ha herbage remained when livestock at the time of livestock removal (mid July) and again in early were removed in mid July, evaluation of the potential for using October. Plant bas81 cover 8nd composition were taken in e8rly higher stocking rates with intensivecarly stocking appeared to be a June the year prior to the study and annuaily thereafter. Overall logical follow-up study. growing season precipitation during tbe study period was below We studied the effects of stocking densities on Kansas Flint Hills norm81,with late-summer precipit8tion much below normal in the bluestem range at 2,2.5, and 3 times the recommended season-long second and third years of the study. Grass standing crop (GSC) in rate from 1 May to mid July on plant populations, herbage yield, mid July decreased witb increased stocking rate, but by early and animal gains. October GSC was similar under the 2.5X and 3X stocking rates, Materials and Methods but continued to be lower than that under the 2X rate. There was no consistent response in mid July forb standing crop (FSC) with The study area was the Kansas State University Experimental respect to stocking rate. In early October, FSC was either not Range Unit located in the northern Kansas Flint Hills near Man- affected by stocking rate (1983, 1986, and 1987) or was greater hattan, Kans. The warm-season, perennial grasses, big bluestem under tbe bigbest stocking rate (1982,1984, and 1985). Tbe major (Andropogon gerardii Vitman.) and indiangrass (Sorghastrum changes in botanical composition and basal cover were a reduction nutonsNash), were the dominant forage species and little bluestem in Indiangrass (Sorgh&run# nutuns Nash) and an increase in (A. scoparius Michx.) and sideoats grama [Bouteloua curtipen- Kentucky bluegrass (Poa prutennb L.) as &eking rate increased. dulu Michx.) Torr.], both warm-season, perennial midgrasses, Botanical composition of big bluestem (Anhopogon gerar& were subdominants. Numerous grass, forb, and woody species Vitman) increased under tbe 2X rate but did not change under the constituted the remainder (Anderson and Fly 1955). Soils were bigber rates. Individual steer gains were similar under the different transitional from Ustolls to Udolls. The principal range sites in the stocking rates, but livestock breed appeared to affect magnitude of study area were loamy upland, breaks, and clay upland (Anderson the gain. Since individual gains did not differ, gains per ba were and Fly 1955). The study pastures had similar amounts of each of substantiaiiy increased by tbe bigber stocking rates. the principal range sites. Each year from 1982 through 1987 on 1 May, six 24.3-ha pas- Key Words: Andropogon, grazing system, stocking rate, iiveatock tures were stocked with yearling steers weighing 200 to 225 kg. gh Stocking rates of 0.47,0.57, and 0.70 ha per steer (3X, 2.5X, and The highquality forage period for warm-season perennial @GE- 2X, respectively) were applied in duplicate on the 6 study pastures lands is limited to the first 2 1/ 2 months of the growing season in which were burned in late April each year. The recommended the Kansas Flint Hills. Grazing at times other than the highquality season-long stocking rate (X) for steers of that size is 1.41 ha for the period results in livestock gains that are sub-optimal. The goal of 5-month grazing season which begins in late April to early May. any rangeland-based program for growing livestock should be Steers were identified individually and weighed on 30 April and 16 maximum efficiency in converting forage to animal product. Smith July each year. Steers were confined without feed and water from and Gwensby (1978) showed an increased conversion efficiency the previous afternoon until they were weighed the following with intensivezarly stocking of Kansas Flint Hills range, with morning. twice the recommended stocking density for the first half of the Each year grass and forb standing crop was estimated at live- growing season and no grazing during the latter half of the season. stock removal in mid July and again in early October by clipping The increased conversion efficiency manifests itselfin increased net IO, O&m2 plots to ground level in both loamy upland and breaks returns when compared to season-long stocking (Bernard0 and range sites. Clipped samples were dried to moisture-free. McCollum 1987). Livestock removed in mid-July usually are Species composition was determined from basal cover estimates placed directly into the feedlot. Posler et al. (1985) studied alterna- (percent of ground covered by plant bases) using the modified tives for managing beefyearlings after intensive-early stocking that step-point method (Owensby 1973). Pastures were sampled annu- included direct placement into the feedlot, grazing sudangrass ally during the first half of June. Within each pasture, 1,500 points (Sorghum bicolor var. sudcrnense Hitchc.), and grazing alfalfa were read along a predetermined grid. Each point was recorded as (Medicugo sutiva L.). They concluded that direct placement into to range site. Pre-treatment species composition and basal cover the feedlot in mid summer was the management strategy of choice. were determined in 198 1, and analysis of variance was conducted Auen and Gwensby (1988) studied winter removal of dormant on the difference between pre-treatment percentages and percent- Flint Hills bluestem herbage and indicated that Flint Hills ranchers ages at the end of the study. could utilize herbage remaining after intensiveearly stocking Plant standing crop data were analyzed as a randomized complete block design using SAS-ANOVA (SAS Institute, Cary, NC) Authors arc professor of range management, assistant professor of animal science, with replication, stocking rate, range site, and year as the variables. and profcaaor emeritus of animal science, Kansas State University, Manhattan 66506. Contribution No. 88488-5, Kansas State University Agricultural Experiment Sta- Plant census data were analyzed similarly except year was not in tion, Manhattan 66506. the model. Probabilities of a significant difference are reported and Mantipt acccptcd I August 1988. means are separated using least significant difference (fi.10).

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 483 320 1

JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND 5 150 1983 1984 1985 1986 1987 E 100 0 75 = 50 5 25 k 0 -25 s -50 G= -75 ij$ -100125 : Q) A JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND JFMAMJJASOND Y Month

Fig. 1. Monthly precipitation and deviation from normalprecipitation from 1982 through 1987 at Manhattan, Kansas. Livestock gains were analyzed using SAS-GLM (SAS Institute, Cary, NC) with replication, stocking rate, and year as classes. Means were separated using least significant difference (P = 0.01). July

Results Precipitation The 30-year average annual precipitation is 84 cm with 52 cm occurring during the growing season. Growing season precipitation during the initial year of the study was above the 30-year average (Fig. 1). Precipitation during the latter half of the growing season in 1983 and 1984 was well below average, and in 1985 was below average until late summer. In 1986, growing season precipitation was above average, but in 1987 it was below average. Overall precipitation throughout the study was below average. r Grass Standing Crop Grass standing crop (GSC) at the time of livestock removal varied among treatments and years but not between range sites (Fig. 2). During most years, GSC was greater for the 2X rate than L for the 2.5X and 3X rates on both major range sites. In 2 of the 6 1986 1987 LSD 0.10 years, GSC was greater on the 2.5X rate pastures than on the 3X, Year but was similar during the other 4 years. In 1985, there appeared to be no difference in GSC among pastures stocked at different rates. Fig. 2. Grass standing crop (kg/ha) in mid July for indicated years and During the latter half of the 1983 and 1984 growing seasons, there stocking rates on intensive-early stocked Flint Hills range. Data represent an average over loamy upland and breaks range sites. was essentially no regrowth on any treatment pasture because of insufficient precipitation. That likely accounted for the reduced that sustained herbage production was likely at any of the rates herbage production during the early season 1985. Even though tested. there was substantial yearly variation in GSC at livestock removal, Average GSC on the loamy upland and breaks range sites in no downward trend was apparent at any stocking rate, indicating early October was greater (KO.01) under the 2X rate (2,132

484 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 700. July

600. Loamy Upland Breaks - u -f, lD 500. Y -

F 400. 0

P z 300, 6 z ~ 200,

b G loo-

0. 1982 1983 1985 387 1982 1983 1984 1985 1986 1987LSD 0.10 Year Fig. 3. Forb standing crop (kg/ha) in mid July for indicated years, stocking rates and range sites on intensive-early stocked Flint Hills range. kg/ ha) than under the 2.5X (1,641 kg/ ha) and 3X (1,721 kg/ ha) rates which were not different. GSC varied among years, but Table 1. Change in percent basal cover from 1981 to 1987 for indicated differences among stocking rate were similar each year. Weather speciea and range sites stocked at 2, 2.5, and 3 times the season-long appeared to be the primary determinant as to the amount of GSC recommended rate from 1 May to 15 July.1 in a given year within each stocking rate. There was no downward trend at any stocking rate. Stocking Rate Forb Standing Crop 2x 2.5x 3x Amount of forb standing crop (FSC) in mid July under different stocking rates was dependent on year and range site (Fig. 3). There Loamy Upland Sorghasrrum nutans was no consistent response to stocking rate. Apparently, there was 1981 1.42 1.83 1.78 an interaction with weather fluctuations that resulted in a variable 1987 1.25 1.51 1.27 response to stocking rate on the different range sites. Change -0.17a -0.32ab -0.5lb In early October FSC either was not affected by stocking rate (1983, 1986, and 1987) or was greater under the higher stocking BWkS Andropogon scoparius rates (1982, 1984, and 1985). Forb standing crop under the 2.5X 1981 1.26 0.91 0.77 rate was greater than under the 2X rate only in 1984. Stocking rate 1987 0.80 0.90 0.33 effects on FSC were similar among years on the 2 principal range Change -O&a -O.Olb -0.44a sites. Boureloua curtipendula Basal Cover and Composition 1981 1.96 1.03 1.16 Basal Cover 1987 1.10 0.94 0.82 There was little change in total basal cover during the course of Change -0.86a -0.09b 4.34a the study. However, basal cover of indiangrass on the loamy Perennial Forbs upland site was reduced more under the 3X stocking rate than 1981 0.63 0.15 0.16 under the 2X rate, but the change under the 2.5X rate did not differ 1987 0.30 0.20 0.16 from that of either the 2X or 3X rates (Table 1). On the breaks Change -0.33a +O.OSb O.OOa range site, reduction in basal cover of little bluestem was greater under the 2X and 3X stocking rates than under 2.5X where there ‘Values with a common letter do not differ (P = 0.10). was no change from pre-treatment levels (Table 1). Sideoats grama basal cover was reduced more under the 2X and 3X stocking rates Botanical Composition than under the 2.5X stocking rate. Perennial forb basal cover was Only 3 species changed in botanical composition from pre- essentially unchanged under the higher stocking rates but was treatment levels. On the loamy upland site, stocking at the 3X rate reduced under the 2X rate. resulted in a greater reduction in indiangrass compared to pretreatment levels than the lower stocking rates (Table 2). Prior to

JOURNAL OF RANGE MANAGEMENT 41(6), November 1999 485 T8bic 2. Ch8ngc in percent composition from 19111to 1987 for indierted

species 8nd r8np sitea stocked 8t 2,2.5,8nd 3 tima tbe wson-long 600 recommended r8te from 1 M8y to 15 July.1 October

I Stocking Rate 2x 2.5X 3x Loamy Upland Sorghastrum nutans 1981 21.1 26.8 21.5 1987 19.5 23.0 17.0 Change -1.6a -3.8b -4Sb Poa pratensis 1981 0.2 0.8 4.4 1987 0.9 2.0 9.4 Change +0.7a +i.& +5.Ob

Breaks .G 8 Andropogon gerardii *i 1981 31.2 32.4 27.8 ‘* :t. 1987 36.8 32.1 27.9 .: Change +5.6a -0.3b +O.ib 1985 llal-1987 LSD0.10 1fear Sorghastrum nuians 1981 18.8 22.9 18.5 Fig. 4. Forb standing (kg/ ha) and 1987 17.8 18.2 17.0 stocking Data repres- Change -i.Oa 4.7b -1.5a - ent over loamy upland and breaks

Walues with a common letter do not differ (P q 0.10). from a single source and were British X Zebu crosses. Since individual animal gains were the same among stocking rates, gain per the study, indiangrass made up 21.5% of the stand compared to hectare was increased linearly as stocking rate increased (Table 3). 17.0% at the end. Kentucky bluegrass (Poaprurensis L.) increased from 4.4% to 9.4% of the composition on the loamy upland site Discussion under the 3X stocking rate. There was no change in Kentucky bluegrass under the lower stocking rates (Table 2). Under intensive-early stocking, precipitation during July and On the breaks range site, big bluestem increased from the pre- August is particularly important for regrowth, and only in 2 of the treatment level of 3 1.2% of the composition to 36.8% at the end of study years was precipitation average or above average. Owensby the study under the 2X stocking rate, whereas percent big bluestem et al. (1969) reported earlier that defoliation during the growing at the higher rates was essentially unchanged from pre-treatment season reduced water usage. Intense defoliation during the early levels (Table 2). Percent indiangrass on the breaks site was essen- growing season may conserve sufficient soil moisture to allow for tially unchanged from pm-treatment levels under the 2X and 3X adequate growth during the late season even during years with stocking rates, but under the 2.5X rate there was a reduction from below average late-season precipitation. Since there was no down- 22.9% to 18.2% (Table 2). ward trend trend in GSC in October as evidenced by the lack of a significant year X stocking rate interaction, apparently Flint Hills Steer Gains range can sustain even the 3X stocking rate for a prolonged period. Steer gains (kg/head) varied among years, but within a given For any stocking rate we tested, no grazing from mid July to frost year, they were similar on pastures stocked at different rates (Table likely allows for sufficient time to store adequate food reserves for 3). Differences in steer gains among years were likely due to vigorous growth the following spring. Changes in composition from pre-treatment levels primarily T8bie 3. Steer pins (kg/steer 8nd kg/h8) on Kansas Flint Hlhs range involved a decrease in indiangrass and an increase in Kentucky intenaivcc8rly stocked from 1 M8y to 15 July 8t 2,2.5,8nd 3 ttmed the bluegrass at the higher stocking rates. Apparently, indiangrass, recommended serson-long stocking r8te.l unlike big bluestem (Smith and Owensby 1978), is not favored by intensive use early in the season followed by a late-season rest. Stocking Rate McKendrick et al. (1975) indicated that new tillers for the next 2X 2.5X 3X Mean 2X 2.5X 3X growing season in indiangrass arise in late summer. Perhaps, the intensive use during early summer does not allow for sufficient Year (kg/steer) (kg/ ha) photosynthetic area to generate tillers following livestock removal. 1982 63 58 62 61 a 86 96 128 The increase in Kentucky bluegrass at higher stocking rates most 1983 60 55 62 59a 83 91 127 likely resulted from reduced fire intensity because of a lower fuel 1984 75 75 76 76b 103 124 156 load (Anderson et al. 1970). 1985 94 84 79 86 c 129 138 163 Earlier studies with intensiveearly stocking (Smith and Owensby 1986 84 86 89 86 c ii5 142 182 1978, Bernard0 and McCollum 1987) indicated no reduction in 1987 81 83 85 83 c iii 136 175 Mean 76 74 75 105 121 155 individual animal performance during the first half of the grazing season compared to season-long stocking at stocking rates twice ‘Year means followed by a common letter am not different (KO.01). season-long rates. We found that the increasing the stocking rates to 2.5 and 3 times season-long rates did not reduce individual changes in type of cattle used in the study. In 1982and 1983, steers animal performance compared to 2 times season-long rates. Con- were purchased from local auction facilities and were mostly sequently, gain/ ha increased linearly, resulting in a greater conver- average-frame British (Hereford and Angus) crossbred cattle. Dur- sion efficiency than that found in previous studies. ing the remainder of the study, steers with larger frame size came

486 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 concIusions Auan, L.M., and C.E. Owensby. 19gg. Effects of dormant+cason herbage removal on Flint Hills rang&ml. J. Range Manage. 41:483-484. 1. Grass standing crop at the end of the growing season under Bernudo, D.J., and F.T. McCoBurn. HM7. An economic analysis of 2.5X and 3X stocking rates was less than that under the 2X rate. intcnsivecarly stocking. Oklahoma Agr. Exp. Sta. Rcs. Rep. P-887. There was no downward tnnd in amount of grass standing crop McKeddck, J-D., C.E. Owen&y, and R.M. Hyde. 1975. Big bluestem and over the course of the study at any stocking rate. Apparently, indiangrass vegetative reproduction and annual reserve carbohydrate herbage production can be sustained in all stocking rates tested. and nitrogen cycles. Agro-Bcosptcms 275-93. 2. Percent of composition and basal cover of the major domi- Owensby, C.E., R.M. Hyde, and K.L. And-. 1969. Effects of clipping nants changed little during the study period. Indiangrass appeared and supplemental nitrogen and water on loamy upland bluestem range. J. Range Manage. 23:341-346. to be adversely affected, particularly at the higher stocking rates. Owensby, C.E. 1973. Modifiid step-point system for botanical composi- Kentucky bluegrass was favored by the 3X rate. tion and basal cover estimates. J. Range Manage. 26~302-303. 3. Stocking rate did not affect individual steer gains overall. Owensby, C.E., E.F. Smith, and J.R. Rains. 1977. Carbohydrate and Because of the higher stocking rates, gain per hectare was signifi- nitrogen reserve cycles for continuous, season-long and int&sivc-carly cantly increased. Livestock type apparently had a significant stocked Flint Hills bluestem range. J. Ranac Manaac. 30~258-260. impact on steer gains. Posler, C.L., GM. Ward, J.G. I&y, C.E.?lwensb< E.F. Smltb, K.K. Bolsen, R.M. Helsal, and N.S. IEB. 1985. Alternatives for managing beef Literature Cited yearlings after intensivezarly stocking of Kansas Flint Hills range. p. 1161-1163. In:Proc. 15thIntemationalGrasslandCong. (ads. T. Okubo Andcrum,K.L., and C.L. Fly. 19%. Vegetation-soil relationships in Flint and M. Shiyomi). Kyoto, Japan. Hills bluestem pastures. J. Range Manage. 8:163-169. Smith, E.F., and C.E. Owensby. 1978. Intensive-early stocking and mason- Anderson, K.L., E.F. Smith, and C.E. Owensby. 1970. Burning blue&cm long stocking of Flint Hills blue&cm range. J. Range Manage. 3131418. range. J. Range Manage. 23:81-92.

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 487 Determination of root mass ratios in alfalfa-grass mixtures using near infrared reflectance spectroscopy

M.D. RUMBAUGH, D.H. CLARK, AND B.M. PENDERY

H8nd sep8r8tion of roots of 2 or more plants species from soil quality of several different types of forage (Clark et al. 1987, cores is a tedious 8nd labo~intensive task. Our objective was to Minsonet al. 1983, Shenk et al. 1979, Ward et al. 1982). NIRS was determine whether neu-infrared refierturce spectroscopy (NIRS) used to correctly determine the species composition of artificially could be employed to estimate root biomass proportions in bbmry created forage mixtures that contained as many as 4 species (Coie- mixture8 of 8lfaifa (IlYe&ugo sativu L.) with each of 4 grasses. man et al. 1985). Botanical composition of tail fescue (Festucu Grasses chosen for experimentation were crested wheatgrass arundinacea Schreb.)-white clover (Trifolium repens L.) pasture (Agropyron cr&t&m L.), intermedi8te wbe8tgr8ss [Tltbropynun mixtures was predicted by NIRS with excellent precision and Mefure&uu (Host) Bukwortb & DR. Dewey& an intergeneric accuracy following proper selection of calibration samples, wave- hybrid [E@rigia repens (L.) Nevski X Pseudimegneria spicata lengths, and data transformation (Petersen et al. 1987). Brink and (Pumh) Love], urd Russiur wildrye [Psathyrostachys junea Marten (1986) showed that NIRS could be used to analyze the total (Fiscb.) New&i]. In tbe first experiment, roots from single-species nonstructural carbohydrate content of alfalfa (Medicago sutiva L.) field plots were washed from soil cores, dried, ground, and roots; the technique was more convenient and accurate than con- mechanically mixed in preselected 8lfaKa-grassratios in which the ventional laboratory techniques. Our objective was to determine if percentage of grass varied from 0 to 100. Equ8tions to measure the NIRS could be used to quantify the root biomass ratios of alfalfa proportion of llf8lfa or grass were developed from nar btfr8red and 4 grass species when grown in binary mixtures. reflectance data using 84 r8ndomiy selected sample8 In the second Material8 and Methods experiment, the 5 plant species were grown in greenhouse pots in pure stands urd in birmry mixtures that included all combinations Experiment 1 of the gr8sses. Root systems were seprr8ted while 8tt8cbed to the ‘NC-83-l ‘alfalfa (Kehr et al. 1975), ‘Fairway’crested wheatgrass topgrowth, dried, 8nd ground. Tissues from single species trat- (Rgropyron cristatum L), ‘Oahe’ intermediate wheatgrass [ntin- ments were mixed md calibration eqtmtions developed from these opyrum intermedium (Host) Barkworth & D.R. Dewey], ‘RS-2’ mixture8 were used to estimate the proportion of alf8if8 md the [Elytrigia repens (L.) Nevski X Pseudoroegnerka spicutu (Pursh) proportion of grass in samples. Sunpies contained either one type Love], and ‘Vinail’ Russian wiidrye [ Psu~hyrostuchys juncea of root or 8 mixture of roots in proportions sbnil8r to those th8t (Fisch.) Nevski] plants were started in the greenhouse and trans- occurred n8tur8lly in the pots. Coefficients of determination (rZ) planted to the field near Logan, Utah, in May 1983. All plants were between tbe emtimrtedand the lcttml root mass r8tlos ranged from spaced 0.5 m apart in a rectangular grid pattern. Details of the 0.92 to 0.99. Determination of tbe proportion of grass in the nursery plan were described by Rumbaugh and Pendery (1986). samples was more 8ccunte and precise than determbmtlon of the The experimental field is at an elevation of 1,369 m and receives an proportion of 8lf8If8.After the approprirte crlibrrtion equations average of 440 mm precipitation annually. The soil is a coarse, silty have been developed, NlRS is more efficient than hand separrtion caronatic, mesic, Typic Haplozeroll. for estimrting rlfaifa-grass root mass ratios. The utility of the Soil blocks 25 cm X 20 cm, extending 25-cm deep, and centered techniques can be increased by developing equations th8t encom- between 2 plants of the same population were extracted in July pass more complex mixtures and 8 wider range of environment8l 1986. Two hundred samples were obtained for each of the 5 plant circumstances. species. The samples were stored in plastic bags at 4” C until processed. Soil samples were soaked in water for 1 hour and the Key Words: Agropyron cr&tatwn L., crested wheatgrass, E&M- roots were extracted by careful washing with a fine, pressurized gia repens (L.) Nevski, intermedi8te whatgram, Medicago sativa spray of cold water over a fine screen. Detritus was removed with a L., legume, NIRS, Psathyrostuchys juncea (Fiscb.) Nevski, Rus- forceps. Root samples for each species were bulked, dried in a sian wildrye, Psewforoegneria spicata (Punh) Love, quackgrass, forced-air oven at 40’ C, and ground with a UDY mill to pass species composition, Th&topyrum intennedium (Host) Barkworth through a l-mm screen. & D.R. Dewey, wheatgmss Samples for NIRS analysis were prepared from pure root tissue. The relationships among plants in mixed species swards are of Mixtures of legume-grass root tissue were created by mixing concern to ecologists, physiologists, agronomists, and range scien- known weights of alfalfa and grass. This was performed for all tists. Many investigations involve the quantification of plant densi- legume-grass combinations and the mixtures ranged from lOO% ties and top growth biomass ratios of the components of species alfalfa to 100% grass in increments of 5%. With these combinations, mixtures. Similar measurements of root biomass often are not a total of 100 samples (25 mixtures for each legume-grass species attempted because of the difficulty of correctly identifying and combination) were generated. Of the 100 samples, 84 were ran- separating roots of different species. domly selected for calibration (equation development) and I6 were Since the use of near infrared reflectance spectroscopy (NIRS) used to test the equation. Variables used for equation development to measure forage quality was first described (Norris et al. 1976), and testing were percentage alfalfa and grass root mass. the technique has been employed to determine the amounts and Equations were developed by collecting the spectra on each sample with a scanning monochromator Pacific Scientific model Authors are research geneticist, research scientist, and range scientist, respectively, L;&%AgS Forage and Range Research Laboratory, Utah State Umv., Logan 63501 coupled to a DECi PDP I l/23 computer. Spectra were

Cooperative investigations of the USDA-Agricultural Research Service and the ‘Mention of a trademark or proprietary product does not constitute a guarantee or Utah Agricultural Experiment Station. Approved as paper No. 3575. warranty of the product by the USDA or the Utah Agricultural Experiment Station Manuscript accepted 25 July 1988. and does not imply its approval to the exclusion of the other products that may also be suitable.

488 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1988 collected from l,lOO-2,500 nm at every 2 nm for a total of 700 data cients of variation [CV=(SEP/ mean of the validation data set)X lOO] points per sample. Percent root mass was entered into the compu- were 2.5% for alfalfa and 4.3% for grass. Both values were as low as ter for each sample and multiple linear regression equations were comparable estimates for forage biomass usually found in most developed by regressing the spectral and root mass data. Equations agronomic and ecological studies. Discrepancies between the were selected based on lowest standard error of calibration (SEC), actual and determined means (bias) were -3.7% and 1.9% of the highest R2, and all wavelengths in the equation having an F value > actual means of the alfalfa and grass samples, respectively. 10.0 (Abrams et al. 1987). Selected equations were tested on the Estimated root biomass ratios for alfalfa and the 5 individual remaining 16 samples and success was measured by a low standard grasses were less accurate and precise than simply distinguishing error of performance (SEP), bias, and high rr. between alfalfa and any grass species. SEC were less than 2% and the corresponding R2 values exceeded 0.99 (Table 2). When the Experiment 2 The 5 species used in experiment 1 were established in U-cm Table 2. The proportions of roots of field-grown elfah end 4 greeem in diameter pots in the greenhouse in January 1987. The 15 treat- utiflchl Iegmncgrass mixtures (Experiment 1): a comparieon of actual ments consisted of 4 plants of 1 species or 2 plants of each 2 species percentages vemoe those ddermined by NIRS. in binary mixtures. Alfalfa seedlings were treated with a commercial inoculum as they were transplanted into the pots. The rooting medium was a sandy-loam soil enriched with a nitrogen-free fertil- Actual (n=50) Determined (n=50) izer. The experimental design was a randomized complete block crop Mean SD Mean SD Bias r2 SEP with 25 replications of the 15 treatments. Alfalfa plants were at the -_(%)- d%)- (%) initial-bloom stage and the grasses vegetative when harvested dur- Alfalfa 51.1 29.7 50.4 29.2 -0.7 0.995 2.09 ing the last week of March 1987. Roots from the binary mixtures Fairway 12.4 25.9 10.1 25.1 -2.3 0.965 5.34 were carefully separated by species while still attached to the Oahe 14.4 26.4 16.1 25.7 1.7 0.964 5.25 topgrowth. Roots from plants of a single species were bulked RS-2 6.2 18.5 8.5 20.2 2.3 0.920 6.14 within each treatment, dried in a forced-air oven at 34’ C, and Vinall 15.8 30.6 16.2 29.9 0.4 0.962 5.96 ground as previously described. SD = Standard deviation. Only root tissues from pots containing a single species were used SEP = Standard error of performance. to develop the NIRS equations. Ground roots from each of the 5 Bias = NIRS mean minus the actual mean. species were used alone or were blended in binary mixtures in 25% increments to create samples ranging from 0 to 100% composition equations were tested on 16 other samples. SEP values were for each species. These 35 samples were used to develop 2 multiple approximately 5% and produced high coefficients of determina- linear regression equations. The first equation quantified the pro- tion. However, coefficients of variation ranged from 36.5% when portion of alfalfa tissue in the samples and the second the propor- estimating Oahe intermediate wheatgrass to 99.0% for the RS-2 tion of grass. Not enough grass roots were available to develop species hybrid. The bias in the estimated proportion of alfalfa root equations to estimate individual grass populations. Fifteen test mass was -1.4% of the actual mean. Analogous values for the 4 samples were formulated. Five consisted of roots from the single grasses ranged from 2.5% (Vinall Russian wildrye) to 37.1% (RS-2 species treatments, Ten other test samples were created by reconsti- hybrid). However, in no case was the r* less than 0.92. tuting the binary-mixtures, using root tissues from the same binary Experiment 2 mixture treatments, in proportions identical to the averages of NIRS successfully discriminated between greenhouse-grown those that occurred naturally in the pots. alfalfa root tissue and grass root tissue although the estimates of the proportions of the 2 species were neither as accurate nor as Results precise as in the field study (Table 1). The standard errors of Experiment 1 analysis were 3-fold greater than in Experiment 1 and the biases In many pastures investigations, it is necessary only to distin- were -6.0% and -2.6% of the actual means of the alfalfa and grass guish the root mass of one species, such as alfalfa, from the roots of samples, respectively. However, the coefficients of determination another crop species category, such as grass. Results involving (0.95 or above) indicated very acceptable levels of precision. artificial mixtures show that NIRS could be used to distinguish Determination of the proportion of alfalfa in a sample contain- between root masses with a degree of precision and accuracy that ing only alfalf4 was very accurate as indicated by a bias amounting would be appropriate for most experimentation (Table 1). Coeff- to only 0.3% (Table 3). When NIRS was used to determine the proportion of grass in a sample that contained only alfalfa, the Table 1. Calibration and analysis statistles for NIRS measurement of the resulting estimate was negative. When NIRS was used to deter- percentage8 of atfalfa 8nd gram roota IOertiflchl mixtures. Plant8wed in mine the amount of alfalfa in samples that contained both alfalfa Experiment 1 were grown in tbe field and those ueed in Experiment 2 and grass root tissues, biases ranged from 5.6 to -19.9% of the were grown in the gnenbouse. actual alfalfa means and averaged -8.9%. When the same samples were used to estimate the amounts of the grasses, biases ranged from -4.6 to -28.4% of the actual grass means and the average bias Calibration Analysis was -13.8%. When 2 grasses were included in the samples, biases Crop Mean Terms R* SEC Bias r* SEP ranged from -9.3% for crested and intermediate wheatgrasses to (%) (%) (%) (%) 4.8% for the crested wheatgrass and Russian wildrye mixture. The Experiment I : mean absolute bias of binary grass mixtures was 4.4%. The mean Alfalfa 50.1 5 0.99 1.48 -1.25 0.99 2.15 absolute bias for all NIRS determinations of the proportion of GraSS 49.9 5 0.99 1.48 1.25 0.99 2.15 alfalfa was 6.1% which was the equivalent of 26.5% of the actual Experiment 2: mean proportion of alfalfa in the 15 samples. Similarly, the mean Alfalfa 53.8 2 0.99 1.82 4.14 0.95 6.38 absolute bias for grass was 3.7% which was 5.0% of the mean Grass 84.1 2 0.99 1.73 -2.14 0.98 4.52 proportion of grass. Thus, NIRS determinations of the proportion of grass were more accurate than those of alfalfa in Experiment 2. SEC = Standard error of calibration. SEP = standarderror of jWfomulnce. Terms = number of wavelengths in equation. Discussion Bias = NIRS mean minus the actual mean. Hand separation of roots of 2 or more species is very tedious,

JOURNAL OF RANGE MANAGEMENT 41 (S), November 1999 489 broader range of experimentation if root calibration equations Table 3. Actwl wl determined proportions of 8lhlfa ti gras in bhry root tiaue ample4 (Experiment 2). were developed for specific components of more complex mixtures. The components of such mixtures would have to differ sufficiently so that NIRS procedures could measure differences in Actual proportions(%) Analyzed proportions (%) molecular motion of the chemical bonds associated with their Alfalfa Grass respective organic compounds. Additional research also will be Alfalfa Crested Intermediate RS-2 Vinall Mean Bias Mean Bias required to determine whether it is feasible to use a common equation to measure the proportions of root biomass in samples 100.0 100.3 0.3 -4.3 4.3 collected from different soils with varying fertility levels and pre- 0.0 100.0 9.9 9.9 101.9 1.9 cipitation regimes. Data from the 2 experiments described lead us 0.0 100.0 7.0 7.0 100.6 0.6 to conclude that NIRS technology can be used to develop proce- 0.0 100.0 6.9 6.9 97.2 -2.8 0.0 100.0 7.1 7.1 101.4 1.4 dures to estimate the proportions of root biomass of different plant 78.9 21.1 71.7 -7.2 15.1 -6.0 40.7 59.3 32.6 -8.1 56.6 -2.7 Literature Cited 53.8 46.2 56.9 3.0 38.7 -7.5 71.3 28.7 62.5 -8.8 26.8 -1.8 Abmms, S.M., J.S. Sbcnk, M.O. Waterhu, and F.E. Buton II. 1987. 61.4 38.6 8.2 8.2 90.7 -9.3 Determination of forage quality by near infrared reflectance spectros- g 42.2 57.8 10.7 10.7 93.1 -6.9 copy: efficacy of broad-based calibration equations. J. Dairy Sci. 0:o 75.0 25.0 1.2 1.2 104.8 4.8 70:806-813. 0.0 67.9 32.1 0.3 0.3 101.2 1.2 Brink, GE., and C.C. Muten. 1986. Analysis of alfalfa root carbohydrate 0.0 88.5 11.5 6.2 6.2 100.6 0.6 concentration by near infrared reflectance spectroscopy. Crop Sci. 0.0 67.8 32.2 6.7 6.7 96.6 -3.4 26:159-161. Clark, D.H., H.F. Mayland, md R.C. hmb. 1987. Mineral analysis of Bii = NIRS mean minus the actual mean. forages with near infrared reflectance spectroscopy. Agron. J. 79~485490. Coleman, S.W., F.E. Barton II, and R.D. Meyer. 1985. The use of near- infrared reflectance spectroscopy to predict species composition of for- labor intensive, and subject to major identification errors if the age mixtures. Crop Sci. 25:834-837. roots are not attached to the aboveground growth. For these Keb;, W.R., D.K. boa, E.L. Sorensen, W.H. Skrdla, C.H. Hwnon, reasons, the numbers of samples taken in root-core studies often D.A. Miller. T.E. Thornwon, I.T. Carlson, LJ. EIIInc. R.L. Taylor. are more limited than desired. Less time and labor consuming M.D. Rum&k, E.T. B-hg&m, D.E. Bro&n, and M.il Miller. i975: techniques, such as the use of NIRS to measure the proportions of Registration of alfalfa germplasm pools NC%3-1 and NC-83-2. Crop legume and grass roots, may facilitate more extensive sampling of Sci. 15:604-605. Minson, D.J., K.L. Butler, N. Grummltt, and D.P. hr. 1983. Bias when research plots in certain types of pasture and range management predicting crude protein, dry matter digestibility and voluntary intake of experiments. While the root mass still must be washed free of soil, tropical grasses by near-infrared reflectance. Anim. Feed Sci. Tech. dried, and ground prior to NIRS determination, hand separation 9~221-237. of the species in the sample would not be necessary once approp- Norris, K.H., R.F. hrnes, J.E. Moore, and J.S. Shenk. 1976. Predicting riate calibration equations have been developed. forage quality by infrared reflectance spectroscopy. J. Anim. Sci. Our procedures and results were analogous to those of Coleman 43:889-897. et al. (1985) and Petersen et al. (1987) who used NIRS to measure Petersen, J.C., F.E. Bwton, II, W.R. Windkam,8nd C.S. Hovehad. 1987. the species composition of forage mixtures. The calibration data Botanical composition definition of tall fescue-white clover mixtures by near infrared reflectance spectroscopy. Crop Sci. 27:1077-1080. using root tissues were subject to the same composition, machine, Rumba@, M.D., md B.M. Peodery. 1986. Interspecific relations and the sampling, and methodology errors that affected forage tissues. breeding of pasture plants for semiarid regions. p. 285-290. In: T.A. Similarly r2 values in these previous studies were slightly higher Williams and G.S. Wratt (cd.). Plant Breeding Symp., DSIR, 1986. when estimating the proportion of alfalfa tissue than when estimat- Agron. Sot. New Zealand Spe-c. Pub. 5. Christchu&h,New Zealand. ing the proportion of grass tissue in mixed species samples. Sbenk. J.E.. M.O. Waterlmua and M.R. Hoover. 1979. Analvsis of for- Our study involved only samples that contained no more than 2 ages’by i&red reflectance. j. Dairy Sci. 62807-812. * species whereas the mixtures studied by Coleman et al. (1985) Ward, R.G., G.S. Smith, J.D. Ware, N.S. Urqulurt, 8nd J.S. Shenk. contained as many as 4 species. Seeded pastures often contain only 1982. Estimates of intake and quality of grazed range forage. by near 2 species but the NIRS technique would have application to a infrared reflectance spectroscopy. J. Anim. Sci. w399-402.

JOURNAL OF RANGE MANAGEMENT 41(g), November 1988 Germination of green and gray rubber rabbitbrush and their establishment on coal mined land

J.T. ROM0 AND L.E. FDDLEMAN

The objective8 of tbia study were to: (1) determine tbe effecta of dominated by bluebunch wheatgrass (Agropyron spicatum (Pursh) temperature and water stress on germination, and; (2) evaluate Scrib. and Smith) and plains muhly (Muhlenbergia cuspid&u effects of seeding date on emergence and auvival of green and gray (Torr.) Rydb.). rubbermbbitbNeb(Clvysollromnc nauwanu(PaUas)Britt. sulmp. The basic requirements for germination and establishment must graveole~ (Nutt.) Piper. md Chrysothanmus nauseosla (P8UM) be elucidated for green and gray rubber rabbitbrush before they Britt. subep. nuuscosus (Nutt.) Piper.). Seeds of botb sbruba were can be used to their fullest potential in reclaiming drastically incubated at 10,20, and 300 C in a gradient of osmotic potentials disturbed lands. Objectives of this study were to: (1) determine the ranging from 0.0 to -1.5 MPa. Seedings were also made in tht field effects of temperature and water stress on germination of their on reeding dates over a period of 3 years. Total germination and seeds, and; (2) evaluate emergence and survival of these 2 subspe- germination rate declined aa temperaturen end osmotic potential8 cies of rubber rabbitbrush as influenced by timing of field plantings decreased; they wtre bigbest for both sbruk at 20 and 300 C and on reshaped coal mined land. lowest at 10” C. Under field conditions seedling populationa wert limited by low emergence and survival relative to viable seed Materials and Methods / planted. Emergence and survival of aeedlinga were highest in an Germination Experiments yttn that exceptionally wet year, declining in 8ubeequtnt were Seeds (achenes) of green and gray rubber rabbitbrush were drier. Emtrgenct rnnged from 0 to 6.9% and 0 to 7.1% and surviv81 harvested at the same site in October, 1979 and 1980, from several of emerged seedling8ranged fkom 6.6 to 5% and 0 to 60% for green naturally established plants on unleveled spoilbanks at the West- and gray rubber rabbitbrush, respectivtly. Survival of green em Energy Coal Mine near Colstrip, Montana. Seeds were taken rubber rabbitbrusb was bigbest from mid-spring plantings, but no to the laboratory, dried at room temperature, cleaned as described distinctively favortblt seeding date was found for gray rubber by Eddleman (1977), and stored in paper envelopes in darkness in nbbitbrusb. Results suggest that seeds of these shrubs should be the laboratory at approximately 20” C. A seedblower was used to wbtn planted prior to or during periods seedbed temperature8 are remove empty seeds and provide uniform seed weights between in tbe 20 to 39 C range and soil moisture is txpttttd near its years of collection. seasonal bigb. Germination tests were conducted 14 and 4 months after collect- Key Words: Chrysothamnus nauseosus, reclamation, wed ecol- ing the seeds in 1979 and 1980, respectively. Solutions of polyethy- ogy, w8ttr stress, nativt species, shrubs lene glycol6,OOO (PEG) were prepared to depress osmotic potentials to -0.3, -0.6, -0.9, -1.2, and -1.5 MPa by adding PEG to Green rubber rabbitbrush (Chrysorhamnus nuuseosus (PaUas) distilled water (Michel and Kaufmann 1973). Distilled water was Britt. subsp. gruveoiens (Nutt.) Piper.) and gray rubber rabbit- used as the control (0.0 MPa). Osmotic potentials of these solu- brush (Chrysothamnus IuIuseosI(s (Pallas) Britt. subsp. nol(seosI(s tions, determined by psychrometric procedures, were within 0.05 (Nutt.) Piper.) are morphologically and ecologically distinct, but to 0.18 MPa of those desired. Seeds were dusted with captan sympatric, subspecies that grow on the northern Great Plains. In (N_I(trichloromethyl)-thiol4cyclohexene-l,2dicarboximide) tocon- many areas rabbitbrush is classified as undesirable by range man- trol fungi. Twenty-five of these seeds were placed in each petri dish agers. However, green and gray rubber rabbitbrush are important on a S-cm square of germination paper and 30 ml of PEG solution components of rangeland, adding structural and functional diver- were added. Petri dishes were covered and placed on trays on a sity. They show potential for use in reclaiming disturbed land by water-saturated layer of germination paper, and the trays were providing cover, stabilizing soils, moderating temperature and enclosed in plastic bags to limit dessication. Seeds were incubated moisture regimes, providing forage for mule deer and pronghom, at constant temperatures of 10, 20, and 30° C without light. Ger- and adding to the aesthetics. Most rabbitbrushes grow rapidly, mination was recorded at 2day intervals through 12 days and at produce abundant seed, but are not aggressive competitors with 6day intervals thereafter through 30 days. Seeds were considered herbaceous plants (McArthur et al. 1979). germinated when cotyledons of seedlings were reflexed; seeds that Green rubber rabbitbrush, a large shrub growing to 2m tall, is failed to meet thii criterion were jud&d abnormal. common on south and southwest slopes of badland plant commun- Within subspecies, treatments were applied factorially to 4 repli- ities in southeastern Montana (Brown 1971). This shrub is an early cates arranged in a randomized complete-block design using year successional species on a wide variety of soil textures on severely of collection, temperature, and osmotic potentials as main treat- disturbed sites including borrow pits, road cuts, and spoil banks of ments. Average germination rates (90 day-‘) were calculated coal and bentonite mines. according to Maguire (1%2), and total germination was expressed A smaller shrub, gray rubber rabbitbrush grows to 0.3 to 0.8 m as a percent. Percentage values were transformed with arcsin (6) tall. It is limited in its colonizing potential on drastically disturbed and analyzed within subspecies using analysis of variance (Snede- sites. Gray rubber rabbitbrush grows mainly on tine textured soils car and Cochran 1980). Regression analysis was then used for of ridges and slopes in good condition grassland communities untransformed data to develop response curves for each tempera- Authors are assistant professor, Department of Crop Science and Plant Ecology, ture and year of collection to determine the best fit regression University of Saskatchewan, Saskatoon, Saskatchewan S7N OWO(Canada), and equation (Steel and Torrie 1980). Because of the high probability associate professor, Department of Rangeland Resources. Oregon State University, Corvallis 97331. At the time of this research the authors were research assistant and of great variation in germination potential of different seedlots of professor, University of Montana, Missoula, respectively. native plants, responses to water stress between years and tempera- Research was funded by the U.S. Department of Energy Grant EY-76-S46-2232. tures were compared. Confidence limits were computed for regres- Manuscript accepted 21 July 1988. sion coefficients, allowing both comparison of water stress effects

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 491 100 100 a) a) 1979 1980_ .-.-. jO”C---- 80 80 _- 2OOC-.-- h s Y .’ g 60

4 E t \ p 40 \ b c L \

20 20

0 0 40l--- 40 b) b)

\ -; 30

0” S 0, G \ = 20 c \ 0 3

2 E . t o 10

0 0 0 -0.4 -0.8 -1.2 -1.6 -0.4 -0.8 -1.2 6 Osmotic Potential (MPa) Osmotic Potential (MPa)

Fig. 1. Response curvesfor (a) totalgermination and (b)germination rate Fig. 2. Response curves for (a) totalgermination and (b)germination rate of green rubber rabbitbrush seea!scollected in 1979 and 1980 and incur- of gray rubber rabbitbrush seeds collected in 1979 and 1980 and incubated at 10.20, and 300 C in a gradienr of osmotic potentials (derived bated at 10,20, and 30” C in a gradient of osmotic potentials (&rived from regression analysis; see Table 1). from regression analysis; see Table 1).

492 JOURNAL OF RANGE MANAGEMENT 41(e), November 1gs8 on total germination and germination rate among temperatures RtdtS across years of collection. Germination Response to Temperature and Water Stress Emergence and Survival of Seedlings in the Field Total germination and germination rate of both shrubs were Emergence and survival were evaluated by sowing seeds of both highest at 20 and 30” C and lowest at lo” C, and usually declined as shrubs on coal mined land that had been reshaped and topsoiled at osmotic potential decreased (Figs. 1 and 2). Germination of the the Western Energy Mine near Colstrip, Montana. Seeds were 1979 collections was higher at 10’ C than the 1980 collections, but sown in early-spring (3 May 1979, 15 April 1980), mid-spring (10 at 20 and 30” C total germination of seeds from 1979 collections May 1978 and 30 May 1979), and fall (25 October 1978 and 12 was lower than that for seeds collected in 1980. October 1979). Seeds collected in 1977,1978, and 1979 were hand Total germination and germination rate for green rubber rab- sown in 2-m rows at a depth of approximately 1 cm the year bitbrush were related to the interaction of year of collection, following collection. In 1978,40 and 48 pure live seed (PLS) were temperature, and osmotic potential. The coefftcients of regression sown per m for gray and green rubber rabbitbrush, respectively. equations for total germination were similar and germination Seeding rates were increased to 200 PLS per m for both subspecies declined in similar fashion within 10 and 30° C across years of in 1979 and 1980 in light of the low emergence recorded for the collection (Table 1 and Fig. 1). Decline in total germination as 1978 seedlings. Seedings were evaluated during their fast growing osmotic potential decreased at 20’ C was significantly different season to determine seedling emergence; all plots were reevaluated (KO,OS) between years. Regression coefftcients for the total in mid-summer 1981 to determine survival. Because seeding rates germination-osmotic potential relationships were significantly dif- were not uniform among years, emergence was converted to a ferent (KO.05) between temperatures for the 1979 collection, but percentage of PLS and survival was expressed as a percent of were similar for seeds collected in 1980 and incubated at 10 and 20° seedlings that emerged. Data were transformed with arcsin (fi) C. Within temperatures, shapes of the germination rate-osmotic and analyzed within subspecies using analysis of variance with 5 potential relationships were similar across years (Table 1 and Fig. replications in a randomized complete block design (Snedecor and 1). When all temperatures were compared within years, slopes of Cochran 1980). Differences in significantly differing means for germination rate versus osmotic potential were similar at 10 and seedling emergence and survival were separated using Tukey’s 20“ C, but 300 C was distinct from the lower temperatures. W-procedure (Steel and Torrie 1980). Total germination for gray rubber rabbitbrush was related to interaction of year, temperature, and osmotic potential; germina-

Table 1. Rcgreedon equation@ and coeffkknb of determination lor totd genuination and germination rate for green and gray rubber rabbitbrush cdkcted 2 yean aad incubated at 10.20, and Jo” C in a gradknt of amotk potentkk.

YC%Ib Temperature (OC) Regression equation R* Green rubber rabbitbrush Total germination (%) 1979 10 Y = 40.1 + 54.3X + 17.5X2 a’ 0.77

1979 20 Y q 45.5 + 32.5X 0.85 1979 30 Y=59.6- 109.4X-264.7X*- 111.1X’ b 0.84

1980 Y = 29.6 + 70.8X + 41.7X* a Ad 0.91 1980 iti Y = 73.8 + 77.8X + 18.7X* A 0.92

1980 30 Y q 74.5 - ii 1.5X - 363.5X* - 182.1X’ b 0.97 Germination rate (A day-‘) 1979 Y = il.3 +9.4x c B 0.76 1979 Y = 18.4 + 12.4X d B 0.91 1979 Y = 28.5 + 43.0X + 30.8X* + 9.9X’ e 0.97

1980 Y q 11.7+ 12.9X c c 0.88 1980 Y = 20.4 + 13.2X d C 0.88

1980 Y q 33.7 + 75.9X + 87.6X* + 39.7X’ e 0.95

Gray rubber rabbitbrush Total germination (%) 1979 10 Y = 23.7 + 26.0X 0.78 1979 20 Y = 63.2 + 44.5X 0.88 1979 30 Y = 23.5 - 205.7X - 432.5X* -203.7X’ 0.75 1980 10 Y = 16.4 + 17.0X 0.52 1980 Y = 70.9 +48.1x 0.83 1980 E Y = 75.1 - 60.7X - 287.3X* - 154.3X’ 0.92 Germination rate ($ day-‘) 1979 10 Y = 7.5 + 7.1x a A 0.76 1979 20 Y = 15.1 I +9.6X b A 0.92 1979 30 Y = 27.9 + 30.7X + 6.6X* C 0.90 1980 10 Y = 7.9 + 8.6X a B 0.68 1980 20 Y = 16.4 + 11.2X b B 0.92 1980 30 Y = 32.0 + 43.8X + 14.8X* C 0.95 ‘Y is total prmiaation or germination rate; X is osmotic potential. bindicates P year of seed colkction. ‘a similar lower case letter within subspeck indicates no signiticamtdiffcrcnce (EO.05) in rqression cocCieats among years within tempemtura and parameters. % similar upper case letter within subspecies indiites no signifwant diicrencc (PSO.05) in regression cocffkients between temperatures within yeam and parametera.

JOURNAL OF RANGE MANAGEMENT 41(e), November 1968 493 tion rate was affected by the interactions of temperature and stress on germination have also been shown for other shrubs with osmotic potential and year and osmotic potential (Table 1). Regres- germination under water stress highest at optimal temperatures sion coefficients for relationships between total germination and (Weldon et al. 1959, Springfield 1968, Romo and Eddleman 1985). germination rate and osmotic potential were similar within Eddleman (1977) concluded that temperatures higher than 20° C temperatures and across years (Table 1 and Fig. 2). However, appeared optimal for germination of these shrubs; germination reduction in total germination as osmotic potential decreased was was reduced and slower at lower temperatures. Similarly, Sabo et significantly different (KO.05) between temperatures. Reduction al. (1979) reported germination was best at warm temperatures for in rate of germination as osmotic potentials declined was similar Chrysothamnusnauseosus (Pallas) B&t. subsp. consimilis (Greene) within temperatures across years; within years the response at 30 H.M. Hall and Chrysothamnus nauseosus (Pallas) Britt. subsp. C was distinct from 10 and 200 C. bigelovii (Gray) H.M. Hall from New Mexico. Germination of C. nauseosus subsp. consimilis was highest at alternating tempera- Seedling Emergence and Establishment in the Field tures between 13 and 27.5O C, and C. nauseosus subsp. bigelovii Precipitation received in 1978, 1979, and 1980 was reflected in germinated best at alternating temperatures between 20.5 and 30° emergence and survival of both shrubs (Table 2). Growing season C. Results of the germination experiments and field plantings Tebk 2. Mean percent emergence in the year of seeding green and gray rubber rebbltbrusb in 1978,1979, and 1980 and rurvivel of the seed- emphasize the importance of maintaining high levels of soil mois- linga at Coktdp, Montana. ture for maximizing germination and emergence of green and gray rubber rabbitbrush. Their emergence and establishment were severely limited in dry years. A sharp contrast in establishment of Emergence in Survival in several shrubs, grasses, and forbs in wet and dry years was also year of seeding 1981 reported by Eddleman (1980). In the present study seedling popu- Year Seeding date’ (%) (%) lations were limited in the field by low emergence and survival a Green rubber rabbitbrush characteristic also common to Chrysothamnus nauseosus (Pallas) 1978 MS2 2.8 br 55.0 a Britt. subsp. albicaulis (Nutt.) Rydb. (Stevens et al. 1986). 1978 F’ 6.9 a 11.5 b The present tests indicate that germination, emergence, and 1979 ES’ 0.9 bc 6.6 b survival can be expected from any season of planting. However, to 1979 MS 0.9 bc 55.0 a maximize germination of green and gray rubber rabbitbrush, seeds F 1979 0.0 c 0.0 b should probably be sown prior to times when seedbed tempera- 1980 ES 0.0 c 0.0 b tures are in the 20 to 30’ C range when soil moisture is expected Gray rubber rabbitbrush near field capacity. On the northern Great Plains, April through 1978 MS 2.0 b 22.6 b June is the period with highest amount and the greatest probability 1978 F 2.3 b 60.0 a of precipitation being received (Caprio et al. 1980). Generally June 1979 ES 7.1 a 22.0 b receives the most precipitation and July is the hottest month. 1979 MS 0.9 b 5.0 b 1979 F 0.2 b 0.0 b Seedlings of rubber rabbitbrush are sensitive to frost and drought 1980 ES 0.0 b 0.0 b (Plummer et al. 1968) thus, the window for germination, emergence, and establishment may be limited by low spring tempera- 50~ text for seeding dates. *MS indicates mid-spring seeding. tures and by summer drought. These relationships therefore sug- JF indic$cs fall seed@. gest that seeds should be sown in May through early June to take *ES indcatcs early sprmg seeding. advantage of favorable temperatures and soil moisture to maxim- %Isimilar letter wthm a subspecies and parameter indicates no significant difference (pIo.05) between seeding dates using Tukey’s W-procedure. ize germination and the period of growth for seedlings. Mackey and Depuit (1985) suggested that rabbitbrush has a precipitation received in 1978 and 1979 was among the highest and broad ecological tolerance and should be considered for use in lowest, respectively, on record for Colstrip, Montana. From April reclamation. Reclamation specialists have ready access to plant through June 1978, 287 mm of precipitation were received, 59% materials in green and gray rubber rabbitbrush since seeds of higher than the longterm mean. Conversely, during the same locally adapted ecotypes are available near most mines on the period in 1979 and 1980, precipitation totaled 80 mm and 109 mm, northern Great Plains and seeds can be easily collected. Rabbit- 55% and 39% lower than the longterm mean, respectively. Carry- brush seeds also germinate well (Plummer et al. 1968, Eddleman over effects of the above-average precipitation in 1978 and low 1977, Sabo et al. 1979), and those collected in the fall can be sown precipitation in 1979 and 1980 were presumably mirrored in declin- the following year (Eddleman 1977). These shrubs should be con- ing emergence and survival in 1979 and 1980 as compared to 1978. sidered for use in reclaiming drastically disturbed land because of Emergence of green rubber rabbitbrush ranged from 0 to 6.9% their rapid growth and reseeding characteristics, and because they and survival of these seedlings from 6.6 to 55% (Table 2). Although may ameliorate moisture and temperature regimes and provide emergence was inconsistent among seeding dates, survival was late season forage for wildlife. They are native species that can add significantly (m.05) higher from mid-spring seedings than from to the structural and functional diversity on reclaimed land that is sowings in early spring and fall. There was no clear trend estab- required by law. lished in the most favorable time for emergence and survival of gray rubber rabbitbrush (Table 2). Emergence ranged from 0 to Literature Cited 7.1% and survival of these seedlings ranged from 0 to 60%. Brown, R.W. 1971. Distribution of plant communities in southeastern Discussion Montana badlands. Amer. Midl. N&r. 85:458477. Green and gray rubber rabbitbrush germinated over a broad Ca~rlo. J.M.. Snyder. R.D.. end C.K. Grunrald. 1980. Precinitation probabilities in Montana. Montana Agr. Exp. Sta. Bull. 712. _ range of temperatures but a narrow range of water stress. When Eddkmen, L.E. 1977. Indigenous plants of southeastern Montana. I. Via- germination under water stress is considered in the context of total bility and suitability for reclamation in the Fort Union Basin. Montana and rate, results of this study indicate that temperatures between 20 Forest and Conserv. Exp. Sta. Spec. Pub. 4. and 30” C were near optimal, At 10 and 20’ C total germination Eddkmen, L.E. 1980. Coal mine reclamation with native plants. Proc. of and germination rate declined as water stress increased. Although Symp. on Watershed Management, ASCE, p. 80-90, Boise. total germination at 300 C was higher at low levels of water stress M8ckey, C.V., 8nd EJ. Depuit. 1985. Natural tevegetation of surface- relative to the control, germination rate subsequently declined as deposited spent oil shale in Colorado. Reclam. Reveg. Res. 4: 1-16. water stress increased. Interacting effects of temperature and water

494 JOURNAL OF RANGE MANAGEMENT 41(e), November 1999 Maguire, J.D. 1962. Speed of germination-aid in selection and evaluation Sntdttor, C.W., tad W.C. Cotluu~. 1980. Statistical methods. The Iowa of seedling emergence and vigor. Crop Sci. 23176177. State University Press, Ames. McArtbur, E.D., A.C. Bltutr, A.P. Plummtr, tnd R. Sttvtnr. 1979. Char- SpringfIeld, H.W. l%g. Germination of winterfat seeds under diiercnt acteristics and hybridization of important Intermountain shrubs. III. levclsofmoisturcstrcssandtempcraturcs. J. RangcManagc.21:314-316. Sunflower family. USDA Intmtn. Forest and Range Exp. Sta. Res. Steel, R.G.D., and J.H. Torrle. l%O. Principles and procedures of statis- Paper INT-220. tics. McGraw-Hill Book Co., New York. _ - Ml&l, B.E., and MR. Kaufman. 1973. The osmotic potential of polycthy- Sttvtnt. R.. Jorcensen. K.R.. Davis. J.N.. and S.B. Monm. 19%. Seed lcne glycol6,ooO. Plant Physiol. 51:914-916. papp& aid pl&mc~t infl&nccs on whiic rubber rabbitbrush establish- Plummtr, A-P., Chritttmtn, D.R., and S.B. Mom. 19&B.Restoring big ment. p. 353-357. In: Proc. Symp. on the Biology of Artemisia and game range in Utah. Utah Division of Fish and Game. Pub. 68-3. Chrysothamnus. USDA Intermountain Forest and Range Exp. Sta. Romo, J.T., nnd L.E. Eddkmtn. 1985. Germination response of grcase- Gen. Tech. Rep. INT-200. wood (surcobarus vermiculatus) to temperature, water potential, and W&ion, LW., Bobmont, D.W., and H.P. Alley. 1959. The interrelation of specific ions. J. Range Manage. 38: 117-120. three environmental factors affecting germination of sagebrush seed. J. Sabo, D.G., Johnton, G.V., Mtrtin, W.C., tnd E.F. Aldon. 1979. Gcrmi- Range Manage. 12236-238. nation requirements of 19 species of arid land plants. USDA Rocky Mtn. Forest and Range Exp. Sta. Res. Pap. RM-210. Floristic changes induced by flooding on grazed and ungrazed lowland grasslands in Argentina

ENRIQUE J. CHANETON, JOSE M. FACELLI, AND ROLAND0 J.C. LEON

Abetnet Ch8nges in community composition of 2 gmssl8nd sites exposed vegetational heterogeneity (Leon et al. 1984). Stands with different to a flood of unusual intensity and duntion were investig8ted in grazing history can be viewed as the units of a landscape mosaic the Flooding Puap8. These gr8sslurds ue subject almost 8nnurlly widely affected by the pattern of land use (Forman and Godron to floodings of lesser m8gnitudc. The study site8 were ldj8cent to 1981, Leon et al. 1984, Urban et al. 1987). The manifold effects of e8ch other, urd differed in veget8tion structure urd composition. livestock grazing involve changes in grassland floristic composi- One hid been grazed continuously by cattle 8nd ~8s showing signs tion leading to displacement of native grasses by exotic species and of intense deteriontion. The other h8d rem8ined ungr8xed during native spreading dicots (Sala et al. 1986, Facelli et al. 1988) of low 15 years. Brul cover by species WM meuured in summer, before forage quality. 8nd 8fter the flooding event. Compositionrrl difference between These grasslands are subject periodically to flooding events of sites decreased with flooding from 68.9 to 39.1%. In the grued site diverse magnitude (i.e., intensity and duration). Flood regime, the cover of alien forbs ~88 reduced by 4g%. After the flooding which is controlled by topographical features, may interact in native gruninoids represented 99.7 8nd 86.7% of the cover, inside complex ways with cattle grazing in determining the spatial pattern 8nd outside the exclosure respectively. Tot81 b8s8l cover ~8s not of vegetation throughout the landscapes of the region (Leon 1975, 8ffected but ~8s redistributed 8mong species 8lre8dy present Sala et al. 1986). A better understanding of the ecological relation- before the flood. Floristic changes would h8ve led to 8n improve- ships between these environmental agents, and their influence on ment of the forage source. We conclude that plant community community structure, is needed in order to improve the manage- response to the event ~8s influenced by the previous gruing his- ment of the grassland resource. tory of the site. The large flood 8cted n 8n overriding environmen- Within a broad context, floodings have been considered as tal frctor which p8rti8lly reverted the effects of gr8xing upon physical “disturbances” which disrupt the structure of plant com- gr8ssl8nd composition. munities (White 1979). Since flood tolerance varies widely among species, a substantial impact on community composition might be Key Words: c8ttle gmxing, community dynunics, disturbance, expected (Kozlowski 1984). There is some evidence suggesting the Flooding Pamp8, for8ge qu8lity, stress organization of flood-prone grasslands in the Flooding Pampa can Grazing by domestic ungulates has well-documented effects be affected by large, less frequent events (Insausti and Soriano upon structure and composition of plant communities (Risser et al. 1988). 1981, MackandThompson 1982, Rakker 1985, Salaet al. 1986). In Here, we report the changes in grassland composition observed the grasslands of the Flooding Pampa, Argentina, the introduction after the occurrence of an unusually large flood on 2 different sites, of cattle by European settlers added a major causal factor for one subject to continuous grazing, and the other protected by fencing during the last IS years, Specifically, we addressed the Authors are graduate student, research assistant, and professor, Dcpartamento de Ecolonia. Fact&ad de Anronomla. Univcrsidad de Buenos Airea. Av. San MartIn following questions: (1) Does a flood of unusual magnitude alter 4453.11417) Buenos A&s, Argeniina. Facelli’s current address is Department of theflo&iccompositionanddo minancerelationsoftheplantcommuni~ Biological Sciences, Rutgers Unwxsity, P.O. Box 1059, Piscataway, NJ 08854, USA. (2) How does previous grazing history influence community The authors wish to thank V.A. Dcngibus, M. Ocsterheld, W.B. Batista, and A. Soriano for helpful comments. The people from Estancia “Las Chilcas” gave v$uablc response to the flooding event? (3) What is the consequence for the assistance. Financial support was prowded by the National Council of Sclentdlc and grassland as a forage resource? Technical Research from Argentina (CONICET). Manuscript accepted 21 July 1988.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 495 Study Areaand Methodc tion lines, 5 m long each (Brown 1954). Previous studies have shown the structural divergence that arose among sites since the The study was carried out in a floodplain rangeland located at establishment of the fence (Sala et al. 1986, Facclli et al. 1988). the center of the Salado River Basin (36O 30’S, 58” 3O’W),the Fin Inside the exclosure native tall grasses dominated the community portion of the Flooding Pampa (58,000 km2). Topographical flat- while the grazed site was dominated by forbs, mainly exotics. ness and absence of integrated drainage systems are the most Between September and late November 1985 (spring) the study conspicuous features of the landscape. Soils, predominantly area remained waterlogged with - 10 to 40 cm of standing water. Natraquolls, are poorly drained because of low hydraulic conduc- This flood had unusual intensity and duration and was the result of tivity and lack of slope. The weather is mild with mean annual intense rainfalls (595 mm) occurring over a 3-month period. Only rainfall around 900 mm, evenly distributed through the year. The short-term (l-3 wk) and shallow (

Table 1. Cbenges ln percent cover (E f SE, n = 4) between Feb. 1985 end JUI. Mti (befom sod after tbe flood respectively) for some of the spedee recorded lo tbe gresed end ungrued &es. Species included nude up about 95% of the tote1 cova. For the”otber spedes”sectlon, mean covaend speclee number (in brackets) ere shown.

Grazed Ungrazcd

Species PdlOOd Post-flood PrC-flood Post-flood Dicots Warm-season Spiionthes stolonifero 0.18 f 0.06 0.11 f 0.02 Aster squamatus 0.05 f 0.02 0.21 f 0.07 Cool-season Mentha pulegiwn’ 11.14 f 2.90 0.68 f 0.26 Leontodon toraxocoidest 2.14 f 0.30 0.22 f 0.09 Plontago lonceolatal 1.61 f 0.74 0.13 f 0.04 Phyla canescens 0.31 f 0.12 0.46 f 0.08 0.15 f 0.15 0.02 f 0.02 Erynqiwn ebrocteatum 0.16 f 0.04 0.44 f 0.12 0.02 f 0.02 0.03 f 0.02 Grasses Warn-season Ponicum gouinii 1.46 f 0.08 1.04 f 0.66 0.01 f 0.02 Setorio geniculata 0.42 f 0.13 0.03 f 0.02 0.02 f 0.04 0.01 f 0.02 Ponicum milioides 4.36 f 2.06 2.88 f 0.23 3.27 f 1.13 2.92 f 0.48 Paspolidium poludivagum 0.61 f 0.10 0.78 f 0.19 1.49 f 0.62 3.01 f 0.66 Leersio hexondro 0.19 f 0.07 2.56 f 0.40 0.87 f 0.78 2.70 f 1.58 Paspalum diiatatum 0.05 f 0.10 4.86 f 2.44 2.99 f 1.20 Stenotaphrum secundotum 1.28 f 0.74 0.33 f 0.29 Deyeuxia viridtjlovescens 0.59 f 0.55 0.25 f 0.19 Ponicum bergii 0.04 f 0.08 0.45 f 0.29 0.07 f 0.06 Cool-season Donthonia montevidensis 2.71 f 0.84 4.64 f 1.12 1.64 f 0.82 1.82 f 1.15 St@ philippii 5.38 f 4.62 0.60 f 0.60 Sedges Warm-season Juncus microcepholus 0.05 f 0.05 0.51 f 0.16 Cool-season Eleocharis viridans 1.50 f 0.75 5.88 f 0.66 2.00 f 0.09 4.28 f 0.39 Other species Dicots 0.33 (11) 0.21 (4) 0.08 (4) Grasses 0.59 (4) 0.31 (6) 0.75 (5) 0.20 (3) Sedges 0.48 (2) 0.05 (2) 0.43 (3) 0.02 (1) Total hasal cover 28.31 f 1.64 21.60 f 1.02 23.28 f 1.74 19.26 f 2.58 ‘exotic species

496 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1966 pendent. Instead, we based the analysis on the confidence intervals of the means (a = 0.05). Floristic differences among sites for POST-FLOOD post-flood observations were examined with the Mann-Whitney U PRE- FLOOD test (two-tailed). Contribution to total cover of native and exotic species, and of cool- and warm-season species were also considered. 1.0 I a The Percentage Similarity index (Z’S, Gauch 1982) was employed I for comparing the overall species composition among sites and between pre- and post-flood conditions. Because our aim was to 0.8 I emphasize the degree of change at 2 different sites, results were n expressed as Percent Dissimilarities (PD = 100 - RS). The average values of basal cover per species standardized by sample total cover 0.6 were used in the computations. The consequences of floristic changes for grassland forage quality were evaluated by the Grazing Value Index (G VZJproposed by m 0.4 Daget and Poissonet (Poissonet et al. 1981): W 2 GYZ= 0.2 *F, (CSi l $1,) 0 0.2 where n is the species number GSi is the relative contribution of species i to the total cover for a given situation, and SZi is an index of the forage value of the species i, ranking from 0 to 5. The SZtakes 3 O OM csws II M CSWS into account nutritional value, cattle preferences, and some plant 2 1.0 traits like growing form, height, and biomass concentration, which b contribute to determine the actual forage offered by a species. Tabulations from Cauhepe et al. (1984) were used to assign SZ g 0.8 values. This index proved to be suitable in previous works in - grasslands and pastures of the region (Cauhepe et al. 1984, Gester- 2 held and Leon 1987). “w 0.6 c Results 0.4 Total basal cover was similar in both sites. Differences (Table 1) were not significant, both before (U = 2, p>o. 1) and after the flood (U = 8,ZQO.5). Total cover does not appear to be shifted in either of the study sites (c.i. broadly overlapped between sampling dates, Table l), although the ranges for the values of the grazed area (1985 = 25.2 - 3 1.1,1986 = 20.2 - 24.6) would suggest a slight reduction of 0 - its plant cover. Floristic changes were primarily related to changes DM cs ws 0 M CS W: in the cover of species present before the flood. grasses ; exotics The floristic composition of the grazed site was substantially IEI modified (Fig. la, 2). The basal cover of alien dicots diminished from 14.94 (f3.82) to 1.07 (f0.25). Before flooding, Mertthapule- Fig. 1. proportional basal cover for the grased (a) and tatgrazed (b) areas gium Leontodon taraxacoides, and Plantago lanceolata, consti- of dicots (D)and monocots (M), and cool-season (CS)and warm-season tuted about 520/, of the total cover. Their contribution declined to (WS) species. 5% in 1986 (Table 1). Conversely, the importance of monocots was enhanced by the increased cover of some grasses and sedges (Fig. la, Table 1). After the flood, native species became the main UNGRAZED component of the grazed site and continued dominating the grass- 66.9 land inside the exclosure. The cover of exotics and forbs remained 1965 higher in the grazed area than in the ungrazed one (U-test, KO.05) (Fig. 1). The cover of cool-season species decreased by 14% in the grazed area (Fig. la), although the aforementioned reduction of cool-season forbs was in part compensated by the increase in the -_-37,6-e ------FOG-----_------55.6--- cover of Danthonia montevidensis and Eleocharis viridans, 2 native monocots (Table 1). Community composition of the ungrazed area varied slightly. + + The cover of sedges increased (Fig. lb), mainly due to the incre- ment in the basal cover of Eleocharis sp. Some grasses also had 1966 39.1 higher cover values after flooding (Table l), but the whole of the grass compartment showed a minor reduction from 20.61 (f2.01) to 14.91 (f2.47), i.e., 11% over the total cover (Fig. lb). Grass Fig. 2. Comparisons between sites anddates of the overallfloristic compo- cover did not differ significantly among sites after the flood (U ~4, sition. Values represent Percentage Dissimilarities (PD = 100 - PS) p>o.3), although its proportion was larger inside the exclosure (Gauch 1982)for each case. (Fig. 1). The relationship between the cover of cool-season and graminoids) did not show values over 0.06%. warm-season species remained almost unchanged in the ungrazed Different species responded positively (increasing cover) to the site (Fig. lb). flooding event. Some species augmented only in the grazed site Several rare species recorded in 1985 were not found in 1986. (e.g., Danthonia sp.), some only inside the enclosed area (e.g., Nevertheless, these species (mostly dicots and annuals) represented Paspolidium paludivogum), and others like Eleocharis sp. and a small fraction of the total cover (<1.590) in both sites. Those Leersia hexandro increased in both sites (Table 1). The latter species that appeared in the samples only after the flood (all response together with the diminution in the cover of dicots for the

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 497 cattle (SZ= 0). The aerial and belowground portions of that species GRAZED UNGRAZED were observed to be strongly affected by a prolonged flood (Insausti and Soriano 1988). Two native cool-season graminoids highly preferred by cattle, Danthonia sp. and Eleocharis sp. (Cauhb* et al. 1984), were promoted. Regardless that flooding occurred during their period of maximum annual productivity (Sala et al. 1981), these species cl1 resisted the stress and dominated the grazed community after the flood (Table 1). The results suggest that the effect of a large flood EC could not be related to the phenological behavior of the species involved. The slight changes of dominance recorded within the exclosure were caused by variations in the cover of individual species, not in any particular floristic group. They had the same direction to those observed in the grazed site, indicating that the changes of species abundances in both sites mainly reflected unequal abilities to cope with waterlogging stress. In the ungrazed area, competition and tolerance due to the small-scale light shortages would play an important role in structuring the community (Tilman 1982, Facelli et al. 1988). The inundation caused the depletion of the available oxygen in the soil (Ponnamperuma 1984), temporarily adding a limiting resource to the plant environment. Community composition is thought to reflect alterations in the resource availability (Tilman 1982). Thus, species which endured the new constraint 0 such as Eleocharis sp. and non-tussocking grasses like Paspalidium 1985 1986 1985 1986 sp. and Leersiu sp., shared the dominance with tall grasses (e.g. Paspalum dilatatum, Ponicum milioides) which dominated before Fig. 3. Percentage contribution lo rhe rota1 cover of the grazed and flooding (Table 1). ungrazed sites in each sampling akre (before and after rheflood) of all The community of the ungrazed site was more resistant to com- recorded species, grouped accordingly to its forage values (SI).27~ positional modifications. Community response to prolonged water- following are theforage valuesfor some representative speciesfrom rhe logging seemed to be largely influenced by the identity of the study sires in the Flooding Pampa: 4 q Paspalum dilatatum, Paspalidium dominant species as determined by the preceding management paludivagum; 3 = Leersia hexandra. Danthonia montevidcnsis; 2 = Stipa history of the site. The response of a community to external per- philippi, Panicum milioides, Panicum gouinii. Stenotaphrum sccunda- turn, Elcocharis viridans, Plantago lanceolata, Leontodon taraxacoides; turbations derives in part from the adaptive properties of its component species resulting from the evolutionary history of the eco- 1 q Juncus microcephalus, Eryngium ebracteatum; 0 = Mentha pulegium, Phyla canesccns. system (Rapport et al. 1985). Exotic eurasian weeds are more successful on sites disturbed by cattle (Mack and Thompson 1982) grazed area contributed to reduce previous site differences (Fig. 2). probably due to a long history with domestic grazing (Knapp The comparison of the overall community composition indicates 1979), but they have not developed resistance to floods of large that both sites converged floristically, i.e., percentage dissimilarity magnitude. On the other hand, native graminoids dominant inside among sites decreased towards 1986 (Fig. 2). the exclosure presumably have evolved under this constraint but Alterations in grassland floristic composition shifted the G VZof not subjected to grazing by large ungulates (Burkart 1975). the grazed area from 26.2 in 1985 to 44.6 in 1986. The contribution Because of flooding, grazing was suspended for 3 months. How- of non-forage species (species with SZ = 0) decreased markedly as ever, the changes observed in the grazed area resembled those they were replaced by native graminoids of more forage value (Fig. which usually occur following 4 to 5 years of cattle exclusion (Sala 3). In the ungrazed area, some native grasses having a S122 et al. 1986, Rusch and L&n unpubl.). Floristic changes induced by reduced their cover (Table 1). The slight variation of GVZ in this the flood were the opposite of those shaped by livestock grazing site (53.9 to 57.3) was accounted for by the increase of species with (Le6n et al. 1984). Flooding promoted the replacement of exotics SZ= 3 or 4 (Fig. 3). Species with very low forage value (0 or 1) were of low forage value by more valuable, native species (Fig. 3). After nearly absent inside the exclosure (Fig. 3) thus the ungrazed com- the flood, a large proportion of the available forage was made up of munity always maintained a higher forage quality. preferred graminoids, a factor which may influence herbage intake of grazers (Allison 1985, see also Senft et al. 1985). McNaughton Discussion and Conclusions (1983) pointed out that many environmental factors may override The flood altered the floristic structure of the grassland, but the the effects of herbivores on vegetation. In our case, the unusually extent of the impact was different in the 2 studied sites. The grazed large flood functioned as the overwhelming agent (coarse-grain site was very susceptible to the infrequent magnitude of the flood signal sensu Allen and Starr 1982, p. 135), masking the ecological (Fig. 2). differences among sites and partially reversing grazing effects on Cattle grazing deeply affects the structure of the native grass- grassland composition. A group of indigenous species responded land, disrupting the canopy and promoting colonization by cool- similarly to this factor in both sites leading to floristic convergence season weedy forbs (Sala et al. 1986, Facelli et al. 1988). These (Fig. 2). species are able to tolerate light seasonal floodings, yet the large In these grasslands, flood&s are recurrent events of varying flood of 1985 drastically reduced their cover. The success of exotic magnitude. Hence it would be appropriate to consider them as a species likely depends on the duration and depth of floodings. consistent driving force of the system (White 1979, Vogl 1980, Recurrent, light floods interacting with continuous grazing might Rapport et al. 1985), which constitutes an essential part of the create advantageous conditions for invaders like Mentha sp. and environmental setting as is fire in other grassland ecosystems (e.g., Leontodon sp., whereas less frequent, more intense events would Risser et al. 1981, McNaughton 1983). Moreover, periodic floods impair their performance (Panetta 1985). Insausti and Soriano may contribute to mitigate the action of continuous domestic (1988) found a similar response in Ambrosia tenuifolia, a common grazing upon range condition (Taboada and Lavado 1988). Never- forb of the “upland” communities which is usually avoided by theless, our results suggest that the outcome of infrequent large

498 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 phenomena for a given community would depend on the initial Mack, R.N., and J.N. Thornpeon. 191)2.Evolution in steppe with few large, conditions, in terms of the abundance of the species present. hoovcd mammals. Amer. Nat. 119757-773.._. McNaugbton, S.J. 191)3.Sercngeti grassland ecology: the role of composite Liter8ture cited environmental factors and contingency in community organization. Ecol. Monog. 53291-320. Allen, T.F.H., and T.B. Starr. 19(12.Hierarchy. Perspectivesfor ecological Oestarbald, M., and R.J.C. Le6n. 1987. El cnvejccimiento de las pasturas complexity. University of Chicago Press, Chicago. implantadasz su cfecto sobre la productividad ~rimaria. Turrialba AlHum, C.D. 1985. Factors affecting forage intake by range ruminants: a 37-29-35. review. J. Range Manage. 38305-311. Panstta, F.D. 19115.Population studies on Pennyroyal Mint (Mentho Bakker, J.P. 19115.The impact of grazing on plant communities, plant puleghm L.). I. Germination and sccdlina establishment. Weed Res. populations and soil conditions on salt marshes. Vegetatio 62391-398. i5:u)l309. Brown, D. 1954. Methods of surveying and measuring vegetation. Comm. Polssonet, J., P. Pobeonet, and M. Tbtaplt. 19(11.Development of flora, Bureau of Pastures and Field Crops, Bull. 42. Hurley, Berkshire, vegetation and grazing value in experimental plots of Quercus coccifeu England. garrigue. Vegetatio 46:93-104. Burkart, A. 1975. Evolution of grasses and grasslands in South America. Ponnamparuma, F.N. 19M. Effects of flooding on soils, p. 945. In: T.T. Taxon 245366. Koxlowski (cd), Flooding and plant growth. Academic Press, N.Y. Cabrera, A.L., and EM. Eardbd 1978. Manual de la flora de 10s alredc Rapport,D.J.,H.A. Regia,and T.C. Hutchlmon. 1985. Ecosystembchav- dorm de Buenos Aims. ACME, Buenos Aires. ior under stress. Amer. Natur. 125:617-640. CauhLp5, M.A., L.C. Hidalgo, and A. Gelatoire. 19M. Valoraci6n mot&- Risser,P.G.,E.C.Birnay,H.D.Blocker,S.W.May, W.J.Parton,and J.A. nice de paetizales de la Depreei6n dd S&do. Rev. Arg. Prod. A&u. Wiens. 1981. The true prairie cocsystem. Hutchinson Ross Publ. Co., 443-44. Pennsylvania. Facelll, J.M., R.J.C. L&n, and V.A. Deregibw. 191w.Community struc- Sata,O.E.,V.A.Deregibus,T. Scblichter,andH.AUppe.1981.ProdU~iV- ture in grazed and ungraxed grassland sites in the Flooding Pampa, ity dynamics of a native temperate grassland in Argentina. J. Range Argentina. Amer. Midc Naturr (in press). Manage. 34:48-51. Forman. R.T.T.. and M. Godron. 19111.Patches and structural compo- Sala, O.E., M. Oeeterbeld, R.J.C. Le6n, and A. Sortaao. 1986. Grazing nents for a landscape ecology. BioSciencc 31:733-740. effects upon plant community structure in subhumid grasslands of Cauch, H.G. 19112.Multivariate analysis in community ecology. Cam- Argentina. Vegetatio 67~27-32. bridge Univ. Press, Cambridge. Sanft, R.L., L.R. Rlttenhouse, and R.G. Woodmansee. 19(15. Factors Hurlbert. S.H. 19M. Pseudomplication and the design of ecological field influencing patterns of grazing behavior on shortgrass steppe. J. Range experiments. Ecol. Monog. 54: 187-211. Manage. 38:81-87. InmwtI. Efccto de1 anegamiento prolongado en P.. and A. Soriano. 1988. Taboada, M.A., and R.S. Lavado. 198. Grazing effects on soil bulk un past&l de la Dcpresi6n de1 Salado (Piov. de Buenos- Aires- density in the Flooding Pampa of Argentina. J. Range Manage. Argentina): dirtamica de1 pastizal en conjunto y de Ambrosia rem&oh 41502-505. (Asteraccae). Darwiniana 28397403. Tllman, D. 1982. Resource competition and community structure. Prince- Knapp,R. 1979. Climates and soil of grasslands areas in Europe. p. 4956. ton Univ. Press, Princeton, N.J. In: M. Numata (cd), Ecology of grasslands and bamboolands in the Urban, D.L., R.V. O’NeiU, and H.H. Shugart, Jr. 1987. Landscape ccol- world. Verlag-Jena, Berlin. ogy. Bioscience 37:119-127. Kozloweki, T.T. 19M. Extent, causes, and impacts of flooding, p. l-7. In: Vogl, RJ. 19(10. The ecological factors that produce pcrturbation- T.T. Koxlowski (cd), Flooding and plant growth. Academic Press, N.Y. dependent ecosystems, p. 6394. In: J. Cairns (cd), The recovery process La&r, R.J.C. 1975. Las comunidades herb&cas de la r&n Castelli-Pi, in damaged ecosystems. Ann Arbor Science, Ann Arbor, Michigan. p. 75-107. In: Productividad primaria ncta de sistcmas herbaceos. White, P.S. 1979. Pattern, process, and natural disturbance in vegetation. Monografii 5, C.I.C., La Plats. Bot. Rev. 45~229299. L&n, R.J.C., G.M. Ruech, and M. Oasterheld. 19M. Pastixales pam- peanos-impact0 agropccuatio. Phytococnologia 12:2Ol-218.

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 499 Grazing effects of the bulk density in a Natraquoll of the Flooding Pampa of Argentina

MIGUEL A. TABOADAAND RAirL S. LAVADO

Abstract The lnfiuence of grazing by cattle on soll bulk density was another. In the case of the Flooding Pampa, despite the economic studied in a typic Natraquoll of the Ploodlng Pampa of Argentina importance of grazing, local literature is very scarce. Rusch and for a period of 33 months, by comparing a grazed sltuatlon to an Leon (1983), for instance, attributed changes in the floristic com- enclosure deferred from grazing for 7 years. Pioods took piace ln position of a natural grassland to soil compaction by trampling. this period as usual. Bulk den&y (BD at -33.3 kPa of water Bulk density is a commonly used parameter of soil porosity, retention varled from 1.00 to 1.11 Mg m’ d ln the ungraaed soil and giving an effective indication of the compaction-regeneration ln the grued soil from 1.04 to 1.16 Mg mm3. processes (Bullock et al. 1985). The objective of this study was to Environmental factors were the primary agent controlling BD; evaluate the effects of cattle grazing upon soil bulk density in the only ln some periods were there sign&ant differences between Flooding Pampa region of Argentina. The implications for stock treatments. Slight increases in BD occurred under graalng after the management are also considered. recession of the flood water, and sign&ant decreases occurred in Materials and Methods the ungrazed soil during the large and sudden falls ln water content. In this case the effect of trampling, therefore, would con&t Study Area malnly of impeding the decrease in BD. No compaction was Research was carried out in an area located in the middle of the observed ln periods when no flood occurred or while soli remahred Flooding Pampa near the town of Casalins (Buenos Aires Pro- submerged ln water. vince), in which floods occur most years. In the period under study The results indicated that the variations of buik density caused (from October 1983 to July 1986) floods submerged the soils with a by cattle trampilng were superimposed on those produced by few centimeters of water in winter-spring 1984 and 1985 and in floods and showed an interaction between the effects of land-use autumn 1986. There were no floods in 1983. and the particular environmental condltlons of the region. Vegetation of the area is a natural grazed grassland community characterized by Rptochaetium montevidensis, Briza subaristata, The Flooding Pampa is a subhumid low, very flat plain. Its Eclipta bellidioides, and Menthapulegium (Perelman et al. 1982). 90.000 km* are mostly covered by halo-hydromorphic soils, the It covers more than 7% of soil surface. Grasses are sparsely most conspicuous being Natraquolls (INTA 1977). The region is distributed over the soil and are stratified in the first 10 cm of plant affected by recurrent floods caused by an excess of water: percola- canopy (Sala et al. 1986). The range is mainly devoted to the tion, surface movement, and evaporation are less than rainfall. Soils remain saturated and ponded most years from winter to late production of beef cattle. Dominant soil is a typic Natraquoll, spring, and exceptionally in autumn; they are dry in some General Guido series, moderately saline phases. The soil is charac- summers. Floods and droughts sometimes follow each other. terized by a tough clayey and natric B2t horizon. Conversely, the Because of several constraints, agriculture is limited and the domi- 0.12 m thick Al horizon is loamy (clay %: 23.60), slightly acid (pH nant land-use in the region is the production of beef cattle on 6.20) with high organic carbon percentage (3.20). This horizon has natural grasslands. Cattle remain over the field all year, including a high swell-shrink capacity as its clay fraction has 30 to 50% of those periods when soil is very wet. Trampling in these conditions smectitic materials, causing bulk density to fluctuate with changes is usually considered a severe damage factor in surface structure of in soil water content. Taboada et al. (1988) related soil bulk density grassland soils (Davies 1985, Scholefield and Hall 1986). to moisture content by the function: Trampling by cattle is reported by the literature as causing BD = 1.20 + 0.0008GW - 0.00016 GW* r = -0.95 (1) increases in bulk density by compaction in the soil surface (Heady where BD is soil bulk density and GW is gravimetric soil water 1975, Lull 1959, Van Haveren 1983, Warren et al. 1986, Willat and percentage. Pullar 1983). Nevertheless some results are in disagreement with Other morphological, mineralogical, and chemical properties of this opinion. Laycock and Conrad (1967) found that some this soil have been published elsewhere (Lavado and Taboada increases in bulk density may be ascribed to the lower water 1985, Taboada and Lavado 1986, Taboada et al. 1987). content that grazed soils usually have (Gifford and Hawkins 1978); Van Haveren (1983) showed that not ail soils may be compacted by Treatments trampling because of the high sand content of some soils; and The bulk density and water content of soil were measured in a Lagocki (1978) even found that bulk density decreased when tram- grazed location and at a location excluded from grazing for 7 years pling was performed on a soil with high water table. Another prior to initiation of the study. The locations were not replicated . subject under controversy is the time required for the recovery of The grazed field was a natural grassland which had never been soil structure, as measured by the decrease of bulk density after ploughed, but had been grazed year-round for more than a cen- compaction. For example, Braunack and Waiter (1985) found long tury. During the study period the mean stocking rate was 1.06 periods for recovery, while Warren et al. (1986) showed that short- cattle ha-’ year-‘. This rate is representative of most ranges of the er terms were required. region. The ungrazed enclosure was a 4-ha field fenced and sur- The factors causing these differences are obviously related to rounded by the cattle range; grazing had been excluded from it different soil properties, environmental conditions, and grazing since 1976. Grazing exclusion resulted in the replacement of a large intensities. They make it difficult to transfer results from one site to number of small tussocks by a few large ones. Total biomass, Authors arc wit! the Dcpartamento de Ecolo ia, Facukad de A$ronomia, Univer- mainly standing dead, increased steadily; litter accumulated. sidad de Buenos Ames, Avenida San Martin 44 f 3,1417 Buenos Awes, Argentina. Another major effect was the disappearance of some native plano- Manuscript accepted 24 May 1988. phile species and most of the exotics from the enclosure (Sala et al.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 1986). Nevertheless the soil only showed significant reductions in Results were statistically appraised by analysis of variances salt content, but not in organic matter, total nitrogen, and avail- (ANOVA). When significant differences in BD were found be- able phosphorus contents (Lavado and Taboada 1985, 1987). tween dates, the Tukey test was used to separate the means. Sampling and Analysis Results and Discussion Sampling was carried out in the Al horizon. Five undisturbed soil cores (9 cm in diameter and 10 cm in depth) were taken at Volumetric water content and bulk density at -33.3 kPa in the random from each treatment on 22 dates during the period under Al horizon are shown in Figure 1. Flood periods are also included. study. Sampling was performed carefully because of the extreme Mean VW ranged from 17.06 to 40.79% in the ungrazed soil, and soil water content found along the time in each core. Bulk density from 19.09 to 42.44% in the grazed soil. Only in the first summer of was determined by the core method (Blake 1965) and gravimetric the studied period (November 1983 to January 1984) was soil dry. soil water content by the oven-dry method. Volumetric soil water The rest of the time it remained with moderate to very high water content (VW) was calculated from them. The depth affected by contents. According to Scholefield and Hall (1986), under these animal hooves was evaluated by additional bulk density samples conditions soil surface damage by cattle trampling is likely to taken at the O-4,4-8, and 8-12 cm depths of the Al horizon using 6 occur. cm diameter and 4 cm depth cores. Five samples were taken at Mean BD ranged from 1.00 to 1.11 Mg me3 in the soil of the random from each treatment in November 1984 and June 1985. enclosure where cattle were removed from 1976, and from 1.04 to In order to separate the effect of compaction by animal hooves 1.16 Mg me3 in the soil under grazing. These values were somewhat from that caused by decreases in soil water content (Laycock and low, but consistent (De Kimpe et al. 1982) with the high organic Conrad 1967), measured values of bulk density were adjusted to a carbon content of the Al horizon. Different behavior was observed fixed soil water content (Perrier et al. 1959). In this case bulk in BD with time in both treatments. In the ungrazed soil there were density was standardized at -33.3 kilo Pascal (kPa) of water reten- lower and significant values on 3 dates (November 1983, January tion (29.96% in this horizon), by means of the slopes of equation and December 1985). They were reached after the soil underwent (1). This procedure was judged to be reliable because of the very great loss of water as its VW showed a sudden fall from a previous high and statistically significant (a10.001) correlation coefficient very wet condition (it was saturated with water in 1983 and flooded found for equation (1). both in 1984 and 1985). The sudden decreases in soil water content Soil surface strength was measured in September, November, seemed to be associated with the parallel decreases in BD. This and December 1984, using a Proctor penetrometer (Davidson relationship cannot be explained easily, and further research is 1965). In each month, 20 measurements were performed following needed; but the sequence of events after the autumn 1986 flood a zig zag path over uncovered soil surfaces in each treatment. provides indirect evidence of the active role of the sudden falls in

UNGRAZEO ENCLOSURE GRAZING FLOOD PERIOOS

10’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ’ ONDJFMAMJJASONDJFMAMJ JASONDJFMAMJJ -19 83’- 1984 --1985-- 1986 -

Fig. 1. Bulk density at -33.3 kPa and volumetric soil water content in the Al horizos andfloods during the studiedperiod. Statistical dfferences between dares are indicated by a (0.05 probability level) and between treatments by a * or a ** (0.05 or 0.01 probability levels respectively).

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 501 water content. After the flood water receded, there was no decrease T8blc 2. Bulk dendty in 4 cmhyen of the Al horizon. in volumetric water content as winter began. Presumably, as a result, BD did not decrease either. The grazed soil did not show Mean Bulk Density (Mg m-7 significant decreases in BD with time, despite its water content behaving similarly to that of the ungrazed soil. Dates Ung& Grazed

Only in some dates (for instance November 1984, September and 04 0.96 l 1.00 June 1985) were processes of compaction reported by most of the Nov. 1984 4-8 1.20 * 1.29 literature (Heady 1975, Lull 1959, Van Haveren 1983; Warren et al. 8-12 1.28 1.30 1986, Willatt and Pullar 1983). These processes were reflected by O-4 0.97 0.99 slight increases in BD in the grazed soil, which appeared after the June 1985 48 1.23 1.24 receding of the flood water. Soil compaction by cattle trampling 8-12 1.29 1.28 was also verified by measurements of soil surface strength (Table ;~rcncca between mean values in same row are significant at the 0.05 probability 1). Approximately two-fold increases were observed in the grazed

T8ble 1. Soil rurf8ce etrengtb 8e m-red by the proctor penetromhr in significant differences between treatments (Fig. l), they appeared 1984. only in the O-4 and 4-8 cm layers. In June 1985 when there were no differences, none of Al horizon layers presented differences in BD either. Trampling influence occurred, then, within the first 8 cen- Mean Soil Strength @PA) timeters of soil depth. Dates Ungrazed Grazed Conclusions September 450 l * 940 November 470 l * 690 The environmental conditions of the Flooding Pampa, particu- December 3710 l * 6440 larly its floods and the water regime of its soils, were associated l+Diifennccs bctwccn mean values arc signiricantat the 0.01 probabilitylevel. with variation in soil bulk density and were shown to be the main cause for the lack of severe and long-term damage by cattle trampling. It is difficult to attribute to soil compaction a major influ- soil as compared to the ungrazed. On the other hand, a highly ence on the floristic composition of this natural grassland com- significant decrease of soil strength occurred in the soil under munity. Other already observed modifications in soil properties grazing from September 1984, when a flood just ended, to (Lavado and Taboada 1987) could be more important. November 1984. The loss of soil strength was not attributable to The results showed that the variations in BD caused by cattle differences in water content between dates (Fig. l), but probably to trampling were superimposed on those produced by climatic and a process of mellowing or softening of soil structure (McKenzie soil characteristics. There were times when grazing effects were and Dexter 1985, Utomo and Dexter 1981). This process consists detectable, but for a longer time frame, grazing effects were com- in partial slaking of soil aggregates which weaken but do not pletely overriden by inherent characteristics operating in the completely fragment into smaller particles. It takes place after the Flooding Pampa. This finding may be applicable in selecting graz- soil is carried to very high water content, as occurred in this study ing systems appropriate for flooding soils; stocking rates must be after flooding. Besides, soil softening at surface is largely enhanced regulated in periods when it is expected that trampling may affect by the remoulding action of animal hooves, thus reflecting the the soil. failure of wet soil to accommodate a load (Mullins and Fraser 1980). Both mechanisms-soil mellowing and remoulding by Literature Cited trampling-may be pointed out as the direct causes of the compaction found in the grazed soil after flooding. Interesting to note, no BI8ke, G.R. 1965. Bulk density. In: Black, C.A. (ed.). Methods of Soil flood occurred in 1983 and soil BD did not increase under grazing. Analysis”. Agronomy Series 9, Part 1, Madison, Wis. Braunack, M.V., 8nd J. Wlter. 1985. Recovery of surface soil properties of This lack of difference between treatments was maintained for ecological interest after sheep grazing in a semi-arid woodland. Aust. J. several months, until the recession of the next flood. Ecol. lOz451460. The significant decrease in BD in the soil of the ungrazed enclo- Bullock, P., A.C.D. Newm8n, 8nd A J. Tbom8eson. 1985. Porosity aspects sure was the other source of differences between treatments. In this of the regeneration of soil structure. after compaction. Soil Till. Res. case the effect of trampling would consist mainly of impeding these 5:325-342. decreases during the sudden falls in water content. On the other D8vidson. D.T. 1965. Penetrometer measurements. In: Black, CA. (cd.). hand no significant differences in BD were found while the soil Methods of Soil Analysis”. Agronomy Series 9, Part 1. Madison, Wis. remained submerged by water. This accords with results of Saini et Dar& P. 1985. Influence of organic matter content, moisture status and al. (1984), who found that soil compactibility under a given load time after reworking on soil shear strength. J. Soil Sci. 36:299-306. De Kimpe, C.R., M. Bender-Cudou, 8nd P. Joliceur. 1982. Compaction tends to decrease at the highest water content, and subsequent and settling of Quebec soils in relation to their soil water properties. Can. increases in bulk density are unlikely to occur. J. Soil Sci. 62165-175. Periods with differences in BD never lasted more than a few Doll, U.M., 8nd V.A. Deregibue. 19%. Efecto de la exclusi6n de1pastoreo months, being shorter than those long-time processes shown by sobre el subsistema subterrhneo de un nastizal temnlado. htimedo. Turri- Braunack and Walter (1985). The cycles of wetting and drying of alba 36337-344. this soil probably played an active role in the recovery of porosity Cifford, G.F., 8nd R.H. H8wkins. 1978. Hydrologic impact of grazing on (Bullock et al. 1985, Newman and Thomasson 1979, Oades 1984), infiltration. Water Res. Resour. 14:305-313. but the disappearance of differences in BD seemed to be more He;ldy, H.F. 1975. Rangeland Management. McGraw-Hill, N.Y. specifically related to the soil rewetting after the summer (Fig. 1). INTA. 1977.La Pampa Deprimida. Condiciones de Drenaje de sus Suelos. INTA, Departamento de Suelos, Public No. 154. Bs. As. In addition, several properties of this soil that maximize the aggre- Lagockl, H.F.R. 1978. Surface soil stability as defmed and controlled by a gation of soil particles (Oades 1984) would prevent deleterious drainage criterion. In: Emerson, W.W.; R.D. Bond and A.R. Dexter effects due to this kind of land-use: its unploughed condition, its (eds.). Modification of Soil Structure. John Wiley & Sons. very high content of organic carbon, and according to Doll and L8V8d0, R.S., 8nd M.A. T8bo8d8. 1985. Influencia de la exclusi6n de1 Deregibus (1986), the great amount of below-ground biomass pastoreo sobre algunas propiedades quimicas de un Natracuol de la concentrated in the Al horizon. Pampa Deprimida. Ciencia de1 Suelo. 3:102-108. Bulk density in 4-cm layers of the Al horizon is shown in Table hv8do. R.S., 8nd M.A. T8bo8d8.1987. Soil salinization as an effect of 2. In November 1984 when BD of the whole horizon showed grazing in a native grassland soil in the Flooding Pampa of Argentina. Soil Use Manage. 3:143-148.

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 Laycock, WA., and P.W. Conrad. 1%7. Effect of grazing on soil compac- Safni, G.L.; T.L. Chow, and I. Gbanem. 1964. Compactibiity indexes of tion as measured by bulk density on a high cattle range. J. Range some agricultural soils of New Brunswick, Canada. Soil Sci., 137~33-38. Manage. 20:lM140. a O.E.;M.Oehrb&& RJ.C. L&n, and A. Soriam. 19g6. Grazing Lull, H.W. 1959. Soil compaction on forest and range lands. Forest Serv. effect upon plant community structure in subhumid grasslands of Argcn- USDA. Misc. Pub. No. 768. Washington D.C. tina. Vegctatio. 6727-32. McKenzk, B.M., and AR. Dexter. 1985. Mellowing and anisotropy Scholefleld, D., and D.M. Hall. DM6. A recording pcnctromcter to mea- induced by wetting of mouldcd soil sampks. Aust. J. Soil Rcs., 23:3747. sure the strength of soil in relation to the stresses exerted by a walking Mdlios, C.E., and A. Fraser. 19g& Urn of the drop pcnctromctcr on cow. J. Soil Sci. 37~165-176. undisturbed and rcmouldcd soils at a range of soil water tensions. J. Soil Taboada, MA., and R.S. Lavado. 1986. Caractcrlsticaa de1r&um a&co Scf. 31:25-32. de un Natracuol de la Pampa Deprimida. Ciencia de1 Suelo. 466-71. Newman, A.C.D., and A.J. Thonuno n. 1979. Rothamstcd studies on soil Taboada, M.A.; R.S. Lavado, and MC. Camlli6n. 1988. Cambios volum& StNCtUrC. III. Pore size distributions and shrinkage processes. J. Soil tricos en un Natrocuol tipico. Ciencia de1 Sue10 (in press). Sci., 30~415440. Utomo, WM., and A.R. Dexter. 1981. Tilth mellowing. J. Soil Sci. Gada, J.M. 19g4. Soil organic matterand structural stability: mechanisms 32187202. and implications for management. Plant Soil. 76319-338. Van Haveren, B.P. 1983. Soil bulk density as intluenced by grazing inten- Perelman,S.B.;RJ.C.Le6n,andV.A.De+bus.1%2.Aplicaci6ndcun sity and soil type on a shortgrass prairie site. J. Range Manage. mcltodo objetivo al estudio de Rs comunidadcs vegctalcs de pa&al de la 36586588. Dcprcsi6n de1 Salado (Provincia de Buenos Aims). Rev. Fat. Agron. Warren, S.D., MB. Nevill, W.H. Blackburn, and N.E. Garaa. 19%. Soil 3:274. response to trampling under intensive rotation grazing. Soil Sci. Sot. Per&r, EA., D.R. Nielnen, and J.E. Doan. 1959. Adjustment of bulk Amer. J. 50:11361340. density to an oven-dry volume basii. Soil Sci. 88291-293. Willatt, S.T., and D.M. Pullar. 1983. Changes in soil physical properties Rus~b, GM., and RJ.C. L&I. 19g3. Sucesi6n inducida por pastorco en under grazed pastures. Aust. J. Soil Rcs. 22343-348. un pastizal de la Dcprcsi6n de1 Salado (Provincia de Buenos Aires). Rcsumencs XI Reuniba Argentina de Fcologla. Villa Giardino, Cbr- doba, Argentina. Stability of grazed patches on rough fescue grasslands

WALTER D. WILLMS, JOHN F. DORMAAR, AND G. BRUCE SCHMLJE

cootiouousstockiog usu8Ily leads to the fonlutioo of grued determine avoidance or selection. Grazed patches are maintained patclkea.However, the effect of p&bee 00 the grazelMd c0mmun- by repeated grazing of regrowth, which is preferred to more mature ity is related to their stability. Therefore, we ztudied the spatial vegetation. stability of grazed pat&e oo Rough Fezcue Gramlande by map Grazed patches may form early in the grazing season when ping forage remov8l clmuee 00 10 site8 over 8 Cyeu period, tezting forage supply is high (Ring et al. 1985) but their formation and stability uzing the Kappa index (H), and characterizing the soila stability are affected by the intensity and time of grazing. Intensive and vegetatioo of overgrazed and undergrazed patches. Spotlal stocking early in the season results in patches that are less well stability of grazed patches between comecutivc yeam wm good defined than those on pastures stocked season-long (Ring et al. (lEO.26) oo &en experieming low grazing prwsure. However, oo 1985). Furthermore, both the size and number of grazed patches dtea having high grazing preezure,spatial stability waz Iesecomizt- increase as the season progresses, presumably in response to eot between comecutive yeuz (DKSO.45) ud low over 8 Cyeu increased grazing pressure. period (K10.10). Overgrazed patches were dominated by graziog- Grazed patches are characteristic of range stocked season-long resistant zeral speck, but undergrazed patcher were dominated by at a moderate rate intended to maintain plant vigor and to allow climu species. Rough feecue (%cstuco SC&&~ 8nd P8rry oat for carryover to the following year. However, the presence of grazs @u&to& porryi) plantz were 50% shorter, and forage patches indicates that some areas are overgrazed while others are production waz about 35% lees, on overgrazed than on under- underutilized. This defeats the purpose of the recommended stock- grazed patches. Soil organic matter, carbohydratea, and depth of ing rate and results in less efficient use of the range resource. It also Ah horizon were zignificantly greater oo undergrazed patches but suggests the possibility that the grassland might deteriorate in ureaze activity, NO& NH,, and available phosphorus were patches despite a moderate stocking rate. greater oo overgrazed patchez. Overgrazed and undagrazed Deterioration in the plant community is largely contingent on patches were stable in the long term, although patch bouodariez heavy regrazing of the patches over many years, which could cause fluctu8ted. the loss of the cliix species and the increase of seral species. Within a field, rough fescue (Fesrucu scubrella Torr. var. major Key Words: grazing pressure, zpatial &ability, organic matter, Vasey), the dominant species in the Rough Fescue Grassland, was speck composition, herbqe yield8 nearly eliminated after 5 years of high grazing pressure (Willms et Patchy grazing occurs when forage supply exceeds livestock al. 1985). After 17 years, the Ah horizon of the soils had deterio- demand (Spedding 1971) and cattle have the opportunity to graze rated, with a reduction of organic matter and a change of color selectively. Patches occur as a result of selective grazing caused by from black to dark brown (Johnston et al. 1971). palatability differences within microsites (Bakker et al. 1984) that Patchy grazing may be eliminated with high grazing pressure or by utilizing a system that uses high animal densities for short Authors are range ecologist, soil scientist, and statistician, Agriculture Canada, Research Station, Lethbrid8e, Alberta ‘II J 4B I. periods of time. The first option is undesirable, because it inevita- Manuscriptaccepted 13 June 1988.

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 503 Fig. 1. Site locations (I-10) on jiwla3 L., M. and Hstockedat 1.2, 1.6, and 2.4 A &U/ha, respectively.

bly causes grasslands to deteriorate, while the second option Methods requires extensive fencing. Season-long grazing at a moderate Ten sites were selected in 1982 among 3 fields (Fig. 1) that stocking rate is the least costly alternative but inevitably leads to contained grazed and ungrazed patches. The fields had been the formation of patches and possible range degradation. There- stocked since 1949 at fixed rates over a dmonth grazing period fore, in order to assess the long-term effects of such grazing man- which extended from mid-May to mid-November: light (field L), agement we studied the stability of patches and their vegetation 1.2 AUM/ ha; moderate (field M), 1.6 AUM/ ha; and heavy (field and soil characteristics. We hypothesized that the spatial distribu- H), 2.4 AUM/ ha. Sites 1 to 4 were located in field M, sites 5 to 8 in tion of patches is stable over the short term, but over a long period field L, and sites 9 and 10 in field H. However, because of localized is dynamic so that degradation of the grassland does not occur. If distribution differences of cattle within each field, the actual patches are stable in the short term but dynamic in the long term, grazing pressure at each site was not known. The sites could, then their only effect might be a reduction of forage yield caused by nevertheless, be categorized into 2 grazing pressure groups (GPG), the removal of litter. However, long-term stability would be low or high, based on stocking rate and their location within a field. reflected by a change in species composition and, possibly, in soil As a result, site 4 was grouped with 5 to 8 to represent the low GPG characteristics. and sites 1 to 3 were grouped with sites 9 and 10 for form the high GPG. Sites 1 to 3 were in an area that was heavily utilized by cattle Materials and Methods because of close proximity to a gate, a water source, and a grove of trees. Although site 4 was nearby, it experienced less grazing Site Description The study was conducted at the Agriculture Canada Range pressure because it was partly isolated by a small gully and a fence Research Substation located in the Rough Fescue Grasslands in that diverted cattle movement. Each site was 5 X 5 m and the comers were marked with wooden the foothills region of southwestern Alberta near Stavely, about 85 stakes which extended about 10 cm above the ground surface. A 1 km NW of Lethbridge. The topography is undulating, varying in elevation from 1,280 to 1,420 m (Fig. 1). Average seasonal precipi- X 1 m grid was placed over each site and the forage removal classes tation at the Substation, over 34 years for the period April to (FRC) were mapped onto graph paper. Removal classes were, initially: (1) no evidence of grazing, (2) some evidence of grazing August, was 348 mm (Table 1). Annual precipitation at 2 similar denoted by the removal of tips of leaves within about 20% from the sites within 65 km (Pincher Creek and Pekisko) averaged 614 mm. The vegetation is representative of the Rough Fescue Association top and representing less than 5% forage removal, (3) the broadest range representing 5% to 75% of use, and (4) 75% or more of forage described by Moss and Campbell (1947). Rough fescue was the removed. Mapping was usually done in October near the end of the dominant species in the community with Parry oat grass (Dantho- grazing season to avoid snowfall. Snowfall delayed the mapping nia purryi Scribn.) the co-dominant. The soils are classified as scheduled for 1984 until April, 1985. Orthic Black Chemozemic (Udic Haploboroll) developed on till overlying sandstone.

504 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 A composite map of the FRC was copied onto tracing paper for T8ble 1. Preciplt8tlon (mm) dwlng the growing se8son 8t the St8vely each comparison representing the consecutive years 1982/83, 83184, and 84185, as well as for the initial and final years, 1982/ 85. The results was a map showing a maximum of 16 subclasses, of a 4 April June July August Total X 4 matrix, representing shifts among the 4 grazing classes between MaY years. However, since FRC 1 and 2 were similar and sometimes 1982 43 104 21 27 228 diflicult to distinguish, they were combined into one class prior to 1983 ;: 21 28 57 32 168 1984 23 46 71 27 168 final analysis. The resulting 3 FRC represented conditions of (1) 1 20 23 0 99 173 undergrazing, (2) intermediate grazing, and (3) overgrazing. The 1985 31 1986 11 52 77 38 256 proportional area of each subclass was determined by cutting the 34-year 64 70 ; 55 60 348 paper and combining and weighing common subclasses. Propor- 8VCIas tional area was assumed to be equivalent to the proportional weight of paper of each subclass to the total. Precautions were taken to avoid contaminating the paper prior to weighing. Paired plots were located within 1 m of each other and were The Kappa index (Fleiss 1981) was used to quantify the spatial protected from grazing with conical exclosures each having a stability of the FRC between consecutive years and between the diameter of 70 cm. Exclosures were erected prior to the grazing years 1982 and 1985. Comparisons were made for individual sites season and were moved to different locations in 1986. In August, as well as for sites grouped by GPG. The Z value was calculated for forage under the exclosures was harvested from 0.25-m2 plots and each comparison and the probability determined in a test of the separated into litter and current production. Each forage comPo- hypothesis that the spatial distribution of FRC between years was nent was dried and weighed to determine yield. Prior to CliPPing in independent. KM denotes that overlap in common FRC between 1985, species cover was determined in each plot, using 2dm2 years was f chance and indicates patch stability, whereas K = 1 subplots, according to methods described by Daubenmire (1968). shows a complete agreement in overlap of FRC from one year to In addition, average plant heights, measured by the 1onRest leaves the next (Fleiss 1981), and therefore perfect stability; K

T8blc 2. Sp8tl8l stability of FRC between consecutive ye8rs 8nd tbc first hypothesis th8t the distrlbotion of FRC between ycus ls lodependent. K 8nd fbul ycu lor lndivkhul sites 8nd combirmtion of siteagrouped = K8ppr Index, SE = St8adud Error. 8ccordhg to grazing p ressure, 8nd probabilltke (P) in 8 test of the

Grazing Site prcssurc number 1982-83 1983-84 1984-85 1982-85 High 1 K (SE) 0.39 (0.07)

3 K 0.34 (0.06) 0.09 (0.06) 0.08 (0.09) 0.02 (0.04) P

10 K 0.45 (0.07) 0.28 (0.06) 0.15 (0.06) 0.07 (0.05) P

Combined K 0.32 (0.03) 0.24 (0.03) 0.16 (0.03) 0.01 (0.02) P

Low 4 K 0.43 (0.08) 0.47 (0.09) 0.44 (0.08) 0.42 (0.08) P

5 K 0.50 (0.06) 0.46 (0.07) 0.46 (0.08) 0. I8 (0.06) P

Combined K 0.50 (0.04) 0.42 (0.04) 0.44 (0.04) 0.29 (0.04) P

Other 7 K 0.08 (0.06) 0.10 (0.04) 0.00 (0.00) 0.02 (0.01) P x.05 0.011 x.05 0.050

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 505 weighed. Urease activity was determined at pH 9.0 by incubating ties from FRC 1 to 3. Grass cover was greater and forb cover less in with tris(hydroxymethyl)minomethane buffer (0.05 M), urea FRC 1. However, the combined cover of species that increase with solution, and toluene at 37’ C for 2 hours, and measuring the grazing was only marginally less (111 vs. 136%) in FRC 1 but the ammonium released after steam distillation (Tabatabai and Brem- cover of species that decrease with grazing was greater (51 vs. 18%) ner 1972). Carbohydrates were determined by the phenol-sulfuric (Table 4). The species composition between GPgroups was similar acid method of Dubois et al. (1956) as modified by Doutre et al. (Table 4). (1978). NaHCOs-soluble phosphorus, or available phosphorus, The heights of rough fescue, Idaho fescue, and Parry oats @~SS was obtained as described by Olsen et al. (1954) while NOS-N and were greater (PCO.01) in FRC 1 than in FRC 3. Heights in FRC NH,-N were obtained by KC1 and steam distillation (Bremner 1 and 3 averaged, for rough fescue, 39.5 and 18.1 cm respectively; 1965). The plant species, forage yields, and soils data were analyzed Undergrazed using least squares analysis of variance. Proportions were trans- 80 formed by the log transformation prior to analysis (Steel and Torrie 1980). Analysis was performed on individual species and groups of species. In each analysis, variation was partitioned by (1) GPG, (2) site nested in GPG, (3) FRC, and (4) the interaction of 70 GPG and FRC. The effect of (1) was tested using (2) as the error term while the remaining sources were tested against the residual. Forage yield was analyxed as above but with year as another factor. In this case, year was tested against the term of year by site nested in 60 (1) while the interaction of year and (3) was tested against the residual. 50 Results The study was made over a period characterized by below- average precipitation during the growing season (Table 1). In 1984 40 and 1985, precioitation was about half of the mean. Spatial stability among FRC as tested with the Kappa index was significant (X0.01) for most comparisons (Table 2). However, spatial stability among individual sites, between consecutive years 30 from 1982 to 1985, decreased within the high GPG but remained $ nearly constant within the low GPG. Spatial stability between 1982 ._ and 1985 was insignificant for individual sites in the high GPG but 5 20 significant for sites in the low GPG. These trends were also con- firmed with an analysis of sites combined by GPG (Table 2). Q From 1982 to 1985, the proportion area of FRC 1 decreased from 53 to 3% in the high GPG but remained nearly constant in the 8 10 low GPG (Fig. 2). The trends for FRC 3 were inversely propor- 2 tional to these. Spatial stability among FRC between years was low in site 7 5 a5 (Table 2). The proportion of FRC 1 present on this site was 23,72, 0 L 98, and 100% in the years 1982 to 1985, respectively, whereas the EE proportion of FRC 3 was 12% in 1982 but 0% in every other year. Overgrazed The proportion of common area occupied between 2 consecutive .- years to area occupied in the second year for FRC 1 averaged 0.73 r Grazing Pressure and 0.85 in the high and low GPG, respectively and, for class 4, averaged 0.43 and 0.53 for the same GP groups, respectively (Table 8 40 3). However, in the 1982/85 comparison, the proportion of FRC 1 was similar between high and low GPG but the proportion of FRC & 3 was considerably greater in the low GPG than in the high. 30 Composition of most grass species, as defined by the foliage ground cover occupied by species differed (KO.01) between FRC 1 and 3 (Table 4). The greatest effect was in the decrease of rough fescue and the increase of bluegrass (Poa) and sedge (Carex) spe-

Table 3. Proportion of common area to area to second year occupied by undeqrued (FRC 1) or overgrazed (FRC 3) pat&a for dtee averaged 10 by bigb vs. low grszing pressure group.

Undergrazed Overgrazed 0 Years compared High Low High LOW 1982183 1983/84 1984/85 1982185 1982-1983 0.85 0.83 0.35 0.81 1983-1984 0.71 0.82 0.51 0.43 Years compared 1984-1985 0.56 0.91 0.42 0.36 Fig.2. Proportion (%) of undergrazed (FRC 1) or overgrazed (FRC 3) Average 0.73 0.85 0.43 0.53 patches in consecutiveyears and in thefirst andfinalyear of the study in 1982-1985 0.80 0.76 0.21 0.73 relation to grazingpressure.

506 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 T8blc 4. C8nopy cover (96) of selected specks or specia groupr, with common cb8r8cteristla, in rehtion to lmdergr88ed (FRC 1) vs. over- Discussion gwed (FRC 3) p8tcben mod to sitea repreaentin@igb vs. low CPG, wltb The areas occupied by the FRC fluctuated as indicated by prob8biUtiee (P) in 8 test tb8t tbe precedio~ me8ne 8m the a8me. Kappa values which were 10.57 for all comparisons (Table 2) and the variable proportions of common area occupied by the under- Grazing grazed and overgrazed classes (Table 3). However, except for the Pressure overgrazed class in the high GPG, the common area occupied Patches Group between the fust and final year of the study was similar to the Under Gver- average of common areas occupied between consecutive years grazed grazed P High Low P (Table 3). Furthermore, within the undergrazed and overgrazed patches, stable conditions were maintained over many years as Specie8 (or species group) indicated by their species composition and soil characteristics. Festuca scabrella 43.7 3.7

T8bk 5. Rerkp production 8nd litter quntity (g/mz) on ondergrrzed (FRC 1) 8nd overgr8zed (FRC 3) p8tche8 across both high 8nd low gr8zing p-Eroulw.

1985 1986 Effect(P)

Under- Over Under- Gver- Grazing Year grazed grazed grazed gl=d (a) (b) aXb Production 235.3 137.9 402.3 266.2 CO.01

T8ble 6. Soil churcteristics in rehtion to gluing prueure group 8nd ondergaed (FRC 1) 8nd overyrzed (FRC 3) patches.

Grazing Pressure Patch

Soil characteristic High LOW EffectP(l,adf) Undergrazed Overgrazed Effect P (1,8 df) Organic matter (%) 12.08 12.53 0.530 13.13 11.48 0.001 Solvent-extractable OM (g/kg) 3.52 3.66 0.680 3.86 3.32 co.001 Carbohydrate (mg/g) 8.12 3.87 0.086 6.69 5.30 0.044 Urease activity (g of N/kg) 0.15 0.15 0.649 0.14 0.16 0.041 Available phosphorus (mg/ kg) 4.41 6.37 0.321 3.85 6.94 0.095 N&-N (mg/ kg) 8.80 9.03 0.913 7.82 10.01 CO.001 N&N (mgl kg) 6.49 6.16 0.875 5.14 a.46 0.019 Ah (cm) 20.80 17.70 0.394 22.40 16.10

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 507 to remove 8% production resulted in uniformly heavy use (Willms Balph, D.F., and J.C. Malechek. 1985. Cattle trampling of crested wheat- 1987). grass under short-duration grazing. J. Range Manage. 38:226-227. Differences in depth of the Ah horizon found between the over- Bremner, J.M. 1965. Inorganic forms of nitrogen. In: Black, C.A. (ed.), grazed and undergrazed patches may have been caused by grazing Methods of soil analysis. Part 2. Chemical and microbiological proper- or trampling. Overgrazing will favour shallow rooted plants which ties. Agron. 91179-1237. Daubemnire, R. 1968. Plant communities. A textbook of plant synecology. might alter soil organic constituents and perhaps depth of soil by Harper and Row, Pub]., London. reducing the amount of organic matter added. Dormaar, J.F., A. Johnston, and S. Smoliak. 1984. Seasonal changes in It is unlikely that the correlation of shallow Ah with overgrazed carbon content, dehydrogenase, phosphatase, and urease activities in patch was due to chance. Within the Rough Fescue Grasslands, mixed prairie and fescue grassland Ah horizons. J. Range Manage. depth of the Ah horizon varied within the range found in our 37:31-35. samples (Shantz 1967); however, that variation is randomly dis- Doutre, D.A., G.W. Hay, A. Hood, and G.W. van Loon. 1978. Spectro- tributed in relation to the vegetation and cannot be related to the photometric methods to determine carbohydrates in soil. Soil Biol. species composition found on overgrazed patches. In an ungrazed B&hem. 10~457462. community, the vegetation is uniformly dominated by rough Dubois, M., K.A. Gilles, J.K. Hamilton, P.A. Rebers, and F. Smith. 19%. Calorimetric method for determination of sugars and related substances. fescue tussocks and litter. It is likely, therefore, that the present Anal. Chem. 24:141l-1414. characteristics of the soils are related to the previous grazing Fleiss, J.L. 1981. Statistical methods for rates and proportions. 2nd ed. pressure on the patches. John Wiley and Sons, Inc., Toronto. Patch stability between years was related to grazing pressure on Johnston, A., J.F. Dormur, and S. Smoliak. 1971. Long-term grazing the site (Table 2). Grazing pressure was also increased by successive effects on Fescue Grassland soil. J. Range Manage. 24:185-188. years of below-average precipitation, which reduced plant produc- Jones, R.M., and D. Ratellff. 1983. Patchy grazing and its relation to tivity. Because stocking rates were held constant, grazing pressure deposition of cattle dung pats in pastures in coastal subtropical Queens- was related directly to productivity and indirectly to precipitation. land. J. Australian Inst. Agr. Sci. 49109-I 11. Low rainfall from 1983 to 1985 (Table 1) was associated with a MacDiarmid, B.N., and B.R. Watkln. 1972. The cattle dung patch. 2. Effect of a dung patch on the chemical status of the soil, and ammonia continuing decline in the proportion of undergrazed patches but an losses from the soil. J. Brit. Grassl. Sot. 27~43-48. increase in the proportion of overgrazed patches on sites with high Moss, E.H., and J.A. Campbell. 1947. The fescue grassland of Alberta. grazing pressure (Fig. 2), whereas the proportion of undergrazed Can. J. Res. C25:209-227. patches was only slightly reduced on sites with low grazing pressure Olsen, S.R., C.V. Cole, F.S. Watanabe, and L.A. Dean. 1954. Estimation (Fig. 2). The consequence was greater spatial stability on sites of available phosphorus in soils by extraction with sodium bicarbonate. experiencing low grazing pressure (Table 2). USDA, Circ. 939. U.S. Government Printing Office, Washington, D.C. The undergrazed and overgrazed patches persisted for longer Ring, C.B. II, R.A. Nicholson, and J.L. Launchbau8h. 1985. Vegetational than the 4 years over which they were mapped. This is indicated by traits of patch-grazed rangeland in west-central Kansas. J. Range Man- the dominance of set-al species on overgrazed patches and the large age. 3851~55. Shantz, B.R. 1967. Rodent-watershed relationships. Proj. Prog. Rep. 82-5- Ah horizon differences in soil chemistry and depth of the between 5-132. Can. Wildl. Serv. overgrazed and undergrazed patches. Speddht8, C.R.W. 1971. Grassland ecology. Oxford Univ. Press, Ely The high degree of stability of grazed patches ensures that some House, London. areas will be overgrazed while others will be undergrazed. The Steel, R.G.D., and J.H. Torrie. 1989. Principles and procedures of statis- effects of such heterogeneity in the grassland ecosystem are varied. tics. A biometrical approach. 2nd ed. McGraw-Hill Book Co., Toronto. Grazed and ungrazed patches provide more diverse habitat for Tabatabai, M.A., and J.M. Bremner. 1972. Assay of urease activity in soils. animals. Undergrazed patches ensure the presence of climax spe- Soil Biol. Biochem. 4479487. cies in the community and the potential for recolonizing over- Walkley, A., and LA. Black. 1934. An examination of the Degtjareff grazed areas. They also provide emergency forage during years of method for determining soil organic matter, and a proposed modification of the chromic acid titration method. Soil Sci. 3729-38. below average precipitation and enable uninterrupted grazing. Willms, W.D. 1987. Forage production and utilization in various topogra- However, undergrazed patches represent unused production, in phic zones of the Fescue Grasslands. Can. J. Anim. Sci. 68:21 l-223. most years, which deteriorates with weathering and is a cost of Willms, W.D., S. Smollak, and A.W. Bailey. 1986. Herbage production season-long grazing which permits their development. following litter removal on Alberta native grasslands. J. Range Manage. 39536539. Literature Cited Willms, W.D., S. Smohak, and J.F. Dormaar. 1985. Effects of stocking rate on a Rough Fescue Grassland vegetation. J. Range Manage. Bakker, J.P., J. de Leeuw, and S.E. van Wicren. 1984. Micro-patterns in 38:220-225. grassland vegetation created and sustained by sheep grazing. Vegetatio 55:153-161.

508 JOURNAL OF RANGE MANAGEMENT 41(8), November 1988 Effects of phenology, site, and rumen fill on tall larkspur consumption by cattle

JAMESA. PFISTER,GARY D. MANNERS,MICHAEL H. RALPHS,ZHAO XIAO HONG, AND MARK A. LANE

AbStlBCt Tall I8rkspur(De~hinh burbeyj)is a m8jor cause of livestock study, but cattle consumed tall larkspur only from flowering plants death on mount8in ranges. The influence of plant phenoiogy, (Pfister et al. 1988). grazing site, 8nd rumen fill on tall lukapur consumption ~8s Numerous reports from ranchers and researchers (e.g. Marsh et evaluated during July 8nd August, 19%7.Livestock consumption al. 1929) indicate that cattle losses are more severe immediately of l8rkspur ~8s determined using bite counts during 4 phenologierrl after rain or snow showers. Several plant and animal factors may st8ges: bud, e8riy flower, flower, and pod. Further, we ex8mined interact to increase losses under these circumstances. Marsh et al. i8rkspur ingestion in a shaded tree site 8nd in UI open sun site 8t 0, (1929) believed these increased losses were probably because cattle 50, urd 100% rumen fill levels using rumiarlly c8nnuUed steers. congregated under groves of trees where larkspur was abundant. Steers on the 0, 50,8nd 100% fill levels consumed 9,15, and 17% Hunger may be aroused (Toates 1980) after cattle spend several Irrkspur, respectively (fiO.15). ,There ~8s 8 site effect (-0.06) hours under tree shelter during storms. Larkspur alkaloids are with steers eating 17 8nd 11% lukspur in the shade and sun sites, N-based secondary metabolites, and the fertilizing effect of anim- respectively. Over the summer, iukspur comprised 6% of cattle als bedding in groves of trees may influence both the palatability diets. No llukspur ~8s consumed during the bud st8ge. Larkspur and toxicity of adjacent tall larkspur plants. Gershenzon (1984) consumption pe8ked at 10% of c8ttie diets during the pod strge. noted that supplementary N promoted alkaloid production in Lepves of t8ii larkspur contrined >3% total 8lk8loids (dry weight) numerous species. Further, differences in shading can alter the in erriy July, but declined greatly with matwtion. Lukspur ~8s chemistry of forage plants compared to others grown in full sun- very nutritious, with crude protein levels 12 to 20%, and fiber levels light (Majak et al. 1977, Wilson 1982). <20% during most of the summer. C8ttie diets, as determined with We wished to test 2 hypotheses: first, based on previous work esopbnge8lly flstul8ted animals, were also high in crude prottin (Ester et al. 1988), we felt that cattle did not consume D. burbeyi 8nd low ia fiber during the summer. We propose a toxic window before the flowering stage. Therefore, we investigated tall larkspur hypdbesisn~g~~~~~mdtoxidty.Thishypo~~ consumption during 4 phenological stages: bud, early flower, full predicts th8t most c8ttie losses will occur during the flowering flower, and pod. Second, we hypothesized that cattle selection for st8ge. We found tbrt t8ll iukspur ~8s unp8l8t8bie to c8ttie from tall larkspur would differ in exposed areas receiving full light the bud st8ge until the flowering r8cemes hrd eiong8ted, md then compared to selection for partially shaded plants within and adja- consumption generally incre8sed with pl8nt nmtmntion. Even cent to a grove of trees. Cattle used the tree area as a source of though p8l8tPbility md consumption lncrersc during the gr8zing shade while loafing and as a bedding ground many nights. Further, season, c8ttle c8n graze tall I8rkspur with a much lower risk of we wished to determine if rumen fill influenced tall larkspur con- toxicosis when toxicity is low lrter in the gruing se8eon. sumption at these 2 sites.

Key Words: poisonous plurts, attie diets, De~hinium bmbcyi, Materials md Methods 8lk8loids, gr8zing behnvior The study site was located at the head of Six Mile Canyon, near Tall larkspur (Delphinium spp.) has been characterized as the Manti, Utah, at 3,200 m elevation. Precipitation at the study site most serious poisonous plant problem on western U.S. mountain occurs mainly as winter snowfall and as summer thundershowers; rangelands (Ralphs et al. 1988). The toxin(s) in larkspur are diter- precipitation totaled 12.7 cm during July and August, 1987. A penoid alkaloids (Pelletier and Keith 1970). Total alkaloid content description of the study area and the dominant vegetation was may be related to plant toxicity (Olsen 1977a); however, individual given by Pfister et al. (1988). Briefly, tall larkspur patches domi- alkaloids, or combinations thereof probably provide more reliable nate areas where snow accumulates during winter. Dense currant predictions of toxicity (Olsen 1983). (Ribes spp.) mottes and open grass-forb associations dominate the There is little information on when and how much tall larkspur is remainder of the non-wooded landscape. consumed by grazing livestock. Early poisonous plant researchers commented that tall larkspur plants were ,unpalatable to cattle Phenoiogial St8ge after flowering, but they gave no data to substantiate their remarks The effect of phenological stage of tall larkspur on consumption (see Glover 1906, Marsh et al. 1916, and Marsh et al. 1929). Recent by grazing cattle was determined using 4 experimental periods. research has not supported these comments. Pfiiter et al. (1988) Period I (bud stage) was from 2 July to 11 July. Virtually all reported that cattle grazing mountain ranges ate large, but not larkspur plants within the pasture were in the bud stage, and fatal, quantities of D. burbeyi during summer. Larkspur plants of elongation of flowering racemes had not begun for most plants. all phenological stages were present on the pasture during the Larkspur plants were classified as flowering after elongation of the racemes. After elongation, flowers typically require at least 2 to 3 Pfiiter and Ralphs are range scientists, USDA-ARS Poisonous Plant Research days to open fully. Period 2 (early flower) was from 20 July to 30 Lab., Logan, Utah 84321; Manners is research chemist, Western Regional Research Center, Albany, C&i. 94710; Hong is research associate, Grassland Institute, July. Approximately 5% of larkspur plants had elongated flower- Urumqi, Xmjiang, China; and Lane was graduate research assistant, Dept. of Range ing racemes at the beginning of this period, and by the end of the Science, Utah State Uoivcrsity, Logan. We thank Larry Mickelsen for valuable assistance with the field work, Dr. John period about 80-m of the larkspur plants had elongated OlFn,for surgery, Dr. Dave Clark for laboratory analyses and Dr. Don Sisson for racemes. Period 3 (full flower) was from 5 August to 14 August. statlstlcal advtcc. Dr. Jim Bowns, Dr. George Ruylc and Dr. Walter Majak reviewed an earlier draft, and we thank them for their helpful comments. Most larkspur plants had flowered early in this period, and about Manuscript accepted 25 July 1988. 1% were beginning to set seed. By the end of this trial, about 50%

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 509 of larkspur plants were in the pod stage. Period 4 (pod stage) was ethanol, and later freeze dried and ground through a l-mm screen. from 19 August to 30 August. Over 95% of larkspur plants were in Diet and clipped larkspur samples were analyzed for crude protein the early to late pod stage at the beginning of period 4. When (CP) using a macroKjeldah1 procedure (Harris 1967), and for period 4 ended, some larkspur pods were shattering, and many neutral detergent fiber (NDF) or acid detergent fiber (ADF) using plants were yellow and senescent. the procedures of Goering and Van Soest (1975). The ADF proce- Different contiguous pastures with abundant larkspur popula- dure was substituted for NDF for diet samples because of unresol- tions were used for each period. Pastures varied in size from about vable filtering difficulties. 2 to 3.5 ha, and were enclosed with electric fence. Five yearling Measurements of temperature and relative humidity were made Hereford heifers (318 kg body weight) that had grazed on the area using a continuous recording hygrothermograph. Barometric the previous summer as calves were used to determine larkspur pressure was measured using a fortin type mercurial barometer. consumption. A bite count technique was used as described by Precipitation was measured in a wide-mouth rain gauge to accom- Pfister et al. (1988) to determine relative amounts of tall larkspur odate snow or sleet. Larkspur consumption was correlated with eaten by cattle. temperature, barometric pressure (corrected for gravity and ambient Larkspur plant parts were collected for analysis of total alkaloid temperature), previous 12- and 24hour precipitation, relative concentration (TAC) at the beginning and end of each period. humidity, and temperature-humidity index (THI). THI was calcu- Samples were frozen in the field using dry ice and ethanol, and later lated as (T&,)+(0.55-0.55 RH)(Tab-58)], where Tab is dry bulb freeze dried and ground through a l-mm screen with a Wiley mill. temperature, and RH is relative humidity (Ingraham et al. 1979). TAC was determined using a modification of the procedure of THI has been found useful as an index of stress in cattle (Ingraham Pelletier et al. (1981) as reported by Ptister et al. (1988). et al. 1979). Standing crop was determined at the beginning and end of each period as detailed by Plister et al. (1988), by clipping 30 to 40, .25m Site and Rumca Fill X lm plots. Tall larkspur biomass was estimated using a double The second experiment compared the selectivity of cattle for sampling procedure as described by Pfister et al. (1988). Tall partially shaded larkspur growing within and adjacent (within 10 larkspur biomass was determined once during each period on each m) to a grove of trees (shade site) with larkspur growing in an open, pasture. unshaded area over 100 m from cattle bedding grounds (sun site). Tall larkspur utilization was estimated at the end of periods 2 to The influence of rumen fill was also evaluated using 3 ruminally 4 using 20 to 30, lmr plots along transect lines as detailed in Pfister cannulated Holstein steers (330 kg body weight) that had grazed an et al. (1988). No utilization estimates were made at the end of adjacent area from 20 July as an adaptation period. The Holstein period 1 because no larkspur had been eaten. steers and the heifers selected virtually identical diets (p>o. 1) after The proportion of time the heifers spent grazing in each of the the first 5 days of this adaptation period (Pfister, unpublished dominant vegetation areas (larkspur, grass-forb, currant) was cal- data). This experiment was done during the flower stage (7 to 12 culated based on behavioral observations for each period. During August) and the pod stage (24 to 30 August). focal-animal sampling for bite counts the vegetation area grazed During each trial rumen fill treatments were imposed by remov- and time of entry and departure was noted for each focal animal. ing rumen contents via rumen cannulae. The 0% till treatment Preference for a particular vegetation area (Pi) during active graz- resulted in complete removal of all rumen contents into a large ing times was calculated as P i= log (percentage occupation of metal container. For the 50% fill treatment, all rumen contents vegetation areai/percentage of study area covered by vegetation were removed, and 50% by volume was immediately replaced. areai) (Putman et al. 1987). No preference is indicated by Pi=O; Residual contents from the 50% treatment were used to completely positive or negative numbers mean selection or avoidance of areas, fill the rumen of the 100% treatment animal. Treatments were respectively (Putman et al. 1987). As used here, this index does not imposed each day at 1000, 1030, and 1100, 2 hours before the consider area preferences for activities other than grazing. animals were observed, to allow any physiological feedback mech- Diet quality was determined at the beginning and end of periods anisms to operate. Animals were rotated through the treatments in 1,2, and 4 using 3 esophageally fistulated Hereford heifers that had a systematic fashion during each 6day period. At the designated also grazed this area the previous summer. Diet samples were time, each animal was led to one site and observed for 15 min. collected in pasture areas representative of the tall larkspur vegeta- Steers were secured with a 10 m tether to a steel post, and allowed tion area, currant area, and the grass-forb area. The fiitulated to graze while bite counts were taken. Standing crop measurements heifers were tethered with a 10 m rope to a steel post, and allowed were not taken at the sites, but each site had adequate larkspur and to graze for 30 to 40 min. Heifers were rotated through the 3 other forage for selective grazing. All animals were very tractable sampling areas during the morning feeding period on a random and thoroughly trained in the procedure before trial 1 began. To basis. Diet samples were frozen in the field using dry ice and minimize the possibility of carryover effects (i.e. destruction of microbial populations from cooling and aeration of rumen con-

Table 1. Mean forage wailability (kg/ha) on the experimental pastwa durtng swmer, 1987.

P&Xi 1 2 3’ 4 Jul2 Jul 11 Jul20 Jul30 Aug 5 Aug 14 Aug 19 Aug 30 Graminoids IlW 116a 183a 248a 172 148 236a 131a Taraxacum offiinaie 56a 42a 1Ola 126a 72a Artemisia ludoviciana 137a 176a 245a 2z 108a 96a Lupinus spp. 38a 59a 58a 2Oa 57a 25a Mertensia leonardi 216a 23b 30a Oa 125a Ob Other forbs 237a 274a 368a 354a 852 564 400a 224b Forb subtotal 684 594 775 658 789 417 D. barbeyi 165 676 1291 1533 ‘Data adapted from Lane (1987); no species separation was done except for D. borbqi. *Meansin the same row for the same period followed by the same letter arc not different using a 95% codiience interval.

510 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 tents), 6 to 8 liters of rumen contents from a fourth ruminally Larkspur Utilization cannulated steer were given to the 0% till animal when its rumen No larkspur was consumed during period 1. During periods 2 to contents were restored each day at 1330. No inappetence following 4, 54 to 60% of the flower and pod apexes showed evidence of the treatments was noted. Between the grazing periods the steers grazing. Utilization of leaves (at least one grazed leaf) varied from were kept corralled, and all animals were fed sufficient amounts of 12% of stalks during period 2 to 70% at the end of period 4. alfalfa hay in separate feedings at 1400 and 2000 to allow adlibifum Total Alkaloid Content of Larkspur consumption during the evening. Usually only small amounts of Leaves of tall larkspur were highest in TAC during the early residual feed were present at 0630. growing season (Table 2), but leaf TAC declined rapidly with plant StatistleaI Analysis maturation. Flowers and pods maintained different but relatively All data were analyzed using the Statistical Analysis System constant TAC during the early and latter part of the summer, (SAS 1985). Bite count data were transformed using an arcsine- respectively. Larkspur plant parts grown under partial shade in the square root transformation (Steel and Torrie 1980). To compare tree site were consistently lower in TAC than were parts collected percent of bites of major diet components, animals were consi- from plants grown in full sunlight. dered as blocks, and periods as treatments, with repeated mea- Phenology and Larkspur Consumption surements over days within periods. Chi-square analysis was used Averaged over the summer, larkspur comprised 6.1% of cattle to compare cattle distribution in the currant, grass-forb and lark- diets. The amount of larkspur consumed was different among spur vegetation areas, compared to the proportion of each area in periods (X0.05). Cattle ate no larkspur during period 1, the bud the study pastures. If cattle grazed at random, the number of stage (Table 3). During the early flower stage (period 2), the heifers minutes of observation time in an area (observed values) would be consumed only a trace amount of larkspur buds, always from similar to the total minutes of observation time multiplied by the plants with some flowering racemes. Cattle selected nearly all proportion of the pasture occupied by that vegetation (expected flowers during this early flowering stage. Larkspur consumption values). Changes in nutritive value of cattle diets and forage avail- decreased during the flower stage (period 3) to about 6% (Table 3). ability were evaluated using 95% confidence intervals. To evaluate the influence of site and rumen fill, the 3 ruminally cannulated Table 3. Dietary componente ($ of biter) and bite rate @ita/mh) by Holstein steers were used in a pooled Latin square, with sites, cattle grazing moooWn nogee in central Utah during 1987. animals and rumen fill levels as main effects; days were nested within site, and the trial was repeated during the flower and pod stages of growth. Item 1 2 3 4 Mean P value Results Larkspur bud 0 0.2 2.:b 0 0.05 x.05 Forage and Larkspur Availability flower Oar 6.5b 0.04a 2.4 co.01 Forage was not limiting during any of the periods (Table 1). leaf 1.2a l.la 8.Ob 2.7

Table 2. Total alkaloid concentration (96 dry wt.) of tall hrkepur during Cattle selected nearly equivalent amounts of flowers and pods as 1987. pods were beginning to form. Consumption of larkspur peaked at 10% during period 4, the pod stage. There was a day effect Plant Part (KO.05) and a day by period interaction (X0.05) for consump- Date stem Leaf Bud Flower Pod tion of total larkspur, and larkspur flowers and pods. There was no difference (m.05) between periods for grass, 712187 1.46 3.11 1.48 shrub, or forb consumption (excluding larkspur). Cattle diets were 7; lb/871 1.03 3.09 1.24 about 70% forbs and 23% grass over the summer. Turuxucum 7/10/872 0.59 1.46 1.26 7/21/87 0.41 1.65 0.77 officinule was highly preferred by cattle in each period, as was 7125187 0.41 1.89 0.85 Mertensiu leonurdi when available. Shrubs were a minor dietary 7131187 0.39 1.84 0.87 component. Correlation coefficients (r) between larkspur con- 8/ 10187 0.49 0.98 0.78 1.15 sumption and environmental variables were low; the highest corre- 8/ 12/87 (shade)’ 0.92 1.08 lation (~-0.3 1) was with THI. In a multiple regression analysis, no 8112187 (sun)4 0.96 1.31 subset of variables had a rr value above 0.16. B/19/87 0.39 0.63 0.86 0.96 9/l/87 0.14 0.23 1.02 Cattle Distribution and Diets 9/ 1 / 87 (shade) 1.14 Cattle did not graze randomly in the pastures. The animals 9/l/87 (sun) 1.23 preferentially grazed the currant and larkspur areas, and avoided 1Samplcsfrom tall larkspur plants before elongation of racemes. the grass-forb area (Table 4). Cattle concentrated (KO.001) in the *Samples from tall larkspur plants after elongation of racemes. larkspur areas during period 2 (Pi = .37) and 3 (Pi=JZ). There was a Samples from tall larkspur plants grown in partial shade. marked influence of vegetation area on cattle diets (Table 4). Cattle Samples from tall larkspur plants grow in full sun.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 511 Table 4. CWle distribution ($6 of gruing tme) eomparsd to vc8station area, and percent of bita in ueb vegetation area.

% of bites3 Time %of % of Brazing Sig. Vegetation area area time Pi’ level* Larkspur 0. Forbs Grass Period 1 currant 10 20 .30 0.001 0 81.1 18.2 grass-forb 79 74 -.03 NS 8 70.8 29.2 larkspur 11 6 -.26 0.001 81.3 18.7 Period 2 currant 27 16 -.23 0.001 87.6 7.7 grass-forb 65 66 0.0 NS E 68.1 31.5 larkspur 8 18 .37 0.01 46:9 43.3 9.8 Period 3 currant 16 33 .32 0.001 1.1 73.8 23.7 grass-forb 74 E -.35 0.001 0.4 74.0 25.6 larkspur 10 .52 0.001 19.2 68.7 11.8 Period 4 currant z 35 .13 0.001 16.0 68.0 15.6 grass-forb 44 0.0 NS 3.0 61.7 32.3 larkspur 30 21 -.17 0.001 32.4 55.9 11.2 overall currant 20 .I1 0.001 6.3 76.4 16.7 grass-forb 65 :: -.06 0.001 0.8 z:: 29.8 larkspur 15 18 .09 0.001 24.9 12.5 IA is ptefennce index indicating animal preference for vegetation areas. Values leas than or greater than zero indicate avoidance or selection of an area, respectively. *Significance level of oneaample x2 test comparing the minutes of gmxing time cattle spent in each arca with the total minutes of observation multipled by the %of area; NS = non-sikficant. Way Got total to 100%because shrubs are not included. diets were 25% larkspur on average when grazing in the larkspur areas, compared to 6 and 1% for the currant and grass-forb areas, respectively. Cattle consumed 47% larkspur during period 2 when grazing in larkspur areas. Nutritive Value of Larkspur and Cattle Diets Tall larkspur leaves, buds, flowers and pods contained high levels (12 to 30%) of CP and low levels (12 to 20%) of NDF (Fig. 1). Cattle diets were very nutritious early in summer, but generally declined in CP and increased in ADF levels with forage maturation (Fig. 2). Few differences (X0.05) in dietary CP or ADF content were noted among vegetation areas on the same date. o’...... ‘.....‘...... “.““’.””””..”” Site and Rumen Fill Trial 2 10 21 25 31 10 19 1 6oJULY AUGUST YEPT Larkspur consumption tended to vary with rumen fill (P = 0.15). T The steers on the 0,SO and 100% levels of rumen fill consumed 9.3, 14.6, and 17.3% larkspur, respectively over both periods. There was a site effect (@0.06), with steers taking 16.9 and 10.6% larkspur bites in the shade and sun areas, respectively (Table 5). Peri- ods were not different (ZQO.1). There was no site by rumen fill L JO-- interaction (P=O. 1). n ~. Biting rate decreased (KO.05) with increasing levels of rumen Z 20-m FLOWER ,‘:,‘- POD fill (Table 5). Overall, the 0,50, and 100% fill treatments averaged -( BUD --__A ----> 29.6, 25.4, and 21.2 bites/mm. Biting rate also was significantly __Zt- lo-- l - (KO.05) higher in the sun area (29.2 bites/min) compared to the LEAF shade area (21.6). No period effect (ZQO.l) was noted for biting 0* rate. 2 10 21 25 31 10 19 1 Discussion JULY AUGUST SEPT This study confirms previous observations (Pfiiter et al. 1988) Fig. 1. Crude protein and neutral detergent j?ber (NDF) levels in tall that D. barbeyi is generally unpalatable to cattle before flowering. larkspur during summer, 1987. These findings are in contrast to early reports indicating that cattle selected more tall larkspur before flowering than after (Glover selected young leaves and stem tips of D. barbeyi early in the 1906, Marsh et al. 1929), but these early reports did not state the season, but did not indicate if the racemes had elongated. Knowles Delphinium species. In our studies, D. barbeyi was not consumed (1974) reported cattle deaths from consumption of D. glaucescens by cattle until the flowering racemes began elongation. Consump and D. occi&rt?ufe early in the graxing season, but no phenological tion of flowering parts and leaves began then and increased with stages were specified. plant maturation. Cronin and Nielsen (1979) noted that cattle Between 50 and 60% of the flower and pod apexes were grazed

512 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1966 Table 5. Percent of larkspur bites and biting rate @ites/min) at two sitea = for steers witk different level8 of men fill. p x 9%Larkspur bites Biting rate .z Level of ._0 rumen fill shade sun shade sun (%) site site SE’ site site SE ; 0 12.7 5.9 2.6 25.0 2.1 al 50 14.2 15.1 3.4 21.7 z:, 1.8 .; .- 100SE 23.73.2 10.82.7 4.5 18.01.2 24:01.8 l6 -z u [r 0 ‘SE= Standardermrofthemean. 3 .’ __-- during each period after the bud stage. During this and the pre- s _,J f -v. I I vious summer (Pfister et al. 1988), pastures were stocked at a :IBud ’ Flower Pod comparable level (27 to 33 cattle days/ ha), and larkspur utilization estimates were similar. Logan (1973) found that cattle utilization of Phenological Stage Fig. 3. Hypothetical relationship between Dalatabilitv and toxicitv of D. zo- -35 Grass-Forb Site (A) barbeyi indicating aprimary toxic window when m&t cattle dea&-my be predicted.

ADF __30 D. occidentale in Idaho was highest in late summer at maturity. In the Manti area, stocking rate does not appear to be a contributing s factor to losses from tall larkspur as pastures are normally grazed --25: at a moderate level. CP Alkaloid levels of tall larkspur during 1987 were simiir to 1986 __ 2 A results (Pfiiter et al. 1988). Williams and Cronin (1966) reported 20 comparable TAC in leaves of D. barbeyi. However, TAC of D. 0 occidentale plants were substantially higher than D. barbeyi for both apexes and leaves (Williams and Cronin 1966), even though 0, k15 D. occidentale is 10 times less toxic to rats than D. barbeyi (Olsen Jul 2 Jul 11 Jul21 Jul31 Aug 19 Sept 1 1977b). Data on TAC without knowledge of toxicity of individual Per 1 Per 2 Per 4 alkaloids may be misleading (Olsen 1983). D. occidentale differs from D. barbeyi in presence or absence of several unidentified 20 Larkspur Site (6) mF 35 individual alkaloids (Manners, unpublished data). For example, T ;r T the toxic alkaloid in the tall larkspur D. brownii is methyllycaconitine, a potent neuromuscular blocking agent (Nation et al. 1982) 30 but the plant contains other alkaloids (Aiyar et al. 1979). Cattle tended to consume more larkspur with increasing levels of s? rumen fill. This suggests that even though larkspur is palatable, Kx$.. 25; hungry cattle will generally select other forbs and grasses over A 9 larkpur. This may be due to reduced selectivity by hungry animals. Steers took more larkspur bites in the shade site compared to the It20 sun site. TAC of flowers and pods were slightly higher in the sun plants than in the shaded plants, but these small differences could be due to chance. The influences of proximity to cattle bedding "ol15 Jul 2 Julll Jul21 Jul31 Aug 19 Septl grounds (i.e. N fertilization effect) and partial shading on TAC are unknown. Plant levels of CP and NDF were simiir between the 2 Per 1 Per 2 Per 4 sites (data not given). Louda et al. (1987) reported that insect herbivory of sun-grown D. nelsoni was greater than for shaded 20 Currant Site (C) plants, but no information on TAC was given. T The nutritive value of tall larkspur is high, and cattle can consume large amounts of D. barbeyi with no apparent ill effects. ~15 i During this and a previous summer, cattle ate large quantities of C .- tall larkspur at times, with no death losses among the experimental 3 0 animals. In the areas surrounding the experimental pastures, cattle I lo-- n losses attributed by ranchers to tall larkspur were about 5 to 10%. No fatal episodes of gluttonous consumption (Cronin and Nielsen -0” 1979) were observed, so we still have little information on factors 2 5-- 20 responsible for triggering this feeding behavior. On 2 occasions 0 t A (late July and mid-August), cattle in this study displayed unusual feeding behavior for 15 to 20 min periods by rapidly consuming Oi 15 larkspur, then running from larkspur plant to plant, and occasion- Jul 2 Jul 11 Jul21 Jul31 Aug 19 Septl ally eating plants to ground level. Each episode followed a cold Per 1 Per 2 Per 4 rainshower, and the cattle were playful and active. Increased excit- ability is symptomatic of larkspur toxicosis (Nation et al. 1982), Fig. 2. Crudeprotein (CP) and acid detergent fiber (ADF) levels in diets and increased movement may potent&e toxic effects (J.D. Olsen, selected by esophageallyfistulated cattle in (a)grass-forb sites, (b) larks- personal communication). We suspect that weather-induced chem- pur sites, and(c) currant sites during periods 1.2, and4 corresponding to the bud, early flower, andpod stages of tall larkspur in 1987. ical changes in the plants may be responsible for these alterations in

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 513 feeding behavior. Plant samples taken after the late July incident Knowles, K. 1974. An evaluation of larkspur poisoning in cattle and the were slightly higher in TAC compared to levels several days earlier trampling damage that occurs during grazing on a summer range in (Pflstcr, unpublished data). Keeler and Laycock (1988) noted that eastern Idaho. MS. thesis, Univ. Idaho, Moscow. rapid diurnal fluctuations in levels oftoxinshave been reported for Lana, M.A. 1987. Conditioned taste aversion: potential for reducing cattle other alkaloid-containing plants. losses to larkspur. M.S. Thesis, Utah State Univ., Logan. Logan, J.R. 1973. Evaluation of a specifIcally formulated supplement for these grazing studies of tall larkspur indicate that there is a the prevention of larkspur poisoning. M.S. Thesis, Utah State Univ., primary toxic window, during which D. barbeyi is both palatable Logan. and toxic, due to levels of one or several individual alkaloids. This Louda, SM., P.M. Dixon and NJ. Huntly. 1987. Herbivory in sun versus hypothesis predicts that most cases of fatal consumption will occur shade at a natural meadow-woodland ecotone in the Rocky Mountains. just after flowering, and this period may last for several weeks (Fig. Vegetatio 72141-149. 3). Observations of fatal gluttonous consumption (Cronin and Ma& W., P.D. Parkinson, R.J. Williams, N.E. Looney, and A.L. Van Nielsen 1979) were not seen during our grazing studies, and are Ryswyk. 1977. The effect of light and moistwe on Columbia milkvetch problematic for the toxic window hypothesis. During such epi- toxicity in lodgepole pine forests. J. Range Manage. 303423-427. Marsh C.D., A.B. Clawso~~,and H. Marsh. 1929. Larkspur or poison weed. sodes, cattle ingest fatal quantities of entire plants (Cronin and USDA Bull. 988. Nielsen 1979) when larkspur presumably contains high total alka- Marsh, CD., A.B. Clawson, and H. Marsh. 1916. Larkspur poisoning of loid levels (Williams and Cronin 1966). The plant fraction that livestock. USDA Bull. 365. deters herbivory of immature larkspur is thought to be alkaloids Nation, P.N., M.H. Berm, S.H. Roth, and J.L. Wiikens. 1982. Clinical (Ralphs and Olsen 1987), because of a bitter taste (Bate-Smith signs and studies of the site of action of purified larkspur alkaloid, 1972). If one or several highly toxic alkaloids are not the deterrent methyllycaconitine, administered parenterally to calves. Can. Vet. J. fraction, then reductions in levels of other, less toxic alkaloids (or 233264-266. some other deterrent fraction such as resins or phenols) may trigger Olsen, J.D. 19778. Rat bioassy for estimating toxicity of plant material gluttonous consumption by cattle. For example, although tannin from larkspur (Delphinium sp.). Amer. J. Vet. Res. 38:277-279. Olsen, J.D. 1977b. Toxicity of extract from 3 larkspur species (Delphinium levels have been strongly implicated in chemical defense of birch barbeyi, D. glaucescens, D. occidentale) measured by rat bioassay. J. (Betula resinifera) against herbivory by snowshoe hares (Bryant Range Manage. 30:237-238. and Kuropat 1980), current evidence indicates that other com- Olsen, J.D. 1983. Relationship of relative total alkaloid concentration and pounds are in fact rendering the plant unpalatable (Reichardt et al. toxicity of duncecap larkspur during growth. J. Range Manage. 1984). Bioassays to determine acceptability of chemical extracts of 36550-552. larkspur are currently underway in our laboratory, and may help Peiletier, S.W., and L.H. Keith. 1970. Diterpene alkaloids from Aconitum, to answer some questions about larkspur/ livestock interactions. Delohinium. and Garrva species. u. l-206. In: The Alkaloids, Vol. XII, (R.H.F. Manske ed.). Academic Press, New York. Peiietier, S.W., O.D. Daiiey, Jr., N.V. Mody, and JD. Olsen. 1981. Isola- Literature Cited tion and structure elucidation of alkaloids of Delphinium glaucescens Aiyu, V.N., M.H. Berm, T. Harma, J. Jacyno, S.H. Roth, and J.L. Rybd. J. Org. Chem. 46:3284-3293. Wilkens. 1979. The principal toxin of De$hinium brownii Rydb., and its Pfister, James A., Michael H. Raiphs, and Gary D. Manners. 1988. Cattle mode of action. Experientia351367-1368. grazing tall larkspur on Utah mountain ranges. J. Range Manage. BatcSmith, E.C. 1972. Attractions and repellents in higher animals. p. 41:118-122. 45-56. In: Phytochcmical ecology, (J.B. Harbome cd.). Academic Press, Putman, R.J., R.M. Pratt, J.R. Ekins, and PJ. Edwards. 1987. Food and New York. feeding behaviour of cattle and ponies in the New Forest, Hampshire. J. Bryant, J.P., and P J. Kuropat. 1980. Selection of winter forage by subaitic Appl. Ecol. 24369374. browsing vertebrates: the role of plant chemistry. Annu. Rev. Ecol. Syst. Ralph%, M.H., and J.D. Olsen. 1987. Alkaloids and palatability of poison- 11:261-285. ous plants. p. 78-83. In: Proc-Symposium on plant-herbivore interac- Cronin, E.H., and D.B. Nielsen. 1981. Larkspurs and livestock on the tions. (F.D. Provenra, ED. McArthur, and J.T. Flinders ed.). USDA rangelands of western North America. Down to Earth 37:11-16. For. Ser. Int. Gen. Tech. Rep. INT-222. Gersbenzon, J. 19&o. Changes in the levels of plant secondary metabolites Ralphs, M.H., J.D. Olsen, J.A. Pfister, and G.D. Manners. 1988. Plant- under water and nutrient stress. p. 273-320. In: Recent advances in animal interactions in larkspur poisoning in cattle. J. Anim. Sci. phytochemistry, Vol. 18, (B.N. Timmermann, C. Steelink, and F.A. 66:2334-2342. Loewus eds.), Plenum Press, New York. Rekhardt, P.B., J.P. Bryant, T.P. Clausen, and G.D. Wieland. 1984. Glover, G.H. 1906. Larkspur and other poisonous plants. Colo. Agr. Exp. Defense of winter-dormant Alaska paper birch against snowshoe hares. Sta. Bull. 113. Decologia 6558-59. Garing, H.K., and P.J. Van Soest. 1970. Forage fiber analyses. USDA SAS Institute, Inc. 1985. SAS/ STAT guide for personal computers, Ver- Agr. Hand. 379. sion 6 Edition. Gary, North Carolina. Harris, L.E. 1967. Nutrition research techniques for domestic and wild Steel, R.G.D., and J.H. Torric. 1980. Principles and procedures of statis- animals. Vol. 1. Logan, Utah. tics. 2nd Edition. McGraw-Hill Book Co., New York. Ingraham, R.H., R.W. Stanley, and W.C. Wagner. 1979. Seasonal effects Tortes, F.W. 1980. Animal behavior-a systems approach. John Wiley and of tropical climate on shaded and nonshaded cows as measured by rectal Sons. New York. temperature, adrenal cortex hormones, and milk production. Amer. J. Willia&, M.C., and E.H. Cronin. 1966. Five poisonous range weeds-when Vet. Res. 40:1792-1797. and whv thev are dangerous. J. Range Manage. 19274-279. Keelcr, RF., and W.A. Laycock. 19gg. Use of plant information in man- Wiison, JIR. 1982. Envir%nmental and &tritio&l factors affecting herbage agement decisions. p. 347-362. In: The ecology and economic impact of quality. p. 11I-131. In: Nutritional limitations to animal production poisonous plants on livestock production, (L.F. James, M.H. Ralphs, from pastures, (J.B. Hacker ed.). Commonwealth Agricultural Bureaux, and D.B. Nielsen eds.) Westview Press, Denver, Colo. Famham Royal, United Kingdom.

514 JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 Habitat selection and activity patterns of female mule deer in the Front Range, Colorado

ROLAND C. KUFELD, DAVID C. BOWDEN, AND DONALD L. SCHRUPP

AbStlWt Twenty-two adult, female mule dttr (O&coiltus hemionus measure the relative degree of preference deer exhibit, during hemionus) were radio-collared with activity stnsors and monitortd winter, for certain major Colorado Front Range vegetation types with ground triangulation from mid-Novtmbtr through Much, for feeding and resting; to determine if the degree of preference for 3 years (19824985) in the foothills wtst of Fort Collins, Colo- varies by time of day; and to determine if other environmental rado, to ttst 4 general hypothtsts about habitat s&&on and factors influence deer activity patterns. Such information will be activity: (1) The proportion of time deer spend fttding and resting useful to range and wildlife personnel concerned with habitat varits with time of day. (2) Dttr alter their activity patterns in evaluation and improvement or with mitigation of habitat losses. response to environmental influences. (3) Selection of sptciflc In the study we examined 4 general hypotheses. (1) The propor- vegetation typts for feeding and rtsting varies with time of day. (4) tion of time deer spend feeding and resting varies among 4 daily Ecotonts are preferred habitats. Dar were monitored during 6-hr periods: sunrise, daytime, sunset, and night. (2) Deer alter their sampMng ptriods: sunrist, daytime, sunset, ad night. Deer ftd activity patterns in response to environmental influences; specifi- most during suns&, night, and sunrist periods and ltast during the tally, snow depth, temperature, wind velocity, cloud cover, and day. Fttding occupied similar proportions of an average deer’s moon phase. (3) Deer select specific vegetation types for feeding time during sunstt, night, and sunrise periods. They preferred the and resting, and selectivity varies among 4 daily periods: sunrise, grassland typt for fttding and resting at night and the mount&n daytime, sunset, and night. (4) Deer select vegetation ecotones for mahogany (Cercocarpusmontanus) type for both activities during feeding and resting. Within these 4 genera1 hypotheses, we exam- all other periods. Preference dttr showtd for the pondtrosa pint ined hypotheses for given comparisons among activity, daily (Pinuspondffosa) type for feeding activity was inversely related to period, or environmental factor combinations. canopy cover. Deer rested most during daytime and night periods. Study Area During ptriods of daylight, dttr using the grassland typt showtd preference for tcotonts with certain types offering tscapt cover. The study area, approximately 14.5 kmz, is about 5 km west of No such preference was observed at night. Deer ftd ltss and rested Fort Collins, Colorado, and lies mainly within the boundaries of more when snow depth txcttdtd 36 cm. No significant differences Lory State Park. It is bounded on the east by Horsetooth Reser- (I90.05) in the proportion of time deer devoted to feeding were voir, on the west by the ridge extending north from Horsetooth found in the following comparisons: clear versus cloudy full-moon Mountain, and on the south by the southern park boundary. nighhts(-50 vs. + 50% cloud cover), full-moon versus new-moon, Elevations range from 1,646 m at the reservoir shoreline to 2,138 m low versus high wind spttds (O-32 vs. 32-56 km/hr), and warm on Horsetooth Mountain over a linear distance of about 2.5 km. versus cold ttmptraturts (+18 to -15 vs. -15 to -23“ C). No Steep, rugged, uplifted hogbacks capped by vertical rock outcrops significant relationships were found for the same comparisons in occur along the western shore of the reservoir. These are covered proportion of time devoted to resting. by dense stands of true mountain mahogany (Cercocarpus monta- nus) interspersed with grassland (Bromus secalinus and Stipa spp.) Key Words: Odocoikus hemionus ha&onus, winter habitat prtf- openings and small patches of ponderosa pine (Pinusponderosa). trtnct, daily activity patterns, radio-teltmttry, bthavior, home West of the hogback lies an open valley about 0.6 km wide running range north and south the length of the area. It is mainly grassland with Habitat for mule deer (Odocoileus hemionus hemionus) along very dense thickets of sumac (Rhus aromatica), hawthorn (Cratae- much of the Front Range of Colorado is diminishing due to gus erythropoda), chokecherry (Prunus virginiana), and wild plum expansion of nearby cities and demand for diversion of habitat (Prunus americana) along draws. Horsetooth Mountain is an area areas to human uses. Because decision on land use may have great of rugged, mountainous terrain with numerous rock outcrops, impact on mule deer habitat, better understanding of deer-habitat ridges, and canyons. Lower portions and some south-facing slopes relationships along the Front Range is needed to maintain the deer in higher sections support dense mountain mahogany with patches population. of ponderosa pine and grass. Higher portions are covered by The tendency of mule deer and Columbian blacktailed deer (0. extensive stands of ponderosa pine with varying canopy coverages h. columbianus) to exhibit preference for certain habitat compo- interspersed with Douglas fir (Pseudotsuga menziesii). mountain nents within existing vegetation complexes has been described meadows dominated by Poa spp. and Thermopsis divaricarpa. and (Mackie 1970, Telfer 1978, Barrett 1982, Hanley 1984, Harestad small grassland parks. 1985, Carson and Peek 1987). The purpose of this study is to Materials and Methods Authors are wildlife researcher, Colorado Division of Wildlife, 317 W. Prospect, Fort Collins 80526; professor, Statistics Department, Colorado State University, Fort Radio-telemetry involving a 2-wavelength, precision-null antenna Collins 80523; and wildlife program specialist, Colorado Division of Wildlife, 6060 Broadway, Denver 80216. _ _ system was used to measure deer habitat selection and activity Research was funded by Federal Aid in Fish and Wiidlie Restoration, Colorado patterns. Immediately preceding the study, performance of the Project FW26P. We thank S. McHugh, B. Parmenter, K. Risenhoover, L. Roberts, precision-null antenna system was evaluated on the study area. A and D. Schaad for field assistance; C. Smith, 0. Landahl, P. Strong, and N. Tmg for computer worlr; L. Lovett for assistance in manuscript preparation, and R. Morris, C. description of the system, its directional accuracy on this study Bergman: and B. Mayer of Colorado Division of Parks and Recreation for their area, and procedures used in locating transmitters with it are cooperation in making the study area available. We also thank D. Radtke and K. Davis of the U.S. Fish and Wildlife Service Western Energy and Land Use Team for described by Kufeld et al. (1987). use of computer facilities; and A. Anderson, N.T. Hobbs, and W. Hoffman for Each radio collar used in the study (radio collars were manufac- manuscript review. Manuscript accepted 29 June 1988. tured by Telonics, Inc., of Mesa, Ariz.1) was equipped with an 1Reference to this trade name does not imply endorsement by the State of Colorado.

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 515 activity switch which permitted monitoring whether a deer was in a deer does when not actively feeding or resting. The “other” head up or head down position by instantly changing pulse rate category included walking, running, intermittent feeding, and when the head moved above or below horizontal. Signals were standing. A deer is immobile when standing, but standing could be transmitted to a receiverdigital processor and recorded on a Rus- differentiated from “resting” because resting deer were usually traci, model 388, dual channel, 12-V strip chart recorder. One side immobile for more than 5 minutes. Each time a deer changed its of the chart indicated head position, while the other side simul- activity, a label was placed on the chart to identify the activity. A taneously recorded signal strength in decibels, indicating whether 32-hr record of charted, known activities was assembled. Both the deer was active or inactive. The chart was calibrated in 2- observers watched the deer and studied the charts until they felt minute intervals for time determination. they had attained proficiency in recognizing activity patterns by A preliminary evaluation was also conducted to determine how reading charts. to accurately recognize deer activities from patterns plotted on Tests were then conducted to determine the degree of accuracy in strip charts. A tame deer (4 different individuals were used) wear- recognizing activity patterns on strip charts attained by the ing a radio collar was allowed to roam freely on the study area. Its observer who would be the chart reader during the course of the movements were recorded using telemetry equipment and a strip study (Observer A). One observer (Observer B) accompanied the chart recorder carried by 2 observers who accompanied the deer. instrumented deer and labelled feeding, resting, and other activities Three activity categories (“feeding, resting, and other”) were rec- on his chart. Observer A simultaneously recorded the deer’s activ- ognized. We define “feeding”as actively taking in food (either head ity using a duplicate set of telemetry equipment and chart recorder, up or head down), “resting”as immobile, and “other”as everything but from a position where he could not see the deer or Observer B.

Table 1. Proportion of time deer devoted to feeding, resting, and other activities during 4 daily periods at Lory State Park.1

% of time devoted to activity Feeding Resting Other Total deer Daily period Mean cv Mean cv Mean cv locations Sunrise 36.94 6.74% 33.82 5.85% 29.25 10.29% 1,043 Daytime 23.01 9.86 39.63 5.30 37.36 6.53 1,037 Sunset 39.48 7.24 29.36 6.71 31.16 8.02 1,034 Night 39.98 6.64 38.78 5.39 22.23 10.80 1,041 Calculated signiticance of paired-r tests2 Sunrise vs. daytime 0.00+ 0.01* 0&O* Sunrise vs. sunset 0.30 0.07 0.41 Sunrise vs. night 0.42 0.04, 0.00’ Daytime vs. sunset 0.00+ 0.01* Daytime vs. night 0.00* z 0.00+ Sunset vs. night 0.86 o:OO* 0.00+ ‘Basedon 22 deer.AU 22 were monitoredfor 1winter; 20 were monitored for 2 winters and 13 were monitored for 3 winters. zComparison of tncans. *Denotes a significant differcncc at crlo.05 for individual comparisons. If a simultaneous significance level rrlo.05 for all 6 comparisons within a given activity is dcsind, then the Booferroni inequality requires individual comparisons to bc significant at 0.008 for simultaneous significance level to be 0.05.

Table 2. Effects of various environmental factors on daily activity patterns of deer.*

Percent of time devoted to activity No. of deer Environmental factor Criteria Feeding cv Resting cv in test Snow depth over 36 cm 25.4 18.1% 53.4 6.3% under 36 cm 33.9 7.2 34.5 4.0 Calculated significance 0.03* o.OO* 16 Temperature above -15O C 34.6 6.5 32.0 4.9 below -150 c 32.7 15.3 35.7 11.2 Calculated significance 0.36 0.16 12 Wind over 32 km/ hr 33.1 14.5 37.7 13.6 under 32 km/ hr 32.6 8.9 32.4 4.9 Calculated significance 0.92 0.33 12 Moon phase full moon 9.7 41.7 9.8 new moon :::, 8.1 37.5 7.8 Calculated significance 0.96 0.28 16 Cloud cover over 50% 38.0 11.4 41.1 13.9 under 50% 40.4 16.8 26.1 13.5 Calculated significance 0.71 0.08 10

lComparisonsfor snow depth,temperature, and wind include all daily periods and other environmental factors. Comparisons for moon phase were made during night periods when snow depth, temperature, and wind were under 36 cm, above -I 5OC and under 32 km/ hr, respectively, and cloud cover was under 50%.Comparisons for cloud cover were also made at night when snow depth, temperature, and wind conditions were the same as for moon phase comparisons and when the moon was full. *Denotes a significant diffennces at aSO.05.

516 JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 Observer A then reviewed his chart and labelled patterns he The study was based on 22 adult female deer captured and fitted believed indicated feeding, resting, or other activity. The number with 2 numbered ear tags and radio collars. Ten of these deer were of minutes in time segments Observer A assigned to a given activity instrumented between 19 January and 14 October 1982, and 12 was compared with the actual time devoted to those activities as instrumented between 31 January and 3 February 1983. Beginning recorded on Observer B’s chart. During 2 tests which covered a in 1982, locations of does were monitored extensively from mid- total of 2,239 minutes of recorded deer activity, Observer A cor- November through March for 3 winters by triangulation using 2, rectly classified 95 and 98% of minutes deer devoted to feeding, 99 2-wavelength precision-null antenna systems mounted on pickup and 9% of minutes devoted to resting, and 95 and 95% of minutes campers. Mortality during the 3-yr duration of the study reduced devoted to other activities. The key to identification of feeding, the number of instrumented deer available for monitoring from 22 resting, and other activities lay in recognizing chart patterns pro- during year 1 to 20 and 13 during years 2 and 3, respectively. A total duced over varying lengths of time. A lo-minute period of charting of 4,365 locations was recorded. was deemed adequate to establish a deer’s primary mode of activity Triangulation to locate instrumented deer was accomplished at a particular time. Gillingham and Bunnell (1985) reported less from 2 sites (receiver points) on the east side of Horsetooth Reser- success in recognizing deer activity patterns using tip-switch voir. Preliminary tests (Kufeld et al. 1987) indicated the 2- collars. wavelength precision-null antenna system permitted detecting

Table 3. Preference or lack of preference for vegetetioa typee aa feeding and resting arem during 4 daily periode bawd on 22 deer.

Calculated significance of vegetation type composition in home Composition of Composition of vegetation types ranges vs. Deer preference for vegetation types in in error polygons error polygons vegetation types2 home ranges While feeding While resting While While For For Vegetation type’ Mean% CV Mean% CV Mean % CV feeding resting feeding resting Sunrise Period’ Mtn. Mahogany 33.02 7.47 43.79 8.52% 47.40 6.07% 0.03 0.00 Pt=f Pref Grassland 36.74 5.04 38.86 7.69 35.14 4.77 0.48 0.48 neutral neutral Pine lo-39 1.27 17.78 1.13 30.00 1.23 32.23 0.67 0.91 neutral neutral Pine 40-69 10.40 14.05 3.95 26.36 4.83 24.92 0.00 0.00 unpref unpref Pine 70-100 7.26 22.40 2.24 27.30 3.03 22.38 0.01 0.02 unpref unpref SU-HA-PRU 9.32 4.57 9.60 14.78 7.89 11.57 0.85 0.20 neutral neutral Mtn. Meadow 0.58 30.31 0.15 34.65 0.20 29.84 0.31 0.59 neutral neutral Rock Outcrop 0.99 37.60 0.23 41.62 0.24 56.17 0.06 0.04 neutral unpref TOTAL 99.58 99.95 99.96 Daytime Period3 Mtn. Mahogany 33.02 7.47 50.89 8.34 53.72 7.63 0.00 0.00 pref Pref Grassland 36.74 5.04 28.12 10.02 25.93 10.78 0.03 0.00 unpref unpref Pine 10-39 1.27 17.78 1.96 29.66 1.45 25.58 0.25 0.62 neutral neutral Pine 40-69 10.40 14.05 7.31 20.33 8.13 17.75 0.10 0.21 neutral neutral Pine 70-100 7.26 22.40 5.25 26.33 4.88 22.17 0.36 0.21 neutral neutral SU-HA-PRU 9.32 4.57 5.62 10.09 5.19 11.22 0.00 0.00 unpref unpref Mtn. Meadow 0.58 30.31 0.16 48.11 0.22 35.37 0.04 0.09 unpref neutral Rock Outcrop 0.99 37.60 0.67 53.21 0.41 52.19 0.50 0.11 neutral neutral TOTAL 99.58 99.98 99.93 Sunset Period’ Mtn. Mahogany 33.02 7.47 45.51 8.09 44.40 6.77 0.01 0.00 Pref Pref Grassland 36.74 5.04 35.47 6.45 34.91 5.78 0.63 0.49 neutral neutral Pine lo-39 1.27 17.78 1.41 40.93 1.46 27.79 0.80 0.66 neutral neutral Pine 40-69 10.40 14.05 5.61 23.16 5.72 22.29 0.14 0.18 neutral neutral Pine 70-100 7.26 22.40 3.45 26.83 4.03 25.98 0.05 0.11 unpref. neutral SU-HA-PRU 9.32 4.57 7.52 9.37 8.95 11.08 0.03 0.75 unpref neutral Mtn. meadow 0.58 30.31 0.19 51.96 0.20 41.12 0.03 0.06 unpref neutral Rock Outcrop 0.99 37.60 0.30 57.94 0.22 48.17 0.06 0.06 neutral neutral TOTAL 99.58 99.46 99.89 Nighttime Period’ Mtn. Mahogany 33.02 7.47 31.84 10.27 36.39 7.16 0.79 0.30 neutral neutral Grassland 36.74 5.04 51.10 5.80 46.08 6.11 0.08 0.01 Pref Pref Pine lo-39 1.27 17.78 0.46 60.34 0.44 37.67 0.23 0.00 neutral unpref Pine 40-69 10.40 14.05 3.00 29.40 3.59 32.68 0.08 0.08 unpref unpref Pine 70-100 7.26 22.40 1.81 34.29 3.08 27.37 0.00 0.01 unpref unpref SU-HA-PRU 9.32 4.57 10.81 11.40 9.49 9.95 0.25 0.88 neutral neutral Mtn Meadow 0.58 30.31 0.15 35.88 0.14 38.33 0.03 0.01 unpref unpref Rock Outcrop 0.99 37.60 0.11 47.64 0.18 58.37 0.02 0.02 unpref unpref TOTAL 99.58 99.28 99.31 ‘Pine lo-39,4&69 and 70-100 represent percent canopy covcrap, SWHA-PRU is the sumac-hawthorn-Prunu type. zMean percent composition of a grven vegetation type m error polygons, while deer were engaged in a particular activity, was significantly higher (KO.05) than mean percent confposition of that vegetation type in minimum convex polygon home ranges if thattype was prcfcrrcd by deer for that activity, not significantly different (F70.05) ifprcfemnce for It was neutral, and significantly lower (pIo.05) if it was unprefcrred. ‘A sampling period lasted 6 hrs and extended from 3 hrs before to 3 hrs after sunrise, midday, sunset, midnight.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1968 517 directionality of radio signals on the study area with sufficient ual polygons contained the actual deer location were generated. accuracy to allow triangulation from only 2 sites. The relatively Rationale and methodology for delineation of error polygons small sizes of deer home ranges we estimated using this antenna using nonparametric tolerance limits are described by Kufeld et al. system and procedures described by Kufeld et al. (1987) reflects a (1987). Areas and percent composition of each vegetation type high degree of telemetry accuracy. If deer home ranges were, in within error polygons were computed using various programs with actuality, relatively small but telemetry accuracy was poor, our the SAGIS system created by the U.S. Fish and Wildlife Service estimates of deer home range size would have been expected to be Western Energy and Land Use Team. Habitat usage was based on much larger. Antenna attitude was positioned and calibrated at the percent composition of each vegetation type occurring in error start of each monitoring session by orienting the antenna toward a polygons as described by Kufeld et al. (1987). fixed beacon transmitter located atop Horsetooth Mountain. A Deer preference for habitat types was determined from paired compass rose on the mast of each antenna was then set to coincide observations of mean percent composition of a given vegetation with the surveyed bearing from its receiver point to the beacon. type within the minimum convex polygon home range (Jennrich Operators communicated via 2-way radios to facilitate obtaining and Turner 1969, Southwood 1966) with mean percent composi- simultaneous directional bearings on each instrumented deer. tion of that vegetation type in telemetry error polygons for each Instrumented deer were located during each of 4 daily sampling instrumented deer. Habitat preference comparisons were made for periods; sunrise, daytime, sunset, and night. A sampling period various activities by paired-r test and results labelled in the follow- lasted 6 hrs and extended from 3 hrs before to 3 hrs after sunrise, ing manner. If percent composition of vegetation type A in error mid-day, sunset, and midnight. Since there were more hours of polygons, while deer were feeding, was not significantly different darkness and fewer of light during winter, there was some overlap (calculated 130.05) from percent composition within their corres- between the sunrise and daytime period and daytime and sunset ponding home ranges, preference for type A for feeding is labelled period. Mean overlap was 62, 81, 72, 39, and 2 minutes, respec- as neutral. If mean percent composition of type A in error polygons tively, during November December, January, February, and significantly exceeded (calculated m.05) corresponding mean March. There was no overlap between sunset and night or night home range composition, preference for type A is labelled as and sunrise periods. We assumed that if differences occurred preferred. This test only examines whether averaged over deer between periods in deer activity modes, they would be most pro- there is a habitat preference and not whether an individual deer has nounced at times near period mid-points (i.e., the times closest to a preference. Type A is labelled as unpreferred if percent composi- sunrise vs. times closest to midday), and that deer activity modes tion of type A in error polygons is significantly less (calculated during 2 adjacent periods near the end of one period and beginning m.05) than corresponding home range composition. of another would be similar. Thus, monitoring of deer activity To measure possible affinity of deer to ecotones, mean distances during proportionately short overlap periods was not expected to from deer-selected locations to the edge of the nearest adjacent adversely affect inter-period activity comparisons, and any detri- vegetation type were compared with mean distances for locations mental effects that might be realized would be outweighed by plotted at random. One-hundred locations were plotted randomly consistency provided by maintaining 4 periods of equal (6 hrs) within the minimum convex polygon home range of each instru- length. Only one period was scheduled for deer monitoring per 24 mented deer for comparison with deer-selected locations. We hrs. Twelve sampling periods were scheduled per month (3 of each labelled preference for an ecotone as neutral if there was no differ- kind) except that 8 periods were scheduled during December (2 of ence (lXt.05) between mean deer selected and randomly selected each kind). Thus, during the 3 winters, there were 168 sampling distances. If a difference was found (B.05) and mean deer periods including 42 of each kind. Each deer was located 1 to 3 selected distances were less than randomly selected distances, (a times during a sampling period. At least 1 location of each deer ratio of l.O). Preliminary teleme- study and not with their activity or location. try accuracy tests (Kufeld et al. 1987) indicated that no consistent At the beginning of each monitoring period, the first deer to be bias should be present in estimated deer locations compared with located was selected by obtaining a random starting point on an actual deer locations. Thus, the effect on statistical tests due to ordered list of deer identifications and then proceeding down the using estimated locations rather than actual locations is to reduce list from the starting point. Immediately after a deer was located, power of the tests; yet many significant differences were still its activity pattern was charted for 10 minutes on a strip chart detected. Only 3 vegetation types, grassland, mountain mahogany, recorder. Activity patterns were categorized according to feeding, and SU-HA-PRU, were used for ecotone comparisons because resting, and other, based on the activity which occupied most of the sample size for them was adequate, but inadequate for other vege- deer’s time during the 10 minutes. tation types. Comparisons were made without regard for activity. A 1:14,049 scale, high resolution, color infrared aerial photo For statistical analysis, deer were viewed as the basicexperimen- taken by NASA from a U-2 plane at an altitude of 21,336 m formed tal unit. Thus, standard errors were based on differences among the basis for analysis of telemetry data. Distances and angles on individual deer means. For each observation, the percentage com- this photo were compared with surveyed distances and angles on position of each vegetation type in the constructed error polygon the ground. There was no apparent photo distortion. The study and deer activity were recorded. Sample means of observations for area was divided into 9 vegetation types, and boundaries of each each vegetation type for each deer within 36 categories (3 years by 4 type were digitized (Gilmer et al. 1973) so they could be regenerated periods by activities) were calculated for those deer which were in a map form by computer. A minimum size of 0.4 ha was present all 3 years. Analysis of variance techniques for these established for considering an area as a distinct type. The 9 vegeta- repeated measures data were used to test for year effects (main tion types were: 3 canopy coverages of ponderosa pine (10 to 39,40 effects and interactions). Since few significant results were found to 69, and 70 to lOO$?Q,grassland, mountain mahogany, sumac- among years for the numerous tests made, sample means for 12 hawthorn-Pnmus (SU-HA-PRU), mountain meadow, agricultu- period by activity categories were calculated for each vegetation ral (only 1 small alfalfa field) and rock outcrop. Rock outcrop was type for each of 22 deer with pooled data for all 3 yrs. In addition, considered a distinct type because areas of exposed rock with very the proportion of observations within activity types were calcu- sparse growths of ponderosa pine exceeding 0.4 ha were encountered. lated for each deer by feeding, resting, and other activity periods. Triangulation data were computerized and error polygons Paired-t tests were used to test for differences between periods in which one could be 90% confident that at least 81% of the individ- proportion of time spent feeding for each of the 6 pairs of periods

518 JOURNAL OF RANGE MANAGEMENT 41(8), November 1988 possible, given 4 periods. Control of the error rate for these com- increases greatly in snow depths exceeding 25 cm (Parker et al. parisons among periods was maintained by use of Bonferroni 1984). Loveless (1967) and Gilbert et al. (1970) reported that snow inequality (Sokal and Rohlf 1981). Similar comparison among depths of 46 and 61 cm, respectively, essentially preclude use of an periods were made for resting and other activities. This same set of area by mule deer. In our study when snow depths receded below tests between pairs of periods using paired-t tests was made for the 36 cm, deer resumed feeding and resting in the same proportions as percentage composition of a given vegetation type within error before the snowfall occurred. Study deer did not appear to be polygon data for a given activity. stressed by excessive snow. November-March precipitation during Percent composition of a given vegetation type within each the 3 winters was near normal, and periods when snow depth deer’s home range was also subtracted from the corresponding exceeded 36 cm did not last more than a week at a time. No vegetation sample mean for each of the 12 periods by activity outward signs of weight loss or poor physical condition were categories. These differences were then used in individual t tests for observed in deer on the area at the end of any winter. each vegetation type for a given period and activity to test equality Air temperatures and wind velocities encountered during 3 win- of the mean percentage composition of a given vegetation type ters of monitoring deer did not appear to influence the proportion within error polygons to the corresponding home range percentage of time deer fed or rested (Table 2). Our lowest recorded air figure. These last tests form the basis for determination of prefer- temperature (-23O C) is below the -20” C operative temperature ence. Difference between mean distance of locations, within a Parker and Robbins (1984) found to be the lower threshold for given vegetation type, selected by each deer and randomly selected thermally critical environment for mule deer during winter. The locations within that deer’s home range, were used in t tests to operative temperature is that temperature actually experienced by evaluate uniform utilization of area within a given vegetation type. the animal, and it can be derived from different environmental Several environmental factors were measured on the study area variables (Parker and Robbins 1984). Since the minimum air during each monitoring session. Snow depths were measured at 6 temperature occurred for only a few hours and since the operative stations representing a north- and south-facing slope at low, mid- temperature for individual deer may have been higher during that dle, and high elevations within the study area. Temperature, wind time because deer can take advantage of microclimates (Loveless velocity, and percent cloud cover were recorded at one of the 1967), operative temperatures likely remained above thermally triangulation points at the beginning, during, and end of each critical levels during the study. Generally, low temperatures and monitoring session. Moon phase was also recorded. Paired-f tests high winds did not occur simultaneously, so the effect of the lowest were used to test for differences in percent time feeding (and/or temperatures on deer was not compounded by wind chill. We resting) on monitoring days when an environmental factor of surmise that deer along foothills of the northern Front Range are interest was above a specified level with percent time spent feeding adapted to temperatures and wind velocities that fall within the (and/ or resting) on days when the environmental factor was below normal range for winter, and do not alter their feeding and resting that level. Because of mortality of some instrumented deer during patterns because of them. Lower temperatures and higher winds the study, some deer had inadequate sample sizes to measure each than we encountered (those that would exceed normal ranges for level of an environmental factor. Only deer with adequate sample the area) or low temperature-high wind conditions may influence sizes for each level of an environmental factor were used in these feeding and resting behavior, however. analyses. Inadequate sample sizes were encountered when more There was no difference between clear, full-moon nights and than 2 levels of an environmental factor were used. (In all statistical cloudy, full-moon nights or between clear, full-moon nights and tests, differences are stated to exist only if they were found signifl- clear, new-moon nights in proportion of time deer spent feeding or cant at KO.05). resting(Table 2). This finding is contrary to the popular belief held by some hunters that deer hunting is more difficult when it is clear ReauIts and Discuaaion and the moon is full because deer feed at night. The instrumented DalIy Aetlvlty Patterns and EnvIronmentaI Influences deer spent a large proportion of their time feeding at night regard- Pronortions of time devoted to feeding (Table 1) are somewhat less of cloud cover or moon phase. conse&ative and proportions of time d&oted to “other” activities HabItat Sekction are somewhat liberal because a deer was considered to be feeding It should be clear that expression of preference (unpreferred, when its main activity mode during a lo-minute period was ingest- preferred, or neutral) by deer for a given vegetation type is only ing food. Intermittent feeding while the main activity mode was made conditional on the choices available in the existing habitat walking, searching for food, ruminating while standing, etc., was complex. Deer used unpreferred types, but less than preferred or included in the “other” activities category. Differences in propor- neutral types, and there were fewer locations (m.05) in unpre- tion of time female deer spent in given activities during specified ferred types than if deer locations within home ranges were ran- periods (general hypothesis number 1) were found in some compar- domly selected. If those particular vegetation types for which deer isons (Table 1). They fed most during sunset, night, and sunrise showed preference had been absent from the study area, then periods and least during the day. Feeding occupied similar propor- neutral types may have become the preferred types and unpre- tions of an average deer’s time during sunset, night, and sunrise ferred types may have been elevated by deer to neutral status. periods. Deer rested most during daytime and night periods. Minimum convex polygon home ranges, constructed from all Precipitation amounts, temperatures, and wind velocities we locations for each deer, averaged 217 ha f 22 ha (Z f 95% confi- encountered during 3 winters of monitoring deer were within nor- dence limit) and ranged from 117 to 323 ha. Yearly minimum mal range for the area. According to long-term U.S. Weather convex polygon home ranges for individual deer overlapped exten- Bureau records for Fort Collins, November through March mean sively. Home ranges were, thus, quite small and fixed over time precipitation during the 3 winters departed from normal by only considering the mobility of which a deer is capable. 4.8 cm, 4.8 cm, and +0.3 cm, respectively, Winds also remained Sixes of error polygons (in which one could be 90% confident within normal ranges for the area. that at least 81% of the individual polygons contained the actual Deer altered their activity patterns in response to some environ- deer location) generated during the study and used for habitat mental influences (Table 2) (general hypothesis number 2). Observed preference determination (as described under “Methods”) were increased resting and decreased feeding activity by deer during distributed as follows: 20.3% - 0.1 to 2 ha; 28.5% - 2.1 to 4 ha; 17% periods of deep snow (>36 cm) (Table 2) was probably for pur- -4.1 to6ha; 12.5%-6.1 to8ha; 10.3%-8.1 to lOha,3.0%- 10.1 to poses of energy conservation. Energy expenditures for locomotion 12 ha; 3.3% - 12.1 to 14 ha, 5.1% > 14 ha. in snow have been shown to increase curvilinearly as a function of Deer selected specific vegetation types for feeding and resting snow depth and density, and net energetic cost of travel to deer and selectivity varied among 4 daily periods (general hypothesis number 3). The most heavily used vegetation types for feeding and

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 519 resting were the mountain mahogany and grassland types. These assume that our instrumented deer used cover primarily for secur- were also the most common types in terms of mean percent of area ity reasons. Sixty-two percent of the area within home ranges of 22 within home ranges (Table 3). instrumented deer consisted of what we call cover types and the Composition of Cover in Home Ranges remaining 38% consisted of noncover types (Table 3). Vegetation types on the study area were categorized according to Preference for Vegetation Types “cover” types and “noncover” types based on how instrumented A range area may contain a variety of vegetation types which deer related to them during the daytime period when, because of individually contain some, or perhaps all, of the habitat needs of high light conditions, deer must rely primarily on vegetation for deer. If an individual type is capable of satisfying all of a deer’s concealment and security. Vegetation types that deer preferred or requirements, habitat diversity may not be necessary. Where types showed neutral preference for feeding or resting during the day- exist that can provide only some of a deer’s requirements or can time period (Table 3) apparently offer adequate security for deer so provide them only during specific time periods, it becomes neces- they were designated as cover types. Those considered cover types sary that a complex of types, which together can satisfy all of a are mountain mahogany, SU-HA-PRU, pine 10-39, pine 40-69, deer’s needs, be present within the home range of each deer. For pine 70-100, and rock outcrop. The SU-HA-PRU type was unpre- example, we found composition of what we defined as cover and ferred by deer for feeding and resting during daytime, but is noncover types within deer home ranges to be 62% and 38%, included as a cover type because it does provide excellent conceal- respectively, but a more productive ratio might IX 100% cover and ment, and because deer preferred the grassland-SU-HA-PRU 0% noncover as long as the cover types also satisfied all of the deer’s ecotone when they were in the grassland type during the daytime needs for feeding. period. The grassland and mountain meadow types were consi- Observed preferences or lack of preference by our instrumented dered noncover types because they were unpreferred by deer for deer for specific vegetation types on the study area appear related feeding and resting during the daytime period. Thus, those types to foliage and its use by deer for feeding and cover. Loveless (1967), apparently do not offer adequate security for deer during daylight working on nearby deer ranges and in similar habitats, also hours. observed that deer were not randomly dispersed over the range but Our definition of cover did not consider insulating qualities concentrated in habitats which provided food and cover. which deer might seek for protection from heat or cold. Freddy Contrary to the hypothesis of Hanley (1982) that deer should (1985) found the absence of thermal cover did not appear to reduce select dicotdominated habitats for feeding, we found that grass- potential for deer to survive winter. Since winter temperatures on land was a preferred feeding type for deer (Table 3). This can be hi study area (Kremmling, Colo.) averaged far lower than ours, we

Table 4. Comparison of distances from deer selected versus randomly selected locations to the edge of the neueat adjacent vegetation type.

Calculated Distance (m) from location to edge of Ratio of significance deer selected No. of instru- nearest adjacent vegetation type deer Vegetation type’ to random mented deer Deer selected location Random location2 selected Type containing Adjacent type nearest to location vs. random averaged Mean cv location location Mean cv distance3 distance_ Sunrise Period’ Grassland Mtn. mahogany 22 22 9% 30 6% 0.72 0.00* Grassland SU-HA-PRU 21 37 8 52 6 0.72 0.00* Mtn. mahogany Grassland 22 2 10 42 7 1.37 0.01* Mtn. mahogany SWHA-PRU 17 IS 32 10 1.05 0.72 SU-HA-PRU Grassland 17 18 19 15 8 1.23 0.36 SU-HA-PRU Mtn. mahogany 11 13 18 11 17 1.20 0.53 Daytime Period4 Grassland Mtn. mahogany 20 22 12 30 6 0.73 0.05. Grassland SU-HA-PRu 19 40 16 51 7 0.78 0.04* Mtn. mahogany Grassland 22 65 12 42 6 1.55 0.00* Mtn. mahogany SU-HA-PRU 19 30 14 35 12 0.86 SU-HA-PRU Grassland I1 20 20 16 8 1.27 z.2 SU-HA-PRU Mtn. mahogany 11 13 18 14 I5 0.91 0173 Sunset Period4 Grassland Mtn. mahogany 22 21 10 30 6 0.71 0.00* Grassland SU-HA-PRU 22 40 11 52 6 0.77 0.02. Mtn. mahogany Grassland 57 10 43 6 1.34 0.01+ Mtn. mahogany SU-HA-PRU 35 12 34 11 1.03 0.87 SU-HA-PRU Grassland 15 12 15 12 1.02 0.95 SU-HA-PRU Mtn. mahogany 12 19 I1 18 1.09 0.75 Night Period4 Grassland Mtn. mahogany 36 11 30 6 1.21 0.07 Grassland SU-HA-PRU 49 7 52 6 0.95 0.44 Mtn. mahogany Grassland 21 57 11 43 6 1.32 0.03* Mtn. mahogany SWHA-PRU 18 37 19 35 13 1.04 0:84 SU-HA-PRU Grassland 17 11 I5 7 1.18 0.27 SU-HA-PRU Mtn. mahogany 12 19 10 19 1.18 0.52 1SU-HA-PRU is the sumac-hawthorn-Phmu type. *One-hundred locations were plotted randomlywithin the minimum convex polygon home range of each instrumented deer for comparison with deer-s&&xl locations. ‘Ratios less than 1.0 suggest preference for cwtoncs. Ratios greater than 1.0 suggest lack of pnference for cc&ones. *A sampling period lasted 6 hrs and extended from 3 hrs before to 3 hrs after sunrise, mid-day, sunset, midnight. *Denotes * agniticant difference at aS4l.05.

520 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1988 exblained as follows: Immature arasses resemble dicots in their cell Preference by deer for the ponderosa pine types as a feeding w&l characteristics, with low leiels of neutral detergent fiber and habitat decreased as canopy cover increased (Table 3). This .is lignin, and consequently allow rapid excretion that is critical to consistent with a strong negative correlation between canopy COV- deer digestion (Spalinger et al. 1986). Because temperatures are erage and understory herbage production reported by Metz (1974) frequently mild on our study area, green grass could be found at in Colorado Front Range pine forests. On our study area the least in small amounts throughout the winter. Others (Carpenter et lO-39% pine canopy coverage type has an understory of primarily al. 1979, and Spowart and Hobbs 1985) have reported high propor- mountain mahogany. The understory gives way to sparse stands of tions of grass in winter diets of Colorado mule deer. According to Vaccinium or no vegetation at all in the pine type with canopy Hobbs and Spowart (1984), grasses in a nearby grassland commun- coverage of 70-100%. Deer apparently find adequate concealment ity were highly nutritious, containing from 6.3-8.1s crude protein in the ponderosa pine types for resting because they used all and from 4346% in vitro digestible dry matter during 2 consecu- densities of the pine type for resting during daytime and sunset and tive winters. The grassland type is very open and without cover, used the pine lO-39% canopy coverage type for resting during and we surmise that deer use these areas only when concealed by sunrise. Preference for pine for resting during those periods, how- darkness because grassland was preferred for feeding and resting at ever, was neutral (Table 3). Deer preferred the mountain maho- night, neutral for feeding and resting during sunrise and sunset gany type for resting during sunrise, daytime, and sunset. In the periods, and unpreferred for both activities during daytime (Table absence of the mountain mahogany and grassland types, the pine 3). This is substantiated by the finding that deer in the grassland with lo-3990 canopy coverage and a mountain mahogany under- type preferred ecotones (general hypothesis number 4) with the story would likely become the main feeding type and its impor- mountain mahogany and SU-HA-PRU types, which offer cover tance for resting would likely also increase. Thus, large monotypes and security during sunrise, daytime, and sunset; but at night deer of pine lo-39% could be considered good deer habitat. Because the apparently did not need the concealment offered by those ecotones 40-69% and 70-100% canopy coverage pine types provide, only when in the grassland type (Table 4). (Table 4 shows that lack of resting cover, however, optimum composition of dense pine in a preference for the grassland-mountain mahogany ecotone at night deer range depends on what vegetation types are available nearby was not significant at EO.05, but calculated P was 0.07). Within for feeding. If adjacent feeding types are primarily open grassland, the habitat complex examined, grassland appears to be a very and the 40-69% and 70-100% canopy coverage pine types provide important deer habitat component. However, because it is used the only available cover offering security for resting, then these mainly at night and lacks adequate cover needed by deer for dense pine types should occupy perhaps as much as 50% of the’ daytime use, we extrapolate that to be considered suitable deer range area. If, however, other higher quality types such as moun- habitat, grassland should occupy less than 50% of a deer’s home tain mahogany or lO-39% canopy coverage pine with a mountain range area and occur in small patches. This recommendation fol- mahogany understory are available to provide both feeding and lows since interior areas of large patches of grassland (more than resting cover, then the composition of 40-69% and 70-lwc can- 100 m from a cover type) received less use than did ecotonal areas. opy coverage pine types can be proportionately less. Thinning and The cover-type patch should be of sufficient size and density to patch cutting in ponderosa pine and mixed conifer stands have provide enough concealment that deer using it appear to feel resulted in substantial increases in grass and forb production and secure. subsequent deer use (Clary and Ffolliott 1966, Patton 1974, Patton Observed deer preference for the mountain mahogany type as a 1976). Thus, where large, dense pine monotypes exist, deer habitat habitat for feeding (Table 3) is supported by Medin and Anderson could be improved by converting pine in excess of that needed for (1979), who found that mountain mahogany comprised 25% (by security cover to openings where forage species could become weight) of the botanical composition in 52 mule deer rumens established through natural regeneration or reseeding. If created collected year long from nearby deer ranges. Another factor that openings are converted to grass and forbs, those openings should may have influenced preference for the mountain mahogany type not exceed 200 m in width so no point in the opening is more than for feeding is that it has an understory of the same grasses that 100 m from cover, as was previously suggested for patch size in the occur in the grassland type. Observed deer preference for the grassland type. mountain mahogany type for both feeding and resting (Table 3) is In the vegetation complex on our study area mountain meadows consistent with higher densities of fecal pellet groups in mountain were among the least utilized vegetation types by deer for both mahogany than in coniferous habitat reported on nearby deer feeding and resting (Table 3). This may have been partly due to ranges by Anderson et al. (1972). The mountain mahogany type availability of more preferred types for both activities. Mountain may supply both food and cover requirements for deer since they meadows may be characterized by various kinds of plant species preferred it for feeding and resting during daylight hours (sunrise, depending on elevation, precipitation, etc. In another situation daytime, and sunset periods) and did not gravitate toward ecotones where other, more preferred types are not readily available or with grassland or SU-HA-PRU types during any period when where different plants occur in mountain meadows, the mountain located in mountain mahogany (Table 4). Thus, even very large meadow type may become of increased importance to deer (Patton monotypes of mountain mahogany should be considered prime and Judd 1970). deer habitat. Neutral status of rock outcrop areas for feeding and resting The SU-HA-PRU type is very dense and should provide excel- during daytime suggests such areas provide adequate midday lent concealment for deer. It occurs as long, narrow configurations habitat for deer. Rock outcrop areas were unpreferred for both of vegetation in draws surrounded by large patches of mountain feeding and resting at night (Table 3). If, however, deer habitat mahogany and/ or grassland. Deer preference for those adjacent needs during midday are satisfied by mountain mahogany or some types for feeding and resting may have contributed to the unpre- other type, then existence of rock outcrop areas in a deer habitat ferred status of the SU-HA-PRU type for feeding and resting complex may not be necessary. during daytime and for feeding during sunset, and its neutral status during other periods (Table 3). Neutral preference for ecotones Conclusions with mountain mahogany and grassland types during all periods We conclude: (1) Deer along the northern Colorado Front while deer were in the SU-HA-PRU type (Table 4) may have been a Range are adapted to temperatures and wind velocities that fall result of the narrow pattern in which the SU-HA-PRU type occurs. within the normal range for winter and do not alter their feeding If resting and escape cover needs of deer using the grassland type and resting patterns because of them, but do adjust to periods of are satisfied by mountain mahogany or some other cover type, then deep snow (>36 cm) by spending more of their time resting to existence of the SU-HA-PRU type in a deer habitat complex may conserve energy. (2) Deer select and concentrate in specific vegeta- not be necessary.

JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 521 tion types that satisfy their needs for food and security, and their Hlreshd, A.S. 1985. Habitat use by black-tailed deer on northern Van- degree of preference for specific habitats varies by periods of day couver Island. J. Wildl. Manage. 49946950. and night. When using vegetation types where security cover was Hobbs, N.T., 8nd R.A.. Spow8rt. 1984. Effects of prescribed tire on nutri- inadequate, deer distributed themselves near ecotones with habi- tion of mountain sheep and mule deer during winter and spring. J. Wildl. tats that offer cover and security. Manage. 48:551-560. We recommend that when evaluating deer habitat or preparing a Jennrich, R.I., 8d F.B. Turner. 1969. Measurement of non-circular home habitat improvement or mitigation plan for deer range, considera- range. J. Theor. Biol. 22227-237. Kufeld, RX., D.C. Borden, urd J.M. Siperek, Jr. 1987. Evaluation of a tion should be given to the juxtaposition, size, and relative contri- telemetry system for measuring habitat usage in mountainous terrain. bution of each patch of existing vegetation in the habitat complex Northwest Sci. 61249-256. to the overall food and cover needs of deer in the area, together Loveless, C.M. 1967. Ecological characteristics of a mule deer winter range, with ways to improve the complex. Colo. Dep. Game, Fish and Parks. Tech. Pub. No. 20. M8eLie, R J. 1970. Range ecology and relations of mule deer, elk and cattle Literature Cited in the Missouri River Breaks. Wildl. Monogr. 20: l-79. Anderson, A.E., 8nd D.E. Media, 8nd D.C. Bowden. 1972. Mule deer Medin, D.E., md A.E. Anderson. 1979. Modeling the dynamics of a numbers and shrub yield-utilization on winter range. J. Wildl. Manage. Colorado mule deer population. Wildl. Mono. 68:1-77. 36571-578. Metz, H.E. 1974. Relationship between Ponderosa pine and understory Burett, R. 1982. Habitat preferences of feral hogs, deer and cattle on a herbage production. MS Thesis, Colorado State Univ., Ft. Collins. Sierra foothill range. J. Range Manage. 35342-346. Puker, K.L., 8nd C.T. Robbins. 1983. Thermoregulation in mule deer and Cupenter, L.H., O.C. W8llmo, 8nd R.B. Gill. 1979. Forage diversity and elk. Can. J. Zool. 621409-1422. dietary selection by wintering mule deer. J. Range Manage. 32:226-229. Parker, K.L., C.T. Robblns, 8nd T.A. Hanley. 1984. Energy expenditures Carson, R.G., 8nd J.M. Peek. 1987. Mule deer habitat selection patterns in for locomotion by mule deer and elk. J. Wildl. Manage. 483474488. northcentral Washington. J. Wildl. Manage. 5146-51. P8tiO0, D.R. 1974. Patch cutting increases deer and elk use of a pine forest Clary, W.P., 8nd P.F. Ffolliot. 1966. Diiferences in he&age-timber rela- in Arizona. J. Forest. 72764-766. tionships between thinned and unthinned Ponderosa pine stands. USDA P8tiOn, D.R. 1976. Timber harvesting increases deer and elk use of mixed Forest Serv. Res. Note RM-74. conifer forest. USDA Forest Serv. Res. Note RM-329. Freddy, D J. 1985. Quantifying capacity of winter ranges to support deer- Patton, D.R., and B.1. Judd. 1970. The role of wet meadows as a wildlife evaluation of thermal cover used by deer. Colo. Div. Wildl. Wild. Res. habitat in the southwest. J. Range Manage. 23:272-275. Rep. July:l3-19. Sok81, R.R., and FJ. Rohlf. 1981. Biometry, 2nd cd. W.H. Freeman, San Gilbert, P.F., O.C. W8Umo, 8nd R.B. Gill. 1970. Effect of snow depth on Francisco. mule deer in Middle Park. Colorado. J. Wildl. Manaae. 34~15-23. Southwood, T.R.E. 1966. Ecological methods. Methuen, London. Gilmer, D.S., S.E. Miller, 8&l L.M. Cowrrdin. 1973. hlysis of radio- Spalinger, D.E., C.T. RobbIns, 8nd T.A. H8nley. 19%. The assessment of tracking data using digitized habitat maps. J. Wildl. Manage. 37:404409. handling time in ruminants: the effect of plant chemical and physical Gilhgh8m, M.P., 8nd F.L. Bunnell. 1985. Reliability of motion sensitive StNCtUre on the rate of breakdown of plant particles in the rumen of mule radio collars for estimating activity of black-tailed deer. J. Wildl. Man- deer and elk. Can. J. Zool. 64~312-321. age. 49951-958. Spowart, R.A., and N.T. Hobbs. 1985. Effects of fire on diet overlap H8rdey, T.A. 1982. The nutritional basis for food selection by ungulates. J. between mule deer and mountain sheep. J. Wildl. Manage. 49942-946. Range Manage. 35146-151. Telfer, E.S. 1978. Cervid distribution, browse and snow cover in Alberta. J. H8o& T.A. 1984. Habitat patches and their selection by wapiti and Wildl. Manage. 42352-361. black-tailed deer in a coastal montane coniferous forest. J. Appl. Ecol. 21:423436. Technical Notes

Separating leaves from browse for use in nutritional studies with herbivores

J.F. GALLAGHER, T.G. BARNES, AND L.W. VARNER

Abstract occidentalis), alfalfa (Medicago saliva), and grasses (Ullrey et al. A technique has ban developed that facilitates removal of green 1964, Ullrey et al. 1967, Ullrey et al. 1968, Mautz and Petrides leafy material from stems of shrub speciea using a thresher. Use of 1971, Robbins et al. 1975, Ullrey et al. 1987). To date, few studies this technique makes possible the rapid removal of leaves from have been conducted with woody plants because of the time neces- woody species that would otherwise require excessive band labor. sary to collect sufficient forage to maintain theanimals for a period Key Words: sampling, feeding trials, brush collection, white-tailed of 10 to 14 days. This is especially true when the leaves are small or deer surrounded by thorns. We developed a technique which allows large quantities of green leafy material to he separated from shrub Nutritional studies of white-tailed deer and other ungulates have stems in a relatively short period of time. largely been conducted using pelleted rations or natural forages that are easily collected such as aspen (Populus spp.), cedar (Z%uia Methods Authon are graduatestudents, Department of Wildlife and Fisheries Sciences Tcxa~ A&M Unwcrsity, College Station,Texas 77843; and associate professor, Texa; Samples of brush species known to he eaten by deer (Vamer and Agricultural Expcrimcnt Station, Uvaldc, Texas 78801. Blankenship 1987) were collected about 15 km southwest of WCthank J.R. Mulkey for use of the threshing machine. This study was funded by Uvalde, Texas. Cenizo (LeucophyZlumf~rescens) was collected by the Caesar Klebcrg program in Wildlife Ecology at Texas A & M University, Welder Wildlife Foundation, Texas Agricultural Experiment Station, and National Rie hand, grasping each branch approximately 15 cm from the end and Association Grants in Aid program. Welder Wildliie Foundation Contribution #326. stripping the leaves in a single pull. Guajillo (Acacia berlandieri) Manuscript accepted 14 June 1988. stems were clipped using small clippers and leaves were removed by 522 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 hand. Blackbrush (Acacia rigid&J, brazil (Condalia obowtah num pans and subsequently separated from any remaining woody and guayacan (Porlierio ongusrifolin)stems were clipped in the materials by sifting through 6- or 12-mm hardware cloth screens, field using small hand clippers. depending on leaf size. Approximately 100 kg of stems less than 20 mm in diameter were collected and transported daily to the Texas Agricultural Experi- Results and Discussion ment Station located in Uvalde, Texas, for threshing. Blackbrush is characterized by small compound leaves sw- Fresh material was then fed into a Vogel Nursery Thresher (Fig. rounded by large thorns. Brazil and guayacan are characterized by 1)’ that separated the leaves from stems. The thresher, which is a complex multistemmed growth form. Both conditions make it nom~ally used to thresh small grains, is an overshot spiked cylinder difficult and time consuming to remove leaves by hand. The thresh- and concave type harvester. The cylinder is constructed of 30-x1 ing technique allowed us to collect and process enough forage to steel pipe with 10 rows of heavyduty alternating teeth. The concave feed 1 deer (1,600 grams as fed) one day in one-half to one man- is constructed of 6-mm steel plate with 6 rows of teeth extending hour, including time for clipping, threshing, and sifting. We esti- approximately 5-cm from the surface. The final separation was mate that manual separation of blackbrush leaves would require 6 achieved by the movement of materials across the shaker pan. to 7 man-hours per deer/day. By comparison, cenizo and guajillo, Forced air from the blower discharged leaves forward and allowed both of which are relatively easy to collect without threshing the stems to drop in front of the tail chute. The blower fan speed because of lack of thorns or complex branches, still required over was adjusted to vary the amount of forced air passing across the 2.0 man-hours to feed each deer/day. plant material which allowed control over the distance the leaves The technique appears to work well on species with small leaves were blown. or single stems. Species with large leaves did not sift away from To catch leaves, a 3X4-m plastic tarp was suspended approxi- stems as well, however. Multistemmed species such as brazil were mately I-m from the tail chute opening at the same height as the more difficult to thresh, especially if there was abundant terminal thresher opening. Stems dropped to floor in front of the tarp. growth. Separation of leaves from stems was improved if the Leaves and a small amount of sterns were then collected in alumi- material was frozen before threshing. This technique does have a tendency to select for older, more mature leaves which are smaller and sift easier. The thresher did not remove leaves from the stems of young succulent growth with the same efficiency as from mature growth since stems from new growth are more flexible and less brittle than mature growth. This technique enabled researchers to collect large amounts of browse material in a short period of time. The final product was largely leaves with relatively little stem material.

Literature Cited

Mnutz, W.W., snd GA. P&ides. 1971. Food passage rate in the white- tailed deer. J. Wild]. Manage. 35:723-731. Bobbins, CT., P.J. Van So&, W.W. Mautz, and A.N. Moen. 1975. Feed analysis and digestion with reference to white-tailed deer. J. Wildl. Manage. 39x57-79. Ullrcy, D.E., LT. Nellist, J.P. Duvenderk, P.A. Whetter, snd L.D. Fsy. 1987. Digestibility ofvegetative rye for white-tailed deer. J. Wildl. Man- age. 51:51-53. “llrey, ED., W.G. Youatt, H.E. Johnson, P.K. Ku,nnd L.D. F.y. 1964. Digestibility of cedar and aspen browse for the white-tailed deer. J. Wild,. Manage. 28:791-797. Ullrey, D.E., W.G. Youatt,L.D. Fsy, andB.E.Brent. 1967. Digestibility of cedar and jack pine browse for the white-tailed deer. J. Wildl. Manage. 31:44844. “llrey, D.E., W.G. Yalutt, L.D. Fay, B.E. Brent, and K.E. Kemp. 1968. Digestibility ofcedar and balsam fir browse for the white-tailed deer. J. Wild,. Manage. 32:162-171. Vmw, L.W., and L.H. BIankmsbtp. 1987. South Texas shrub: Nutritive value and utilization by herbivores. USDA Forest Serv. Gen. Tech. Rep. INT--222. 179 p.

JOVRNAL OF RANGE MANAGEMENT 4,(6), November 1988 523 publisher suggests), then I find it difficult to believe that they would Book Reviews take the trouble to go through a key, full of technical words, with the probable exception of the “botanizer’*who more likely than not will know already what the plant is. Photosynthesis Volume I. Energy Conversion by Plants and The body of the book is formed by descriptions of each species Bacteria. Photosynthesis Volume II. Development, Car- and sometimes, in the case of large families, a commentary. Each bon Metabolism and Plant Productivity, Both ~01s. description is, in most cases, accompanied by a drawing with a caption summarizing the structure of the species, such as general Edited by Govindjee. Academic Press Inc. 111 Fifth size, leaves, flowers and fruit characteristics. The dimensions, quite Avenue, New York, New York 10003. Price Vol. I. $79; surprisingly and happily do go by the metric system. The quality of Vol. II. $65. the figures varies very much, from the most precise detail as in the These two volumes summarize clearly a great amount of macpalxochitl (Chirunthodendron pentaductylon Larr.) to the research and information and impress the reader with the scope of almost flat character-less appreciation of the gobemadora (Larrea photosynthesis research and the wide range of technologies that tridentata (DC.) Coville), which unfortunately closes the descrip- have been involved in establishing a not yet complete knowledge of tion section. This greater body of the book is consistently devoid of the process. “typos”; I only detected two very minor ones. Needless to say about These books are aimed at the specialist in photosynthesis, but any compendium on flowers, there are grave omissions, some although they deal in depth with a topic fundamental to range worse than others: Buuhiniu spp. is not mentioned although is seen productivity, much of the subject matter is far removed from the in most Mexican towns; neither Gliricidiu sepium nor Custillou practical issues of range management. However, some topics elusticu is included, perhaps due to the natural bias of an aridland covered (mostly in volume II.) are of particular significance to taxonomist. Each description is headed by the full scientific name productivity of range plants, and to those involved in their adapta- (including authority, a good point) and one Spanish vernacular tion to diverse environments. name. Vernacular names in Mexico are quite useless, unless you Many of the contributing authors indicate where they see oppor- say where you heard the same. The case in point is toloache tunities for improvement in our knowledge, and why promising (Dutrtru cundidu (Pets.) Safford), which is called Ploripondio in prospects of changing photosynthesis to increase yield have not the book; such a name will probably be recognized only by people been realized; or where they see avenues for research which may from the central and southern states. pay dividends. Of particular interest to the range scientist will be The last sections are made of an illustrated glossary (unnecessary those chapters in Volume 2 dealing with photosynthetic considera- for the botanizer, but fair for the layman), a rich (but not complete) tions in the adaptation of species to various environments and the selected references section, and finally a very good index that likely changes in productivity of species as the carbon dioxide includes family and species names, vernacular and scientific concentration increases. together with the Spanish vulgar names for the species. The integration of this multi-disciplinary knowledge into simu- It would have been a more interesting book if the Masons lation models is advocated as a powerful tool to improve the extended their rich comments, which we only see here and there, productivity of photosynthesis research and to enhance the chan- about the uses, origin, folk, etc. of the species. I think that those ces of success in future attempts at manipulation. Models such as “tidbits” of information were the most enjoyable parts in my these should be of particular importance to those deciding on reading of the book. I also guess the book will be enjoyed by a select research funding priorities. With the prospect of utilizing genetic number of people who have a secret love for plants and also know engineering to change the processes of photosynthesis this approach something about botany, but by far I would not recommend it to also seems likely to be of key importance to allow the genetic any unqualified traveller to Mexico.-Abraham de Albu, Jalpa, engineers to evaluate the likely outcome of their efforts before Zacatecas. major investments are attempted.-Jtl. WiZliums,Pullman, Wash- By ington. Ecology: Individuals, Populations and Communities. Michael Begon, John L. Harper, and Colin R. Townsend. 1986. Blackwell Scientific Publications. Published in the A Handbook of Mexican Roadside Flora. By Charles T. United States and Canada by Sinauer Associates Inc., Mason Jr. and Patricia B. Mason. 1987. University of Publishers, Sunderland, Massachusetts 01375.876 + xii p. Arizona Press, 1615 E. Speedway, Tucson, AZ, 85719. illust. $33.95. 380~. $19.95 (paper). Begon, Harper and Townsend focus on three levels of organiza- The book is perhaps one of the first efforts to make available to tion as indicated by the title. Rather than seek an arbitrary balance the English speaking people some knowledge of the rich flora among the topics of contemporary ecological research, the authors evident as they travel along Mexican roads. Most likely the effort is use the levels of organization theme to produce a comprehensive bound to be ambitious considering how rich the country is in treatment of population ecology which is well integrated with climates, due mainly to its variability in topography and latitude. topics from the other two levels. The book does an excellent job of Nonetheless the Mansons have managed to work with 68 families repesenting the complexity of ecological systems and of developing and illustrate almost every species described. The book is quite ecology as an intellectually challenging field. adequately introduced with a concise description of the vegetation Coverage is designed to reach students whatever their strengths-in of the country; unfortunately, this does not conform at times with field or laboratory, in theory or in practice. Theory is given a names and descriptions made by botanists and ecologists, such as careful treatment, the best of any ecology text I know. The authors the so-called short-tree forest. This section is crying out for a good are especially careful to be clear about assumptions. The interpre- map; it is no use to talk about places when a foreigner does not tations of experimental results from the field and the laboratory know their location. Thereafter we are invited to use a 6 page key in are equally clear and include discussions of limitations due to order to know exactly what plant we have in our hands. The key, experimental design. obviously, is specific for the plants described in the book and The book is organized into four main parts. “Part 1: Organisms” although not tested by me, hopefully only will lead to the correct includes chapters covering “The Match between Organisms and species. It seems that if the book was made for all those “travellets their Environments’*, “Conditions”, “Resources”, “Life and Death returning from Mexico,* which encompasses the “curious tra- in Unitary and Modular Organisms”, and “Migration and Disper- veller, the amateur naturalist and the armchair botanize? (as the sal in Space and Time ‘*.“ Part 2: Interactions” includes two chap-

624 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1966 ters on competitipn, three on predation, including herbivory, and philosophy and the beginnings of range research were initiated in one chapter each on “Decomposers and Detritivores”, “Parasitism Region 4 of the Forest Service. In fact, this is the story of the and Disease”, and “Mutualism”. Part 3 includes “Two Overviews”: birthplace of range management on a global perspective. “Lie-Histroy Variation”, and “Abundance”. “Part 4: Communi- The authors have literally dug through the files of the Region, ties” concludes the book with chapters on “The Nature of the and compiled a monumental record of the day-today problems Community”, “The Flux of Energy and Matter through Commun- facing federal land managers, and the solutions successfully (some- ities”, “The Influence of Competition on Community Structure’*, times) worked out with users, politicians, and the general public. The Influence of Predation and Disturbance on Community Struc- The record begins with a brief description of settlement of the ture “, “Islands, Areas and Colonization”, “Community Stability intermountain region, and development of the livestock and and Community Structure”, and “Patterns of Species Diversity”. timber uses on public lands, and continues in almost 300 pages of The authors do not distinguish an ecosystem level of organiza- detail to the “Sagebrush Rebellion” and beyond. Details from tion, averring that “No ecological system, whether individual, personal diaries and official memoranda along with a well-selected population or community, can be studied in isolation from the abundance of photographs gives the reader an unusually clear environment in which it exists.” (pp. 591-2) For this reason matter understanding of how it all happened. Activities of General Land and energy exchange are treated as supplementary approaches to Office personnel in management of the Forest Reserves, beginning understanding community structure. in 1891 with the establishment of the Yellowstone Forest Reserve, The book is written to engage the serious student and mature to the transfer of authority to the Forest Service in 1905, is approach to the subject is assumed. The book is usefully and explained from a local perspective that was new to me. Struggles of extensively illustrated on nearly every page. The wide left margin these early Rangers and Supervisors to bring order and reason into of each page contains brief summary statements which serve to range stocking rates puts present pressures from the public into highlight the important concepts. perspective. This is a very contemporary book and outdated statements such One of my favorite chapters tells of the wisdom of administra- as the assertions that the carbon dioxide concentration of the tors in establishing the Great Basin Research Station (the first), atmosphere “ . ..varies little around 300 parts per million,..” (p. 85) where the truths and limits of proper stocking were clearly demon- or that “The soil water content at the wilting point does not differ strated, and gave a scientific foundation to theories of proper significantly among plant species.” (p. 89) are exceptions and not management of rangelands. We are still extrapolating and verify- the rule. ing results of some of the early experiments conducted there. Though I would prefer to see ecosystems treated differently, I do The history of the Region is told in a series of chapters organized not fault the authors for their focus on individuals, populations, in chronological order and focused on prevailing philosophical and communities. This book stands out for its excellent coverage eras. For example, titles are: Beginnings of resource administra- of population ecology. It is not an ecosystem text. That it includes tion: the Intermountain West under the Forest Service, 1905-1909, substantial treatments of matter and energy exchange is simply a Forest protection and management: 1910-1929; Forest manage- bonus. This text deserves a careful examination for use in any ment in a depression era: 1930-1941; Organization and planning general ecology course with a strong population emphasis.--E&n for intensive management: 1942-1949; Toward stewardship and H. Franz, Pullman, Washington. multiple use management; 1950- 1959; An era of intensive multiple-use management: 1960-1969; Forest planning and management under The Rise of Multiple-use Management in the Intermountain pressure: 1970-1986; and The Intermountain Region in retrospect. West: a History of Region 4 of the Forest Service. 1987. People will read this book for various reasons. For me, it is practically the story of my life, from boyhood hiking in the Thomas G. Alexander et al. Superintendent of Docu- Wasatch Mountains, to teaching and researching in a university ments, Washington D.C. (order no. 001-001-0631-9) $13. environment struggling with problems parallel to those of the History is not only interesting and exciting, but it is a record of Forest Service. Many will undoubtedly find in it a resource base for the victories and failures of the past by which we can learn to plan future study and reference. Others may seek data for praising or for success in the future. (Somebody else has said this same thing condemning the Forest Service. But best of all, read it if you want previously and better than I have, but I am convinced that it is to renew your faith in the wisdom and dedication of the men and true!). This history should be of particular interest and value to women who have laid the foundation in the past and set the goals range managers, because a significant part of range management for the future.-Grant A. Harris, Washington State University

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 525 Index, Volume 41 A Big bluestem (Andropogon gerardii) seed Cattle distribution, 73 Abandoned farmland, 22 production, 253 Cattle grazing rates, 78 Abemethy, Rollin H., 30,35 Big sacaton (Sporobolus wrightii), 12 Cattle grazing (behaviour), 118 Above-ground biomass, 477 Big sagebrush (Artemisia tridentata), 58, Cattle grazing tall larkspur on Utah Adams, L.D., 435 63, 172,395 mountain rangeland, 118 Additional soil moisture (irrigation), 387 Biological insect control, 100 Cattle vegetation, and economic responses Aerial seed application, 53 Biomass, 197,227 to grazing systems and grazing pres- Agrostology, 350 Bionomics of patterned herbicide appli- sure, 282 Alabama, 274 cation for wildlife habitat enhancement, Cecidomyiid midge (Contarinia wattsi Alfalfa (Medicago sativa), and cicer milk- 317 Gagne), 253 vetch (Astragalus cicer). 378,383 Bitterbrush (Purshia tridentata), 63, 122 Cell&se vs. rumen fluid, 337 Alfalfa-grass mixtures, 488 Blackbum, Wilbert H., 108,296 Cellulase vs rumen fluid for in vitro Ali, Mohamed Ben, 167 Blackgrass bugs (Labops hesperius and digestibility of mixed diets, 337 Alkali sacaton [Sporobolur airoides-‘Salt- Irbisia brachycera), 100 Central Colorado, 391 talk,’ ‘Salado’(NM-I 84, and ‘Durango Black-tailed jackrabbits (Lepus caltfor- Central Great Plain, 86, 239, 253, 264, 84”-J, 216 nicus), 162 481,483 Alkaloids, 118 Bluebunch wheatgrass (Agropyron spic- Central Great Plain (Sandhills), 4 16 Allelopathic effects of sandbur leachate atum), 88, 232 Central Kansas, 48 1,483 on switchgrass germination, observa- Blauer, A. Clyde, 114 Central Texas, 108,296,401 tions, 86 Blue grama (Boutelouagracilis), 287,391 Chambers, Jeanne C., 267 Allelopathic potential, 86 Bobcat (Felts rufus), 322 Chaneton, Enrique J., 495 Allred, Kelly W., 350 Bode, R.P., 287 Changes in grass basal area and forb den- Alsike clever (trifolium hybridum), 53 Boe, A., 342 sities over a 64-year period on grass- American licorice (Glycyrrhiza lepidota), Book Reviews, 94,184,270,359,447,524 land types of the Jomada Experimen- 342 Booth, D. Terrance, 335,351 tal Range, 186 Analysis of Russian thistle (Salsolo spe- Borman, M.M., 167 Chemical analysis, 401 cies) selections for factors affecting Bowden, David C., 70,515 Chemical analysis (swainsonine), 399 forage nutritional value, 155 Box, T.W., 66,477 Chemical composition of old world blue- Anderson, D.M., 78 Bradley, Lisa C., 322 stem grasses as affected by cultivar and Animal gains, 282 Brakensiek, Donald L., 307 maturity, 40 Animal performance, 259 Bransby, D.I., 274 Chemotaxonomy, 395 Annual and perennial, 88 Brejda, J.J., 416 Chewings, V.H., 430 Annual flooding, 495,500 Breed comparisons and characteristics of Clark, D.H., 488 Ansley, R.J., 83,90, 227, 355 use of livestock guarding dogs, 249 Clark, Robert G., 407 Araya, Jaime E., 100 British Columbia (Canada), 503 Clawson, W. James, 193 Argentina, 220,495, 500 B&ton, Carlton M., 407 Claypool, P.L., 40 Arizona, 216,307 Brooke, B.M., 53 Clearcut-logged sites, 53 Armentrout, Susan M., 139 Broom snakeweed (Gutierreziasarothrae), Climatic-edaphic relationships, 127 Arthropod predation of black grass bugs 92 Cluff, Greg J., 150 (Hemiptera:Miridae) in Utah ranges, Browse, 522 Coal mine reclamation, 491 100 Bruchid beetle [Acanthoscelides aureo- Coastal bermudagrass and Renner love- Arthur III, W. John, 162 lus (Horn) Coleoptera: Bruchidae], 342 grass for fertilization responses in a Association of the wheat stem sawfly Brush removal, 172,470 subtropical climate, 7 with basin wildrye, 328 Brush seed germination, 49 1 Coastal bermudagrass (Cynodon dacty- Atrazine dissipation and off-plot move- Brown, Ray W., 267 ion). 7 ment in a Nebraska sandhills subirri- Browning, James A., 239 Coastal Plain, 441 gated meadow, 416 Brummer, J.E., 264 Cochran, Robert, 483 Buffelgrass (Cenchrus ciliaris), 127 Coffey, K.P., 426 B Bulk density, 500 Coleman, S.W., 40 Barnes, T.G., Buman, Robert A., 30,35 Collecting, drying, and preserving feces Bamett, Donald A., 340 Bunting, Stephen C., 232,410 for chemical and microhistological anal- Barren-a, M.G., 410 Burning, 172 ysis, 168 Basal area, 186 Burning (tire), 232, 239, 404, 407, 410, Colorado, 332,378,383,515 Basal cover, 282 413 Columbia milkvetch (Astragalus miser), Bassiri, M., 378,383 Burning tallgrass prairie, 178 26 Bastin, G.N., 430 Burritt, E.A., 346 Columbia needlegrass (Stipa columbiana), Beck, Reldon F., 186 232 C Bedunah, Donald J., 144 Columbus, J. Travis, 350 Behaviour, 5 15 California, 25 1 Common garden, 155,391 Beetle, Alan A., 370 Canopy estimation (photo digital), 355 Comparative chemicl composition of armed Berg, William A., 22 Carbohydrates (nonstructural), 235,287 saltbush and founving saltbush, 40 Bermudagrass (Cynodon dactylon-‘Callie, ’ Carbohydrates (TNC), 481 Comparison of water use by Artemisia ‘Coastal,* ‘Brazos,’ and hybrids S-54 Carter, M.R., 253 triokntata spp. wyomingensisand Chryso- and S-16), 274 Catingueira (Caesalpinia pyramidalis), thamnus viscidtflorus spp. viscidtflo. Bemardo, D.J., 178 66,477 rus. 58 Big bluestem (Andropogon gerardii), 239, Cattle, 522 Conrad, B.E., 274 481,483 Cattle (behaviour), 509 Consumption, 509

526 JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 Contamination, 416 Drying, 168 False broomweed (Ericameria austrotex- Control of threadleaf rubber rabbitbrush Drying method, 346 ana) perennial grasses, 2 with herbicides, 470 Duncan, Don A., 193 Farming, 155 Continuous and rotation grazing, 78 Dunn, T.G., 435 Fecal analyses, 168 Cool-season grasses, 488 Fecal output (predication), 426 Correlation of degreedays with annual Feces, collecting, 168 herbage yields and livestock gains, 193 East African ecosystems, 450 Female mule deer (Odocoileus hemio- Cover, 245 Economic analyses, 172, 178 nus), 70 Coyote and bobcat responses to inte- Economic assessment of risk and returns Fences (electric), 25 1 grated ranch management practices in from prescribed burning on tallgrass Femandez, Robert J., 220 south Texas, 322 prairie, 178 Fertilization (N-P-K), 235 Coyote (Curtis latrans), 249,251,322 Economic comparison of aerial and ground Fertilization (nitrogen and phosphorus), Cox, Jerry R., 12, 127,216,404 ignition for rangeland prescribed fires, 7 Crane, R.A., 317 413 Fertilization (soil nitrogen), 22 Creosotebush (Larrea tridentata), 387 Economic optimum big sagebrush con- Fertilization (IV), 387 Crested wheatgrass (Agropyron crista- trol for increasing crested wheatgrass Fiber, 346 turn), 488 production, 172 Field, 100 Crested wheatgrass (Agropyron deserto- Economics, 317,413,435 Field colonies, 104 rum), 73, 172,378,383,421 Eddleman, L.E., 491 Fire effects on tobosagrass and weeping Crested wheatgrass (Agropyron a’eserto- Edmunds, D.A., 391 lovegrass, 407 rum), and (Agropyron crbtatum), 259, Edwards Plateau, 108,296,313 Fischbach, David A., 365 421 Effect of burning on seed production of Fisher, F.M., 387 Crested wheatgrass (Agropyron spp. ‘Hy- bluebunch wheatgrass, Idaho fescue, Flameleaf sumac (Rhus lanceolata), 48 crest?, 30, 35 and Columbia needlegrass, 232 Flinn, R.C., 317 Crude protein, 235,401 Effect of clipping on the growth and mis- Flint Hills, 481,483 Cumberland Island National Seashore, erotoxin content of Columbia milk- Florida, 460 441 vetch, 26 Floristic changes, 495 Cunningham, G.L. 387 Effect of defoliation frequency and N-P- Floristic changes induced by flooding on K fertilization on maidencane, 235 grazed and ungrazed lowland grass- D Effect of drying method on the nutritive lands in Argentina, 495 Dabo, S.M., composition of esophageal listula for- Forage and livestock production, 193 Dahl, Bill E., 337 age samples: influence of maturity, 346 Forage estimates, 167 Day/night temperature regimes, 35 Effects of burning on germinability of Forage production, 108 De Alba-Avila, Abraham, 216 Lehmann lovegrass, 404 Forage production (yield), 430 Deer 522 Effects of phenology, site, and rumen fill Forage quality, 40, 155 Defoliation impacts on coppicing browse on tall larkspur consumption by cattle, Forage yield, 7,26,78 species in northeast Brazil, 66 509 Forage yield (grasses, forbs, shrubs), 16 Defoliation of Thurber needlegrass: her- Effects of prescribed fire on Chamaes- Forage yield (green and dead), 12 bage and root responses, 472 partium tridentatum ((L.) P. Gibbs) in Forb, 291 Degreeday correlations, 193 Pinuspinaster (Aiton) forests, 410 Forest, 63, 144 Dehulling, 335 Effects of temperature, water potential, Forest management, 466 Dehydration, 378 and sodium chloride on Indiangrass Forests, 26 Dehydration effects on seedling devel- germination, 207 Fourie, J.H., 127 opment of four range species, 383 Electric fences for reducing sheep losses Fowler, James L., 155 Demarais, Stephen, 340 to predators, 25 1 Frasier, Gary W., 366 Demography, 267 Elk sedge (Carex geyeri), 144 Frequency, 2 Depuit, E.J., 372 Ellis, James E., 450 Friedel, M.H., 430 Desert saltgrass (Distichlis spicata), 150 Emergence, 216 Front Range, 515 Design of rain shelters for studying water Engelsjord, Michael, 224 Fruit, 351 relations of rangeland shrubs, 83 Engle, D.M., 178 Frye, D.O., 73 Determination of root mass ratios in Erosion, 296 Fulbright, Timothy E., 207,401 alfalfa-grass mixtures using near infrared Esophageal (cannula), 269 reflectance spectroscopy, 488 Esophageal fistula (forage samples), 346 G Developing countries, 167 Establishment of winter versus spring Gains, 178 Dickerson, Jr., Roy Lee, 337 aerial seedings of domestic grasses and Gains (cattle), 435 Dicks, H.M., 274 legumes on logged sites, 53 Gallagher, J.F., 522 Digestibility, 426 Estimating digestibility of oak browse Ganskopp, David, 472 Digestibility (in vitro), 255,337,346,401 diets for goats by in vitro techniques, Garza Jr., Andre, 401 Dissipation, 416 255 Gas exchange, 239 Diterpenoid alkaloids, 224 Estimation of phytomass for ungrazed Gates, Robert J., 162 Dormaar, John F., 503 crested wheatgrass plants using allo- George, Melvin R., 193 Dormant, 48 1 metric equations, 421 Georgia, 441 Douwes, H.E., 26 External markers, 426 Genetic variability, 391 Dowhower, S.L., 227 Germination, 335 F Downy brome (Bromus tectorum), 30,35 Germination as affected by temperature, Drane, J.W., 274 Facelli, Jose M., 495 water potential, and salt, 207 Drought tolerance, and plant selection, Factors intluencing infiltrability of semi- Germination of green and gray rubber 383 arid mountain slopes, 197 rabbitbrush and their establishment on Dry-weight (estimates), 430 Fagre, Daniel B., 322 coal mined land, 491

JOURNAL OF RANGE MANAGEMENT 41(e), November 1988 527 Germination requirements of flameleaf H Indian ricegrass seed damage and germi- sumac, 48 Habitat selection and activity patterns of nation response to mechanical treat- Germination responses of desert saltgrass female mule deer in the Front Range, ments, 335 to temperature and osmotic potential, Colorado, 5 15 Indian ricegrass [Sorghastrum nutans 150 Hacker, R.B., 73 (L.) Nash-‘Cheyenne’, ‘Llano’, ‘Lo- Gibbens, Robert P., 186 Hageman, James H., 155 meta’, ‘Oto’, and ‘Tejas’J, 207 Gifford, Gerald F., 159 Hairy mountain mahogany (Cercocar- Inexpensive alternative esophageal can- Gillen, R.L., 264 pus breviflorus), 153 nula for growing steers and wethers, Goat grazing, 66 Hall, J.W., 26 269 Goats (Capra hircus), 255 Hamilton, Wayne T., 2,317 Infiltration, 197,296,303 Goebel, Carl J., 88 Hand removal, 66,477 Infiltration an interrill erosion responses Grami, B., 378,383 Hanson, Clayton L., 307 to selected livestock grazing strategies, Gramineae, 350 Hardesty, L.H., 66,477 Edwards Plateau, Texas, 2% Grass and legume seeding, 53 Harris, Grant A., 88 Infiltration equations, 159 Grass and shrub densities, 291 Hart, Richard H., 282,362,435 Infiltration rate, 307 Grasslands, 503 Harvest efficiency, 279 Influence of climate and soils on the dis- Grass-legume establishment, 210 Harvester ants (Pogonomyr?nex owyheei), tribution of four African grasses, 127 Grass seed germination, 35,86,88, 150 104 Influence of forest site on total nonstruc- Grass seeding, 127 Haws, Austin, 100 tural carbohydate levels of pinegrass, Grass seedling competition, 30 Heifer nutrition and growth on short elk sedge, and snowberry, 144 Grass-shrub, 186 duration grazed crested wheatgrass, Infhtence of hunting on movements of Grass spiklet formula: an aid in teaching 259 female mule deer, 70 and identification, The, 350 Heitschmidt, R.K., 227 Infrared reflectance spectroscopy, 488 Grasses and forbs, 337 Henderson, Douglas M., 395 Insect longevity, 104 Gray horsebush (Tetradymia canescens), Herbage, 235,239,245,259 Insect-plant association, 328 16 Herbage accumulation, 264 Insects, 253 Gray rubber rabbitbrush (Chrysotham- Herbage disappearance, 264 Insect-plant predation, 342 nus nauseosus spp. nauseosus), 49 1 Herbage dynamics, 279 Integrated brush management, 317 Grazing, 441,460,495,500 Herbage dynamics of tallgrass prairie Integrated wildlife, 322 Grazing effects of the bulk in a Natra- under shoot duration grazing, 264 Intermediate wheat grass (Elytrigia rep- quo11 of the Flooding Pampa of Argen- Herbage (estimations), 421 ens), 488 tina, 500 Herbage growth (basal area, height cub), Intermountain area, 30, 35, 58, 73, 88, Grazing intensity, 274 407 100, 104, 114, 122, 150, 159, 162, 172, Grazing of crested wheatgrass, with par- Herbage production, 282 232,249,259,309,332,335,372,378, ticular reference to effects of pasture Herbage removal, 48 1 383,395,421,435,470,472,488,491 size, 73 Herbicidal control of pricklypear cactus Intermountain Forest-Meadows, 118,509 Grazing management, 279 in western Texas, 3 13 .Introduced grasses, 7,22 Grazing pressure, 503 Herbicide application, 3 17 Irrigated pasture, 274 Grazing, stocking, and production effi- Herbicide (atrazine), 416 .Irrigation, 155,210 ciencies in grazing research, 279 Herbicide control, 3 13 Irrigation water for vegetation establish- Grazing system, 264,287,303,483 Herbicides, 470 ment, 210 Grazing systems, 259,279,282,291,296, Herd size, 73 J 450 High intensity-low frequency, 108 Jacoby, P.W., 83,90,227,355 Grazing tolerance, 472 Hinnant, Ray T., 168 Gray rabbitbrush (Chrysothamnur nau- Jacobson, Tracy L.C., 332 Hironaka, M., 232 Johnson, D.E., 167 seosus), 63 Hobbs, N.T., 368 Great Basin, 30, 35,58, 73, 88, 100, 104, Johnson, Craig L., 421 Holl, F.B., 53 Johnson, James B., 328 114, 118, 122, 150, 159, 162, 172,232, Home range, 515 249,259,309,332,335,372,378,383, Johnson, Patricia S., 421 Honey mesquite (Prosopis glandulosa), Jorgensen, Clive D., 104 395,421,435,470,472,488,491,509 83,90 Green, Jeffrey S., 249 Jurema preta (Minosa acutistipula), 477 Hong, Zhao Xiao, 509 Justification for grazing intensity exper- Green rabbitbrush (Chrysothamnus vis- Horn, F.P., 40 cidjlorus), 58 iments:analyzingandinterpletinggmzing Horses (wild-feral), 441 data,274 Greenhouse, 30 Hunt, C.W., 426 Greenleaf manzanita (Arctostaphylospa- Hunting, 70 K t&a),, 63 Hutten, Neil C., 159 Kalmbacher, R.S., 235,245 Griffith, Brad, 122 Hydrologic impacts of sheep grazing on Kaltenbach, C.C., 435 Griffith, Larry W., 335 steep slopes in semiarid rangelands, Kansas and Oklahoma, 178 Growth and gas exchange of Andro- 303 Kenya, 450 pogon gerardii as influenced by bum- I Kleingrass (Panicum coloratum), 127 ing, 239 Koerth, B.H., 317 Growth and production, 216 Ibarra-F., F.A., 127 Kothmann, M.M., 168 Growth of Gutierrezia sarothrae seed- Idaho, 16, 104, 162,232,307 Krueger, Janice K., 144 Idaho fescue lings in the field, 92 (Festuca idahoensis), 232 Kufeld, Roland C., 70,515 Growth stage, 40 Improved pasture, 7,22 Guard dogs (livestock), 249 In vitro digestibility, 235 L Gulf Coastal Plain, 322 Indian ricegrass (Oryzopsis hymenoides) Laboratory,48,86,88,100,150,207,255, ‘Nezpar’), 335 335,337,346

528 JOURNAL OF RANGE MANAGEMENT 41(8), November 1988 Lane, Mark A., 509 Mesquite (Prosopis glandulosa) trees, Nutritional studies, 522 Large and small rainfall additional soil 355 moisture (irrigation), 387 Methods of ytterbium analysis for pre- 0 Large and small rainfall events, 387 dicting fecal output and flow rate con- Oak (Quercus gambellii), 255 Lavado, Ratil S., 500 stants in cattle, 426 Obligations and expectations of your Lawrence, B.K., 83,90,355 Miller, Jack R., 98 peers: manuscript review at the Journal Laycock, W.A., 391 Miller, Richard F., 58 of Range Management, 368 Leachate, 86 Miller, S.J., 426 Oil and gas, 372 Lecaptain, Roxanne, 155 Mine spoils, 210 Old world bluestem (Bothriochloa cau- Legume (Medicago.sativa), 488 Mixed diets, 337 casica), 22 Lehmann lovegrass (Eragrostis lehman- Mixed-grasses, 282,287 Old world bluestems (Bothriochloa spp.- niana), 127, 404 Mobile screen, 355 Caucasian’; ‘Ganada’; ‘Plains’; ‘WW- Leon, Roland0 J.C., 495 Moina, Randy, 309 Spar?, 40 Levels of a neurotoxic alkaloid in a spe- Molyneux, Russell J., 399 Olson, Kenneth C., 259 cies of low larkspur, 224 Monsen, Stephen B., 30 One-seed juniper (Jumjwus monosperma), Lewis, Clifford E., 460,466 Montana, 144 139 Livestock gains, 483 Mortality, 267 Optimal stocking rate for cow-calf enter- Livestock grazing strategies, 108 Moser, L.E., 416 prises on native range and complemen- Livestock poisoning, 224 Mountain big sagebrush (Artemisia tri- tary improved pastures, 435 Longevity of harvester ant colonies in dentata), 114 Orchard grass (Dactylis glomerata), 53 southern Idaho, 104 Mountain rye (Secale montanum). 30,35 Oregon, 58,251,309 Louisiana, 274 Movement, 416 Orphan fawns, 340 Low larkspur (Delphinium nuttalianum Muir, A.D., 26 Ortiz, Melchor, 153 P&z), 224 Mule deer females (Odocoileus hemio- Osman, Abdelgader, 92 Lowland grasslands, 495,500 nus), 515 Osmotic potential, 48, 150 Owens, M.K., 73 M Mule deer-induced mortality of mountain big sagebrush, 114 Owensby, Clenton B., 483 Mahgoub, El Fatih, 153 Mule deer (Odocoileus hemionus), 122 P Maidencane (Panicum hemitomon Mule deer (Odocoileus hemionw) brows- Schult.), 235 ing, 114 PH, 48 Paruelo, Jose M., 220 Majak, W., 26,224 Murphy, Alfred H., 193 Pastoral societies, 450 Malechek, John C., 255,259,346,477 Murray, Robert B., 16 Pasture size, 73 Management systems, 460 Mycorrhizal fungi, 309 Manglitz, G.R., 253 Pastures (native and improved), 435 Manners, Gary D., 118,509 N Patagonion Semidesert Grassland, 220 Mapping (table), 267 Nass, Roger, D., 25 Paterson, J.A., 426 Mapping table for obtaining plant popu- Nastis, Anastasios, S., 255 Patterns of American licorice seed preda- lation data, A, 267 Native woodland, 66,477 tion by Acanthosceliaks aureolus (Horn) Marmeliero (Croton hemiargyreus), 66, Neal, Donald L., 353 (Coleoptera:Bruchidae) in South Dakota, 477 Near infrared reflectance spectroscopy 342 Martin, F.G., 235, 245 (NIRS), 488 Patton, Bob D., 232 Martin, S. Clark, 291 Nebraska, 416 Paw branco (Auxemma OncocaIyx), 66, Martin-R., M.H., 127 Nebraska Sandhills, 86 477 Maser, Chris, 309 Needleandthread (Stipa comata), 287 Peek, James M., 122 Maser, Zane, 309 Nelson, M.L., 269 Peer review of technical manuscripts, 366 Maturity, 346 Nevada, 150,307,332 Pellets, 16 Mayeux, Jr., Herman, 2 New Mexico, 216 Pendrey, B.M., 488 McArthur, E. Durant, 114 Night-time applications, 313 Perennial warm-season grasses, 127 McCollum, E.T., 178 Nonactive growth, 48 1 Petersen, Joseph L., 3 13 McCollum, F.T., 264 Nonlinear regression, 42 1 Ptister, James A., 118, 346, 509 McDaniel, B., 342 North American Forests, 224 Phenological stages, 472 McDougald, Neil K., 193 North Dakota and Montana, 210 Physical development, 340 McGinnies, W.J., 391 Northcentral Australia, 430 Physical development of orphaned white- McPherson, Guy R., 413 Northcentral Oklahoma, 40 tailed deer fawns in southern Texas, Mechanical, 172 Northcentral Texas, 227 340 Mechanical brush control, 245 Northcentral Washington, 122 Pickett, E.E., 426 Mechanical brush removal, 159 Northeastern Brazil, 66,477 Pieper, Rex D., 92, 139, 153 Mechanical shrub control on flatwoods Northeastern California, 63 Pine-wiregrass vegetation, 460,466 range in south Florida, 245 Northeastern Oklahoma, 264 Pinegrass (Calamagrostis rubescens), 144 Mechanical treatments, 460 Northeastern Mexico, 7 Pinyon pine (pinus edulis), 139 Medusahead (Taeniatherum asperum), Northern and Southern hemispheres, 127 Pivot balance, 167 88 Northern California, 193 Plant-animal mutualism, 309 Mesquite (Prosopis glandulosa), 227 Northern Colorado, 70 Plant community development on petro- Northern Great Plains, 210 leum drill sites in northwestern Wyo- Northern Mexico, 12,216 ming, 372 Northern Texas, 48 Plant distribution surrounding Rocky Norton, M.K., 73 Mountain pinyon pine and oneseed Nutrition, 410 juniper in south-central New Mexico, 139

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 529 Plant populations, 267 Richardson, B.Z., 372 Seasonal burning, 2 Plant responses to pine management and Ries, R.E., 210 Seasonal burning and mowing, 12 deferred-rotation grazing in north Rio Grande Plain, 2, 7,3 17, 340 Seasonal burning and mowing impacts Florida, 460 Riparian (moist) valleys, 399 on Sporobolur wrightii grasslands, 12 Plant succession, 309,372 Robbins, K., 342 Seasonal carbohydrate levels, 144 Plant taxonomy, 350 Roberts, Fred H., 407 Seasonal clipping, 26 Planting depth, 216 Rock (soil profile), 307 Seasonal grazing, 78 Planting depth and soil texture effects on Rock (soil surface), 307 Seasonal stocking of tobosa managed emergence and production of three Rocks, 197 under continuous and rotation graz- alkali sacaton accessions, 216 Rocky Mountains, 70,515 ing, 78 Plowing (web plow), 245 Roder, W., 86 Secar bluebunch wheatgrass as a compet- Poaceae, 350 Roller chopper, 245 itor to medusahead, 83 Poisonous plants, 118,509 Rolling Plains, 227,313,401 Sedimentation, 303 Poisonous plants (miserotoxin), 26 Romo, J.T., 491 Seed dormancy, 335 Porter, Sanford D., 104 Root and shoot biomass, 92 Seed germination, 404 Potential biological predators, 100 Root and shoot growth, 472 Seed midge pest of big bluestem, 253 Potter, Robert L., 3 13 Root and shoot production, 387 Seed morphology, 35 1 Power, J.F., 210 Root containerization for physiological Seed production, 232 Precipitation, 186 studies of shrubs and trees on range- Seeded wheatgrass pastures (Agropyron Predator control, 249,251 land, 90 SPP.), 100 Prediction of soil cover and soil rock for Root distribution, 220,227 Seeding depth, 48 rangeland infiltration, 307 Root excision, 378 Seedling competition between mountain Predicting biomass of five shrub species Root excision and dehydration effects on rye, ‘Hycrest’ crested wheatgrass, and in northeastern California, 63 water uptake in four range species, 378 downy brome, 30 Preserving, 168 Root growth box Secalemontanum GUSS., Seedling development, 378,383 President’s Address: Furthering the Range 30 Seedling root and shoot production, 30 Management Profession, 98 Root growth of Artemtkia tridentata, 332 Seedling root growth, 332 Price, D.L., 227,355 Root mass ratios, 488 Semidesert Grassland, 12, 78, 92, 186, Pricklypear cactus (Opuntiu spp.), 313 Root separations, 488 216,291,404,407,413,430 Production, 269,477 Root systems of two Patagonian shrubs: Semidesert Shrubland, 387 Production (calves), 435 A quantitative description using a geo- Semidesert Woodland, 139,153,197,303 Production (grasses and shrubs), 410 metrical method, 220 Separating leaves, 522 Production (green biomass), 66 Rough fescue (Festuca scabrella), 503 Separating leaves from browse for use in Protein, 12 Roundy, Bruce A., 150,404 nutritional studies with herbivores, 522 Ruby Valley pointvetch (Oxytropis ripa- Severson, Kieth E., 291 Q ria), 399 Shannon index, 466 Quackgrass (Pseudoroegneria spicata), Rumbaugh, M.D., 488 Shea, P.J., 416 488 Rumen fill, 509 Sheep losses, 251 Quadrat frame for back country vegeta- Russian thistle (Salsolu spp.), 155 Short duration, 108 tion sampling, 353 Russian wildrye (Elymus junceus), 378, Shrub above-ground biomass, 66 Quantitative description (geometrical 383 Shrub above-ground biomass estimates, methods), 220 Russian wildrye (Psathyrostachysjunceu), 63 Quinton, D.A., 26 488 Shrub (Chamaespartium tridentatum), R Ruyle, G.B., 404 410 Shrub mortality, 66,114 Raguse, Charles A., 193 S Shrub production and utilization, 153 Rain shelter construction, 83 Sabii’ (Mimosa caesalpinia),, 66 Shrub root distribution, 90 Ralphs, Michael H., 118, 509 Sagebrush (Artemisia spp.), 16 Shrub seed germination, 48 Range condition, 282 Sagebrush (Artemisia triakntata)and (Ar- Shrub seeding growth, 92 Rasmussen, G. Allen, 48,413 temisia wyomingensis), 332 Shrub succession, 122 Ratliff, Raymond D., 353 Sagebrush-forest, 372 Shrub water relations, 83,90 Rawls, Walter J., 307 Salas, Virginia, 155 Shrub water use, 58 Reader expectations: how important to Saltbush, 401 Shrubs (Mulinum spinosum) and (Sene- meet?, 365 ygras&mb (Spartinaaltem~ra), cio filaginoides), 220 Reclamation, 210,372 Simanton, J. Roger, 307 Recruitment, 267 Salts, 150 Simple pivot balance for measuring phy- Redberry juniper (Juniperus pinchotii), Samuel, Marilyn J., 282 tomass in quadrats, 167 413 Sandbur (Cenchrus longtspinus), 86 Simpson index, 466 Reece, P.E., 287 Sanderson, Stewart, C., 114 Simulation and management implications Rego, Francisco, C., 410 Sandoval, F.M., 210 of feral horse grazing on Cumberland Remote areas, 353 Scarification, 48,335 Island, Georgia, 441 Response of false broomweed and asso- Scamecchia, David L., 279 Site., 509 ciated herbaceous species to fire, 2 Schaalje, G. Bruce, 503 Slash and longleaf pine Pinus elliottiand Response of three shrub communities in Schrupp, Donald L., 70,515 P. palustris). 460,466 southeastern Idaho to spring-applied Scifres, C.J., 317 Slope, 197 tebuthiuron, 16 Scott, Gretchen, 337 Small-mammal mycophagy in rangelands Rethman, N.F.G., 127 Season of cutting affects biomass proof central and southeastern Oregon, Rethwisch, M.D., 253 duction by coppicing browse species of 309 Revegetation, 30,210,216 the Brazilian caatinga, 477

530 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1666 Smith, Michael A., 282 Switchgrass (Panicum virgatum). 86 Varner, L.W., 522 Smith, Ed F., 483 Vegetation and livestock management, T Smith, P.W., 372 322 Smooth brome (Bromus inermis), 53 Taboada, Miguel A., 500 Vegetation canopy, 307 Snake River Plain, 328 Taliaferro, C.M., 40 Vegetation changes, 186 Snowberry (Symphoricapos albus), 144 Tall larkspur (Delphinium barbeyi), 118, Vegetation cover, 197 Snowbrush (Ceanothus velutinus), 63 509 Vegetation patterns, 139 Soil depth, 197 Tanaka, John A., 172 Vegetation response to the Santa River Soil development, 372 Tanner, George W., 245,460,466 grazing system, 291 Soil ingestion, 162 Taylor, Jr., Charles A., 108,2% Vegetation sampling (framequadrat), 353 Soil nitrogen accumulation in fertilized Tazi, Mohammed, 88 Viewpoint: Who are those Smiths, 370 pastures of the soil-plant water poten- Teaching aids, 350 Vigor of needleandthread and blue grama tial, 58 Tebuthiuron (N-[5-(l,l-dime- after short duration grazing, 287 Soil texture, 216 thylethyl)-1,3,4-thiadiazol-26+N,N’di- Vogel, K.P., 253 Soil water potential, 30 methylurea), 16 Volume-weight comparisons, 421 Soils, 159 Temperature, 48,88, 150,491 Vora, Robin S., 63 Temperature effects, 35 Some observations from the excavation W of honey mesquite root systems, 227 Temperature requirements for mountain Some vegetation responses to selected rye, ‘Hycrest’ crested wheatgrass, and Waggoner, Jr., J.W., 435 livestock grazing strategies, Edwards downy brome germination, 35 Waller, S.S., 86,287,416 Plateau, Texas, 108 Terry, W. Stephen, 460 Washington, 88,251 South America, 66,220,477,495,500 Test, Peter S., 282 Water and nitrogen effects on growth South Dakota, 342 Theade, John, 251 and allocation patterns of creosote- Southcentral New Mexico, 139,153,155 Thermogradient plate, 35 bush in the northern Chihuahuan Desert, Southeastern Arizona, 12,291,404 Trace elements, 162 387 Southeastern Florida, 235 Threadleaf rubber rabbitbrush (Chryso- Water potential, 239 Southeastern New Mexico, 303 thamnus nauseosus spp. consimilis), Water potential and salt, 207 Southeastern Texas, 274 470 Water stress, 491 Southeastern Wyoming, 282 Threeawn (Aristida spp.), 460,466 Water uptake, 378 Southern British Columbia (Canada), Thurber needlegrass (Stipa thruberiana), Weeping lovegrass (Eragrostis curvula), 26,53 472 22,127,407 Southern Coastal flatwoods, 460 Thurow, Thomas L., 108,296 Weeping lovegrass (Eragrostis curvula Southern Coastal Plain (flatwoods), 245 Tiller growth, 287 ‘Renner >,7 Southern Coastal Plains, 235,274 Timothy (Phleum pratense), 53 Welch, Bruce L., 332 Southern Florida, 245,466 Tobosagrass (Hilaria mutica), 78,407 West, Neil E., 421 Southern Great Plain, 22,40,48,178,207 Toxicological investigations on Ruby Val- West Texas, 216,407,413 Southern Idaho, 159,249, 328 ley pointvetch, 399 Westcentral Utah, 259 Southern New Mexico, 78,92, 186, 197, Trace element intake via soil ingestion in Wester, David B., 407 387 pronghorns and in black-tailed jack- Western Great Plain (shortgrass prairie), Southern Texas, 2,7,317,322,340 rabbits, 162 391 Spatial distribution, 503 Trees, 477 Western Nebraska, 287 Species composition, 282 Tromble, John M., 197 Western Oklahoma, 22,239 Species diversity, 466 Truck-mounted mobile screen for pho- Western Texas, 3 13 Species (plant) composition changes, 108 todigital estimation of whole plant leaf Westfall, Stanley E., 353 Split-tee cannulas, 269 area, A, 355 What do researchers like to read?, 367 Spomer, George G., 395 Tsuchiya, T., 391 Wheat stem sawfly (Cephus cinctus, Hy- Spore dispersal, 309 Turner, Monica Goigel, 441 menoptera: Cephidae) basin wildrye (Elymus cinereus), 328 Spring herbicide application, 16 U Stability of African pastoral ecosystems: Wheatgrass (l’hinopyrum intermedium). UV identification (absorption), 395 488 Alternate paradigms and implications Ueckert, Darrell, N., 3 13 for development, 450 Whisenant, Steven G., 470 Understory plants, 139 White clover (T. repens), 53 Stability of grazed patches on rough Upper Snake River Plain, 16 fescue grasslands, 503 White-tail deer (Odocoileus virginianus). Use of comparative yield and dry-weight- 340 Standing crop, 2,279 rank techniques for monitoring arid Steep slopes, 303 Whitford, W.G., 387 rangeland, 430 Whole plant, 355 Stocking rate, 264,435,483 Use of leader lengths and diameters to Stocking rate effects on intensive early Wiedenfeld, Robert P., 7 estimate production and utilization of Wilcox, Bradford P., 197,303 stocked Flint Hills bluestem range, 483 Cercocarpus breviforus, 153 Stocking rates (cattle), 178 Wilcox, D.G., 127 Use of UV absorption for identifying Wildlife, 114, 162, 340 Stubbendieck, J.L., 86 subspecies of Artemisia tridentata, 395 Subspecies, 395 Wildlife activity, 5 15 Using the Green and Ampt infiltration Wildlife habitat, 122, 3 17 Successional patterns in bitterbrush hab- equation on native and plowed range- itat types in north-central Washington, Wildlife management, 322 land soils, 159 Wildlife movement, 70 122 Utah, 73,100,114,118,307,332,421,509 Suzukida, Margaret, 155 Williams, M. Cobum, 399 Svejcar, Tony J., 239 V Willoughby, Robert L., 193 Swift, David M., 450 Variability within a native stand blue Willms, Walter D., 503 Winter-annual grasslands, 193 Swindel, Benee F., 466 grama, 391 Wilson, A.M., 378,383

JOURNAL OF RANGE MANAGEMENT 41(e), November 1968 531 Wilson, Charles B. 193 Woodruff, Roger A., 249 Y,Z Winter and spring speeding, 53 Workman, John P., 172 Yonker, C.M., 391 Winterfat diaspore morphology, 351 Wright, Henry A., 48,367,413 Young cattle and sheep, 269 Winterfat (Eurotia lanata), 351 Writing for the reader, 362 Young, James A., 364 Wood, John M., 245 Writing for your audience, 364 Youtie, Berta A., 122, 328 Wood, M. Karl, 197,303 Wyoming, 332,435 Ytterbium analysis, 426 Woodland, 410 Zaiglin, Robert E., 340 Zak, J.C., 387

Thanks to Jerry Cox for compiling the Index for Volume 41.

532 JOURNAL OF RANGE MANAGEMENT 41(6), November 1988 Journal of Range Management

Official bimonthly publication of Society for Range Management

Volume 41,1988

533 JOURNAL OF RANGE MANAGEMENT 41(6), November 1966 Table of Contents Volume 41,1988

Number 1, January Number 2, March Response of false broomweed and associated herbaceous species to President’s Address: Furthering the range management profession- fin-Hennan S. Mayeux, Jr.. and Wayne T. Hamilton . , . . . . 2 Jack R. Miller ...... *...... 98 Coastal bermudagrass and Renner lovegrass fertilization responses in Arthropod predation of black grass bugs (Hemiptera: Miridae) in a subtropical climate-Robert P. Wieder&ld ...... 7 Utah ranges-Jaime E. Araya and B. Austin Haws ...... 100 Seasonal burning and mowing impacts on Sporobolus wrightii Longevity of harvester ant colonies in southern Idaho-Sanford D. grasslands-Jerry R. Cox ...... 12 Porter and Clive D. Jorgensen ...... 104 Response of three shrub communities in southeastern Idaho to springapplied tebuthiuron-Robert B. Murray ...... 16 Some vegetation responses to selected livestock grazing strategies, Soil nitrogen accumulation in fertilized pastures of the Southern Edwards Plateau, Texas-Thomas L. Thurow, Wilbert H. Black- Plains- William A. Berg ...... 22 bum, and Charles A. Taylor, Jr...... 108 The effect of clipping on the growth and miserotoxin content of Mule deer-induced mortality of mountain bii sagebrush-E. Durant Columbia milkvetch- W. Majak, D.A. Quinton, HE Douwes, McArthur, A. Clyde Blauer, and Stewart C. Sanderson ...... 114 J. W. Hall, and A. D. Muir ...... *...... 26 Cattle grazing tall larkspur on Utah mountain rangeland-James A. Seedling competition between mountain rye, ‘Hycrest’crested wheat- Pfiiter, Michael H. Ralphs, and Gary D. Manners . . . . . *... 118 grass, and downy brome-Robert A. Buman, Stephen B. Monsen, Successional patterns in bitterbrush habitat types in northcentral and Rollin H. Abemethy ...... 30 Washington-Lierta A. Youtie. Brad Grvflth. and James M. Temperature requirements for mountain rye, ‘Hycrest’crested wheat- 122 grass, and downy brome germination-Robert A. Buman and Peek ...... Rolh’n H. Abernethy ...... *.... 35 The influence of climate and soils on the distribution of four African Chemical composition of old world bluestem grasses as affected by grasses-Jerry B. Cox, M.H. Martin-R,, F.A. Ibarra-F.. J.H. cultivar and maturity-S. M. Dabo, C. M. Taliaferro, S. W. Cok- Fourie, N.F.G.Rethman,andD.G. Wilcox ...... 127 man, F.P. H0m.andP.L Claypool , ...... 40 Plant distribution surrounding Rocky Mountain pinyon pine and Germination requirements of flameleaf sumac-G. Allen Rasmussen oneseed juniper in south-central New Mexico-Susan M. Armen- and Henry A. Wright ...... 48 trout and Rex D. Pieper ...... a...... 139 Establishment of winter versus spring aerial seedings of domestic Influence of forest site on total nonstructural carbohydrate levels of grasses and legumes on logged sites-R.M. Brooke and FB. pinegrass, elk sedge, and snowberry-Janice K Krueger and Ho11 ...... 53 ...... 144 Comparison of water use by Artemisia tridentata spp. wyomingensis Donald J. Bedunah and Chrysothamnus viscidijlotus spp. viscidtflorus-Richard F. Germination responses of desert saltgrass to temperature and osmotic Miller ...... a... 58 potential-Greg J. Cluffand Bruce A. Roundy ...... 150 Predicting biomass of five shrub species in northeastern Califomia- Use of leader lengths and diameters to estimate production and utili- Robin S. Vora ...... *...... 63 zation of Cercocarpus breviflorus-El Fatih Mahgoub, Rex D. Defoliation impacts on coppicing browse species in northeast Brazil- Pteper, and Melchor Ortiz ...... 153 L.H. Hardesty and T W. Box ...... 66 Analysis of Russian thistle (Salsola species) selections for factors Influence of hunting on movements of female mule deer-Roland C. affecting forage nutritional value-James H. Hagernan. James L 70 Kufeld, David C. Bowden and Donald L. Schrupp ...... Fowler, Margaret Suzukida, Virginia Salas, and Roxanne Le- Grazing of crested wheatgrass with particular reference to effects of captain ...... 155 pasture size-R.B. Hacker, B.E. Norton, M.K. Owens, and D.O. Frye ...... 73 Using the Greenand Ampt infiltration equation on native and plowed Seasonal stocking of tobosa managed under continuous and rotation rangeland soils-Neil C. Hutten and Gerald F. Gtfford ...... 159 grazing-D. M. Anderson ...... 78 Trace-element intake via soil ingestion in pronghoms and in black- Design of rain shelters for studying water relations of rangeland tailed jackrabbits-W. John Arthur, III, and Robert J. Gates . . . . shrubs-Pete W. Jacoby, R. James Ansley, and B.K. Law- 162 rence ...... 83 Simple pivot balance for measuring phytomass in quadrats-D. E. Allelopathic effects of sandbur leachate on switchgrass germination: Johnson, M. M. Barman, and Mohamed Ben Ali ...... 167 observations-W. Roder, S.S. Wailer, and J.L. Stub- Collecting, drying, and preserving feces for chemical and microhisto- bendieck ...... *...... 86 logical analysis-Ruy T. Hinnant and M. M. Kothmann . . . . . 168 Sccar bluebunch wheatgrass as a competitor to medusahead-Carl J. Goebel, Mohammed Taxi, and Grant A. Harris ...... 88 Economic optimum big sagebrush control for increasing crested Root containerization for physiological studies of shrubs and trees on wheatgrass production-John A. Tanaka and John P. Work- rangeland-R.J. Ansley. P. W. Jacoby, and B. K. Lawrence . . . 90 man ...... 172 Growth of Gutierrezia sarothrae seedlings in the field-Abdelgader An economic assessment of risk and returns from prescribed burning Osman and Rex D. Pteper ...... *...... 92 on tallgrass prairie-D. J. Bernardo. D. M. Engle, and E. T Book Reviews ...... 94 McCollum ...... 178 Book Reviews ...... 184

534 JOURNAL OF RANGE MANAGEMENT 41 (S), November 1988 Number 3, May Number 4, July Changes in grass basal area and forb densities over a 64-year period on Justification for grazing intensity experiments: Analysing and inter- grassland types of the Jomada Experimental Range-Robert P. preting graxing data-D.I. Bransby, B.E. Conrad, H.M. Dicks, Gibbens and Rekion F. Beck ...... 186 and J. W. Drane ...... 274 Correlation of degreodays with annual herbage yields and livestock Grazing, stocking, and production efficiencies in graxing research- gains-M&in R. George, Charles A. Raguse, W. James Clawson, David L Scarnecchia ...... 279 Charles B. Wilson, Robert L Willoughby, Neil K. McDougald, Cattle, vegetation, and economic responses to grazing systems and Don A. Duncon, and A&red H. Murphy ...... 193 grazing pressure-Richard H. Hart, Marilyn J. Samuel, Peter S. Factors influencing infiltrability of semiarid mountain slopes- Test, and Michael A. Smith ...... 282 Bradford P. Wilcox, M. Karl Wood, and John M. Domble Vigor of needleandthread and blue grama after short duration ...... 197 grazing-p. E. Reece, R P. Bode, and S. S. Weller ...... 287 Effects of temperature, water potential, and sodium chloride on Vegetation response to the Santa Bita grazing system-S. Cktrk Indiangrass germination--Timothy E. Fulbright . . . ..a..... 207 Martin and Kieth E. Severson .**...... 291 Irrigation water for vegetation establishment-RR Ries, F.M. San- Infiltration and interrill erosion responses to selected livestock grax- doval. and J. F. Power ...... *. 210 ing strategies, Edwards Plateau, Texas-lhonms L Thurow, Planting depth and soil texture effects on emergence and production Wilbert H. Blackburn, ond Charles A. Taylor Jr...... 296 Hydrologic impacts of sheep graxing on steep slopes in semiarid Jerry R. Cox ...... *....*.... 216 rangelands-Bradford P. Wilcox and M. K. Wood ...... 303 Root systems of two Patagonian shrubs: A quantitative description Prediction of soil cover and soil rock for rangeland infiitration- using a geometrical method-Robert J. Fernimdea A. and JoircM. Walter J. Rowls, Donald L. Brakensiek, J. Roger Simanton. and Paruelo ...... a 220 Clayton L Hanson ...... 307 Levels of a ncurotoxic alkaloid in a species of low larkspur- Walter Small-mammal mycophagy in rangelands of central and southeastern Majak and Michael Engelsjord ...... 224 Oregon-Chris Marer, Zane M4ser, and Randy Molina . . . . . 309 Some observations from the excavation of honey mesquite root Herbicidal control of pricklypear cactus in western Texas-Joseph L systems-RK. Heitschmidt, R.J. Ansky, S.L Dowhower, P. W. Petersen, Darrell N. Ueckert. and Robert L Potter ...... 313 Jacoby, and D. L Rice ...... 227 Bionomics of patterned herbicidal application for wildlife habitat Effect of burning on seed production of bluebunch wheatgrass, Idaho enhancement-C.J. ScrQres,W. T. Homilton. B.H. Koerth, R. C. fescue, and Columbia needlegrass-Rob D. Putton, M. Hironaku, Flinn. ond R. A. Crane ...... *...... 317 ond Stephen C. Bunting ...... 232 Coyote and bobcat responses to integrated ranch management practi- Effect of defoliation frequency and N-P-K fertilization on maiden- ces in south Texas-Liscr C. Brodky and Daniel B. Fagre . . . . 322 cane-R.S. Kalmbacher and F. G. Martin ...... 235 Association of the wheat stem sawtly with basin wildrye--Berm A. Growth and gas exchange of Andropogon gerardii as infIuenced by YoutieandJamesB. Johnson ...... , . . . . . 328 burning- Tony J. Svejcar and James A. &owning ...... 239 Root growth of Artemisia tridentata-Bruce L. Welch and Tracy Mechanical shrub control on flatwoods range in south Florida- L C. Jacobson ...... 332 George W. Tanner, John M. Wood, Robert S. Kalmbacher, and Indian ricegrass seed damage and germination responses to mechani- Frank G. Martin ...... *.*....a...... 245 cal treatments--lorry W. Griflirh and D. Terrance Booth . . . . 335 Breed comparisons and characteristics of use of livestock guarding Cellulase vs rumcn fluid for in vitro digestibility of mixed diets-Roy dogs-Jeffrey S. Green and Roger A. Woo&uff ...... 249 Lee Dickerson, Jr., Bill E. Dohl. and Gretchen Scott ...... 337 Electric fences for reducing sheep losses to predators-Roger D. Nass Physical development of orphaned white-tailed deer fawns in south- and John Theade ...... 251 em Texas-Stephen Demarais, Robert E. Zaiglin. and Donald A. A seed midge pest of big bluestem-M.R. Carter, G.R. Manglitz, Bamett ...... 340 M.D. Rethwisch. and K. P. Vogel ...... 253 Patterns of American licorice seed predation by Aconthoscelides Estimating digestibility of oak browse diets for goats by in vitro aureolus (Horn) (Colcoptera:Bruchidae) in South Dakota-A. techniques-Amrstarios S. Nastis and John C. Malechek . . . . . 255 Roe, B. McDaniel, and K. Robbins . . ..*...... 342 Heifer nutrition and growth on short duration grazed crested Effect of drying method on the nutritive composition of esophageal wheatgrass-Kenneth C. Olson and John C. Male&k ...... 259 tistula forage samples: influence of maturity-E A. Burritt, J.A. Hcrbage dynamics of tallgrass prairie under short duration graxing- viler, and J. C. Male&k ...... 346 J. E. Brummer. R. L Gillen, and F. T McCollum ...... 264 The grass spiklet formula: an aid in teaching and identification- A mapping table for obtaining plant population data-Jeanne C. Kelly W. Allred and J. Travis Columbus ...... 350 Chambers and R. W. Brown ...... 267 Winterfat diaspore morphology-D. Terrance Booth ...... 351 An inexpensive alternative csophagealcannula for growing steers and A quadrat frame for backcountry vegetation sampling-Donald L wethers-M. L Nelson ...... 269 Neal, Raymond D. Rotliff. and Stanley E. Westfall 0...... 353 Book Reviews ...... *...... 270 A tNCk-mounted mobile screen for photodigital estimation of whole plant leaf area-R.J. Ansley, D. L Rice, B. K. Lawrence, and P. W. Jucoby ...... *...... 355 Book Reviews ...... 359

JOURNAL OF RANGE MANAGEMENT 41 (S), November 1988 535 Number 5, September Number 6, November Writing for the reader-Richard H. Hart ...... 362 Stability of African pastoral ecosystems: Alternate paradigms and Writing for your audienceJames A. Young ...... 364 implications for development-James E. Ellis and David M. Swtft Reader expectations: How important to meet?-David A. Fisch- ...... 450 bath ...... 365 Plant responses to pine management and deferred-rotation grazing in north Florida-Clifford E. Lewis, George W. Tanner, and W. Peer review of technical manuscripts-Gary W. Frasier ...... 366 Stephen Terry ...... 460 What do researchers like to read?--Henry A. Wright 367 ...... Species diversity and diversity profiles: Concept, measurement, and Obligations and expectations of your peers: manuscript review at the application to timber and range management-Clifford E. Lewis, Journal of Range Management-N. T Hobbs ...... 368 Benee F. Swindel, and George W. Tanner ...... 466 Viewpoint: Who are those Smiths?-A&n A. Beetle ...... 370 Control of threadleaf rubber rabbitbrush with herbicides-Steven G. Plant community development on petroleum drill sites in northwest- Whisenant ...... 470 em Wyoming-P. W. Smith, E.J. Depuit, and B.Z. Richard- Defoliation of Thurber needlegrass:herbage and root responses- son ...... *...... 372 David Ganskopp ...... 472 Root excision and dehydration effects on water uptake in four range Season of cutting affects biomass production by coppicing browse species of the Brazilian caatinga-Linda H. Hardesty, Thadis W. species-M. Bassiri, A.M. Wilson, and R. Grami ...... 378 Box, and John C. Malechek ...... 477 Dehydration effects on seedling development of four range species- Effects of dormant-season herbage removal on Flint Hills rangeland- M. Bass&i, A.M. Wilson, andB. Grami ...... , . . . . 383 Lisa M. Awn and Clenton E. Gwensby ...... 481 Water and nitrogen effects on growth and allocation patterns of Stocking rate effects on intensive-early stocked Flint Hills bluestem creosotebush in northern Chihuahuan Desert-F. M. Fisher, J. C. range-Clenton E. Owensby, Robert Cochran, and Ed F. Zak, G.L Cunningham, and W.G. Whitford ...... 387 Smith ...... 483 Variability within a native stand of blue grama- W.J. McGinnies, Determination of root mass ratios in alfalfa-grass mixtures using near W.A. Laycock, T. Tsuchiya, C.M. Yonker, and D.A. Ed- infrared reflectance spectroscopy-M. D. Rumbaugh, D. H. Clark, munds ...... 391 and B. M. Pendery ...... 488 Use of UV absorption for identifying subspecies of Artemisia Germination of green and gray rubber rabbitbrush and their establishment on coal mined land-J. T. Romo and L.E. Eddle- tridentata-George G. Spomer and Douglass M. Hender- man ...... 491 son ...... 395 Floristic changes induced by flooding on grazed and ungrazed low- Toxicological investigations on Ruby Valley pointvetch-M. Coburn land grasslands in Argentina-Enriqw J. Chaneton, Jose M. Wilhams and Russell J. Molyneux . . . , ...... 399 Facelli. and Rolando J. C. Leon ...... 495 Comparative chemical composition of armed saltbush and fourwing Grazing effects of the bulk density in a Natraquoll of the Flooding saltbush-Andres Garsa, Jr., and Dmothy E. Fulbright . . . . . 401 Pampa of Argentina-Miguel A. Toboada and Ratil S. Effects of burning on germinability of Lehmann lovegrass-G.B. Lavado- ...... 500 Ruyle. B.A. Rouna’y, and J. R. Cox ...... 404 Stability of grazed patches on rough fescue grasslands- Walter D. Willms, John F. Dormaar, and G. Bruce Schaabe 503 Fire effects on tobosagrass and weeping lovegrass-Fred H. Roberts, ...... Effects of phenology, site, and rumen fill on tall larkspur consumption Carlton M. Britton, David B. Wester, and Robert G. Clark . . . 407 by cattle-James A. pfiter. Gary D. Manners, Michael H. Effects of prescribed fire on Chamaespartium tridentatum (L.)P. Ra$hs, Zhaol Xiao Hong, and Mark A. Lane ...... 509 Gibbs) in Pinus pinaster (Aiton) forests-Francisco C. Rego. Habitat selection and activity patterns of female mule deer in the Stephen C. Bunting, and M. G. Barreira ...... 410 Front Range, Colorado-Roland C. Kufeld. David C. Bowden, Economic comparison of aerial and ground ignition for rangeland and Donald L. Schrupp ...... 515 prescribed tires-G. Allen Rasmussen, Guy R. McPherson, and Separating leaves from browse for use in nutritional studies with Henry A. Wright ...... 413 herbivores-J. F. Gallagher, T G. Barnes, and L. W. Vamer . . . 522 Atrazine dissipation and off-plot movement in a Nebraska sandhills Book Reviews ...... subirrigated meadow-J.J. Brejda, P.J. Shea, LE. Moser, and SS. Waller ...... *... 416 Estimation of phytomass for ungrazed created wheatgrass plants using allometric equations-Patricia S. Johnson, Craig L John- son, and Neil E. West ...... 421 Methods of ytterbium analysis for predicting fecal output and flow rate constants in cattle--k P Coffey, E. E. Pickett, J.A. Paterson, C. W. Hunt, and S. J. Miller ...... 426 The use of comparative yield and dry-weight-rank techniques for monitoring arid rangeland-M. H. Friedel, V.H. Chewings, and G. N. Bastin ...... 430 Optimal stocking rate for cowsalf enterprises on native range and complementary improved pastures-R. H. Hart, J W. Waggoner, 435Jr., TG. Dunn, CC. Kaltenbach, and L.D. Adams ...... Simulation and management implications of feral horse grazing on Cumberland Island, George-Monica Goigel Turner , ...... 441 Book Reviews ...... 447

JOURNAL OF RANGE MANAGEMENT 41(6), November 1988