The Impact of Cattle Introduction on Arid Sheeplands

7^ ro -?'5

TI{E IMPACT OF CATTI.,E TNTRODUCT]ON CN ARIÐ SHEEPLANDS

Timothy James Fatchen, B. Sc. (tlons . ) ,

Department of Botany, Univer"sity of Adelaide

Thesis submitted fon the degree of Doctor of PhiJ-osophy Uiriversity of Ade-laide

Febr"uary, 1975, SUMMARY

This thesis investigates the effect on vegetation of recent cattle introduction to the ar"id sheeplands of South Austnalia. While the effects of sheep on these nangelands are wel-l documented, those of cattl-e a:re not known. The study thus nepnesents the initial examination of the cattle impact.

The main aim wä.s to dete:rmine the extent to which catt-le nodified the existing sheep-induced vegetation pattenn. For the sho::t tenm, thíswas achieved by a two-yealr companison of vegetations, separatei-y grazed by cattl-e and sheep, on a rep::esentative station. The companison was begun soon after the first introduction of cattle. Stocking effects were isolated fi:om background vegetation pattern by appJ-ication of the piospher"e concept. Detection of pattenn in density and estimated biomass of major" pastu:re species was emphasized, with negression and Pnincipal

Components analyses being employed. The ::esults were complicated by the exceptionalJ-y heavy and extended rains which fel-l- duning much of the obse::vation period. The nesul-ts norretheless demonstrated that shor"t-tenm stocking effects diffened fon sheep and cattle, but became obscured in periods of pastune g::owth.

No basis was found for the pnediction of the long-term trend fnom the shor.t-term obsenvations. Less intensive comparisons, with both sheep and vingin sites, wer"e used to infen long-term changes in vegetation nesulting fnom cattle stocking. Stocking effects were iso.l-ated by examination of mapped nelative abundances of specíes, following a caneful matching of site landscapes. The nesufts indicated

changes simil-a-n to those neponted to occun with sheep stocking. The causes of vegetation pattenn and change hlelle considened,

but the t?eseanch needed to aflot^¡ definitive statement of these

causes was beyond the scope of the study. Reponts on possible factons such as stock diet and behavioun are' therefone, re-' viewed in detail. on the basis of these reponts, it was concluded that the rnajonity of the stocking effects obsenved did not nesult fnom the pr:eferential selection of species by the grazing animal. This is contnany to common assumption.

Mechanical- distr::rbance of soil and plants ín the cou::se of gnazing was pnoposed as the majon detenminant of stock-induced vegetation change. This hypothesis r.Ias pantially tested, but - with equivocal nesuft. The implications of the outcome of this study ar"e discussed,

as negands both the future of the south Aust:ralían sheeplands and the requi:rements fo:r future study. DECLARATTON

This thesis contains no matenial- which has been accepted for the award. of any other degnee on diploma in any Univensity. To the best of my knowledge and belief this thesis contains no matenial pneviously published on w::itten by anothen penson, except whene due nefe::ence is rnade in the text.

T ,J Fatchen ACKNOtr'ILEDGEMENTS

I wish to thank Dn R.T. Lange fon his suppont during this study. Mn Andrew Bnown and Mr D.H. Cna¡e also gave much assis- tance through both advice and introductions at the outset. 0f the many pastor"alists who tol-erated my pnesence and questions, I eepecially wish to thank Mn Hugh Macfachlan fon his co-operation :reganding the observations on Mt Victon station, and his staff, Mn and Mns Fned Sleep, who kept a weather eye on mel M:r l'lalten Fnench and sons of Pennatty station, and Mn Tom All-Íson a¡d Mn Jack All-íson of Roxby Downs station fon th-eir help in other fiel-d studies; and Leste:: Lond, Ga:rnham Skipper"' and Dudley

Ful-Ien. Al-1 the pastonalists I met showed a wel-come intenest and a great hospitality, and gave me no littl-e education in the intnicacies of management in the pastoral countny.

My thanks must also go to Dr- Geor"ge Ganf and Dz' Susan Ba¡ken who cnitical-ly read pant of tb-e manuscript; to Daphne McEwen who typed the text; and to Debonah, Fatchen, who with patience and pensevenance not only withstood my plreoccupations but also dnafted the final diagnams.

The research was supported by a Commonwealth Postgnaduate

Student Awand and funds fnom the Sandlands Bequest to the Uni- vensity of AdelaiCe. NOTE ON PLANT IDENTIFTCATION AND NOMENCLATURE

Voucher: specimens of all species named in this thesis have been ì-odged at the State Her^ba::ium of South Australia, Adelaide (AD) and at the Botany Depantment, Univensity of Adelaide. Most identifications vlene verified by Dn Hj. Eichl-en and staff at the Henbanium. The remainden, involving vegetative specimens on1y, were ver-ified by comparison with AD collections.

All names are those given by Black (fg+e-t9S7), as amended by Eichler (1965). The appnopniate authon citations ane given in these wonks. CONTENTS

SUMMARY

DECLARATION

ACIO{0T/¡LEDGEMENTS NOTE ON PLANT IDENTIFICATION AND NO}ÍENCLATURE

Chapten Page

t INTRODUCTION t

t-T CAUSE AND PURPOSE OF RESEARCH t

T_2 PLAN OF RXSEARCH I

1-3 HISTORY AND ANMINISTRÂTION OF RANGELANDS TN SOUTH AUSTR,ALIA 4 1- 31 Intz,oduction 4 7-32 Hístozg 4 1-33 AdmLnistration 6 1-34 Ihe r.ecent íntxoduction of eattLe on æíd z'øtgeLands I

1-4 DEMARCAT]ON AND CLIMATE OF SOUTH AUSTRALIAN RANGELANDS 10 1-41 Demarcation 10 7-42 CLimate 10

1-5 VEGETATION OF SOUTH AUSTRALIAN RANGELANDS t2 t-sr Main diuisíons t2 7-52 Trend ín grazed nøtgeLand uegetat'ion 13 1-53 Regeneration of chenopod shrubLøtds foLLouing 16 ouez'grazing 1-54 Contrasted AnerLcan øtd AustraLíæt sítuations 16

1_6 ASPECTS OF SHEEP AND CATTLE PERFORMANCE ON ARID RANGELANDS !7 1-61 fntroduetion t7 t-62 Aspects of physíology 18 1-63 Dietæy seleetion 20 1-64 Aspects of z,anging beharsíoz.tt 25

7-7 THE PIOSPHERX: STOCK-]NDUCED VEGETATION PATTERN 26 t-77 The concept of the piosphere 26 7-72 Tlte fsyn of pLont/distøtce reLationships in the piosphere 29

1-B THE CAUSAT'ION OF PATTERN AND TREND 31 Chapter Page

2 CATTLE ON SOUTH AUSTRALIAN SHEEPLANDS 34

2_T VEGETATIONS GRAZED BY CATTLE I^I]THIN THE DOG FENCE 34

2_2 MANAGEI'ENT SYSTEMS 35 2-21 Møtagement north qnd south of the dog føtce 35 2-22 Pz,obLems in ehartgíng to eattLe from sheep 39 2-23 CattLe husbandry on former sheepLm.d 4t

2_3 PASTORALISTSI OPINIONS ON THE EFFECT OF CATTLE +2

2_4 ECONOMICS OF CATTLE INTRODUCTION 43

3 COMPARISON OF SHEEP AND CATTLE PIOSPHERES OVER A SHORT TERM: METFIODS 46

3_1 TNTRODUCTTON 46

3-2 APPROACH TO TrrE MAJOR STUDY (SAUplrne BY QUADRATS) 50

3_3 PASTURE POPULATIONS STUDIED AND PARAMETERS ESTIMATED 52 3-31 Mqjor species 52 3-32 Minor speeies 59

3_4 SAMPLING SYSTEM 62 3-41 SatnpLing for catbLe/sheep cornparison 62 3-42 SønpLíng for background dynønics 64 3-+3 Quaárat size and shape 66

3-5 ANALYSTS I'ÍETHODS 67 3-51 Influence øtaLysís 67 3-52 CozveLation øtd regression 69 3-53 PrineípaL Conrponents AnaLysis 7t

3_6 M]NOR STUDY (SEUPITNE BY STEREOPHOTOPOINTS) 73 3-61 GeneraL approacLt 73 3-62 Photogtaphíc methods 73 3-63 PLant popuLatiorts studied øtd paxønetez's estimated 75 3-64 SarnpLing system 76 3-65 AnaLysis nethod 17 Chapten Page

3_7 RAINFALL MONITORTNG 78

3_8 STOCK MONITORING 78 3-81 Stoeking LeueLs 78 3-82 Mouements 79

4 COMPARISON OF SHEEP AND CATTLE P]OSPHERES OVER A SHORT TERM: RESULTS 80'

4-t RAINFALL OVER THE STUDY PER]OD 80

+_2 STOCK MONITORING 81 4-2t Stoekíng v,ates 81 +-22 Stock mouements 83

4-3 ABUNDANCE AND DYNAMICS OF MAJOR PASTURE SPEC]ES 83 4-31 In the ahsqnee of grazing 83 4-32 Grazed cornpæed uith wtgrazed 86 4-33 The effect of the 1973-1-974 rains 87 4-34 CorreLatíons üithin qnd betueen speeies B7

4_4 RËGRESSION ANALYSES: MAJOR SPECTES 89 4-41 Intz,oductíon 89 4-42 Lineæ and non-Linear negz.essions of totaL bíons,ss on distutce 91- 4-43 Regnessíon øtaLyses foz'Stipa nitida 94 4-44 Regresston øtlaLyses fon Enneapogon spp. 97 4-45 Regz,ession øtaLgses foz' Arístida contonta 103 4.46 Regression anaLgses foz'Kochia astnotnicha aduLts 108 +-47 Regressíoù artaLyses foz' Kochia asfi:otricha iuueniLes !12 4-tr'B Regression anaLyses foz' Kochia pyramidata 113 4-49 Ge-nexal discussion on vegression øtalyses 115

4-5 ABUNDANCE OF MINOR SPECIES 120

4_6 INFLUENCE ANALYSES !23

4-7 PRINCIPAL COMPONENTS ANALYSES 724 4-77 ResuLts t24 4-72 Discussion 126 Chapten Page

4-8 ANALYSES OF PHOTOPOINT DATA (UTNON STUDY) t3r 4-81 The effects of cattLe øtd sheep gtazing on pLøtt habits 131 4-82 Brousing by cattLe 131 4-83 Shruh øtd grass dynønícs 133 4-84 Píosphere patterms 13s

4-9 DISCUSSION AND CONCLUSIONS 135 4-91 The reLatiue intportance of stocking and oth,er influences 135 4-92 Linea.r. relationships uithtin the piosphez'e 737 4-93 PossibLe causes of stock-induced uegetatíon pattern 138 4-94 0n monítoring stock-induced chønges 139 4-95 The shot't term inrpaet of eattLe inttoduetion 741

5 LONG TERM CHANGE IN VEGETATION GRAZED BY CATTLE 144

5-1 INTRODUCTION 144

5-2 METHODS 145 5-2t SeLection of sarnPLe æed.s 145 5-22 Pavametev's used ín sønPLíng 146 5-23 SørcpLíng sgstem ]-47 5-24 Data presentation and øLaLysis 148 5-25 CriterLa for inferring Long tetrn ehange 148 5-3 RESULTS: VTRGIN/CATTLE COMPARISON 150 5-31 Stocking hóstory of sønpLe æeq,s 150 5-32 Descriptíon of matched Løndscapes 151 5-33 Abundøtce of pasture species 1s4 5-34 Assoeiation of pastu.re speeíes 1sB

5-4 DISCUSSION: VIRGIN/CATTLE COMPARISON 159

5-5 RESULTS: CATTLE/SHEEP COMPARISON 762 5-51 Stoeking hístory of sønpLe areas t62 5-52 Deseription of matched Løtdscapes 163 5-53 Abmdønce of Pasture sPecíes 165 5-54 Associatídn of pashtre specíes 167

5-6 DISCUSSI0I'T: CATTLE/SHEEP COMPARISON !67

5_7 SUMMARY AND CONCLUSIONS t77 Chapten Page

6 RELAT]ONSHIPS BET}IEEN SPECIES ABUNDANCES AND SOIL DISTURBANCE 1-73

6_1 INTRODUCTION t73

6_2 DESCR]PT]ON OF THE DESERT 174 6-2t CLinrzte t74 6-22 Physiogz,aphy øtd soíLs L76 6-23 Vegetation 176 6-24 Domestíe grazing in the Desert 177

6_3 COMPARABILITY OF THE DESERT AND OTHER STUDY AREAS t78

6-4 T,IETHODS !78 6-41 lot obseruation of species íncidences acrloss dwte profiles !78 6-+2 lor obseruatíon of species densíties aeross t79 dune profiLes

6-5 RESULTS 180 6-51 Influence øtaLgses on speeies incidenees 180 6-52 Density distributíons aLong topographic profiles 181

6-6 DISCUSSION t82 6- 61 The effect of soiL dístwbartce !82 6-62 Conrpar,íson uíth z'esuLts from Petnatty aztd Rorhy Dot'ms sur¿ueAs 183 6-63 ConeLusíons 187

7 SO}IE EFFECTS OF CATTLE ON PERENNIAL CHBIOPODS 188

7-7 INTRODUCTION 188

7-2 POPULATIO].I CiÏA]{GE IN ,{. VESTCARIA AND .K. ASTROTRTCHA 188 7-27 Methods 188 7-22 Stoekíng histoty of stttÅy æea 189 7-23 ResuLts 190 7-24 Discztssion 192

7-3 A CROSS_FENCE COMPARISON OF BUSH DENSTW: CATTLE VS UNGRAZED t92

7_4 CONCLUSIONS 194 Chapten Page

B GENERAL DTSCUSSTON 195

8-1 INTRODUCTION 195

8-2 TTIE VALIDITY OF INFERENCES REGARDING LONG TERM PASTUR.E CHANGE 196

B-3 DISPARUTY BET!.IEEN SHORT-TERM AND LONG-TERM OUTCOMES 198

B-4 PIOSPHERE PATTERN 200

8-5 THE CAUSÉ OF STOCK-INDUCED VEGETAT]ON CHANGE 20t

8-6 PASTORAL ]MPLICATIONS 204

B-7 CONCLUSIONS 207

BIBLIOGRAPHY 2lo

APPEÑDTCES CHAPTER 1

TNTRODUCTTON

]--1 CAUSE AND PURPOSE OF RESEARCH.

Significant cattl-e hends we:re introduced on South- Austnalian arid sheeplands in 1970, after a centuny of almost exclusive sheep gnazing (see fig. 1.1). The ef,fect of sheep on the vegetation of these ::angelands was weff documented, but l-ittle was known of the effect of cattle. Ignorance of th-e anid environment in gener:al, and the effect of sheep in pa:rticulan, had l-ed to injudicious stocking in the past. This in tunn nesul-ted in widespr:ead pastune degnadation and soiJ enosion, wittr- concommittant losses i¡¡ animat pnoduction. Tlre introduction of cattfe was neganded as having much the same potential fon destnuction (".g. by Lange, 1971-), in the absence of information about the animals t effect on the vegetations concerned. The nesear"ch, reponted in this thesis, represented the finst attempt to pnovide the missing information.

The main aims wene: 1. To identify the inpact of cattle on the vegetations of the arid forme:: sheeplands; and 2. To pr"edict the long-term trends in such vegetations

und.en cattle grazing.

The secondany aim was: to dete::mine the pninary causes of the changes in vegetation nesulting f::om cattl-e intnoduction.

I_2 PLAN OF RESEARCH.

The r"eseanch- plan is shown irr figu:re 1.2 A litenatune review' summa-nized in the remainden of this chapten, was undentaken at the -1- FIGUR¡ 1.1

SOUTH AUSTRALTA

STATION LOCALTTY MAP

1. Curnamona 12. Nonning

2. Edeowie 13. Oakden Hills

3. Fnome Downs 14. Pernatty

lr Kolendo 15. Roxby Downs

5. Koonamone 16. Siam

6. Mahanewo t7. South Gap

7. Manunda 18. Teetulpa

8. McCoyrs Ì1e11 19. Thunlga

o Mt Victon 20. ![abricoola

10. Mulyunganie 27. Yudnap.inna tt." Netley Gap .Oodnod o(io

-1 L-

\- 2 \i \ r0

21 5 I 20 t2 PI 'e 19- gu sto 7

ù' de

Arid sheep grozíng oreos

Arid cottle çrozing oreos

Dog fence

0 150 300 km 2 commencement of the study. Extensive neviews exist fon most of the aspects discussed, and there seemed l-ittle point in newniting these neviews. Instead, a condensed outline of the backgr:ound is given, adequate fo:r an undenstanding of the genenal situation, but details ar:e not stnessed unl-ess they tr-ave an inrnediate bearing on the panticuÌan pr"oblem unden investigation.

Thnee courses of neseanch were necessary. Before commencing detail-ed investigations, infonmation was nequired, on the mannen in which cattle were nun on formen sheeplands, the pastrrne t5lpes on which they were nun, and the pastonalistst attitudes. The information was gained duning widespnead travel- through the sheeplands, fnom subjective observation and discussion among pastonalists; ald is summanized in chapter 2. The inpact of cattle on vegetation was detenmined by a d.etailed companison, over a two-year period, of two nepresentative pastunes; one newly under cattfe, the othen stil-l unden sheep. The r"esu-Its were expected to show the beginnings of l-ong-term pastur"e divergences between gnazing regimes, in addition to allowing identification of the immediate impact of cattle. The methods are descr:ibed in chapten 3, and the results ane given in chapte:r 4.

The outcome of long-tenm cattle-stocking !¡as observed' fon isolated situations wher"e cattle had been run on fonmen sheeplands pnion to the majon introduction. Tnends in vegetation unden cattl-e wene infenred langely by companisons of pasture speciest

abundances, between cattle-grazed sites and both sheep-g::azed and vingin sites. The two main observations are neponted in

chr-apten 5. In the course of th-ese observations, mechanical I

Fig. t. 2 STUDY PLAN

INÌROOUCÌORV

Eoclgtound lnlormot¡m

Chopter I

PO'TENÍIAL CAUsEs CONTEXfUAL SHOffI.TERM LONG-lERM IMPACT ON oF sTocK lMlìac1 5fu0tE5 OF CAIILE INfROOUCTION VEGETÀTION Relo¡onshrps Coltle h 5Â rongè- ond tonds hrstory ond ol sheep otrponsff between mtrogÐent ond coltle prosphere. ond cotlle distwbcnce eflecls Chopter 6 CtûÞtq 2 Methods Chopt{ 5 t"'Y _t _ ol cottle ü Resulls chenop.ods

Chopler l. Chopts 'l

c€r.tERAL DtSCUSSION

choptq I distur"bance (tnampling of plants, and distunbance of soil-s ) was pr.oposed as the main determinant of long-term vegetation changes ::elating to stock. The study of pJ-ant distributions in relation to soil stability, nepol?ted in chapte:r 6, provided a par:tial test of this hypothesis. In addition to the obser:vations of chapter 5, two mino:r studies were undentaken, to detenmine the long-term effect of cattl-e stockíng on pel?ennial chenopod species (chapten 7). The thnee counses ane dra\^¡n together: in chapter 8, in which thein overall- l?esults and impJ-ications are discussed, and the final conclusions given.

The neseanch was aimed at p::oblern solution, not elucidation

on compat?ison of techniques. The various techniques and analyses wene employed as me¿ü-ts to this end. Methods we:re selected, firom the gneat vaniety avail-able, on the grounds of familianity' conven- ience and efficiency. In some cases, equally efficient and convenient methods, other than the one used, were avail-abte (..g. methods of biomass estimation; section 3- 3 ). In such cases, justification is given for the eventual choice. The amount of data gathened, and the number of analyses perfonmed, cou]d not be pnesented in ful.l- without making the thesis unwieldy. The procedures which have been adopted ane:

l-. Whene statistical tests wene pe::formed, only those nesu]ts found significant ar:e included; the rest ane stated to be nonsignificant, but no funthen info::mation is given; 2. Where analyses were perfòr"med by subjective app:raisal- of data, the oniginal data is incl-uded. 4

T-3 HISTORY AND ADMINISTRATION OI'RANGELANDS IN SOUTHERN AUSTRALÏA.

I-3I fnfu,oductíon.

The vegetation of the South Austral-ian sheeplands fonms part of a suite of nefated associations, extending east into New South- l'lales and west into !{estern Austnal-ia. The S.A. and N.S.W. vegetations both have been grazed by sheep fon over a centuuy, and much of the nesearch on the effect of sheep in N.S.r¿'I. is dinectly nelevant to the S.A. situation. The Statesr sheeplands nonethe.l-ess have diffenences both in their. histories and in thein administrat- ions, as outlined in the next two sections. These differences alle neflected in diffenences in the pnesent condition of the vegetation.

Hence the reseanch was limited to S.4., despite the simil-ar incr:ease in cattl-e on N.S.l{. arid sheeplands.

I-32 Hístory. The early histony of anid N.S.ll. is more fuJ-ly documented than that of anid S.A. The Majon reviews ane Cain (fg0Z) fon the

Westenn Division (a:rid N.S.W. west of the Darling Riven), Williams (fg0Z) fon the semi-a:rid Rive::ina (the alluvial plain bounded by the Munnay and Lachl-an nivens), and Heathcote (fg6S) fon the Itarnego district (nonthwest N.S.ÌÍ. and southwest Queensland). The two majon neviews fon S.A. ane ThneadgiJ-l (LSZZ) and Bowes (1968), both of which consider the State as a who1e, not only the rangelands.

Pastoral expansion into the anid zone of both States commenced in the 1850ts. Catt.l-e wer"e introduced occasionally, but the main 5 emphasis was crr sheep. The pnesent boundar.ies of gnazing lands were reached by 1870 (Robents, 1924; Th::eadgill, 1922). Penman- ent water was fneely availabl_e only in N.S.W., fnom spnings on wel-ls. In S.4., neliable and accessj-bl_e sounces wer:e pnesent onJ-y in the Fl-inder.s Ranges , the 01ar"y spuu, and on the edge of the Gneat Antesian Basin (fig. 1.3). Thus much of the gnaz]ng in S.A. was inter"mittent until- the tu:r'n of the centu:ry, with stock nemoved at the onset of drought (see Banker, l-970).

Pnovisíon of Ia:rge sunface catchments and sub-anrtesian bones became technical-ly and financial-ly feasibÌe at this time. 0n the othen hand, gnazing irl N.S.W. was continuous. i,tithout pnior exper-ience in such countny, stocking rates wene set too high for" pasture maintenance. I,I-idespr:ead destruction of the perennial shnub pastur.es ensued. 0n1y in the semi-arid Rivenina wene the nesultant pastures aJ-so pe::ennial: elsewhene, the shrubs wene ::eplaced by ephemerals , abl-e neithe:r to suppont f1ocks in d:rought, nor to pnovide adequate pnotection against soil- enosion. Pasture destnuction and soil- enosion reached a peak in the sevelîe dr"ought of the 1890rs. The nesul-ting cnash in sheep population was so ser:ious as to warrant a Royal Commission of enqui:ry (Robentsr 1924). The watened parts of the S.A. nangel-ands suffened simila::ly. In a contemponary account Dixon (fggZ) stated that damage fnom or/el1- stocking was "so gneat as to entail a national- l-oss". Like statements r^rene made even eanlien, by Goyden (Surveyon-General , in S.A. Par-lt. Papen No. 78, l-865-1866). Howeven, lange tr.acts of penennial pastu:res nemained elsewhere, because of the inter.mittent stocking. Stock numbens have neached tb-e level-s of 1BBO-I-B9O Fig. 1 .3

S. A. RANGELAND WATER RESOURCES

GREAT ARTE SIAN BASIN

\ ù

BASIN * EUCLA BASIN T

RIE T* RR ER R A. ¡t

COWELL BASll'l

Ground woter bosins * - mointy subqrtesiorr MURRAY BASIN ADELAI * Hf-J Lond over 300m * --t-- Dog fence

0 150 300 6 in only ten of the subsequent yeans (Bar-nard, 1969). Penny

(1969) consider"ed that this may imply a neduction of the pr"od.- uctivity, nesulting fnom the ea::Iy ove::grazing. ff so, the situation plresents a panallel- r.Iith th_e Gneat plains, U.S.A., whene injudicious use of fi:agile lands also nesulted in sevene damage and lowened pnoductivity (see Bennett et al_, 1938).

By 1920, most of th S.A. r:angelands had impnoved waten resounces which all-owed contj,r¡uation of stocking thnough dr:oughts (..g. as ::eponted by the Mutoonoo Pastonal Co., 1g51); but the earlien experiences l¡Iene not nemembened. On the el?lloneous notion that canrying capacity is dependent only on the level of rrimpnovements'r such as fencing and waten facilíties, much of the stocking between 1920 and 1940 was still- above the optimum fon pastune maintenance. The nesul-t was again degnadation and enosion (Ratctiffe, 1936). Ovenstocking was al-so causing funthen damage in Western N.S.!i. (Beadle, 1948), exacenbated by the administration system. Since 1940, the situation has largely stabil-ised (Pe:rny, t97O), although some cases nemain of continuing degnadation.

7-33 Adn¿nistratíon.

In both arleas, the oniginal holdings wene lange (e.g. in S.4., blocks of 100mi2, with lessees taJcing sevenal contiguous blocks at a time--data fnom Richar.dsonr1925). Whil-e leases nemained la:rge in S.4., the r^restel?n N.S.W. holdings r^rene subdivided on the basis of r'living areasrl eanly this century (Beadle, 1948). War"ing (1969) outl-ined the need for lange holdings on Austnalian nangelands; arid 7 pastoralism is necessarily extensive because of th-e low netunn pen unit anea. Fnom an economic viewpoint, only lange holdings p:rovide adequate netunns on bottr l-aboun and investment: fnom an ecological viewpoÍnt, they ane vital for" landscape stabil-ity, penmitting lighten stocking nates than small holdÍngs fon the same ovenall- netunn. Reid (fgOg) sumnarized the situation, as foIlows. The avelîage station a-neas in N.S.W. and S.A. r:espectively wene 29'OOO acres (l-2,oooha) and 58r70o acnes (24roooha), but sheep numbens were 51460 and 3r4l-0. CHis figunes include both ar:id and semi- anid nanges). The S.A. enter"pnises were the mone efficient, with Iess agistment and fodden irnport, and smallen vaniation in sheep numbens, than in the N.S.W. enterpr.ises. This nesulted f:rom both the mone extensive gnazíng and the highen quality of the pastunes in S.A. Land tenure in most of the S.A. nangelands is by 42 yean Iease, with stocking limits set by Govennment neguÌation, and with pnovision fon nevocation of leases if damaged by over:stocking

(Pastona1 Act, 1936-1969). Some cases exist of leases being fone- cl-osed on these gnounds. On the othen hand, the leases of anid N.S.lI . arle perpetual , with little Gover"nment contnol , and littl-e pnovision against damage (Heathcote, ì-969). The provisions in

S.A. have probably had some effect in maintenance of pastur:es, despite the nebulous chanacten of many of the nestrictions.

(Lange, l-971-, reports the difficulty of policing the negulations). Ihe N.S.W. legisÌation has, if anything, impained pastu.ne consenvation. B

I-3+ The recent íntz.oductíon of eattLe on ayfd sVrceplarLds.

The s.A. anid sheepJ-ands Ì¡/ere sepanated histonically into sheep and cattre ranges, wittr, sheep in the southenn rangerands, and cattle in the nor:ttrer"n r"angeJ_ands. Sheep-stocking was favouned in eanì-y settl-ement: an export manket existed fon woól, but only a small- l-ocal ma:rket for beef (Robents, 1924). Funther, the me::ino sheep, most commonry nun, couì-d rnaintain pnoduction on chenopod pastunes to a g:reaten extent than the available cattle bneeds (section I-62, see also chapten 2). Howeven, viable sheep entenp:rise nequined both the contnol of the wild dog (dingo)r and the maximum util-isation of pastures. The constnuction of dog- pnoof and subdivisional- fences, md of numenous water: facil-ities, was feasible only in the southern nangelands; whene the penennial pastures, if not abused, alJ-owed continuity of stocking, with

:retunns justifying expenditune. whene such development was not feasibl-e, cattle wene lrun. The division between the sheep and cattl-e areas was highlighted by the erection of the Dog Fence (fÍg. 1.1), consolidated fnorn eanl-ien venmin-proof fences by act of Parl-iament (Dog Fence Act, 1946). Until_ l_970, this fence nepnesented a tnue demancation, with most of the significant cattle hends troutsidet' (nonth)and most of the significant sheep hends Itinsider' (south) th-e fence.

The uneconomic wool- pnices of l-970-l-971, and the resultant nunal r:ecession, 1ed to an Austnatian-wide upsulrge in cattle pnoduction, and a connesponding dscreâse in sheep pnoduction

(see Thompson, 1971; also figs, l-.4, 1.5). fn S.A., this was Fig. 1 .4 AVERAGE GREASY WOOL PRICE FOR S.A. 1951 - 1971 (Courtesy Burequ ol Census ond Stotl5tics)

100 200

t50 Cents Cents kq b 100 '. ---\ -

r965 r970 r95l t955 1960 Yeor

Fig. 1. 5 CATTLE ANID SHEEP NUMBERS IN S,4.,

1951 - 1971 ( rncrudes hrsh rornrort districts) (Courtesy Bureou ol Census ond Stot¡stics)

+ Colile Sheep

1t, Sheep 'numbers r 10-6 l0 Cottle numbers I x l0-: + .:- ì t t t ++ t,

t951 t955 1960 1965 1970 Yeor 9 most evident in high nainfall, and rnixed fa:rming al?eas, but was neftected in ttre anid sheeplands. Between l-969 and 1971, pnopenties can:rying cattle inside ttr-e dog fence incneased five- fold (Lange, 1971)o Many of the pastonalists werîe considering a complete switch to cattl-e fnom sheep, and most wer:e considening cattle introduction in orden to divensify thein entenpnise.' 10.

1-4 DEMARCATION AND CLIMATE OF SOUTH AUSTRALIAN RANGELANDS

I*4I Demaz,cation.

Goyder: set th-e eanl-iest bounda::y for the S.A. nangelands, known as tf Goyden I s Line of Rainfallrr, conresponding :roughJ-y to the 12tr (goom) annual- r'-sohyet (s¡pp mo. 78rl-865-66). This was said to divide agnicultr::ral land finom land suitable only fon pastor:alism. Goyden appeans to have used the southenn limí.ts of AtripLen uesieayía (saltbush) as an indicaton.

Between l-930 and 1950, va:rious pnecípitation/evaponation indices we::e used to set the l-imits of the anid areas, as defined by length of gnowing season. For instance, Tr"umble (fg+S) defined anid pastor:al aneas as those with a gnowing season of less than five months; with plant g::owth nequining a monthly P/Eo'75 gou-t"n than 0.3. Meigs (rgsg) used a similan index. Penryrs (fgOZ) definitíon and demarcation of nangelands is used hene. A nangeland is defined by a rainfall inadequate fon economic cr"opping on pastune imp::ovement. This diffens fnom the common Amenican definition, whieh embnaces sub-humid areas capable of pastune impnovement (Box and Perny, 1971). Fon S.4., Perr:y defined the boundary as the 250nrn annua.l- isohyet, with the nangelands occupying 87eo of the Statets total anea (see fig. 1.6).

I-42 CLinate.

Gibbs (fs0g ) and GentiÌli (tglZ) p::esent extensive neviews o.r the climate of anid Austnatia. South Austral-ia in panticulan

is adequately docurnented in the S.A. Yean Books (Buneau of Census Fig. 1 .6 DEMARCATION OF THE S. A, ARID PASTORAL ZONE

\ B.--

G tv') c)

J tJ- v OL APY SPUR GAWLER..RA

Southern boundorY of orid zone, os defined by Perry, 19 67.

Northern timit of winter roins, ofter Leigh ond Noble. 1969.

Ì--Jut- Dog fence

0 r50 300

km tt. and Statistics, l-966- ). Th-e anidity of the S.A. nangel-ands denives fuom the pnesence of dr"y, subsiding air. masses of tnopical or"igin. Invasions by moist aí,:r masses increase in likelihood towa-nds the southenn and nonthenn mangins. Incur"sions fnom the south give a winten maximum :rainfall- in the souttrer"nmost anid areas, while monsoonal nains fnom the nor"th- give a summer maximum nainfal-l in the nonthernmost arleas. (The fonegoing is fuon

Bureau of Census and Statistics, l-974). Between .Iies a bnoad belt with no mar"ked seasonality in rainfall-: r'eponts at?e given by Osbor.n (fgzs) for Koonamone station, and Jackson (fesg) for^

Yudnapinna station. Leigh and Nobl-e (fg6g) give the app:roximate boundary of winten rain a¡d summet? monsoon inf.Iuences (fig. 1.6). The r.ainfall- is low and ennatic. The mean annual isohyets ane shown in figune 1.7. The associated standa::d deviation is mone than 3Oeo of the mean (Gentil-li, J97I, 1972). Sol-a:r nadiat- ion input,s ane high (fig. l-.8) and inc:rease to the nor"th. Consequently, evaponation and day temperatur?es cü'e high (figs. 1.9 and 1.10). The diur"nal nange is of the orden of 2OoC. (Specht,

1972).

The climate of Yudnapinna station is nep::esentative of th-e anid sheeplands. Jackson C:-gSg) neported the folJ-owing. The mean for the peniod 1884-1939 was 226rrn, with minimum and maximum annual nains of 6Omrn in I92t and 459mm in 1928. Over a seventeen- year: per"iod, falls of unden 7.5rrn in 48 houns, negarded by Jackson as trineffectiverr fon plant g:rowth, accounted fon almost half the mean annual r"ainfal-I . Th-e mean annuaf tempenatur:e ove:: this peniod was l8oc., with a mean pan evaponation of ca 2r4Oomm. In Fig. L 7 S.A MEAN ANNUAL RAINFALL mm

340S

1330E

ADEL

200 km

(Courtesy Br-¡reou ol Census ond Stotlstrcs)

Fig 1.8

S.A. MEAN ANNUAL SOLAR RADIATION

ot lhe ground col/cm2. doy

5CC

t,7'¡

45C 4

/.2

/.c,J

3{oS

I 3308

AOEI. ÂIDE

200km

(Courtesy Bureo.l of Meteoroloçy Adelorde) Fig. 1 .9 S.A. MEAN ANNUAL EVAPORATION

325 0

32 50 3000 2750 250

3¿o5 r250 1330E

100 0

ADEL AIDE

F-- 200 km

(Courtesy Bureou ol Census ond Slotrstrcs)

Fig. 1.10

S. A. MEAN MAXIMUM TEMPERATURE for JANUARY o Celsius

yos

13308 22.5

A CEL A IDE

r--- 200km-3 (Courtesy 8¡reo¿ ol Census ond 5tolrstlcs) t2

sunmerl, the mean daily evaporation approached l-2nm.

1-5 VEGETATION OF THE SOUTH AUSTRALIAN RANGELANDS

I-5I lhain díuisions,

Thene ane two main groups of plant communities in the S.A. nangeJ-ands: those cornmunities with semi-succulent shrubs of the

fanily Chenopodiaceae eithen as dominants on in the undenstorey

of woodl-ands, and those communities with tussock on hummock

gï'asses as dominants on in the understo:rey Cspectr-t , J-972). The fonmen extend north to the limits of winten nainfall (l,eigh and

Nobl-e, l-969), and a:re the most common conmunities in the S.A. and N.S.W. sheeplands. The l-atten, showing affinities with the sub-

tnopical summer :rain communities of centnal Aust::alia, a:le also found in the S.A. sheeplands, and extend wefl south into the winter" rai¡fatl- bel-t. Both groups ane consider"ed in this thesis.

The particular associations al?e not of pr ime impontance hene: t where necessary, they a:re discussed in the body of the thesis.

Howeven, the fotlowing points are impor"tant. Of the semi- succulent shrubs, the most significant are the long-Iived penennials Atr,ípleæ uesicæia (saltbush), Kochia sedifoLia,

K. astrotïy\cha and K. pyrønídata (bluebushes ). These species have attracted much ecological nesear:ch. For instance, autecologies a:re described by Osbonn et al- (tSgZ), on the gfowth of A. Ðesiearia. unde:: stocking; Burbidge CIS+S) and Beadle (1952) on genmination nates; K::rowles (l-951-) and Condon and l(nowfes (fgSZ)

on phenology. Much of th-e r:esea:rch has been airned at ,4. uesicæia. 13. alone (see Jones, Ed., 1970). Nonetheless, tlrene nemain senious gaps in knowledge concer"ning tTrese species (Penny 1970). They oniginally compnised tlre majority of pastr.t'es of the finst guoup of communities. Gnasses of th-e genena Stipa, Aristída, E?agrostis'

Triodía and Zygoehloa compnised the pastu::es of the second g:roup.

Iriodia and ZygoehLoa speeies ane per:ennial-, but have never been gnazed to a Ia:rge extent in S.A.: of the r"emainder, only centain Eragrostis species are penennial, but nanely conbribute much to the pastune (Specht, 1972). Much less neseanch, nelevant to the

S.A. situation, has been penfonmed fon such- species.

The finst group of communities is found on caJ-eaneous, gJpseous on sal-ine soifs, with cJ-ay B horizons and, often, ma:rked texture contnasts in the soil profile (ttdupl-exrr soifs of Nonthcote, 1971-). The soils ane highJ-y erodable, and the chenopod shrubs ane very sensitive to overgrazing: such landscapes have suffered most under. stocking (Leigh and Noble, 1969). The second gnoup is

found on less enodable deep on clayey sands; a less easily damaged

landscape (Condon, Newman and Cunningh-am, 1969).

I-52 Trend ín grazed rangeland uegetation.

The or:iginal vegetation of the southenn Austral-ian nangelands, incJ-uding those of S.4., was considered climax, as ì-imited by soils

and micnocl-imates (I'lood, l-937). Most of the pnesent vegetation

can be rega:rded as a grazing disclimax (Moone, l-953b), relatively

stable at pnesent CPenny, l-969). Vanious paths and outcomes of pasture degr"adation, in S.A. and elsewh-ere, have been neponted as foll-ows. 14.

Moone CfSSSa, l-953b) out]íned gr.azing succession in the semi- arid Riverina. Penennial- AtripLeæ pastunes on light-textured

soils showed succession under: heavy sheep gnazingo initially to

KocVtia aphylla-Stipa faLcata (penenniaf ctrenopod*tall grass ) pastune, aften destnuction of tYrc Atr"LpLeæ. Heavier gnazi:ng on such a pasture resulted in a Danthonia (pe::ennial mid-g::ass)

pastune, considered mone productive than the oniginal. Stil-l

heavier gr"azing -l-ed to an annual- short-gnass pasture of low pnoductivity. 0n heavier textured soils, diffenent species

wer:e involved, but the succession r^ras also perennial- chenopod

pasture, to perennial- gnass pasture, to annual pastune unde:r incneasingly heavy stocking. t¿lilson and Leigh (rs6+) and liillians (f g0S ) demonstnated sirnil-a:: successions for" othen -communities in the Rivenina.

On the mor.e a¡rid S.A. and wester"n N.S.W. :rangelands, the stanting points (perennial chenopod pasture) and endpoints (ephemer:al pastune) ane simifan to the fonegoing, but the::e is no + inter:mediate perennial grass stage. Beadle (fg+g) and Ratcl-iffe (fgSO), the two most extensive studies, contain many examples of gnazíng succession. Those hene ane fnom the more local- study of

Cnocken and Skewes (tg+t), on Yudnapinna station. Heavy gnazíng

of an open woodÌand with Koehia sedifoLia (perennial bush) pastune resulted dir"ectly in an ephemeral Bassia-shont grass pasture, of tow fodden vafue and with- l-ittle pnotection against soil enosion. Similarly, overgnazing of an AtrípLeæ-Koe[tia (perennial bush) pastune, causing destnuction of th-e bushes, Ied dinectly to an Atr,ipleæ spongiosa (eptr-emeral- forb ) pastu::e of low value. 15.

Less neseanch, has been d.irected towards gnazing succession in

hummock and tussock g:rasslands, by compa::ison with the clr-enopod

pastures. Resea::ch nelevant to S.A. nangelands has been penfon-

med at Alice Spnings, N.T., and deals I'rith cattle grazing. Whe::e

pastunes wene penennial at the start of gnazing, ovenstocking nesul-ted in an increase of unpalatabl-e species and a decnease in

paÌatable species (Newman and Condon, 1969; Maconochie and Nelson,

OglZ). This is not homologous with the shnubland situation, where the salt- and blue-bushes are not usually the most palatable species. Perennial pastures did not necessanily show a decnease

und.en heavy stocking of the penenniaÌ species: Tr'íodia, an unpalatable penennial, was neported to incnease under stocking

(Newman and Condon, 1969). Overgnazing, particulanly in the first group of communities,

is accompanied by sevene .l-oss of plant nutnients. fn soils carr5/-

ing AtrípLeæ uesícazLa, 279o of the total nitnogen, 2I9o of the total-

phosphor:ous, and 3O9o of the total- organic matten ane in the top

* lOcm of the pnofile (Char"ley and. Cowling, 1968). The consequences of enosion fol-lowing overgnazing of the plant covell ane self- evident. Funther-, the phosphor:ous is la:rgely onganic in or-igin.

A 1evel of gnazing, not resulting in erosion, nonetheless may

nemove a significant pant of the supply. Regenenation folJ-owing overgnazing is ::etarded by the nesultant nutnient deficiencies, in addition to othen soil- modifications which have occur:ned (Beadle

and, Tchan, 1955). Lich-en cnusts ane a pnominent feature of many virgin soifs. The lich-ens, in addition to stabilising the soil sunface ¡ mây contr.i-bute significantly to the nitrogen balance 16.

(Rogens et al-, 1966). Stock impact ís first shown in th-e br.eakup and loss of this cnust (Rogens and Lange, 19?1).

I-53 Regeneration of ehenopod slætúLøLds folLoaíng ouergz.azíng.

The Koonamone Vegetation Reserve (now tlre T.G.B. Osbonn

Vegetation resenve at Koonamone) was establ-ished in l-925, in the

t?wonst eaten out cor"nenrr of a paddock on Koonamore station

\ (see Osbor.n, 1925). Regeneration in the absence of sheep has

been followed since (reports: Osbonn et al-, 1935; I,trood, 1936; Hal-l et al, 1964). Of the penennial chenopods fonming the

oniginal vegetation, negenenation of AtripLeæ uesicaîía and

A, stípitata was evident in 1936, and .langely complete by 1964;

but Koch'ùa sedifoLia lnas not negenerated, alth-ough- some individ-

uals sunvived the oniginal overgnazing. Soil sur"faces have sta-bilised, but equal-Iy sJ-owJ-y. Pidgeon and Ashby (fg+O), investig-

ating negene::ation of ovengrazed commons aften exclosure, at Broken Hil-l-, N.S.W., repo:rted that decades would be nequired fon ti complete regene:ration. On Manunda station (f:-g. f .f ), folJ-owing a long and sevene stocking histony, negenenation was hastened sig-

nificantly with mechanical aids, eombined with excfostre of

paddocks, fuom 1945. Much of the chenopod pastur.e has negenenated,

yet centain paddocks a:re stil-l- largely ephemeral g::asslands, and

ane still- c.l-osed to stock (personal obsenvation ).

I-54 Contrq.sted Aney"Lcayt qnd Austz,aLion situations.

The extensive Amer"ican C1angely U.S.) nangeland neseanch is

fnequently referred to he::e, fon rnettrodol-ogy, companison and intenpnetation of tnends, in str.uctr:r'a11y similar vegetations

grazed by sheep and cattle. However:, there ane impontant differ-

ences in management stnategy. In southenn Austna.lia¡ anid

nangelands, chenopod shru-bs ane of majo:r importance. They pnovide an adequate for"age fon stock in dnought (Wood, l-936; I'[ilson, 1966)

and a majon guand against, soil erosion (e.g. wind enosion, as shown

in the wind tunnel- studies of Marshall, 1973). The ephemenal plants of degraded pastu:res cannot ful-fil such functions adequately.

Management, thenefore, is ained at maintenance of the penennial

chenopods. On the other" hand, management in the U.S. nangelands nay benefit fnom the nemoval of similar shnubs (..g. sagebnushrshad-

scale) as the resultant pastu:re may be penennial grass. Such grass can pnovide tlr-e :requir:ed dr:ought reserve and soil p::otection, while

incneasing domestic secondar:y pr:oduction (..g. as demonstnated by

Shown et at, 1969; Bleak et al, 1965).

1_6 ASPECTS OF SHEEP AND CATTLE PERFORMANCE ON ARID RANGELANDS.

q I-6I Introduetíon.

The sheep indust::y in the S.A. rangelands is based wholly on st:rong-wool production, f::om the S.A. strain of Merino (a lar:gen-

fnamed str.ain than found elsewhene in Austnalia-:B::own and Hutchinson, 1973). The cattl-e industny is based on.beef pr^oduction fnom Hereforcl and Sh-ortlronn cattle. Thene are no significant herds of the Bos indie¿zs bneeds o:: c::ossbreeds, despite thein success and acceptance elsewhe:re in Australian nangelands (see Alexanden and Carnail-I, 1973). Hence the foll-owing discussions concentnate on the Meri¡o and th-e two B::itish cattle b::eeds. 18. I-62 Aspects of physíoLogy,

' Ttre Meni¡ro is better" adapted physiological-Iy to arid conditions, than the Bnitish breed cattle. With the pnotection of wool, sheep can graze comfo::tabJ-y in the sun: although sunface wool tempenatures may neach g2oC., there is l-ittl-e heat penetration

(l.lacfanl-anei; fgSO). Exeess heat is dissipated rnainJ-y by panting, which allows gi:eater control of wate:r l-oss than sweating; waten is efficiently consenved in both the nenal system and the intestine, nesulting in low water tr::c.nover"; and the metabolic rate is l-ow (Macfanlane, l-968). By companison, the Bnitish cattle bneeds often have a coat which collects :radiant eneng1/ but is not a good insulator; heat is dissipated mainly by sweating; the nenal system and intestine ane less efficient in waten conser.vation (e.g. Macfanlane, 1968, cites maximum urine concentnations of 2.6 osmol ./l-itne fon Shortho::ns, companed with 3.8 osmol./litne fon I'feninos); and both the wate:r turnove:: and metabolic r.ate are highen than foi: the Merino (Macfarlane, 1968). Differences in waten tr::r'nover: a::e highlighted in table l-.1-.

On a comparative basis, the waten turnoven of the Menino was only slightly mone than that of the came.l-, wheneas the turnoven in the

Shonthonn appnoached that of the waten buffalo. Even the more efficient beef bneeds we::e not as well adapted as the Merino. Macfanlane (tg6l+) nepor.ted that, at Alice Spnings in summe:r, cattle depnived of waten su¡rvived only foun to five days, but Meninos six to ten days, fu::then il-lustnating the better adapted water balance of the latten. Schmidt-Nielsen (1964) pnesents a mone genenal neview, which supponts and expands the foregoing statements.

?tetal . 19. TABLE 1. ]-

VÍATER TURNOVER TN RUMINANTS

Cfrom Macfarlane, 1968 )

(a) Adul-t ruminants at Al-ice Spnings, in summen after nain.

Body Total Body waten solids body waten tu¡rnoven

90 Litnes ml/kgo't'' uu

Menino 38. 5 46.7 36s

Shonthonn 39. B 267 765

Camel- 47 ,7 385 350

(¡) Cattle at Danwin in wet season.

Zebu 38. 6 6ltr" 361

Santa gentnudis 37.1 629 384

Shonthonn 26.r 739 484 Ìlater buffalo 2I.4 786 524

o.85 (kg"'"" is a measure of metaboÌic body size, which allows companison between species innespective of statu-r-e ). 20.

The sa.l-t content of ch.enopod pastu::es may be veny high.

Lange (fS0Z ) neponted tevels of up to 2Oeo by weight in some chenopods. Ifilson and Leigh (fgOg) neco:rded 6.25eo Na in

A. uesicania., and neponted that sheep gnazing this species had maximum daily sodium intakes of 2OOg. To cope with such a salt intake, the Meninots basic adaptation is to dnink mo:re waten (to a totat of 1O-l-9 litr"es/day--Macfan.Iane et aJ-, 1967). Vlilson and Leigh reponted an inc:rement of 74ml per gram Na ingested.

The sal-t intake may be fu:rthen incr-eased by saline waters. Cattle in such a situation cannot tol-enate the salt levels (Schmidt- Nie.Lsen, 1964, p203) and lose condition as a nesul-t, wheneas the sheep can maintaín pnoduction. This ís a rnajon neason fon the lack of cat'cle-stocking on intact chenopod pastures. Othen aspects of physiotogy, also showing the supe::ionity of the Menino oven cattle in an arid envi::onment, arle discussed by

B::own and Hutchinson (1973).

I-63 Dietarg seLeetion.

The selection of species by the gnazing animal is oft-en considened a prime cause of stock-nel-ated vegetation pattern, and so req- uines an extended discussion. Most reseanch on dietary selection in Aust¡alian nangelands has deaft with eithen sheep on cattle: there is litt1e dinectly companative wo::k neported. Fon this neason, some investigations are quoted hene, f:rom similar" situations ovenseas, and from highen nainfall areas, as examples of the genenal di-fferences and sjmila:rities to be expected in the S.A. sÍtuation. The discussion is concenned mone with stock 21. prefenences than pl-ant palatabilj-ties. The words are often used in neponts as though intenchangeabJ-e; but a her"bivone I s pr-efer'- ence for a panticuJ-an species is dete::rnined by the nelative cincumstances resulting in choice, wheneas palatability is a function of intr^ínsic pnoperties of a plant (Headyr 1964). A senies of shont tenm, intensive sh-eep grazing tnial-s were nun in a variety of associations in the Rivenina (Leigh and

Mulham, 1966a, 1966b, 1967; Robands et al. 7967; Leigh et al 1968; Wilson et al-, 1969). Pattenns of selection wene sirnil-a:r thnoughout. For a penennial bush community of Atripleæ Uesieæia-

Kochía aphglLa, Leigh and Mul-ham (fgooa) demonstrated a high degnee of species selection by sheep, and changes in diet cornel-ating with pastune change. In winten, /. uesicaYia compnised 759o of the avail-able forage, but 3.5eo of the diet. Species of Hordewn (annuaf grlasses) and l'úedícago (legume, annual firom a forage viewpoint) comprised 59o and 72eo respectively of the pasture, but 339o and 22% nespectively of the diet. In summerr hrith die- back and gnazing down of ephemenals, A. ues¿earia compnised

78eo of the pasture, but !O% of the diet, while SporoboLus

(perennial g:rass) comprised 0.59o of the pastr:r'e, but 82eo of 22 the diet. In another situation, Leigh and Mul-b-am CrSGZ) neported that .4. pesieay,ia pnovided 9O9o of the diet, in summen dnought, æd in the absence of most annual species. Howeven, the winten pattenn was simil-an to the above, species nepresenting only 4eo of tire total pastune compnised SOeo of the diet. Likewise, in a Dartthoniq caespitosa (pe::ennial gnass) community, annual species wene prefenred when pnesent, althoughr, only a smafl part of the total fonage. The contribution of Danthonia to the diet was significant only in summer dr"ought (Robards et aI, f967 ) .

Ttre same pattenn is reponted under station conditions, with large paddocks and fnee-nanging sheep. I^leston and Moin (rsog ), observing sheep gnazing penennial Mitche.l-l- grass pastune in anid

Queensland, neponted an equivalent sequential consumption. penennial species foi:med the mainstay of the diets only when annual species had diminished gneatly. A funth-en aspect of the degree of selectivity is tt¡-e highe:r p:ropontion of nutr-ients in the diet, compared with the pnopontion in the ovenall pastu:re. For example , I^lein and Tornell- (fSSg) and Weston and Moi:r (fg$g) demonst:rated highen cnude p:rotein and lowen cnude fibne pnopontions in the diet than in the pasture.

The pneference hienanclr-y of cattle on Australian r"angefands is sinil_a:: in outl_ine, but with manked diffenences in detail. Low

(fSZS) obsenved fnee-ranging Str-orth-onn cattle on grass pastu:res at Al-ice spnings. Perennial-s wene grazed only in the absence of ephemenals. Cattle slr-owed manked pnefe:rences for panticulan 23. cornmunities, depending on th-e lattenrs state of gnowttr- Csee also

Low, 1972). In the s¿tme a¡ea, Cfi-ippendate Ctg08) sþowed the pnefenence of cattle fon epÌremenal guassesr €.9. Enneøpogon species. In both- studies, browse formed a minor cornponent of the diets. Chippendale repo::ted that,, even in dnougirt, bnowse fnom

penenniaÌ sh:rubs and trees compnised onty 7-9eo of tÏre diet (..f .

the sheep br^owse consr:rnption in dnought, above). The same patterns are reponted in divense situations elsewhe::e (u.g. Clary and

Pea:rson (fgOS) in the semi-anid U.S.; Hannington and Pratchett (fsze) in Uganda). As with sheep, cattle intake is of a highen nutritional quality ttr-an the pastur?e as a whole (..g. Galt et a1, 1969; Newman,1969). At pnesent, there are no direct compar^isons of dietary sefection, between sheep and cattfe, on Austnal-ian nangelands. The p:robable differences must be infer":red fnom investigations dealing with one on othen animal. Since sheep and cattle studies dealt with dispanate vegetations, and since sel-ection is highJ-y dependent on local cincumstances, such comparison is adequate onJ-y for the outline of differences. Hence, the folJ-owing examples ane taken fi:om the Amenican lite::atr¡re, to indicate with mone centainty the

ge¡renal- differ ences between sheep and cattle pnefenences, likely to be found on the S.A. sheeplands. The diffening situations pneclude any detail-ed extension of ::esul-ts to the S.A. case' but the genenal pattenn is likely to be similar.

V van Dyne and. Heady (_rgoSa) companed dietany sefection of cattfe and sheep grazing a mature (dny) annual- nange. During 24. observations, pastur"e standing enop decneased by 7O9o. Prefer- ences changed significantly as the for.age decr:eased, but not in the same manner for sheep and catt.l-e. Thein pr"efenences ÌIere significantly different threughout, for half the forage species present. Further, sheep maintained a higher quality diet (Van Dyne and Heady, 1965b). Paral-l-el-:results were reponted by Cook et aI (1967), in a companison of sheep and cattle grazing on sagebrush. Some pneferences wene found equivalent, but on the whole the cattfe preferred and mainly consumed grasses, while the sheep utilised forbs and bnowse to a much greater extent. The repor:ts on the individual- henbivor"es, already discussed, imply that cattle and sheep would show some simila:rity in dietary selection on the S.A, nangelands: ephemenal plants would be gnazed when pr.esent, and perennials only grazed du::ing stness peniods. However, the above indicates that ther"e would be considerable differences in specific prefenences, particuÌanIy with regar"d to shrubby species. This J-ast point is of major" impontance fon the S.A. rangeJ-ands, wíth thein ttigh shnub component. Har"nington and Pnatchett (l,glg) r:eponted that cattl-e avoided non-gnass species

on the Ankol-e rangeJ-and (Uganda): Gal-t et al (l-969) on ar"id U'S. nangeland, and Thetfond et al (tïlt) on semi-ar:id U.S. nangeland

demonstnated the low pnopontion of br.owse in cattle diets by comparison with sheep: and Chippendal-e (fg0e) pnovided an Austral-ian example of the low utilisation of browse by cattle. With such a pattenn r"ecurring in these diverse situations, the same should occun in the S.A. sheeplands. Hence it may be postul-ated that the intr"oduction of cattl-e onto the sheeplands is likely to prove l-ess deleter.ious to the chenopod pa.stures than was the forme:r sheep grazj]ng: the cattle ane l-ess likely to affect the shnub species. 25. I-64 Aspects o¡' ranging behauiout',

On Austnalian rangelands in dry condition, stock a:re dependent on the water.s provided. As indicated in section 1-62rcattfe a:re mone dependent than sh-eep. Lynch CfgZg) neponted that merino sheep on semi-ar:id pastune with some gneen feed drank dail-y only when tempenatures exceeded 41oC, but at ttrnee-day intenvals at l-owen tempenatunes. Catt1e under simil-an conditions nonmal-ly drink daiJ-y (Lynch and Alexanden, 1973), although some individuals may nemain away fnom wate:: for 36 houns (Schmidt, 1969). Hence the pasture utifisation is linited by the travelling ability of the anirnals; in dry periods fon sheep, and most of the time for' cattl-e. This pantly underl-ies the piosphene concept (to be discussed).

Thene is contnoversy as to how fa:: sheep and cattfe graze fi:om waten, under: r"angeland conditions. Cattl-e are often reputed to walk furthen (..g. chapter 2). They ane known to be capable of long traveJ-: Bonsma (fgt+g) dr:ove herefonds and sho::thorns, unden hot conditions' to the point of exhaustion, achieving 2Skm/day for: the smooth coated vanieties. llooly coated animals walked no mone than Bkm, but such vanieties are not common on the souther"n rangelands. Schmidt (1969) obsenved sho:rthonns under: field conditions on the Bankly Tablel-and, N.T., in a tine of high temperatunes and dry, spar.se pastune. 4O9o of cattle walked 6-7km befone commencing gnazing, and t emained wittrin 8. 5km of water. Unde:: ext::eme pasture sca::city , grazistg extended to Ìl-km fnom waten. The najonity, howeven, extended to a maximum distance of only 41cn.

Newman and Condon Ctg6g ) stated the limit of catt.Ie gnazing unden 26.

similar conditjons to be 5ni CSkm). For" compa::ison, Henbe]- et

aI (l-967) r:epor:ted cattl-e to extend only 3.5km fnom water: on a

New Mexico dr.y nange.

Squines and !'lil-son CfgZf ) performed a ser-ies of expe:riments on menino tnavel- under serni-arid conditions (see also Squires, 1970; Squir"es and Hindtey, 1970). Waten and feed--Atz,ipleæ uesícaz.ia--were separated by walkways devoid of fonage. Merinos

walked a daily total- of 12]cn. Fonage and waten intakes wene

unaffected until the J-ength of the walkways exceeded 4km. In field studies, Lynch (fg0Z ) nepor:ted similar" total distances

walked in semi-a:rid pastune, with a maximum distance fr"om waten of 3-4km. Osborn et al (fggZ) neponted that sheep remain langely

within 3km of waten, with vintualJ-y no gnazing beyond 6km.

Ratcliffe (fgg6), on the other hand, descnibed cases whe::e veget-

ation was denuded by sheep, in d::ought, up to 5mi (ekm) fi:om waten.

Hence there is no ce::tain evidence that, under" station conditions,

the commonJ-y ::un cattl-e bneeds wiJ-I graze funthen fnom water than

will sheep.

I-7 THE PIOSPHERE: STOCK INDUCED VEGETAT]ON PATTERN

I-7I The coneept of the piosphere.

The dependence on water supply and the l-imits of ranging of

stock cause radial patterns in nangeland vegetation centred on the

water: point. Reseanch into such pattenns in southenn Austnalian rangeJ-ands has been confined to those caused by sheep. Pant of the contniSution of the present neseanch is the outl-ining of pattenn 27. caused by catt.re.

The sheep-induced pattenn was finst quantified by Osborn'

Wood and Paltridge (tggZ), who examined density pattern of

Atz,ipLeæ uesicæia in South Austnafia. They distinguistred four concentnic densitY zones : 1. The rrArr zolle, immediately aIOund the water to a maximum of 4OOm; sevenely overgrazed and tnampl-ed (the pastonalistsI rrsacrifice zonett ); 2. The rtBrr zone) lying between 400m and l-600m fi:om waten' in which most gr"azing took place; heavily gnazed, but withapparentlymorevigonousgrowththanthellA|lon outen zones; 3. The rrctf zone, Iying between l-6oom and 3600m; occasionally grazedl

4. The I'Dll zo-ne, beyond 36OOm f::om waten; ungnazed, and beyond the no::mal ::ange of sheeP.

They emphasised that these zones were not shanply defined, vanying in extent with stocking rates and paddock topography' Vafentine (fS+Z ) observed a similan radial- attenuation of

gnazing pt?essulre with incneasing distance fuom water in forest range

grazeð, by cattle (U.S. ). G:rass standing crôops incneased with dist-

arice fi:om waten, and valentine summarized. the nadial- pattenns ín

tenms of zones, much as above. (Herbel et al, 1967, did not detect such pattern when obsenving cattle on a New Mexico anid range, however). Lange (fgOe ) fur"then quantified the pattern descnibed by

Osbonn et al, in a scrutiny of sheep tnacking and dung drop on 28.

S.A. nangeland. The term t?piosph-er:ett was introduced to name the system of stock/pJ-ant inte::actions centlred on the water point. Lange stated that the piosphene shoul-d be considened the basic ecological- unit ín nangelancì. g:razing studies. Using the fnamewonk of this papen, Banke:: and Lange (fgOg, l-970) and

Rogens and Lange (fgZf) have since examined the vegetation pattenn in detail. Th-e nesu.l-ts of thein studies display a mone continuous vaniation in panameters than was suggested by

Osbor"n et al . Like pattenns ane described fon other henbivores in diverse situations, whe:re ranging is limited by the need to r:eturn negulanly to a panticul-ar" point. Ogden et al (fgZg) reponted the formation of rthaloestr anound Ïlest Ir¡dian patch reefs as being a g:razíng effect. The standing cnop of benthic algae incneased with distance fnom r:eefs, due to gnazing of the echinoderm Diaã,qna. This herbivore was l-imited in its fonaging by the need to netu:rn to reefs in daylight, to avoid predatons. Anothen example is given by Lock (tSlZ), in Uganda. Hippopotamus grazing maintained tussock glîasses, :replaced by cauline-leaved species in the absence of grazing. G::azing was limited to alleas neall channels and lakes, and attenuation of grazíng pressune with incr.easing distance r"esufted in conrelated vegetation patte::n: increase in cauline-leaved species and decrease in tussock glrasses. In this case, the pattelr.n was not centred on a single point, but

on wall-ow aneas. 29.

I-72 ?he form sf pLant/distance nelationships in the piosphere.

In ttre absence of domestic g:razingr ân even standing c::op (..g. of biomass, densit¡r) is expected, with due allowance fon background influences such- as soil va¡iation. A stocked vrater: point in such a situation intr.oduces a gnazi-:ng pnessune, varying with distance, and causing correlated pattenn in vegetation. The only certain charactenistic of the distnibution of gnazing pllessune is that it asymptotes near zero beyond the nange of stock.

Within the piosphe::e, the possible distnibutions are numenous. The r:esults of Osborn et aI (1932) imply an exponentia.l- decnease in grazing pnessure with distance, and dissímilar vaniation in plant density and condition (peak density in rrcrt zorte, peak condition in rrBtr zone). Banken and Lange (fg6g, 1970) demonstnated pattenns

in species incidence which appear to nelate, molre di:rectì-y, to an exponential decnease in grazing pressune, with concentnic pattern most cl-eanly defined immediately around watens. 0n the othen hand,

Lange (f90g) neponted finean decnease in dung dnop with incneasing distance (though not to the l-imits of grazing). As dung drop may be taken as indicative of time spent in an area (Ares et aÌ 17972, aften Hanksrrlg6O ), Langers nesul-ts irnpJ-y a like decrease. in stocking pressune. Rogens and Lange (fgZf) demonstnated a nea-r- linean increase of lichen species fnequency with distance fnom waten, excluding the area within 200m of waten, to a maximum of 2km.

This is significant, since the lichens ¿Lne not actualJ-y grazed,

but are easily destnoyed by tnampling. They ttrus pl'ovide a

cleaner" measune of ovenalL gnazing pnessuue than fodder" plants, 30. since the latter do not necessa:rily r:efÌect th-e genenal pattern' due to diffenences in the grazing animalsr pneferences. In genenal terms, th-e distr:iSution of grazing pressure ar"ound a waten point is almost impossible to dete:rmine. Reponts on stock r:anging behavioun only give a general indication of the alreas unden heaviest pnessune, and the matten becomes mone complicated when stock show pnefenences for: particulan aneas (..g. Lowt 1972). Nonethefess, distance fuom water pr:ovides an indeæ of stocking pressune, and so the studies in this thesis, which deal- with the piospher:e, considen pJ-ant pa:rametens in refation to this index. The papers using quantitative data, Lange (fSOg) and

Rogens and Lange (1971), show linean relationships between various pa:rametens and distance from waten: in the absence of specific infonmation to the contnany, this linear fr"amewonk has been ernployed hene. The::e are othen anguments in favour of a linea:r app:roach. Fon instance, the vaniation to be expected in a natural population may obscure the undenlying grazing:relationship. l'his

is demonstr"ated in chapten 6, where vanious functions a¡:e fitted to quantitative data. The va-niances invol-ved wene such that curvilinean functions, which were thought to app:roximate mone

closely to the data, wene no mone efficient than simpler linear

functions in accounting for" the vaniances. Patten (1972) , dÍscussing pnair:ie ecosystem modeJ-ling, has pointed out that

trlinearised model-s behave mone like ecosystems than do non- Ìinean onestt. Furttren, a 1Ílean relationship ::emains the simplest hypothesis in the absence of infonnation to the contrary.

Nonetheless, th-e sea:reh, in much of this thesis, fon finear 31. nelationships )etween plant panameters and the index of gr"azing plressune is based on the reponts of Lange (f90g ) and Rogers and

Lange (fgZf), which- have demonstr"ated.the existence of such nelationships.

1_8 ]TIE CAUSATION OF PATTERN AND TREND.

The main components of stocking pnessune on nangelands ane g:razing (in the eating sense) and disturbance in the counse of g::azing. The low stocking r:ates emptoyed (::anely mone than

l sheep to 5ha, I beast to l-Oha) minimise nutrient tr"ansfer effects, pnominent at times in patter"n causation on hígher: nainfall pastu::e (".g. Seans, l-949; Hilden, 1964). A possibte significant exception is the input of nutnient immediately about the wate:: point. In the litenatu:re, little atten-i:ion has been given to the effects of distunbance, and, in particulan, tr:ampÌing. T::ends and

patte:rns in stocked vegetations are often considered to result pur"eJ-y

fuom the eating out of more prefenned species, with an attendant incnease in l-ess p:referned species being due to lowened competition.

For- instarrce, in their. major review on Austnalian nangefand vegetation, Leigh and Noble (fSOg) stated that tr".extensive botanical changes have occunred in these (chenopod) shnr:bfands following gnazíng and i,ndiscniminate lopping fon drought feed. This has nesufted in the disappeal?ance of the palatabte shnub species, and a marked incnease in othen less palatabl-e species.rt In an eanl-ier" review, Beadl-e (fgsg) afso considened changes nesul-t-

ing fr"om stock in these terms. CHowever, Moone (tgSg) in the same 32. publication considered tramplj-ng of equal ímportance to grazfug). Hence, considerable emphasis has been placed on studies of dieta:ry selection, as a means of understanding t} e mechanisms invol-ved in stock-induced vegetation change.

Even in intensive grazlrng tnia1s, the effects of tnampling are often ignor.ed. Fon example, Leigh and Mulham (tgO0a), in a tnial- already diseussed, did not considen trampling at al-l. In simil-an expeniments, Laycock et aI (1972) have shorr'n that t:rampJ-ing may account for" up to 50% of fonage nemoved by sheep.

This is not to say that tnampling is always disnegar:ded as a facton, nor is it to be taken as denig:rating the undoubtedly impontant reseanch into the selection aspect. Nonethe.less, numerous neponts make only passing refenence to the role of tr^ampJ-ing, and basically considen patterns on dynamics in terms of species palata- bilities o:: stock pr"efer.ences (Beadle, 1948, repnesents a pnime example). The revÍew of r:esearch into dietany sel-ection suggests that this is a simpl-istic appnoach: the species consumed vany with time and pasture condition, and such vaniation nendens the palatability on un-pal-atability of species J-angel-y irrelevant.

Two signal papens point to the impontance of distunbance as a pnime facton. Rogens and Lange (fgZf) demonstrated a cfear stock impness on l-ichen distributions, which coul-d only be caused by tnampling. The changes impJ-y a break-up of the soil cr"ust, loss of the nitrogen-fixing lichens, and consequent soil changes. Such soil changes must nesult in cor^related changes in fodder plant populations, due to the distr:::bance caused by stock, but not to thein actual gnazing. Heady CfgO0) investigated the effect of 33. various wild a¡d domestic ungulates on East Af:rican gnassland. Each species infl-uenced composition diffe¡ently to some extent in a manner aceountable by differing pr:eferences. However, the most st:rfüing corr"elations found wene between pastune composition and stocking rate, irrespective of the panticular ungulate run. Thus the spectrum of pnefenences, of the ungulates, had little to do wíth the u.l-timate state of the pastune. The main al-ter"native is the trampling facton. Quinn and Hervey (l-970) have shown that tnanpÌing l-osses, fon cattle on sandhill nalge, inenease geometnic- aÌIy with stocking intensity: hence thene exists the mechanism to explain Headyts observations. This thesis considens both- prefenential grazing selection and disturbance as possible causes of the patterîn in vegetation shown to relate to stocking. The inplications of its data ane discussed in the course of analyses, arld the two factons a::e discussed at length in the final chapten. CHAPTER 2

CATTLE ON SOUTH AUSTRALIAN SIJEEPLANDS

2_I VEGETATIONS GRAZED BY CATTLE WITHIN THE DOG FENCE

Specht (tSlZ) r:ecognised l-5 najon plant associations in the anid l-and systems of South Austral-ia, of which 14 occun within the

Dog Fence. Cattl-e ar.e nun on atl of these, but with vanying success. Associations with a Ìrigh pr,opontion of ephemer:als ane most favour:ed, particulanl-y those found on sand soils and fl-ood- fl-ats. The fonmen contain few perennial chenopods in the pastunes,

The l-atten may be characterised by penennial chenopods (..g. Koehia pyr'ønidatd, bvt al-so contain many highe:r-quality fodden species, and, in addition, pnovide gneen pasture mone fnequently than most othen associations (Wood, 1936). Sand associations ar:e panticulanly impor:tant in the northwest sheeplands. In this a::ea is an extensive sand:ridge system, with alternating nidges and clay-soiJ- flats (Jessup, f95l-). Pnion to l-970, the few significant cattle herds in the sheeplands were nun on this sand/clay system. Veny few cattle ane run on intact chenopod shz'ubland (i.e. high density of Afu"LpLet or Kochía species, minimal- soil- distunbances, and no overt pasture degr.adation. ) Some stations, for: instance Nonning (rig. 1.1), attempted cattle grazing on such pastunes, but found that only sto:re cattle--sold fon canning¡ or? to othen pnoducens for fattening--were pnodueed. Stone cattfe ane less p::ofitable than fat cattl-e, the latten having a t'finisht'suitable for: sal-e as tab.l-e beef. On the other hand, stations nunning cattle on sand soil or floodflat pastulles can expect to pnoduce fat cattl-e one yea:r in thr^ee, on avenage. (fnis is based on reports fr-om Pernatty and

-34- 35.

Roxby Downs stations, fo:: the per-iod (1940-1970). NonetheJ-ess,

where AtripLeæ^or Kochia- dorninated pastur"es have detenionated,

with bushes severely ttr-inned on altogethen nemoved, cattle are nun

pnofitably on the resultant, la::gely ephemer:al pastu::es. The emphasis on pelrennial pastunes, and the production potential of

such pastures, is summanized in tabl-e 2.I, fon a sel-ection of

stations with catt.l-e inside the Dog Fence.

An appanent anomaly in an othenwise cfearcut situation is the

nurnben of stations gnazing cattle on K'ochia astrotrícVta associations. This association, as it appeans now, has a very high

ephemeral component (q..,r. chapter: 4). Despite Spechtrs (1972) napping, showing it as a p:re-settl-ement association, f consider it the nesult of pasture degr:adation due to stocking, fon the following reasons.

t. Lange tnacts of an Atripleæ uesicæòa-K. asttotricha

association occu-1: where thene is no histony of oven- grazing (Jessup, 1951). This may repnesent the b fonebean of the present association. 2. K. astz,otricha can sunvive ovengnazing whictr- will eliminate

A. uesicovía (chapten 7). Hence thene is a mechanism for' neducing .4. uesiearia-IL astrotv'íchn to the present K. astrotv"ieVtø association.

2_2 MANAGEMENT SYSTEMS

2t2l Management noz,th ætd souttt of the Dog Fence

Manked dj-ffenences exist between cattle stations outside, and 36 TABLE 2.1

PASTURT] LANDSCAPES FOR A SELECTION OT SOUTH AUSTRALIAN MIXED BEEF/STIEf,P STATIONS

(a) Stations Station Pasture cl-asses Cattl-e run on First introduced Potential for present fat cattl-e

yes Curnamona AtripLex/Kochia I'l,oodflat 197 r K. astrotríeha Disclimax grass Floodfl-at yes Edeowie Discli max p,rass Disclimax g:rass 19 7l no Kolenda AtrípLeæ/Kochía AtnípLex/Xochia 1970 Tr'íodia no Mahanewo AtripLeæ/Koehia Att iples/Kochia 1970 yes Hanunda Atnipleæ/Xochia Floodflat Ì9 70 Disclimax grass Floodflat !,lt Victor AtrípLet/Xochía K. astrotricha l97l- yes K. astt'ott4cha Disclimax grass Disclinax grass Floodflat

Mulyungarie K. astrott'icha K. astrotrícha l_965 yes Disclimax grass Disclinax grass Floodflat FloodfIat yes Netley Gap Discì-imax grass Disclimax grass 1958 and t{abnicoola

Norrning Atníplea/Kochía AtripLea/Koehía 19 50 no Tyí.odia Oakden Hil,Ls Atr"ípleæ/Koehia Sand/clay 19 70 yes Sand/clay

Perr¡atty Atuípler,/Koehía Sand/clay i93 0? yes Sand/CIay Ffoodflat ia: FLoodflat

Roxby Downs Atríp|.er/Koehía Sand/cIay 1930? yes Sand/cIay

Sian AtnípLer/Koehia Floodfl-at t970 yes ?t,íodía FIoodflat

South Gap AtvipLeæ/Koehía Sand/clay l-930? yes Sand/clay Thurlga Disclimax grass Disclinnx grass 797I yes ,ITLOdLA t

TABLE 2.1 (Cont.)

(b ) Pasture Classes

Class Specht's (1972) associations Major forage

Atriplet/Koehía l'|y op orutn platy eaz,pum Casuæina e?istata Perennial chenopods Aeaeia soudenií- Koehía serLífolia AtripLeæ uesicæia- ræiolaena Leptolepís AtnipLeæ uesícavia Koehía sedifoLia

Disclimax grass The above associations, as Ephemeral grasses, herbs degnaded by overgrazing

Sand,/clay Aeaeia LinophyLLa- Ephemeral. grasses, herbs A. nø¡otlosa AcasLa qneura- A. brachystachya with pockets of othen associations

Floodffat Nítreria schobez,i PenenniaL chenopods, ephemenal gr:asses, herbs EVc alyp tus tni er othee a

K. astrotrieha Koehia a.sttotricha EphemeraJ- grasses, perennial- chenopods

G) T'ríodia Not recognised by Specht Perennial grass \¡ Hills of the Gawler Rd., with Tvodia irnitøts 38. sheep stations inside the Dog Fence; both- in facilities and management strategy. Cattl-e stations outside are unfenced except along boundar"ies: stock movement is contnolled only by the placement of

watens. On th-e otlrer hand, ttre extensive fencing on sheep stations is the most expensive single investment. A century of subdivision of sheep stations has :resulted in a lange number of smalI paddocks; paddocks langen than 5o00ha are uncommon. The watening facil-ities also diffen to some extent, although much of the difference nesults from the subdivision. Many cattle stations have access to the Gneat Antesian on similan basins, and the main watens ane deep, high capacity and ve::y expensive bo:res.

Few sheep stations are ovelr major: basins; instead, water is supplied by sunface catchment, and smali bores into local' sub- antesian aquifers. The l?easons for" the development of facil-ities on sheep stations have ah-eady been rnentioned (section 1-34). In panticulan, the subdivision was aimed at bninging as much pasture as possible within the r.ange of sheep. Little r^ras aimed at * pasture conservation: most of the fence l-ines nrn stnaight, north-south on east-west, anbitnanily with negand to pastu:re types

(..g. figr::re 3.1); while stock waters a-r?e conmonly laid at the supply source, with scant negard for the erosion potential of the site.

on sheep stations, flocks are seg?egated for age and sext with bneeding at fixed tines. This is closefy similar to higher.- nainfal-I sheep husbandny, and considenably mone efficient than the open system of cattle-stocking outside the Dog l-ence. The management tnadition in th-e sheeplands is set stocking, with flock size 39. langely detenmined by ttr-e numben that can be ca:Tied in drought.

Much of the outside cattle countny is stocked inte::mittently' with lar:ge seasonal- fluctuations. In dnougþ-t, most cattl-e ane nemoved: in ptenty, lange numbens may be :run. Ratc.Iiffe (fS+Z) pnesented an account of th-is system of management, still closely applieable to the present situation. The reasons alre biol-ogical and economic' The sheepla¡ds stilf possess much pe::ennial- pastu:re, despite wide-

sp:read degnadation (section 1-3), and so continuous stocking is feasibl-e; but the nonthern pastu¡es are ephemenal to a ì-a::ge extent (section 1-5). Sheep can either maintain pnoduction in dnought, on at l-east survive to produce wool pr"ofitably at the end of the drought, and so, aue wonth rnaintaining. Howeven, cattle cannot maintain or improve their condition ín dr:ought, and the loss of condition means loss of money. Tf numbens above a small nucleus ane netained, the pastor:alist is faced with u¡saleable cattle at the end of the dnought: by the time thein condition impnoves, they are too old to command good pnices'

2-22 Pv'oblems in ehangíng to eattLe from stteep'

Penhaps the most important pnoblem when intnoducing cattle to

sheeplands is the setting of stocking levels. The pastoraÌist usually cannot use past expenience, as can be done in setting

sheep levels. The matter is complicated by the Pasture types involved (panticulanly the salt content) and the quality of stock r¡raters. Legisl-ative guidance is limj-ted to the Pastoraf Act, wh-ich- a::bitraril-y sets one beast equivalent to five sheep; 40.

undefined as to age on sex. This is a much highe:r ratio than woul-d be expected in terms of.nerative fonage or. waten demand. To date,

tTre decision has been ad hoc, wittt the pastonalist using his own

app:raisal of sho:rt-tenm pastune condition to check over"grazing on unden-utilisation of pastr:r'es. Equivalences determined since 1970, by pastor"alists in th-e Gawlen Ranges, and indicative of the gene::al- l-evels e.l-sewhene, alle:

I cow/unit anea Z-B ewes/unjt anea

I2-I4 wether s/unit anea

2 steer"s/unit anea The distnict stocking :rate is 25-30 sheep/mi2 (1 sheep/1oha). Sheep orientated facilities must be adjusted to cope with cattle. Intennal- subdivision fences ane often flimsy--tow, with smalÌ posts, and plain wir.e--and ane banel_y adequate at times to

nestnain sheep, let alone cattl-e. S-i-milan r estn j_ctions apply to water tnoughs and handling ya:rds. The gneatest nuisance, and the

rnost expensive to nectify, is the abÍJ-ity of cattle to walk thnough

sheep fences. Locating stations whene cattl-e wene nêstnicted to

set aneas Vras a majon difficulty in this study: the most common

neply was that the cattl-e "go whene they wantrr. Very few stations have gone to the expense of stnengthening fences. Rathen, the two main solutions have been to shift cattle fnequently to fnesh pastu::e, or r mot?e commonly, to cr"eate semi-open nange by removing, orô al-lowing cattl-e to remove, intennal- sub- division fences. 0n Mulyungar:ie station (fig. 1.1), several

fonmen sheep paddocks wene coal-esced to fi:om one .l-l-Omi2

(ZeSOOtLa) paddock wittr nuJ-tiple watens. On Mt Victor stati.on 41 (chapters 3, 4), cattle broke down intervening fences, and, despite repain effor-ts, cneated a single rôange fnom two lange paddocks. The semi-open l?ange is perhaps .the best solution: some segrlegation of he:rds can be maintained, and the cattl-e, having set thein oivn bounds, do not usually stray funther". A funthen pnobÌem in adjustment is pnovision of water of adequate quality fon cattl-e. Thís may nequíre extensive upgnading of sunface catchments if gr.ound wate:: suppries ane too sa.l-ine. su¡face catchment is not always feasibfe, pa:rticu.La:r1y in sandy countny, and is expensive.

2-23 Cattle husbøtdrg on former sheepløtd.

Cattle husbandny inside the Dog Fence vanies mar.kedly fr"om station to station. At one extneme, the system is no differ.ent fnom that of outside cattle areas, with no segnegation of her"ds. At the othen, it is akin to the sheep system, with herds segregated fon age and sex, and bulls joined at set times. In veny few cases do cattle and sheep shane paddocks. These cases are mostly on stations with minon cattle herds (30-50 head) but lange sheep fl-ocks (sevenal- thousand head). An exception is Nonning statíon, with 500 beasts. 0n this station, cattl-e sha¡e J-ange paddocks with sheep, on a ::atio of 50 beasts to 1000 sheep. The management attempted highen ratios, but found that cattl-e condition detenionated, possibJ-y due to interference between the two ungulates. 42.

2_3 PASTORAL]S..TS t OPINIONS ON THE EFFECT OF CATTLE

Many gnaziers considen cattle to be less des inuctive in thein

grazing behavioun than sheep. The neasons offened by them ëLne as foll-ows.

1. Thene ane fewen hooves per" weigh,t on money unit, anC lhus there should be l-ess overall- tnampling.

2. Cattle tr"acking patterns ane the l-ess intense. The sheep

pattenn is a dense networl< of incised t:racks nadiating fnom waten (".g. ph-otographs in Lange, 1969), but the cattl-e pattenn is a ver.y few, langely unbnanched tnacks, although

mone deeply incised than sheep tnacks. This is al-so considened to indicate l-ess tnarnpling under cattle stockíng . than unden sheep stoeking.

3. The grazing habits of sheep and cattle diffe¡-: sheep a:re

said to be destructive and uneven grazens, nemoving sma1l

plants completely but narel-y touching rank gnowth; wher"eas

cattle graze less closely and mone evenly. This is a

valid point. Bennett et al (fSZO) discuss, in such ter:ms, the ::ole of cattl-e grazing in maintaining short pasture

fon sheep in highen nainfall arleas. 4. Cattle graze funther fnom water than sheep, permitting

a mone even utilisation of pastu:re than would a¡

equivalent nurnber of sheep (but see section 1-64).

It follows that certain pastu:res, easily damaged by sheep, have always been nega:rded as potential-Iy betten suited to cattle stocking.

Ttre drop in wool p::ices pnovided the irnpetus to switch such pastunes fuom sheep-stockÍng to cattle:stoclcing. For example, 43

Edeowie station, on th-e outwash pJ-ains between tlre Flindens Ranges a¡rd Lake Tonnens, cannied sheep f:rom th-e mid-1850rs to 1970, and was heavily overgnazed, The pe:rennial pastures were langeJ-y destnoyed in the counse of th,is grazing, but enosion was more sevene than usual, due to the veny fniable deep loam soil. The pnesent l-essee has nun cattle since 1970, mone in an attempt to stop the sheep-induced enosion than in onden to incr-ease pnofitabí-lity. Since the intnoduction of cattle, the erosion has ceased; but this may have resulted mone fuom abundant ephemeral growth in the subsequent yeans, than from the cattl-e- stocking itself.

Some pastonalists consider cattl-e mone destnuctive than sheep.

On Mulyungar^ie station, the management considered replacing the entine hend (1500 head, Henefonds) with anothen bneed: the

Herefonds woufd not g:raze mone than 2-3km fnom the I¡Iaters, and the pastures wene clearly detenior.ating. 0n Oakden Hills station, cattle showing simil-an behavioun alre moved to new pastu:res fr:eq- uentJ-y, to avoid such damage. 0n McCoyts We.l-l- station, cattle wene intnoduced in l-971, but nemoved in 1973: although no dete:rionation of pastu:res was obsenved, the cattle welle not nanging as fa:: as had the sheep.

2-+ ECONOMICS OF CATTLE ]NTRODUCTION

The economics of tlr-e changeoven to cattle have not been stnessed, as they ane peniplr,enal to the pnesent investigation. Nonetheless, they nemain th-e cause of the situation. fn the period of low wool pnices, prlospects for cattle introduction were buoyant (Thompson, 44.

1971 ), but, even so, a changeover vras not necessarily wonth-- while. Tab1e 2-2 sboro¡s budgets pr"epaned for stations in the

Gawlen Ranges, by th-e S.A. Depa:rtment of Agriculture in 1972;

assuming a mi¡imurn gneasy wool pnice of 35c/1b Q7c/kg). The

Gnoss Mangins for wool- pnoduction were as higtr- as fon beef

pr:oduction. However", a switch to cattle would also requine heaqy investment; in the punchase of stock, and the upgnading

of facil-ities. In view of the compar"ability of r"etunns f:nom cattle and sheep enterpnises, such ínvestment was not necessarily justified. Since 1972, wool pr-ices have nisen to economic l-evel-s. Pastonalists intending a complete switch to cattle sfowed on hafted the cattle buildup when the wool pnice stanted to rise. llith the coll-apse of beef expo::t mankets (Nov., l-974, reponted in The AusttaLíøt, Sydney, I5/II/74), no pastonalist is at pnesent

considening funthen incneases in cattl-e he:rds. Thus it appeans that the switch to cattfe wil-l- not neach the scale anticipated at the stant of tt¡-is study (March , 1972), due to the combination of pastune limitations, pnactical- and economic pr:ob1ems. Nonetheless,

a neduction in the pnesent cattfe nurnbens on for:men sheep-lands is

unlikely. The pastoralists of the S.A. a:rid sheeplands now

necognise the need fon at l-east some diversification, to insul-ate

fnom manket vaganies an industr.y original-Iy dependent on one product. TABLE 2.2

1972 WETHER AND STEER BUDGETS FOR GAI,ILER RANGES AREA (Countesy A. Bnoan, S.A. Dept of AgnieuLttu,e)

vETHER BUDGET STEER BUDGET

Asswnptions. Asswnptions.

hlethens retained for 4 years. Gross mar.gin per: steer = 4% deaths pen annum. net value pen steer minus foregone income fuom weaner:.

Retuzn per uethzx. Net uaLue per steer. 1,. Wool: 1. Gnoss return, 190kg carcass @ SZ+/Ug $ge. eo 6.Skg/hd 77ë/ke net of mar.keting changes. G $s.oo 2. Variabl-e costs: 2. Sale of 27% wethens @$2.00 off shears, c.f.a., Fr:eight, Gawler Ra. to Adelaide $2.OO net of tnanspont and manketing charges. â0.+Z Commission etc. @ 69o gross sales $S.gO lntenest on capital outlay of Tota1 r:etunn pen wether. $S. +Z stock, 89o of $85 average val-ue $0.90 3. Deaths , 29o $l- . Z o Cost pez, uethey,. Total variable costs $Zf.+O t. pr::rchase 25% neplacements @ $9.75 on prope::ty. $O.g+ Net value per steen $zz. +o 2. Openating costs: Sheaning $O.SO Crutching $o.fS lnleaney eaLf uaLue. Jetting. $0. oS - grloss (transport 3. Intenest on cpital outlay of stock, value - + selling charges ) 125ks x 55ë/kg ($o.oo + $4.20) $se. eo Beo of $Z . gS avel?age value. $0. Zg = -

Total va::iable costs $f.gz Gross matgín per steer,: $18.60

Gross margin per uethez, -+- (¡ (r"etu::n - va¡iable costs ) = $3. 55 In the Gawfer" Ranges, 6 to 7 wethers can be run in the area requi::ed for 1 steen. CHAPTER 3

COMPARISON OF SHEEP AND CATTLE PIOSPHXRXS OVER A SHORT TERI'I; METHODS

3-1 INTRODUCTION

This chapten discusses the techniques used in compa::ing vegetation dynarnics in a newly esta-blished cattle piosphene with those Ín a sheep piosphene oven 22 months. The object of the study r¡as to detect any differlences in pattenn o:: tr:end unds: th-e two animal-s.

Two independent approaches wer?e used, one quantitative and dealing with, biomass and density of species, estimated fnom quadnats, the othen semi- quantitative, using steneophotognaphs in an independent sampling system.

Sections 3-2 to 3-5 deaÌ with the fir"st appr.oach Cmajon study) and section 3-6 with the second (minon study).

The nequi:rements fon the reseanch r¡rere: 1. A paddock about to be, or recently, tu-nned oven to cattle;

2. A paddock still- unde:: sheep with a compa:rable vegetation;

3. Both paddocks remaining unden the same station management,

giving an empinical equivalence of stocking rates; and

4. Cattl-e and sheep ::estricted to the par.ticulan aneas and not intenchanged within the study period. In 1972, only two pairs of sites fulfill-ing these requinements wene known, viz.--on Roxby Downs station, whene cattl-e were soon to move into a fonmen sheep area, and. on Mt Victor. station, wher:e cattle had been run on fonmen sheepland since Novemben, 1971- (locations; fig. 1.1). The finst was not acceptable due to difficulties of access combined with a hetenogenous vegetation (see chapter 5). The second offened easien access and a wider. lîange of vegetations, more uniform in stnuctwìe on the sca.l-e of sampJ-ing considered; although thene was the disadvantage that the initial impact of cattl-e could

-46- 47. not be observed di::ect1y. A detailed descniption of the envi:ronment and fandscapes of Mt Victon is unnecessany; the full- discussion on the adjoining

Koonamone station, given by Canrodus et al Cfg6S) applies equally to Mt Victor station. The majon featu:res ane as fol-lows.

Mean annual rainfall (1958-L972) is .l-97mm, rvith a slight summel? maximum. (fa¡fe 3-1). The observation unfortunately coincided with the heaviest rainfall- on recond fon the area (Bur"eau of Census and Statistics, 197t¡). The majon land systems

(fig.3.1) ane gently undulating plains and downs with duplex orl g:radational soifs (nomencl-atune: Northcotertg65); with

Myoporum pLatyca.r'pum, Acaciq ØLelt:Í¿q, Kochia species and Atz'ípleæ uesicar,'La as char:acten plants, in tall- shrubf and o:r lott open woodland fonmations (Specht 1972). Thene ane small- r'ocky nanges bearing chenopod shnubl-and (in southern aneas) or A, qnewa. taI1 open shnubl-and (in northern ar"eas). The lease was taken by the present l-essee in November" f971, and cattl-e wene introduced by him, for the fir"st time. The plan of the station is given in figune 3.2. A bnief ireconnaissance detenrnined the most comparable piospher"es, by refenence to soit types, nelative pnopolrtions of majo:: pasture plants, amount and type of t::ee cover, and topogr"aphy; within the l-imits of sampling anticipated (section I-72). Figur"e 3.3 illustrates thein sal-ient features. The cattle paddock sel-ected was Black HiIl, mainly level with a gentle south slope. Bfack Hil-I itself, a nocl

TABLE 3.1

RATNFALL (mm) er MT vrcroR srATroN, tgsg-!972

L9 5B L9 s9 L9 60 L961_ L962 L963 L964 L965 1_966

J 49 30 23 0 63 13 6 0 7

E t7 16 4 \' 0 4 5 9 49

M 48 1B 0 0 t2 0 0 0 8 A t2 0 6 27 0 2I 14 0 t M 6 25 34 6 2B 44 15 7 t4 J 0 2 4 1 0 40 2 t9 26

J 1B t7 27 2 3 7 + 6 0

A 49 7 I t4 10 1B 34 20 0 o ù 72 I 19 9 0 4 78 15 t 0 55 10 2 2 72 19 23 7 7t N 63 0 2I 15 0 0 7 7 13 D 20 25 5 16 67 72 10 I 75 349 158 1_53 95 L95 1,76 19B 86 295

1967 1968 L969 1-970 797L L972 MonthLy Propontion meqv¿s 0f ømuaL meavl J 11, 87 34 0 0 143 31 .16 F 16 0 44 0 26 74 14 .07 M 2 0 33 2 105 0 15 .08 A 0 48 6 62 29 2 15 .08 M 6 40 16 0 77 10 t7 .09 J 0 19 72 5 16 5 10 .05 J 0 23 3B 0 38 7 t2 .06 A 3 I 3 28 19 1B 16 .08 S 10 0 9 62 26 4 77 .09 o 4 0 I 1B 0 16 t2 .06 N 0 9 5 7B 36 t 77 .08 D 0 10 0 7! 27 2 19 .10 52 244 209 266 339 222 L97 L. 00 F]GURE 3.1

VEGETATION SYSTEMS ON MT VICTOR STATION t HILL SYSTEM: skeletaf sandY soils.

1a Low shnubland: Att"ipLeæ uesicaría

1b Tall open shr"ubÌand'. Acasia ØLeur'a. with ¿. tetra- gonophyLLa

2 FLOODPLAIN SYSTEM: deep Ioams.

2a Low shr:ubland: KocLtia pynatnidat¿ with Nitraîia sehobeYí

2b TaIl shnubland: Acacia u¿etov"Lae with .K. pyt'ønidata' K. bz'euifoLia

3 SOUTHERN PLAINS SYSTEM: texture contnast soils.

3a Low shnubland: 3ai. A. Ðesiearia, 3aii. /. üesicazia and Kochia qstrotr"icha 3aíií. Ko chí a s edi foLia 3aiv. .K. ast!'ot?Lcha

3b Ephemer:al henbland: grôasses and Bassia species.

3c Low open woodland: 3ci. Myoporum pLatycæpun 3cii. Cagua?Lna cz'istata Sciii M. pLatycarpum and A. dneut'a. 3civ. /. ØteuÍ'a and Cassia nemophiLa

4 NORTHERN PLAINS SYSTEM: san<1y gnadational soil-s.

4a Low shnu-bland: 4ai. K, astrotrLcha Ltaii. .4. uesícar"La r¡ith .K. astv'otr"íeha

4b Ephemenal her.bla¡d: grasses .

4c Low open woodland: 4ci. Acacia ØLeura 4cii. Ez,emophila duttonü with E. sturtii 4ciii . Ca,suar"ina cr"is tata lo

3o i

iii

3b 3

¡ii

l0km

-r (--\

;- --1

o I

/.ó /.

l+c

I 3ro¿5 S

I330,5 L

Fig. 3.2 EMU LEG MT V¡CTOR STATIOhI I I PADDOCK PLAN NORfH fowERoo

5prrngs 0om Dom t SOUfH SPRINOS 8orÊ or W.ll + fowERoo Eoglahowk Tonk ond Trough a V 0oñ

Homcsteqd I Stolron Bor. lrqck oLo stAftoN Rocks FrncÊlin? 8ùe

Poddock Nome COTV.S (ltP ¡ Hors. Poddæk) Old Stotoñ Do

þkm

Corp'hd troñ unconlroll.d oe.iol pholoqrophs OOUBLE WELLS

ooubl. Wells BøC

0om Moashès Domç CHIMNEY MARSHES

Chrmñcy ¡ Oom

Oom Clumnry { Mdde Oom ond Bo.. Sora ARKAROOLA

RAM

I(ARLINKA Hul oom IIN HUI

ltN NORI H Oom HUf EINOARIGA wEsf Coñr 8da cor'l's Fig. 3. 3 VEGETATION AND LANDFORM OF STUDY PADDOCKS

CATTLE PADDOCK (Btock Hiil)

Block H'll Borê >/ n:f \ \ eìocli i\H \> ì\ 5toney trSe ^\

\ { Elock Hrll 0om CHARACTER PI-ANTS

Trees / tqll shrubs A. oneuro SHEEP PADDOCK (Emu Les) ,,:,i Cosuorino cristolq EremophÌlo species Myoporum platycorpum

Posture plonts \ ñ K. ostrotricho N ,, /4. vesicorio K pyromidota ,T Ø Ephemerols

\ * Woterpoint --.-'Trock .-'Droin :\ 1mi l-:-_r_-J. lkm 49.

Acaeia qneuta-A, tetragonopvuJLla tarl open shnubland, with a g?ass

undenstoney. The l-evel- areas, with gradational soils, Ìreï'e covened mainry by Kochia astrotricha row open shnubland, with scattered

groves of A, Øreura.. Howeven, the south-west cor"nen had

M. pLatycarpum-A. aneura low open wood..l-and vrit]n Kochia pyratrn dnta understorey. The paddock was ser^ved by two weJ-l-sepanated waters, on the southern and nonthe::n fences.

The deciding factor" in the choice of the paddock was the position of the south,ern waten, Black Hill- dam. Hene, the bush grew to r.¡ithin 200m of ttre wate::, much c.l-osen than oth¡-er" situations observed. This atl-owed analysis of g:radients in this majon pastune species, fuom the immediate vicinity of the water" to the l-imits of sampring. (the importance of the. bluebushes in s.A. rangel_ands was discussed in section l--5. ) Further, the dam was not praced in a watercou:rse, with th-e concomittant soil and vegetation discontinuities, but was fed by J-ange dnains, utilising the genenal souther:ty slope of the anea. Thus it coul-d be r:egarded the same as a pipeline tnough, placed ar.bitnar"ily with nel_ation to soil_s and vegetation (c.f. Banken and Lange, l-969 ).

The sheep paddock , Emu Leg, had unifonm topognaphy, soils and. vegetation. ft was flat with a gentls nonthenì_y s]ope, had g::adat- ional soils the same as those in, Bl-ack Hill, and had a row shr"ubland of K. astz,otz,icha wi-tí gnoves of A. ryLeuya. The density of the tatter incr-eased to the nonth. The paddock had two watens, each with encincring vegetation simil-an to that around B]ack Hil-l dam, though with slightJ-y more Á. aneLffa. Each waten was a tnough suppried by pipeline, anbiüranily placed with negard to soil_ vegetation, and sited l-ess than lkm apant nean the south fence. plates J.rTrg,4 show PLATE 1. Black Hill Paddock, Mt Victor Station. Kochia astyotricTta slnrvb pasture (A), with the gnasses Enneapogon auenaceus and E. cyLinÁrieus fB). Tnees ane Acacia aneura. 1'3/2/74

PLATE 2.

Bl-ack Hil-I Paddock, Mt Victor Station. Kochía pyr'ønídntø shn:b * pastune((e), wíth Acacia anaæa (B), Gnasses are Enneapogon species. t3l2/74 }Í tl. :.

nf's : * Si PLATE 3.

B]ack HiIl Paddock, Mt Victor Station. On Black HilI' looking towand

Bl-ack Hil-I Dam (A). G::ass pastur:e on skeletal soil-, with the tree Acacía tetragonophylLa ß), ß/2/74

PLATE 4.

s]nr'¿b pastu::e Emu Leg Paddock, Mt victo:: station. Ibchía astz'otrieha (B), (¡), witn the grasses Enneapogon a2enqceus atfd E. cyLindricus Stipa nitidn G). Tr:ees a:re Acacia aneurd" !3/2/74

50. raepnesentative ¡¡iews of the vegetation.

The major. problems in companing the two Ìíene: 1. The presence of Black Hil-I in the cattle paddock--the vegetation hetenogeneity this caused was avoided by

nestnicting sampJ-ing to the K. astz'otricha type, but the hill nay have introduced othen variation by influenciñg

cattle movement.

2. The pnesence of juxtaposed r¿ater"s in th-e sheep paddock velrsus a single waten in the equivalent position in the cattle paddock--this situation was furthe:r conplicated by failunes and the eventual closing down of one of the waters. 3. The 1a::ge scale lrete::ogeneity of the cattle paddock vegetation--while this was excluded fr.om tlre actual sampling, it

yet may have influenced util-isation of centain pants of the pastr:re. Necessarily, this woufd alten thre magnitude of stocking effects in a way not present fon the sheep case.

Nonetheless, the tv¡o situations r^rene the most satisfacto:ry availabl-e.

Because of the unavoidabl-e differences between paddocks, emphasis

was placed on detail-ed anal-ysis of effects within paddocks, with a

more general compar"ison between paddocks. The dinect-ì-on of

diffenences r^ras sought nather than absolute rnagnitude.

g-2 APPROACH TO THE MAJOR STUDY (Se¡lpl,ltte BY QUADRATS )

obsenvations wene made oven 22 months, fnom septembe:: L972 to August 1974. Eigh-t obsenvations hfelre made foir each paddock, at approximately thr:ee-month intervafs, following a pilot obsenvation in I'{ay in the cattle paddock. Further observations llelre attempted ' 51

TABLE 3.2

DATES OF OBSERVATIONS ON MT VICTOR

Given as the mid-date fon each tnip, with the intenvaf between observations.

TRÏP DATE TNTERVAL (døys) Cattle Sheep Cattle Sheep

(pilot) o 6-7-72 90 t 4-t0-72 2-9-72 57 94 2 30-tt-72 +-t2-72 80 72 3 tB-2-73 1-5-2-73 7+ 75 + 4-5-73 t-5-73 89 86 5 29-7-73 26-7-73 100 100 6 6-17-73 3-7r-73 104 104 7 79-2-74 t5-2-74 + 165 172 :! I 3-8-74 6-8-74

No access between tnips 7 and 8. 52.

.Tune and ear:Iy Ju]y, L974' but access l^ias impossible due to exceptional_ rains. Tabl-e 3.2 gives the mid-dates for each observation. The fir-st in each paddock wene not at equivalent dates, because of ttre extt-a time nequined fo:: initial setting-up. Fon the limited peniod, in an area of ennatic climate and plant gfowth, it was vital to concentnate on featu:res likely to yield pnecise infor.mation fon the effort expended; as follows. 1. Onty those patte::ns and changes \^Iere sough-t, which could develop in 22 months: pattenns relating to past stocking histony (..g. in the density of tong-lived penennials) could not be reganded as germane to the pnesent pnoblem. 2. The study was nest::icted to the dominant vegetation type, as chanacterisecl by K. astz,otricha, to ninimize va::iation intnoduced by lange scale pattenn.

3. The most time consuming, but most sensitive, measul?es hrere used only on speeies visibly fo::ming the bulk of the

Pasture. {Þ 4. Sampling was nestr-icted to a maximum of 3km fuom the nea::est water:, following the national-e of section I-72.

5. Whil-e the rnajor cont:rol of the obse::vation I^Ias inhenent in companison of the two piosphenes, a limited numben of

exclosulles r^rere planned as separate contr.ol , to check vegetation dynarnics between paddocks in the absence of gnazíng.

3-3 PASTURE POPULATIONS STUDIED AND PARAMETERS ESTIMATED

3-3L Maior specíes.

Majon species wene defined as those which visibJ-y contnibuted 53

the bul-k of the standing pastur.e. In both paddocks, they wene the perennial bluebushes Kochia astnotricha and K, pyz'ønidata, and the ephemenal gnasses Stípa n|tida, Enneapogon QUenQ.ceLß'

E. cyLinfuícus anò. At'istidø contoY'ta, TtIe Enneapogon species contnibuted at l-east half tt¡-e standing b:'-omass of the pastune fon most of the peniod.

These species p::ovided an excellent basis fo:: interpnetation of pattenn due to pnefenentiaf gnazing selection. They al-l-owed a thnee-way comparaison, of bush with gnass, bush with bush, and

gl?ass with gnass. They a:re mone ful-l-y documented than many

othen common anid pasture plants, although such basics as phenology, population dynamics and the l-ike ane still- la:rgeIy unknown. Fon sheep, K. astrol;t'ichø' and K. pgrønidøta are

neganded as highly palatable and unpalatable nespectiveJ-y in

most assessments (.. g. Jessup, 1951) : simila:rly, the grrasses pnovide a sliding scale fnom the supposedly highly palatable Stipa to the narely gnazed Ay,istida (Osborn et af, 1931; Jessup, 1951). a Panameters estimated fon these majon speeies l^Iene density, and

above-gnound feaf biomass. Density is a much used and easily inter:pr"eted par:ameten (Gnieg-Smith, l-964). In the pnesent study, it is def-ined as

the numben of individual-s per" hecta::e, gained fnom an initial- count of individual-s pen quadr:at. It is limited in its use by the r.igoun of defínition of the unit to be counted (Watker", l-970)' Gnieg-Smith quotes several counting methods using both plotless

ancl quad::atting techniques. Estimations may be used nather- than dir"ect counts, e.g. visual assessment from a vehicle tnansect 5q. (Jessup, 1951- ) on fnom li¡e tnansects (.l4cfntyne, l95g ) . Dinect

counts ane obviously supenion to estimates, but, even so, Lyon (rg0g) points out that the vaniances associated with any of the panticulan methods a::e rrl-ess than satisfactonyrr when applied to

shrubl-and siniÌan to th-e pnesent case. Quadr"atting will give l-owen variances than plotless methods, and ncmrohr-nectanguran quad-

nats give a lower: variance than the mone usual squane. (wight,1967,

draws simil-ar concl-usions with- negand to biomass). such a shape wiJ-l

reduce va::iances by cutting acl?oss any pattelrns caused by contag-ious

distnibutions at the level of the individual (discussions on such distlributions ane given by Andenson, 1971, for^ an Austr"alian situation, and Gnieg-Smith and Chadwíck, 1965, fon an Afi:ican situation).

Definition of individual-s of K. pyrønt data, Stipa and

Atístida p::esented no pnoblems in the field; the finst was prîesent as a discrete bush, the l-atter two as discnete bunches, and these

units hrere sconed as individuals. Both of tTrc Enneøpogon species q wene pnesent al-so as disc:rete bunches, but care was needed to dis-

tinguish the separate bunches when p::esent in quantity. Furthen, the two species wene difficult to sepanate without flowenheads,

panticularly aften pnolonged gnazing. As this may have led to

senious confusion, the two species wel?e pooled fon counting and

estimation punposes, as Enneøpogon spp..tl

:F M. Cnisp (Ph.D. Thesis, Botany Dept., Univer"sity of Adel-aide, 1975) has infonmed me that th-e two species share a similan habitat on the Koonamor.e Vegetation Reser:ve; and that cross-fence comparisons between protected and gnazed aneas indicate that stock effects ane of the same or.den and dir.ection fon both. 55.

Individual- definition was difficul-t fon K, astz,ottfcha. fn this study, an individual- was defined as a molre o:: less continuous canopy for"med by one on several shoots on a common mound. The difficul-ties anose because mounds we:re often formed anound stems emanating f:rom diffenent nootstocks, oll separ"ate mounds associated with two stems fnom th-e same nootstock; and a continuous canopy could be formed by stems fnom different rootstocks. The above definition hras considered the best compromise in an untnactable situation. For. considenation of the shont tenm dynamics of the long lived species, sepanate counts wene made fon juvenile bushes. These wene defined as both tnue seedlings, and smal-l- shnubs with secondalry thickening, less than lOcm in height on diameten.

Bíomass is an appnopniate panamete:: to measure, fuom a gnazing viewpoint, as it is usual-l-y mol:e jmmediately affected by g:razing than density. Further, it is mone :readily intenpr-eted than ott¡-er sensitive measunes of abundance, such as aenial cover on basal anea.

The most st::aight-fonwa::d measunement is by destnuctive hanvest; un::easonable beyond a centain point, in tenms of both the time nequined and the distunbance caused in the tanget populations. Fo:r extensive measwement, estimation methods, calibnated by limited hanvest, must be used.

There are numenous tech¡iques availabl-e. Fo:: example,

Hutchings and Schmautz (l-969) discuss a method which- exPnesses the visual estimation of quadnat biomass in nelation to a neanby standard, laten harvested; Mannetje and Haydock (fgOS) describe a system of ranking species by estimated weight; RobeJ- et al (1970) use a visual- obst::uction technique; Robands et aI (1967) 56 descnibe a di:rect visual assessment, based on tnaining by assessment and har:vest; and Campbell a¡d Annold (1973) describe a simil-ai: system, æd also discuss ttre use of capacitance meters. With the possible exeeption of th-e capacitance measures, aÌ1 the methods are lrepolrted to wo:rk successfully within the pastr:re types concenned. The l-ist of methods is by no means exhaustive (fon othens, see Monley et al-, (t904), and Walken, (1970) ). Howeven, the neports indicate that any method is likely to b e effective provided that adequate contnol- and training is maintained. The method used hene was a modification of the eanly and veny simple system of Pechanec and Pickfond (1937). In thein method, the biomass of an individual- is estimated as a multiple of a hand- held specimen of known biomass, and accuracy is maintained by nepeated tnaining (estination followed by hanvest). ft has been used successfuÌly in S.A. nangeland situations by final yean unden- gnaduate students (Dept. of Botany, Univensity of Adel-aide), and was extensively used by me, with acceptable nesu.l-ts , in an eanlien study.:'; Its advantages are the ease, speed and simplicity of estimation. In the pr:esent study, the following pr"ocedures l.Iere used:- 1. Leaf biomass (wet weight) was estimated. Fon grasses this was defined as the total standing cnop, dead o::

alive, ovelr J-cm above the ground (as cut with shear"s ) . Fon the two bush species it was defined as the pnopontion

:t Fatchen, T.J. (1971); Measunes of vegetation and stocking effects on South Australian pasto::âI stations. Honou::s Thesis, Dept. of Botany, Univer"sity of Adel-aide. 57.

of above gr:ound biomass, dead on al-ive, removed by hand-str ipping leaves (thus excÌuding the bul-k of the

woody br"anches firom estimation ) . 2. A single hand-hel-d specimen was used for: each- species on each dayrs run 3. The finst hor:r. of each day was occupied in tnaining (longer" fon th-e initia] observations ) and tr:aining was continued fon each species until three consecutive estimations fel-l- within the desined accunacy. Spot checks on species were

made at intervals during the daY.

4. The maximum er?lror allowable was set at 15%. (Othen studies, panticulanly tlrose of Tnumble and !{ood:roffe (fgS+) and Leigh and Mulham (tg6Oarb) have aimed at highen accuracy.

Howeven, th-e effont expended to gain inc::eased accuracy is veny great, and obse::ven fatigue has a much gneaten as shown by Morley et effect within finer elrrlolr l-imits ' aI (1964). Results indicate that the variations in biomass of target populations both ín time and space wene such as to nenden the enron at each- sampling point insig- nificant by compa:rison. Hence field estimates were tneated in analyses as measurements without ernor:. 5. Estimations r¡rere sconed fon groups and not individuals. This atlowed a carry-oven of estimation f:rom one irdividual to the next, thus reducing nelative ennon when the individ- ualrs biomass was not an exact multiple of the hand-hel-d specimen. Grasses wene estimated in five anbitnany bfocks per: quadnat, bush-es in g::oups of three to five 58.

individuals, and the sco:res for. the panticulan species

summed to give the quad:rat bulk estimate' a calibration nun hlas car::ied 6. Du:ring each observation ' out fon each species, to pr:ovide infonmation on accuracy

oven a numben of measu::ements with no intermediate checks.

The run was made at the end of the day, when fatigue was greatest. Twenty grouPs of individuals of the panticular species we:re estimated in one run, with no pnelimina-ny or intenmediate tnaining and har.vest. At the end of the ::unt all gnoups were harvested and weighed' (X' pArqmidata had to be checked by estÍmating pa:rts of individual-s on small individua]s: the size of most individua].s made stnipping of gnoups fo:r cal-ib:ration impractical in terms of time).

Runs fo:: all species, all times r ane given in figunes 3.4to3.S.Mostestimateslaywittr-int}re15%er-non limits, but theire ÍIere occasionally widely vanying estimates. Morley et al (1964) suggest that th-ese result from fatiguerand ane pnobably unavoidabfe' K' pyt'øn'Ldata estimations showed the most fnequent discrepancies; this is ascnibed to the ,thicken foliage of the species obscrrring theleavesbefowthecanopy'Thisrcombinedr'¡iththe lange size of many individual-s, flâY have led to a signif,-

icant pnopontion of inaccurate estimates ' 7. The hand-Ìreld specimens to be used for a dayrs estimation gene::al wer-e sefected at the beginning of ttr-e day from the aneaoftlrenun.Thisavoidedspecimensfromrfonexample' Fig. 3.4 BIOMASS ESTIMATIONS AND ACTUAL WEIGHTS FOR S. dII]DA

100

.: o o o Time 0 Time I Time 2 ô o (5heep) o uÞ U U F F 50 po EO Ë É þo f¡ 50 þo t5() ts F-: U U U

a Time 1 Time 2 l+ (cottlG) Time

s0 100 Éo 50 þ0 EO 50 r00 r50 ACTUAT (9) (s) ACTUAL (9) ^CTUAL t50

too

o .t o Time 5 Time 7 ô o u U F F 50 l0o 't50 É E 50 100 150 Ø * U U

100

ïime 6 Time I

50 100 rs0 50 r00 ÂCTUAL (g) ACIUÄI (q) Fig. 3.5 BIOMASS ESTIMATlONS AND ACTUAL WEIGHTS FOR ENNEAPOGON SPP.

ó o o Time 0 Time 1 Time 3 o o (SheeP) o U U U ts È Þ 50 þo 1s0 t 50 00 t50 E 25 50 otr È ts U U U

L Time 1 Time 2 Time (Cottle)

50 þ0 150 25 50 '15 25 50 't5 ACTLIAL (9) ACIUAL (g) ACTUAL (s)

r00

o o Time 5 Time 7 o o U U ts F 50 100 150 s0 100 r50 F Þ- U U

Time 6 Time I

50 100 150 50 r00 150 AC'ruaL (q) ACfUÅL {q) Fig. 3. 6 BIOMASS ESTIMATIONS AND ACTUAL WEIGHTS FOÍì NTOR

50 o g o I Time 3 Time 0 Time o ô o (Sheep) U UJ U È k 150 25 50 75 50 r00 r50 50 r00 3 _Ft z oÊ ø gJ ÙJ U

Time I Time 2 Time 4 (cottte) r50 50 75 25 50 75 50 100 25 (q) ,TCTUAL (9) ACTUAL (9) ACTUAL

þo

50 50 o o Time 5 Time 7 o ô U U ts F : 50 þ0 r50 r- 50 r00 t50 F F u U

Time 6 Time I

50 50 r00 150 ÁCIUAL (9) ACIUAL (g) Fig. 3. 7 BIOMASS ESTIMATIONS AND ACTUAL WEIGHTS FOR T

300

ro0

r00 o o o Time 0 Time 1 Time 3 ô o o U U l¡J ts È roo æ0 300 ä r00 æo 300 50 þe 150 5 z ¿ F ts ts rn t^ U U ul

100 b 1 Time 1 Time 2 Time (cottle)

þo 200 300 50 100 t50 ro0 200 300 ACIUAI (s) ACTUAI (s) ACTUAL (s)

300

200

u o Time 5 lrme I ô o U U ts F Ð0 300 300 ts ts l/) a U U

200 200

r00

Time 6 Time B

t00 200 300 200 300 (g) ACIUAL (g) AcluAr Fis.3.B BIOMASS ESTIMATIONS AND ACTUAL WEIGHTS FOR K. PVRAMIDATA

o o Time 0 3 c¡ Time I Uo Time 3 U o (SheeP) F u 3 Éo 2æ æo ä ts Þã I ro0 200 300 roo 200 æ0 6 útr U U U 600

Time 1 Time 2 l. (cottlc) Time

200 t00 3(þ r00 æo 300 æ0 ¿00 600 (9) ACÍUÁL ACÍUAL (9) ACTUAL (9)

r00 o I q do Time 5 Time 7 o U 3 ts à too 200 æo U 100 æ0 3@ øts= U

ro0

Time 6 Time I

t00 200 300 r00 200 300 AclUAt (g) ACIUAL (9) Fig. 3.9 WET AND DRY WEIGHTS OF S. NITIDA HAND HELD SPECIMENS

ttl I l- Time 0 ¡- Time 2 IT I trJ I UJ 5 to t5 F= = 5 r0 l5 l¡J F ì ìUJ t0 r0

Time 1 ¡ cottle Time 3 . sheep

510 t5 510 15 DRY WEIGHT (9) DRY wElGHr (s)

oì Í'l F Time 4 ; Time 6 ö trJ ö ì t¡J 5 t5 3 t-- þ 5 10 15 t¡J F UJ = Ì

5 Time 7 Time 5 ^ Time I

510 l5 510 15 oRY wElGHr (g) DRY WEIGHT (s) Fig. 3.10 WET AND DRY WE¡GHTS OF ENNEAPOGON SPP HAND HELD SPECIMENS

5 0 d 0 Time 0 Time 2 Time 1 F :- ü 12 E g U U 5 ro t5 3 5 t0 I 5 ro ti F F U E ì H

Time 1 Time 3 Time 5 . deod

't5 5 þ t5 5r0 t5 5 t0 MY WEreHT (9) DRY WEËII (9) ORY wErc'HT (9)

6 o ¡- Time 6 I ts a I UI ! 5 f) t5 I ts U Þ ¡ ìt¡J to Time I

5 l0 15 æ Time 7 oRY WEIGHT (9)

5 t0 t5 mY wElGHl (9) Fig. 3.11 WET AND DRV WEIGHTS OF A. CONTORTA HAND HELD SPECIMENS

5 o g )- a Time 0 rF- Time 2 I ü l¡J l¡J 3 5 r0 t5 = 5 r0 l5 t- l- l! td 3 I

, .l ) Time 1 . cqtlle Time 3 ¡ sheep

510 l5 510 t5 (g) DRY WEIGHT (g) DRY WEIGHT t5

ol c'ì F b t- Time I ltme I I 9 t¡J t¡l 5 10 l5 ì 5 10 t5 F.= F trl td È = ¡¡

5 5 .Time 7 Time 5 ,Time I

510 l5 510 t5 (g) DRY WEIGHT (9) DRY WEIGHT Fig. 3 .12 WET AND DRY WEIGHTS OF K. ASTROTRICHA HAND HELD SPECIMENS

s0 50 50

30 30

20 20 .,: l0 t0 Time 0 Time 3 Time 6

510152025 5 r0 5 r0 o 10 t\ t I I c, 30 30 30 t¡¡ I t 20 20 l- n t f s r0 Time 1 l0 t0 r cottte Time 4 Time 7

510152025 5 t0 5 40 40 40

30 30 30

r0 t0 10 Time 2 Time 5 Time I

510 rs 2025 5 10 5 t0 DRY WEIGHT (9) Fig. 3.13 WET AND DRY WEIG}-ITS OF K. PYRAMIDATA I-IAND HELD SPECIMENS

40 s0

30 10

20 30 I

10 10 Time 0 Time 3 Time 6

10 ?o r0 20 r0 20 I l+O

40 ì30 t I I $æ ft n 30 t- Time 1 20 st0 r cottle Time 4 Time 7

10 20 l0 20 10 20 40 50

30 30

20 20 30

10 20 Time 2 Time 5 Time B

t0 20 20 10 20 DRY WETGHT (g) 59.

a lrghtly grazed a::ea being used to estimate plants in a

heavily gnazed auea, with pr:obable diffel?ences in for"m and foliage.

8 Aften tnaining, ten fur.then speeimens of each species,

estimated as equivalent in leaf biomass , r¡ret?e collected, and weighed en masse and stoned fon l-aten oven dnying.

This neduced :relative ennor in measur:ement of wet weight due to insensitivity of field balances. In addition, it

yielded a mol?e accurate measune of the estirnation unit

weight, by reducing the errrlor? implicit in the individual variation of specimensi

Estimations wer.e of wet weights. Thre dry weight equivalents were obtained by oven drying the sto:red samples at SOoC fon 48 hor.:r's. This nequir"es a linean relationship between hret and dr:y weights. Figu:res 3.9 to 3.13 show this to exist for- any given time, aÌthough the nelationship may vary considenably with tine.

3-32 Mínor species.

Density counts of al-l individual-s wene made for alf othen species encounter.ed in quadnats (listing:tabl-e 3.3). Most wer:e epheme:-als. The two majon pnoblems in counting wel?e Ca) visual obstnuction by lange bushes arfd gr:asses, and (¡) tfre numben of vegetatively sinilan species; e.g. Bassia diq.cqnt?n, B, obLiquieuspie,

B. patentícuspis, B. éelev,oLaenoides and MaLacoeerq, tnicorme, highly simil-an in appeanance, and fnequently encountered. Visual obstruction l-ed to spurious nises in appanent abundance of othen species as the grasses trere gnazed down, while the identification 60.

TASLE 3.3

PASTURE SPEC]ES FOUND IN QUADRATS AT MT VICTOR

POACEAE

Atistida contoz'ta Enagrostis dieLsLi

Dactg Lo cteníum raduL æts LørtifLoz'a Enneøpogon auena,ceus Stipa nitída eyLindrLcus Tragus austv'aLtæ¿us Trirøphís molLís

LILIACEAE

BuLbine sewLbarbata

CHENOPODIACEAE

AtrLpLeæ Lintbata Bas sia s cLeroLaenoídes spongiosa Chenop odíum des ez'tozam stipitata EnchyLaena tomentosa uesícarLa Koehía aphylLa Bassia bifLora aßtrotricha brae?typtera eæcansat'a dectuvens georgeí diacantha pyrønLdata diuarícata sedifoLia Løticuspis tomentosa F Lintbata MaLacocera tricorvte obLi-quicuspis Rhagodía spinescens van deLtophyLLa patenticuspis SaLsoLa kaLí porad.oæa.

AMARANTHACEAE ZYGOPHYLLACEAE

PttlLottts obouattæ Tr"LbuLus tetryestvLs

FABACEAE SAPINDACEAE

CLiøtthus forrnosus Dodonaea mLcz'ozyga

GERANIACEAE EUPHORBIACEAE

Erodium cAgnorum Euphonbia drumnondii 61

TABLE 3.3 (Cont.)

MALVACEAE

AbutiLon otocarpum Sida íntu'ieata Sída cormtgata petrophíLa vaÏ1 uirgata Øtg1rs tifolía

CONVOLVULACEAE MYOPORACEAE

ConuoLuuLus er,¡'tb es cqns EremophíLa duttonii

ASTERACEAE

Angíørthus pusiLLus He Lip te rwn floz'ibundum Braehyscome ciLiæís ?Mintæia LeptoPltylLa

C aLo e ephaLus dit trL c\ii

I 62. problem, if not adequately solved, would have pnoduced spu::ious negative connelations between the species confused. Fon the most difficult species, identification was aided, in the field, by Iabelled vouche¡s.

3-4 SAMPLING SYSTEM

To answen even the gener"al question of whett¡-en stock-i-nduced .diffenences existed between the paddoeks, it was necessany both to define wor.king hypotheses and to sel-ect methods of analysis, befor:e commencing sampJ-ing. While studies such as this nequire a broad data base fon fonmulation of specific hypotheses, the sampling system must have a rational- basis to ::eturn usable information. Much wasted effont and intnactabl-e data can result if the appnoach is to alnass data first, then to considen its tneatment.

The general hypottr-eses were:

1. Background vegetation pattern and dynamics wene the same in both paddocks.

2. There wene differences in the effects of sheep and catt.l-e.

3. These effects could be isol-ated f:rom background noise by

the detection of piospher:e symrnetny.

The seanch was thus fon pastune vaniab.les which had a changing nelationship with distance fnom waten: the nature of this nelationship pnovided the basis of companison between the two paddocks.

3-4I SatnpLing fot, eattLe/shrcep cornparísons.

The analyses mainly considered wene neg:ressions of plant 63.

palrametens on distance fnom water, requilsing at l-east some fonm of nandomisation to estimate the variances invol-ved. Ba:rken and Lange, fon investigations of piosphene pattenn, employed regulan sampling. In the context of single surveys with mapping as a main aim a negulan system is appropriate (t^littiams, I97I), a:rd such sampling is used in the late:: chaptens of this thesis on this basis. However, such a system cannot provide a statistically acceptabJ-e estimate of

vaniance (Gnieg-Smith, l-964; Kenshaw, 1973 ). Restricted nather. than full- randomisation was used to inerease

p:recision, as demonstrated by Pechanec and Stewar"t C1940) fon biomass, Goodall (fgSZ¡) for cove¡ and Bordeau (tgSg) for- density and basal- al1ea. Fur"then, ::estnicted randomisation pnovided a mone even spread of quadr.ats fon possible secondary mapping than full r"andomisation, Seventy quadnats were l-aid in each paddock, as shown in figur:e 3.1-4, with a sampling density of-10 / U^2. The numben was

the maximum penmitted by time, and was J-arge enough to al-fow a

wonkabl-e estimate of statistical panametens. + The cattle paddock was laid fírst, tLre shape of ttre sample area

being govenned by the 3km limit and the hete::ogeneit:i-es mentioned ea::Iien. The sheep paddock sampJ-ing anea was set up to app::oxim- atê this shape, in an attempt to avoid any hidden bias (".g. potential I'g::azing into the rn¡indtr effects). Incomplete sampling btocks a¡ound the fringes were al-l-otted quadnats ín p::opontion to thein alrea.

Sunveying was by pnismatic compass and rangefinder', with watens

and the southenn fences pnoviding neference points. Due to the pnoblems of estimating bearings of wire fences r a false angle was ::ead in th,e cattl-e paddock. Duning the fuiitial- quad::at enection' Fig.3.14

(O) CATTLE PADDOCK (BIOCK H¡II) QUADRAT LAYOUT

56 66 57

555/- 65 52

E 64 5l 6a 50 53 69 70

62 63 38æ t& E ns o' 49 5ô ¿l ¿6 59 10 1t 4g 47 60 t. 6l 25 24 21 t9 36 17 29 E 26 3a 33 2e 30 20 35 t2 3r n 23 22 6 lt r5 11 r8 t? 16

I 3 5 ? r2 I I 7 c r0 tl E t E lT Oom H¡tt 0 250 500 7so fooo matrt5

(b) SHEEP PADDocK (Emu L"g) oUADRAT LAYoUT

à ztzz 20 32 \ 2l 36 35 /""""t r9 3r 34 33 J 30 t 29 t6

28 O14 't7 N 21 26 r3 16 68 l5 53 58 25 56

r0 9 57 2t, 1't 12il 7 ô 52 12 a 5l 5¿ 55 56 70 6 50 69 J 39 19 ¿t ¿0 63 65 ¿5 æ 3 16 ¿8 s9 62 64 ¿¿€ 37 E I E L1 E E 6l

0 250 500 750 1000 Limits of K. ostrotricho ----'---- mclacs

Dense mulgo groúes ¡¡¿¿¿¡¡¡¡r

Quodrqt 67

Trough a Fence -x-x-

Exclosure E 64

,trvo quadnats wer"e found to Le the wrong side of th.e fence. As

the fence had r.emoved a pnoportionate area of ttr-e sarnpling blocks

concer"ned, they wene not r"el-ocated in the block but two new

quadnats wene laid in fninge areas to keep N equal to 70.

Samples wene located at the same sites, througtrout the study. Goodall- (f gSZ¡ ) showed tL¡-at this lremoves a large pa:rt of the sampling enlron fi:om the estimate of change. (He pnesented data fon cover: changes o\Þll a yêar, obtained fnom penmanent quadnats,

in companison with data fnorn separ"ate quadr ats for the beginning and end. Var.iance of the change was l-ess when estimated fnom the per.manent quadrats ). However, successive obse::vations at the

same point possibJ-y give nise to cor"rel-ations of the successive

changes, not ne.l-ated to the :real situation (Grieg-Smith, 1964).

As yet there is no centain evidence that this happens. Such an eventuality could be avoided by a stepwise sanpling: measurement at one point at the stant and finish of an obsenvation period, with a simul-taneous measunement el-sewhere fon the stant of the next peniod. In the present context, such an appnoach woul-d have halved the time-nestricted sample number, by foncing the duplication of quadnats, and by nequii:ing sunveying of new quadrats at each time. Funthen, the cer"tain advantage of neduction in ernol?'out- weighed the potential- disadvantage of conrelated changes, if the

same quadnats were nead thnoughout.

3-42 Sønpling fon backgzowtd dynønLcs.

Majon differences in background dynamics were considered.

unJ-ikely; the paddocks had simila:r vegetations and soir-s, and 65 wene sufficientty cfose to renden differ"ing nai¡fall- inputs irnpnobabl-e. Thus onJ-y a general check was ::equined; on dinection of :response to rain and season r"ather than the absolute magnitude of the nesponses. 0n1y a small sample was considered necessany.

Within eaeh paddock, five 4Ox1Om2 exclosunes were erected.

The nunber r^ras limited ]angely by non-statistical factors, viz. a wine shortage, the expense of compì-eteIy stock-proof fencing, t::anspont pnoblems, and the time and effont j¡vol-ved in enection. The panticular. size was sel-ected, because (a) it allowed rnultiple

sampJ-ing within the exc.losure; C¡) it fil-led a camena fname (see section.3-6); (c) it penmitted a sizeable buffen zone anound the excl-osed quadnats; and (d) massive constnuction was needed to exclude cattle negardless of dimensions. Exclosures wene rest:ricted to areas adjoining tnacks, ttr-oughr

sited at least 4Om away to avoid immediate tr-ack effects. They

wene located by numbening each O.l- rnile tr.ack segment within the

sample anea and takíng five segments at nandom. Locations ane given in figr:r'e 3.14.

Two quadrat sets wene laid with-in each exclosune giving N=lO

fon each paddoek. The low numben conbined with- tlre over sampl-ing pnobably gener-ated a lange sampling ernor, and any palrametr"ic test

ni:n between paddocks fon exclosune quadnats, o:r between inside

and outside quadnats would b e likely to give misleading nesults.

Nonetheless the sampJ-e size was considered adequate to give the genenal check :requi:red, fulfilling the punPose of the exclosures. 66.

3-43 Suadrat síze øt'd sVnPe. The sizes of quadnats welre lOx]m2 for- ephemeral species

(ca. 50 species) and 2Ox2m2 fon la:rge penennial species (8 species).

They wene l-aid as shown in figu:re 3.15. The manker post was lightly seated in the gnound, and easily knocked over by stock' though nevelr dislodged entirely. This fo]lowed pnior expenience wíth stock using quadnat pegs fon scnatching posts, and consequently biasing resul-ts by trampling quadrats excessively' The Iong nectangulan shape was used for t}.e lleasons given in section 5-31-. Sizes wene set pnagmatically, with quadnats as la-nge as was feasible. The lange:: quad:rat folfowed the wonk of Bar"ker

and Lange (fgOgrl-970), who used 20x1.5m2 quadrats fon density counts in similar cincumstances, for the perennial chenopods. This size was tnied initiatly, but incneased to 2Ox2m2 to lessen edge er"nons fon K. astz,otrícha, A small-e:: size was considered in the fi:rst place for the ephemer:als. The lange penennials have a much mo:re

open distnibution of individuals than the ephemenals. Flence if

the same sized quadr.at were to be used for both, ttr-e threshol-d of observation woul-d be nelatively highen for the perennials'

because of thein wider sPacing.

Howeven, sconing time was the deciding faetor. Bot., 2ox2m2

and 2OxIm2 quadrats were altempted fon ephemenafs. The formen took two hor¡rs to score completely (in a peniod of refatively low biomass),

the obser^ver was easily l-ost inside it, and undue tr:ampling occr::rned' The l-atter took fonty minutes to scone, and although the disorien- tation and trampling pr.oblems wer"e langely eljminated, the time was still ::egar"ded as too gneat. Fig. 3.15 QUAÜRAT LAYOUT

OUT I IN our OU T IN OUT ïü IN IN 10m 20m morker peg NOT TO SCALE

Fig. 3.16 PHOTOGRAPHIC METHOD

cqmero

1.5m

sighting scqle cctmero post post post

co 20m t Þ

comero += r- 10m rodius bqseline sighting 10 cm -+t post ¡ I Þ scale comero post post 67.

Because or' the diffe:rent sizes used, all quantitative data has been expnessed in the comron units kg/ha and density/ha. llhere association analyses are applied, only incidence in the small quadnat is consider.ed.

3-5 ANALYSIS METHODS.

Ernphasis was placed on the analysis of quantitative data, using connelation, negnession and related techniques. Howeven, a first-onder analysis of species incidence was made, using Infl-uence Anal-ysis.

3-5I fnfLuenee ætalysis.

Tnfl_uence analysis (Lange, 1968) is a method for. detection of envir-onmental- influences on the basis of species association.

AppJ-ications are shown by Bar:ken and Lange (fSOg), Rogers and Lange (fgZf ) and Lange (1971_). Col-l-ectively, these papers express all the nelevant featr::res, and highlight the danger" of injudicious use, of the analysis. Bniefly, independent nodes of species inte:raction, consequent on underlying independent influences, al?e defined by the pattelrns of association. The reliability of the node is detenmined by the

amount and level of neinfo:rcement within the association mat::ix. Fon each node, two of the possibl-e combinations of species incidence are set as the poles of intenaction. fnfluenee Ta,tings (tnts) ane assigned to all possibl-e combinations; the poÌes nepresent

the extnemes of tlre i¡f.l-uence range. This pnocedure is 68

, TABLE 3.4

DETERMTNATION OF INFLUENCE RATTNGS

(a) +-sp- B ,- +-sp. E Nodes. ,- "O\ "Oi *\Ì sp. C sp. F

Node I Node fi

(b) Potes of intenaction.

Node I ABC Node II DEF ++- +++ --+

Node I Node fi Poles Species Recoding IR Pol-es Species Recoding IR combinations conbinations ABC ABC DEF DEF +++++-2 '&++++++3 :l'+++++3 ++++-2 +-++t ++++2 +++2 ++7 -++-+-t -++++2 +++2 :+-+-t :r -++I __l __-; 'Jr 0

I 69. summarized in table 3.4. Quad:rat data is then ne-examj¡red, and the various species cornbinations assigned the i:espective influence nating. Backplotting th-e IRrs reveal-s the extent and st:rength of the influence.

3-52 Cott,eLation øtd z'egression.

The text used for guidelines in the application of connefation and regression analyses was Sokal- and Rohl-f (fgOg, pp 494 et seq.).

Two sets of nefations i¡Iere sought; the degnee of plant/plant inter-action, and the nelationship of plant pa::ameters to stocking pt?essure, as indexed by distance from waten. The fi:rst set was dealt with by correlation and the nelated Pnincipal Components analysis, the second set by linea:r regnession (foll-owing th-e discussÍon in sectíon I-72 on linean nelations within the piosptrer"e)'

Fon negnession analyses, the basic assumptions vlel?e (a) ttre ¿is- tnibution of the Y vaniabl-es--p1ant parameters--apPlroximated nonmality for any val-ue of X; and (b) X variables coufd be tneated as measured without en:ror. Regr-ession analyses wer"e applied to the individual major species, but not to the individual rninor species' It was fe.l-t that the info::mation gained, if the l-atter were considened in detail, would not justify the time expenditune. Instead, the Pnincipaj- Components extnacted we:re used as the basis for summarizing piosphene relation of minon species (discussed bel-ow)' simple r:eg::essions of biomass and density on distance fuom waten we:re cal-cul-ated finst. Standing cnops were used r"athen than changes ir standing clrops, once tlre changes oven a period were found to be highl-y dependent on the standing clroPs at th'e stant. 70

Howeven, sirnple equations did not allow consider"ation of the subtl-eties of the situation, for example, plant competition contnoll-- ing a-bundance independent of stocking. Thenefo::e, multiple regnessions wel?e attempted, to isolate dinect stock effects. The pnoblem of coll-inea:rity of X variables then a.nose. While thene is no absolute ::equi:rement in the nathematics fon complete independence of these variabl-es (ltead, 1971), covaniance between them will lead to nedundancy and inte:rpnetation problems (Yanranton, 1971). The rnost efficient solution is to replace the va::íables with a small-en set showing less covaniation--a complex.model-ting situation beyond the scope of this thesis-- on al-tennatively, to include only one var"iabl_e fi:om the covarying set in the equation (Austin, l-971).

The genenal equations used fo:r majon species in the pnesent study hlere of the form

(e, B)+ b^ I,t B=b o +btD+b2 ï\t -

D=bo+ba (e B) + b W t - 2 whene B and D ane the biomass and density of the species in question; oniginally B.t - B is the biomass of the other species, considened as an index of competition; and W is the stocking index' distance

fnom wate:r. The probabl-e col-l-inea:rity between X va::iabl-es hlas considened unavoidable to a great extent' fon the foll-owing neasons. 1. While both density and biomass of a species are tikeJ-y to var.y with W, the var:iation is not necessanily of the

same onder" fon both- panametens. Hence, when dealing with biomass rthat pant of vaniance due to the nelationsh-ip with density, independent of th-e stocking effects, should be accounted fon; even though this may intnoduce IJ-.

col-llnearity of the X variables. 2. Similanly, th-e abundance of a. spe-cies may be affected by the presence of other species, independentl-y of stocking effects. Such a component in its vaniance

shoul-d also be accounted for. To reduce collineanity difficulties, the compound vaniable of average size was used. This was defined fon each- species as the ratio of biomass to density, B/D, in each quadirat. Ttre equation used was

B/D=bO+bt (s B) + b W t - 2

3-53 PrincipaL C onrponents artaLy sis .

pnineipal Components analysis (pCe) is a sound mettrod fon sumrn- arizing mu1tivaniate data, by means of muftiple axes ordinations. Recent neviews of the method ar"e avai]able in Orloci (f S00 ) , Piel-ou (fSOS ) and Kershaw (1973 ). fts no:rmal use in ecology is the detection of underlying environmentaf infl-uences (as in Gr-ieg-Smith et al, 1967; Austin, .l-968; Austin and Gnieg-Smith, J-968; Kershaw 1968; and Noy-Meir", 1971). Thene a::e two senious l-imitations: 1. The p::incipal components on axes aire mathematical entities which need not have biological- significance; and 2. Being linear functions of the o::iginal data, the

components are unJ-ikely to show simple nelationships with non-l_inean environmental- inf-l-uences. Howeven, the particula:r influence sought, stocking pnessure, may be

appnoximated by a linean model; thus this second l-imit- ation need not obtain for ttre present case)' 72.

Analysis was performed on a CDC 6400 computen in the University of Adelaid-e, using the :tFactor'r Analysis program'

BMDO3M, of the Heal-th Sciences Computing Facility, U'C'L'A'

This pnog:ram penformed a pnincipal components solution of a coruelation matnix, followed by onthogonal notation of axes. The inÍtial solution is not unique: rotation of axes does not

change the info:rmation content (i.e. the configunation of data points), but provides a new sol-ution of diffenent p:roper-ties, alÌowing easien inter:p:retation (Kershawr 1973)' Examples of the effective use of notation in phytosociological- studies ane given (1971)' by Ivimey-Cook and Pnocton (1g07), and. Noy-Meir: Detailsofpnocedurewe]îeasfo]-]-ows.PCAwasper.fonmed with on density data fon individual times and for all species ' distni-butions assumed approximately nonmal' The fi::st five

components, accounting for: the bulk of the totat vaniance' welle notated. Site affinities Con loadings ) for: each rotated

component were calculated next, and negressed on distance fr-om water". This detenmined which component(s ) nelated to the stocking influence. species affinities with the stocking compon' ent(s) wene comÞared subsequently between paddocks, to seek diffe::ences in the species affected under cattle and sheep.

The use of site affinities in regr:ession analyses was valid: the or:iginal data b.l-ock was muftivaniate-normal in distnibution,

the eomponents wene linear functions of this datao and site of affinities were derived fnom the o::iginat data by application ttre components. Hence site affinities wene also nonmally distnibuted, and so, acceptable fo:: use as values of the dependent vaniable in negression. A simil-an justification of sueh 73. method is given by Austin CfgOg, l-971-), and an example of its use by Noy-Mein (rgz+).

3-6 MINOR STUDY (Sampling by stereo-photopoints )

3-6I GeneraL approach.

The photopoi-nt system, dealing with the same tr-ypotheses as the majon study, was set up Ca) to pr"ovide al insurance in the event of difficulties with the major study (particular:ly in measurement of biomass); (b) to give an indication of pÌant dynamics in a:reas beyond the sampling limits of the rnajon study; and (c) to pnovide some information on taII shrub or: tree dynamics not cover:ed by the majon study. rt was designed to be nun as swiftly and easily as possib-le, allowing samples to be taken whil-e passing through the distnict if necessary. Hence the photopoints wene sited near tracks to al-Iow the use of any vehicle in sampling'

3-62 Photographic methods.

steneophotognaph pairs were taken at each photopoint using a standa:rd 35nrn camena, fnom a height of 1.5m, with a baseline of 20cm'

Photognaphs wene oblique, of the middle-distance type, with most information in the area between 5 and 15m fuom the camera. The camena field was centred on a set post ca 2Om away' A ventical scale was placed at any point within the field along a 10m nadius from the camena position Csee fig. 3.16). Strict contnol- was kept onthe position in the fietd of the centering post and the horizon, 74. to prevent wanfering of the fiel-d between tr:ips as much as possible'

In o'Lher" vegetation studies, ventical plro'[ographs have been used nather than obliques. For example Winbush et al (fS6Z ) used -to t:racings fnom steneopairs taken ve::tically estimate changes in coven.

Thre oblique type was chosen fon the following neasons:- 1. It is easien to ínterpr"et an oblique than a vertical- photognaph,inthesensethattheobliquegivesanimmed- iately identifiable Pictu::e; 2. A vertical photograph timits the size of the photosample

much mone than an obJ-ique taken fnom the same height;

3. The epheme::al populations in the study anea were langely upr.ightinhabít,hencechangesj-nbiomasswenemone likely to be conneLated with height than cove::1 4. As with most a:rj-d negions, much of the gr"ound was bare, and this would decr:ease the information on plant biomass

if ventical Photos were usedl 5. The one set of photognaphs could be used for both pasture and tnee dynamics if necessatry; and 6. Simila:r photognaphs, used in othen studies, have pnovided tnactable data (e.g. Hal-1 et aI, 1964; Albentson et al,

1e67 ). As fa:: as !,/as possible, photogTaphs at a point wene taken each time unden similar light conditions to minimise variations in intenpnetation resulting firom shadowing. Dur:ing the study, it was found that overcast conditions produeed a more easily :read stereo pair: than st-rong light, alth-ough- individual pnints hlere .Iess clear 75. through lack o-Î contrast. Btack and white film was used throughout r"athen than coloun, because of the cost of the latter. Howeven, it is my opinion that, had cost penmitted cofour, the information content of photographs would have increased gr"eat1y. Species (panticular"ly grasses) can be separ:ated on the basis of th-ein colour coding, and the state of pnoduction c¿m be partially estimated fi:om tlre col-oun of the plants.

3-63 PLøtt popuLatíons studied artd pa.rØneters estimated.

The type of photognaph combined with the use of black and white film limited both the species identifiabl-e and the parameters which coufd be estimated. Only the majon species and tall shrubs or trees wer?e consider-ed fon analYsis.

G:rass species wene lumped. as one unit (gnass standing biomass).

K. astrotz"Lcha anò. K. pAran¿datd were readily recognisable, and we::e each sconed for standing l-eaf biomass.

No attempt was made to calibrate the r"el-ative biomass on the photognaphs with absolute biomass (photognaplr- followed by full hanvest). This would have been necessany had the majon study not yielded tractable data, but the hanvesting requined was considered too time demanding unless absolutely necessary. The decision whether or not to attempt dinect cal-ibration was delayed until Februany 1973, by which time it was clear that the methods in the major study wene wor:kable. 76.

3-64 SønpLing systen.

Photopoints wer:e sited at negulan intenvals along tnacks, though at feast 50m away to avoid inmediate tnack effects.

ReguJ-a:r nattrer. than random sarnpling l^ras used to pngvide an even sampling with the limited numbe:: of r"eplicates. Sampling along tnacks was necessar"ily biased. Tracks are taid to p:rovide access to wate::s and fences, and usual-ly avoid hills. As a nesult, the sampling in the cattle paddock was nrainly nestricted to the K. astrotz'icha areas, with littl-e satnpling of the othen vegetation, types, and sampÌing in the sheep paddock was langely :restnicted to the southenn fence.

These limitations must be considered in companing paddocks. since the sheep paddock had fewen usable tracks, photopoints wer.e placed at shonter intervals (O.S mi, 0.5km) than in the cattfe paddock (O.S mir 0.8 km) to give an equivalent neplication between paddocks. 23 and 20 photopoints wene set up in the cattle and sheep paddocks r:espectively. In addition, a photopoint 77.

TABLE 3_5 PHOTOPOI¡IT OBSERVAT]ONS CATTLE SHEEP MAJOR STUDY TRIPS (cattte) 7 /7 /72 Pil-ot 4/s /72 4/s /72 1 (SheeP ) r/ro/72 s/ro/72 I (Cattle) 2T/rr/72 z (soth)

13/2/73 3 " 28/4/73 + t'

24/7 /73 s " ULL/73 6 " 13/2/74 7 "

3I/7 /74 B 't

was placed in each encl-osr::re. Tabl-e 3.5 gives obsenvation dates fon the photopoint system, and figur"e 3.17 the sampJ-e distri-bution'

3-65 AnaLysís method.

Analysis of photopoint data was kept elementary and semi- quantitative, once the majon study methods were found effective'

Fon each obser"vation, every photog::aph was enlanged to a standand size. In combination with the control- on camel?a field, this pennitted the overLay, unde:r a stereo viewer, of ptr-otographs taken at diffenent times, but at the same point. The inage nesulting but showed tnees, fences, hills and the nest in thnee dimensions, pastune plants as ovenlaid tl^Io-dimensíonal- images. TtLis allowed a napid estimate of the changes in the pastr:::e between times. Fig. 3.17 PHOTOPOINT LOCATION

CATTLE PADDOCK

Btock Hitt Bore o o o o ,-- ¿¡¡ o U) ØQo .¡'t'\'l'l'\'l'l'/'¿ zt, o ¡' ¡'7¡ 4¡,¡, ¡1,11

o o o WESIERN o SEI o o o o o EASIERN o SET Btock Hitt (token os dotum)

SHEEP PADDOCK o Photopoint

¡- ,v lU .4

Trough lmi rough o o 1km o o o o o oooo WESTERN SËT EASTERN SET

(token os dotum) 78.

Using this technique, the diffe::ences between consecutive times Í¡er1e scol^ed, for each point, as obvious increase/decnease, sJ-ight inc:rease,/decr:ease' or no apparent change. Appendix 1 pnovides worked examples of the full- time senies, for three photopoints from each Paddock. Following the estimation of change, the data scores were inspected for any piosphene pattenn. The pattenns found wene subjectively compared between paddocks.

3_7 RAINFALL MON]TORING.

Rainfal-l was reconded at Mt Victon homestead. Gauges wene not placed in the study areas, due to th-e length of time between obser'- vations. Even with a kenosene on oif film, results for the intended per:iods would be inaccu::ate due to evaponation, and the erron would not be consistent with season. Fu::th-en, such results could give onty the total- rain since th-e pnevious obse::vation'

Fon par:t of the period, :rainfal-I :reconds were obtained fnorn the

adjoining Koonamone and Curnamona stations. The rationafe was that close conrelation in the rains for these stations, relatively widely sepanated, would irnply close cor"relation in th-e nains fon the two paddocks' with thein closer: pr:oximity'

3-8 STOCK MONITORING.

3-BI Sl;ocking LeueLs.

Fortunately, ttr-e station maintained adequate stock l?ecords.

These wene the basis for. follovring stock fl-uctuatiors and levefs. 79.

Stock were mus;er"ed and counted at least twice a yean, with the count reconded on the same day. Difficulties anose after Februarry' l-g73. At this time, cattte bnoke the fence between Black Hill' and the paddock no:rth, old Station, cneating a semi-open lrange. Thene was, howevell, no evidence that cattle then cong::egated in one on othen paddock aften this, but nathen tlrene was a constant inten-

change between paddocks. Fnom this time, entnies in stock ledgers

combined the two paddocks, and- ane given likewise in chapter 6.

3-82 Mouements.

No objective observations of day-to-day movements wer:e made'

The curnulative penetration of stock in both paddocks was subjectively estirnated by obsenvation of pad patte:rn and relative dung dnop'

+. CHAPTER 4

RESULTS CO}.4PAR]SON OF SHEEP AND CATTLE PIOSPHERES OVER A SHORT TFRM:

4_1 RAINFALL OVER THE STUDY PERIOD.

Two-, thnee- and four-month running means for nainfal-l in

1972 anð.1973, at Mt victor, Koonamore and cur:namona stations are plotted in figure 4.1 . The positions of these stations rel-at-

ive to the sampJ-e areas ane shown in figune 4.2. There is cfose connelation in the ::unning means at each, station, both in the patter:n and the rnagnitude of nainfall. Since these stations a-ne r"elatively widely spaced, yet show such connel-ation, nainfall inputs fon the two sampled paddocks, with thein closen spacing, can be considered.equivalent. Funthen, ::ainfall- at Mt Victo:r station can be taken as an accunate indicaton of nainfall in the

sample al?eas. Monthly nainfall at Mt Victor station, fon .ranuany 1972 to July l-974 inclusive, is given in figu-re 4.3, with the two-, three- and four,-month :running means plotted. The majon featu:res were: å 1. Heawy fal-Is at the start of L972, wíth l-ittle for- the

nemainder of that Yean; 2. Heavy falls in the fir"st quar"ter of 1973, fol.ì-owed by two

.months of lighter rains; 3. Consistently heavy fal-l-s fnom rnid-l-973 to the end of observations at the beginning of Àugust l-974'

Th:roughout the s.A. ::angel-ands, the fal-l-s fnom mid-1973 wene the heaviest on necord (Br:r'eau of census and Statisties, 1974). Fon j¡stance, mean annual- nainfal-l at Mt Victor' f958-l-970, -80- Fig. 4 .1 2-,3-, AND /.-MONTH RAINFALL RUNNNNG MËANS FOR STAT¡ONS T"¡EAR STUDY ARËA

350 2 month running meon

250

200

I r00 I t I Mt victor ---- Koonomore Curnomono

3 month running meon E E 200 J J l¡.

300 4 month running meqn

t50

Mt victor 50 ---- Koonomore Curncmono

N D FMA MJJ A S ON D J FM AM J J ASO 1972 1973 TIME Fig' l' '2 LOCAT¡ON OF SAMPLE AREAS IN RELATION TO RAINFALL MONITORING STATIONS

, I To Frome Downs Stqtion I t boundories o CURNAMONA H.S , t .-'Mail roods , , o , HSHomesteods , t :'.'.* t Somple oreos , I , I I EMU LEG 5mi PADDOCK 5km ,

o Kìtloworro o.5.

Gtenorchy o o.s { BLACK HILL PADDOCK

t , t ,ltoouauonE o 'ì-YT VICTOR I H.S I H.S. I I t / I t / I 2 I \ ^å*'J"Vegetotiori Reserve ,r I I \ /To Yunto Fig. /..3

RAINFALL AT MT VICTOR STATION

DURING THE PERIOD OF OBSERVATIONS

MONTHLY RAINFALL 500

t: .tI 400

1972 1973 19?/, Ë -9 I

J) 300 !_z .ts - z-morìth rLnn¡ng meoñ E.

- 3 - montrì runnrnc ñeon 200 ¡ --

4 - month runnlnq meÕn

't00 I t

Jon Apr Jul Jcn Apr Jut Oct Jon APr Jul Oct 19?t, 197?, 1973 81. was 178mm, \^rith a range f::om 51mm (fgOZ) to 34Omm (fSSg). Total- nainfall for 1973 was 4O7mm, and fo:: the fi:rst seven months of

1974 was 64Omm. These atypical rains severely limited the findings of the shont-term observation.

4_2 STOCK MONTTORÏNG

4-2I Stocking rates.

Table 4.1 gives the stocking levels in the sampled paddocks, fon both the study period and th-e pneceding rnonths. Fo:: cattle, Ol-d Station figr.u:es fi:om the sanpled paddock and its neighbour ' paddock are colnbined: as al-:ready stated, these two paddocks for"med-

a semi-open cattle nange fnom FeÞrualay' 1973' senved by five watens' The main featu:res were: . 1. The steady incnease in cattfe nrunben th:noughout the study peniod;

2. The fluctuations in sheep numben, culminating in a

25% mortality between Februar"y and May J-974, due to fly- stnike, æd folfowed by the nemoval of sheep theneafter; 3. The natio of sheep to cattle, of the orden 5:1, and markedly higtr.en than the ratios given in section 2-22.

The f.l-uctuations and absolute diffelrences in stock numbers impty that the paddocks wer:e not equivalently stocked, in terms of numben pen unit area or number Per waten. 82.

TA3LE 4.1 STOCK NUI'ßERS

(a) Cattle: B1ack Hil-l and Ol-d Station Paddockst. 10600ha total al?ea, 5 effective watens.

Peniod Total- stock ha/anima]. animals,/waten

Dec 1971- 300 35 60 .Iuly 1972 July 1972- 350 30 70 Sep 1972 Sep 1972- 390 27 78 Nov l-972 Nov 1972- 470 23 94 Sep 1973 Sep 1973- 500 2I 100 Aug 1974

Sheep pnior" 2400 4.5 480 to Dec 1971

(b) Sheep: Emu Leg Paddock*. 6000ha total ar1ea, 2 effective watens.

Dec 1971- 600 l_0 300 Jun 1972 Jun 1972- 1000 6 500 Aug J-972 Aug 1972-' 600 10 300 Man 1973 Man 1973- 400 15 200 Aug 1973 Aug 1973- 830 7 415 Feb l-974 Feb 1974- ca 200 deaths fi:om fly-sbrike. May 1974 May 1974- stock nemoved. Aug 1974

'* Figu:res given a::e total stock: included in cattle figures are cows and calves, in sheep figu:res are ehle hoggets and wether"s. 83.

4-22 Stock mouements,

. Figunes 4.4 and 4.5 show th-e extent of stock penetration in the paddocks sampled, over the study period. They do not show the seasonal vaniation in stock movement, and should only be taken as a general indication of the areas usual-Iy grazed. The main pattenns we¡e:

l-. The relative concentration of sheep movement in the

southenn Pant of the Paddock; 2. The absence of cattle fnom the hill areas; 3. The number of ver:y long, heavily incised cattl-e pads, suggesting a significant pr"opo::tion of rrwalkersrt (sctrmi¿t, 1969; see section l--64 ) in the cattle herd;

4. The use of ::oads bY cattle.

The cattle nanged over a much J-anger plloPolrtion of the paddock than did the sheep, though not fu:rthen from watelr. Hence absofute statements on the stocking levels of the two paddocks are not possible. Although it was suggested above that the ovenafl- stock- ing nates wene not equivalent, with the sheep stocked mone lightJ-y than the cattle, the concent:ration of the sheep into a srnall pant of the paddock may have nullified this diffenence'

4_3 ABUNDANCES AND DYNAMICS OF MAJOR PASTURE SPECIES

44I fn the absence of g?azing.

Fon both paddocks, figures 4.6, 4.7, and 4.8 dispJ-ay the total- biomass of the majon pastue species, the individual speciesr bio- Fig' I*'Ir CATTLE UTIL¡SATION OF BLACK HILL PADDOCK

cLJ s'rÀircÌr PAf,OOCX Tq Cld Stotroô lo Cld S!olrc. Bore r I Bcre Crd 9lc:,on:cr t wcterÞoint

Deepty incised (to 10cm) Pods or grouPs ol Pods

Not utilised no hoof morks or ,\ dung

Occosionolly ut¡lised hoof morks' some dung

Constonlty util¡sed hoof morks, 1mi # substdlorY trccks, lkm dung droPs cq 1/i0m

Fig ' t* '5 SHEËP UTILIZATION OF EMU LEG PADDOCK

t, Woterpotnt

Not utitised: no trocks or dung rntoct lrchen crust

0ccosionolly utilised foint trocks, some breokttP ol llchen crust, little or no dung

Å Consìonlly utltised cleor, rncrsed t¡ocks. removol of much of lichen crust, dung droPs co 1-3/m2 over 90 doYs

lleovily utrlised cleor, lnctsed trocks, no ltchen crr.rst. dung droPs co l0-25/m' over 90 doys lkm 84. masses, and the individual- speciesr densities in excfosed quadrats'

TotAL biomass showed close cornelation between pacldocks for the first hal_f of the study. From time 5, howeven, a gneaten standing cnop was evident for the sheep paddock exclosur^es :relative to that of tl¡-e cattle paddock exclosunes. Fluctuations in total bíomass, for. both excfosulre sets' connelated most closely with the three-month nainfal.l- means. Although amplitudes diffened. between exclosul?e sets, the directions of responses to rain- falI were the same in bottl- cases. Periods of majon pastu:re gnowth were found between times 3 and 4, and between tfunes 5 and 7'

Erwteapogon spp, compr:ised half the standing cnop in both exclosune sets. Following the finst significant :rains, both its density and Leaf biomass wene gl?eater in sheep paddock exclosu:res than in cattfe paddock excl-osules, but the di:rection of lresponses to nainfall, fon both par"ametelrs, vlere the same for both paddocks .

A?istídø. contoYta anð. Stipa nitídn compnised only a smafl propontion of the total standing clrop oven most of the study per:iod.

The for"men displayed the same pattern as given for Enneapogon spp"

The latter showed very c.Iose connel-ation between paddocks throughout the study: fuom an initially high s'tanding biomass and density , Stipa abundance decneased to nean zero) with new growth appeaning only after the winte:r and spr:ing nains between times

5 and 7. Koehia astl,ot?icha biomass in sheep paddock exclosures was elosely connefated with that in cattte paddock excl-osures, both fo:: standing crops and seasonal changes. The density of adults relîained constant thnoughout. Seedling density fluctuated with Fig. 4.6 MEAN BIOMASS (TOTAL) FOR EXCLOSURE QUADRATS WITH ASSOCIATED STANDARD ERRORS

Cqttte poddock

I eoo 0l J t- I- t¡l

I = n00 oÉ, Sheep poddock tn (n oE õ I i æ0

a,400 3 month rqinfott running meon E E

J Éroz É,

-rul Oct Jqn Apr Jut Oct Jon Apr Jul t972 1973 1971 Fig. 4.7 MEAN BIOMASSES OF MAJOR SPECIES FOR EXCLOSURE QUADRATS WITH ASSOCIATED STANDARD ERRORS

cottt. poddock ræ I Stipo nitido ,i sn."p poddock r00 I I

80 I

li t. It l I I' l þ 1i I'

o E -EN I Enneopogon spp. I 800 l¡J = É o 600

r,t1 ú,

o 400 õ

200 I I

I I ii

I {

Aristidq contorto I i

æ l: I I

Jon AP. Ju Jul Jon Apr Ju Oct 1972 1973 19?4 (Cont Fig. 1.7 (Cont.)

I I I I I I roo I I I i d I I E I Ct¡ I J I I I f t' (,E l¡l Kochio qstrotricho = ôÉ.

(/) vl Kochio pyromidoto Idt (No{e: only two quodrols in eoch exclosure set conloined X. pyromidoto )

E 3 month roinfoll running meon E 400 zÉ a E

Jul Oct APr Jul Oct APr Jut r972 1973 1974 Fig. 4.8 DENS¡TIES OF MAJOR SPHCIES FTR EXCLOSIJRE QUADRATS wlTH ASSOCIATED SIAND/iRD S RROfìS

50 Stipo nitidq

I 25 I 't

I I I I It I I

400- Enneopogon spp Ë È 300 o = -$zooo c o o_

ã,0 Aristidq contorto ãzooro I ¿

KOChiO OstrotriChq (¡uvenrres)

I i

?¿oo E 3 month roinfotl running mectn € J J È-z zoo É.

Jut Ocl Jon Apr Jut Oct -lorì Apt J ,I 1972 1973 1911.

K ostrotrlcho (cCults) onC F. pyf ,lrrlCrltù derì',r'r e; ''::' " Cctll¡: ËiCCc'ck renictl-lod cor:ltcnt, wlitr rtLpecr.rr,, :¡" ::! ' antl C ç: plnnts/quo,lrot rn thP- ¡'¡11[r-' ;rirlCoL/.::: 4 '1 ': rrn.J C 2 pìrrnls/qutl¡1r,:l rn !h,' 'il'.:.':) [:r!1r]:irr'{ -t . ShaeO p,_:..jcjr_'cx r'rrth regí,ecllvr: stonCotcf error! o1 '^.1 ! ti ! .r' arì(l I Ù :l 85. no discer-ni-b.l-e pattern, but counts were consistent]-y higher: in the cattle paddock exclosures.

Kochia pyz,ønid.atø r^ras inadequately sampled on. both excfosune sets. Fþom the bushes sampled, its biomass appeared relatively constant thr:oughout the study per"iod.

Thus the di:rection of nesponses to nainfa.l-l of the najon species wer.e the same for" both paddocks in the absence of gnazing. This was considered to indicate that backg:round dynamics wel?e companable between paddocks. Howeven, clear diffenences welre evident between paddocks; in ttr-e magnitude of these nesponses, and in the nelative standing cnops. Atthough these diffenences may have nesulted in pant fnom th-e limited sampling, they wei:e taken as nepnesentative. The implications were: 1. Stocking effects woul-d be l-ess evident in the sheep paddock than in the cattl-e paddock: th-e Ïr-igtr-er standing

cnops in the for:mer v¡ould mean a relativel-y smallen Pnop- ontion of the total pastu:re removed by stock than in the cattle paddock, and threrefone a masking of the sheep effects by backgnound vaniation; 2. Direct comparisons between paddocks (e.g. analyses of - vaniances) to dete::mine stock impact would misl-ead, since the absolute amounts of the various species diffe::ed in the absence of gnazing. The search for piosphere pattern thus nemained. the only feasible rnethod of detecting stock effects with certainty, and the gene:ral companison of sheep and cattle piosptr-e::es th-e only sound basis for companison between padd.ocks 86. t+-32 G?azed cornpared uit?t ungeazed. . For both paddocks, figr:nes 4.9' 4.lO and 4.11 r"espectively display the total- biomass of majon pastu:re species, the individual- species t biomasses and th-e individual species I densities in open quadnats. FT"om the p::evious section' no comment on relative stocking effects can be made fnom direct companison between paddocks. Howeven, within paddocks, diffenences wei:e obse:rved between grazed and exclosed quadnats potentialJ-y due to stocking.

Tn both sheep and cattl-e paddoeks, Enneapogon density and biomass did not incnease significantly in open quad::ats until- time 4. In exc.l-osu:res, biomass had noticeably incneased by time 3, although density had not. This inplied that the incnease at time 3 was due to negnowth fnom ol-d butts. By comparison with the sheep paddoek, va-niations in Enneapogon density' on open quadnats, was not as grreat in the cattfe paddock, and, in pa:rtic- u1an, it was not neduced to the same extent in the pe::iod of no growth (fnom time O to time 3). This implied that the cattle wene leaving mone butts tha¡ the sheep. Thenefone, if the finst nesponse of Enneaptogon to irain is a neg:nowth fnom the nemaining butts, cattle stocki.ng will al-low molle immediate r:eg::owttr- than sheep stocking, because of the gr:eaten pnoportion of butts nemaining. A less obvious difference between nesponses of open and exclosed quadnats in the sh-eep paddock was shown by K. astt'ott'icLta biomass. In exclosunes, biomass I^Ias not significantly different between times 2 and 3, but in open quadnats, it showed a significant decrease between these tirnes. In the cattle paddock, however, Flg. 4.9 MEAN BICIMASS fiOTAL) FOR OPEN QUADRATS WITH ASSOCIATED STANDARD ERRORS

poddock ¡ Cottte I

200 (t -c. -tol

1600

Sheep poddock OEo 1000 (n (n 800 o= õ

400

æ0

Ê, E 3 month rqinfott running meon

Éz É.

Apr Jut Jul 0ct Jon APr Jul Oct Jon t972 r973 1974 Fig. 4.10 MEAN BIOMASSES OF MAJOR SPECIES FOR OPEN OUADRATS IVITH ASSOCIATED STANDARD ERRORS

I cotte poddock

! I Strep Poddock

I Stþo nitido I I I I I f)0 I

I I t i I 9, I I E

1 -t- mon 1100 (D sE rt20 l¡J = Enneopogon sPP. oÈ

(n Ø 400 = I6 il I t? l' t !i t

Aristido contortc

I I { t ii { ¡

APr Jul Jut Oct Jon Apr Jul Jon 872 1973 1974 (Cont Flg . 4.10 (Cont.)

Kochio qstrotrichq

I I I

I I t I t00 d E a ol J

I I öuo I l¡J t ii I = oÉ. Ø !r) Kochio pyromidoto o

õ I

I + + I t I

E 3 month rqinfott running meqn E lzoo É.

Jut Jul Oct Jon APr Jul APr 19't2 1973 19',11 Fig' l* '11 DENSITIHS OF MAJOR SPECIES FCR OPEN QUADRATS WITH ASSOCIATED STANDARD ERRORS

50 Stipo nitido t 1 { 25 a I I a

a I Enneopogon spp. a I d + IJ 200 d ? ) I q i¿ >vt 100 c l o t + o- Aristidq contorto ñz

H2s I ti

Kochio ostrotrichq (juvenites)

I t I

E 1oo E 3 month rqinfoll running meon

þzoo É.

APr Jul Oct Jon Apr Jul ct Jqn Jut 'r974 197? 1973 Cottle poddock K.ostrotricho (odults) ond K. pyromrdoto densitres remoined effectrvely constont, wilh respeCtive meons in cottle poddock; Sheep poddock of 9.1 -1.2SE ond 2 5-0.5SE tlre ond 8.9-1158 ond 0.5-0 /,SE in the sheep poddock 87.

both open and exclosed quadrats strowed no significant diffe::ence in K. astz,otz,icha biomass between these times. This suggests that the sheep, but not the cattle , r^relle consuming or otherwise nemoving the K. astrotr'ícha dtrr"ing this period.

4-33 The effeet of the L973-L974 rains.

The finst major pastune g:rowttr- during the obsorvation

nesulted fnom fate summen nains which, while heavy' $Iere by no means

unusual (c.f. summer ::ains in pnevious yeans--tabIe 3'1)'

Because the a:rid nangeJ-and soil-s a::e typically low in plant nutnients, a¡d in panticulan nitnogen and phosphorous (Chanley and Cowling, 1968), such a g:rowth br:r'st woufd be expected to lock up a significant pnoportion of the avail-able nutnients, and so l-imit the nagnit- ude of future incr eases; at feast until considenable pasture break-

down had occurlred (see Beadle and Tchan' 1955). Nonethe'Iess, the

pastr:r:e growth following the exceptional nains fnom rnid-1973 was

Ê veny much gneater than the eanly 1973 growth: co:rnelation between pasture biomass arrd rainfal-I was of the same order in the laten periods as in the finst peniod, and not less as woufd be expected if the initial g:rowth bur"st had removed a significant pnopontion of the available nutnients. This impties that the nutrients

may not be as limiting as is normally assumed'

4-34 Cort,eLqtions uíthin artd bett¡een specíes.

Figu:res 4.12 and 4.13 display ttr-e significant cornelations between species and between biomass and density parametens, fon the 88. cattle and sheep paddocks respectively. In both areas, biomass (b) and density (d) were consistently highly co:rnelated within species, as should be expected. Howevell, correlations between species vanied g::eatly with time. It woul-d be u¡wise to infen a great deal fnom these connelations. Analyses were pe::formed pninanily to show the possible cova:riation between species, under- lining the difficufties of coping with cova:riation in negression equations. It can be seen that, wel?e equations fon each time rnodelled to all-ow fon af 1 the demonstnated cova::iation, each equation would be st::uctr:red diffo:ent1y, and so the companability of nesults fuom equations at diffenent times would necessa¡iIy be

::educed. In this study, comparability I^Ias Panamount. Nonetheless, the coruelations between stipa anò. Enneapsogon in the sheep paddock nequi:re some comment, because of the change of sign with time in thei:: nel-ationships. Other cor::el-ations necur.ning with tirne mai¡rtain the same sign. T]ne Stipa/Enneøpogon nel-ationships indicated that two distinct pnocesses wene in openation. Possibly the positive connelation which developed in the no-g::owth peniod was a gnazing effect. Aften a long per:iod without gnowth, grasses would be g:razed on tnampled out over: much of the range. The sunvival of individual-s of a par:ticula:: species in pa:rts of the range would indicate that they had escaped gnazing to some extent. Hence othen species might be expected in the same area, fon the same lleason. A positive con::elation so fonmed shoul-d disappea-n following gnowth, since the species would neappear elsewhe::e on the nange, not only on alleas which had Fig. 4.12 INTERCORRELATIONS BETWEEN PASTURE SPEC IE S (BIOMASS ond DENSITY) l¡l THE CATTLE PADDOCK

Bsn e54 a5o rnfD \.D \g N yt B B¿ t --.\ - --\ - ìì¡Jþ Ea D \. Ko Ër ,å; \ B Ac ñ D

2 TIME O TIME 1 TIME

gsa g$a a5¡ \P D N) - ,-D --, ':t vt ,B { \\ E_9 - l1 - -.\ \ - ì. le, D D '.{D Ko \ B B \ B Ac # # TIME 3 TIME 1 TME 5

6Ð sÐ r:1D \p \P -- B B \ --- I .. l.o Eo D \ --{D D Ka Ko \ \ B B B B Ac ñ D ñ

TIME 6 TIME 7 TIME 8

Positive Negotive .01)p 001)p

Sn = Sttpo ntttdo Eo = Enneopogon spp, e[c - B: BÍomoss D = Densíty Fig. 4.13

INTERCORRELAT¡ONS BETWEEN PASTURE SPECIE S (BIOMASS crnd DENSITY) lN THE SHEEP PADDOCK

85, D

':/ B ':/ B Eo Eo ,l; D i; D

TIME 1 TIME 2

5n B Sn D \g Kp yl yl B B B EJ Ee D D l,* ;; ^15 B ^i"; B Ac D D4s *1 TIME 3 TIME 1 TIME 5

sÐ Sn D D Kp vl \ B vl Ee 'l D D å; Ke Ke B ¿/ B B B DAC DA9 D Ac

TIME 6 TIME 7 TIME 8

Positive Nego trve 0llp 001 )p Fi1. 1.14 ¡NTERCORRELATIONS BETWEEN PASTURE SPECIES (AVERAGE SIZE) IN THE CATTLE PADDOCK

Sn Sn Sn

Kp Kp Kp Eo

Ko /'" Ac Ko Ac

TIME O TIME 1 TIME 2

Sn 5n 5n

Kp Kp Eo Kp Eo

Ko [" Ko Ac Ko Ac

TIME 3 TIME 1 TIME 5

Sn Sn

Kp Eo Kp Kp Eo

Ko Ac Ko Ko Ac

TIME 6 TIME 7 TIME 8 æ Positive Negqtive 01) p .001) p Fig. 4.15

INTERCORREL ATIONS BETWEEN PASTURE SPECIES (AVERAGE SIZE) IN THE SHEEP PADDOCK

Kp Eo Kp

Ko Ac

2 TIME 1 TIME

Sn 5n

Eo lþ Eo Kp Eo

Ko Ac Ko Ko

5 TIME 3 T]ME 4 TIME

5n 5n

a Kp Eo Kp Eo

Ko Ac Ko Ac Ko

T]ME 6 TIME 7 TIME 8

.001)P Positive Negotive O1)P -

à 89. escaped graz:fLg. The positive cor¡.e.lation between stipa and

Enneapogon did in fact disappgar following growth. The negative con:relation between the two species, deveJ-oping f:rom time 6 onwar"d, could be ascniòed to competition between them: Av'istida ger.minated fo]lowing surffnen r"ains (rigr::re 4.11) and was alneady abundant when the Stípa germinated following winter and spning nains. Hence Stipa wou.l-d only establ-ish in aneas whene Arístídn' was low in ablndance, and so give nise to the negative association' Figr:r,es 4.14 and 4.15 display the significant cor':relations in avenage size (biomass/density , B/D) between species, fon the cattle and sheep paddocks r:espectivety. The most stniking feature was tire l-ack of congl?uence with the cornelations obsenved between species for biomass and density. For instance, there was positive co:rnelation between stþa and Enneapogon average size, but negative conrelation between the speciesf biomasses at time 8. Thus the compound va::iable detected relationships

not shown by its constituent variables. Its sepanate use in negression equations can therefo::e be justified on this basis, in addition to the or"iginal justification of neducing collineanity in the equations.

4-4 REGRESSION AÌtrALYSES: MAJOR SPECIES

4-4I fnfu,oduetion,

only linear negnessions of plant parameters on distance fnom

water wene attempted for data fnom individual- species, following

tlr,e argument of section I-72. However, centain pJ-ots of total 90. biomass agai-nst distance from water suggested that cur"vilinean functions migtr-t pnovide a better fit than a linear negnession, and cunve-fitting was accolrdingl¡r attempted for" this data. As stated in the previous chapter, th-e sheep piospher"e was complicated by the pnesence of two waters in close p::oxirnity.

Ttrene þras stnong conr.elation C:: = 0.853) between distances of quad.rats firon the south- waten and distances firom the north waten' and one or other distance pananetel? had to be excl-uded fuom equations to avoid redundancy (see section 3-52). The distance to the nonth waten was th-e term excluded: the nonth waten was Subject to fnequent stoppages due to pipeJ-ine breaks ' and $Ias completely shut down in Octoben, 1973; wheneas the south wate:: remain- ed in service continuously up to the::emoval- of the sheep in 1974'

Anea-dependent units were standandized, to avoid undue confusion in intenpretation, of nesults. Biomass is given as kg/ha and density as individual-s/ha, as estimated fnom quadnats. The thr:eshold of detectíon fon biornass vanied with the size of hand- hel-d estimation units, but was of the order Skg/ha. The thneshold for: density, of one individual- pen quadr.at, was equivalent to

IOOO individual-s/ha fon species counted in smal-l- quadnats, and to 250 individ.uals/ha fon species counted in J-ar:ge quadrats. Results of negression analyses arae pnesented as fol-lows. The regression coeffecients and equation efficiencies are presented in tabl-es, seg?egated for individual species and panametens. The tabtes alre extensive, and resu.Its show intricate vaniation with tlre tirne of obsenvation. Hence, fon each species and each of its parameter^s, summaries of significant :relationships are given 91. in the text, with-out imnediate discussion, in or"den to highligh-t per"tinent aspects. For each species, a discussion on its nelationships follows ttre summa:ries, befone th-e nesults fon othen species are discussed. A general discussion on the regression analyses is presented in section 4-49: this discussion considens al.l species and all rel-ationships. In the tables, only equations displaying eithen a significant nel-ationships or: an R2 val-ue gneater than O.2OO a:re given. A significant nefationshíp is defined by the pl?esence of a par:tial neg:ression coefficient significantly diffenent fuom zeno at the .01 level, as deter:mined by a t-test.

4-+2 Linear Øtd non-Linear, Í'egvessions of totaL biomass on &Lstmee.

Sevenal plots of total- biomass against distance f:rom water:, using cattle pad.dock data, showed the pattenn displayed in figur:e 4.t6. To test whethen the assumption of lineanity in the piosphene was in fact justified in such cases, the foì-lowing equat-

ions wene fitted to cattfe paddock data fnom al-I obsenvation times.

B=t bo + bfll lnBa=bO+brW

Btol_ -b +b- w2

Bt= bo* b vÍ+b w2 1 2

The latten thnee equatíons were Pnoposed, not fuom a theoretical base, but simply as t::ansformations on cunves which potentialÌy p.'esented betten appnoximations to the data curve than the simpJ-e l-inea:r equation. The resufts fon ttre vanious equations are given ',

Fig. 4.16 TOTAL BIOMASS WITH DISTANCE CATTLE PADDOCK, TIME ZERO

r500 a

r000

TOTA L BIOMASS a (kg / ho) a

s00 a

0 0 s00 1000 1500 2000 2500 DISTANC E (m) 92. in table 4.2 Ttre fol-l-owing points wene evident. !. The squane and quadnatic equations did pnovide a bette:: fit than the simple linean equation, but the diffenences in efficiencies were sma1l. 2. All equations wer:e g::oss]y inefficient in accounting fon

the vaniance of Ba at times 4 16 and 7: these times coincided with or foll-owed pe::iods of pastu::e growth' 3. b, fon the fi:rst and third equations decreased, fi:om time 0 (pilot tr:ip) to time 3 incfusive.

Fnom the fir^st point, there was no reason to abandon the seanch for linear nelationships between plant abundances and the index of stocking pressure, distance from water:. Cfearly, consid-

er:abJ-y molre complicated equations would be nequir-ed to pnovide a fit significantly more efficient than that of the simple J-inean

equation. TLre inadequate theor"etical basis cannot justify the use of such equations.

Fnom the second point, peniods of najon gr:owth appa::ently

obscur"ed the stocking impr.ess, as indexed by W. T\nto possiSle neasons are offened to account fon this. Firstly, stock may not

have been watering negularly on the dams at this time, but on casual sunface waters elsewhere in the paddock; hence the shont-

ter-m piospher:e pattenns would become confused. Second]y, the

biomass nemoved by stock may have been negligible by comparison with the pasture g::owth, and the background var"iation would thus

obscune the stocking imPress. The third point suggested that stock nanged further as pasture 93

TABLE q.2

LINEAR AND NONLTI.IEAR REGRESSIONS

FoR ToTAL BIOMASS, CATTLT PADDoCK

Ihe equations: 1 Ba=bo+ lt1 lll

2 2 Ba=bO+ b1 VJ 2 3 Ba =bo* btW+b2l'¡ 4 lnBt=bo + btw 5 l=t bo* bt /w

2 (a) R values.

Time Equation 1 2 ? 4 5

0 0. 380 0. 440 0 .457 0. 353 0.148 t 0. 308 0.322 o.322 0.315 0. 167 2 0. 302 0. 309 0.310 0.266 0.170 3 0.233 0.240 0. 240 0.161 0. 144 4 0. 039 0.036 0. 040 0. 043 0. 017 5 0.212 0.223 o.223 0,202 0.100 6 0. 061 0.052 0.064 0.053 0.040 7 0.077 0.070 0.078 0.777 0. 063 I 0.354 0. 413 0.430 0. 289 0.135

(b) Regnession coefficients with associated standard errors, equations 1,2 and 3.

Time Equation 1. Equation 2. Equation 3. bo bt bo b1 bo br. bz SE SE SE SE

0 78.2 0.230 186. 8.95E-05 299 --0. 204 1. 61E-04 +0. 0357 +1.228-05 +0. 145 +5.238-05 1 80. I 0.7'19 77 3. 6.59E-05 161. o.213 5.84E-05 +0 .0 324 +1.168-05 +0, 140 +5.048-05 2 45. 3 0,7t7 103 4. 0 7E-0 5 85. 0 0. 0332 2.97E-05 +0.0206 +7. 388-06 +0.0887 +3.208-05 3 87.2 0.0850 131 3.12E-05 127. 0.0190 2.45E-05 +0.0187 +6.738-06 +0.0809 +2.928-05 l+ 293. 0.0523 29 3. 1. 80E-05 254. 0.0703 -6.68E-06 +0.0314 +1. 738- 0 5 +0.136 +4.918-05

5 68.8 0.135 138 4.95E-05 13r+. 0.0070 4.758-05 !0 .0315 +1.138-05 +0, 136 L4.908-05 6 23s. 0 .0601 277. 1.99E-05 206 0. 119 -2.778-05 t0 .0286 +1.048-05 L0. 124 +4.478-05 7 494. 0 . t52 580. 5 . 228-0 5 465. 0.208 -2.088-05 +0.0641 +2.318-05 +0.276 +9.978-05 B 152. 0 .295 290. 1.158-04 445. -0. 280 2. 138-04 +0 ,048s +1. 668-05 +0.197 +7.118-05 94 dec::eased. Were this not so, tlre intencept CbO) woutd be expected to decnease and ttre slope (br) to remain constant as the pastune was grazed dovrn within tlr-e nange of ttre animals: but the situation was the reverse of this.

The equations wer.e also fitted to data fuom the sheep paddock' using distance fuom south waten as W in the finst instance. No t significant r"elationships wene detected, and R- values wene below

.1OO in eveny case. Equations were then calcufated using distance fuom the north V¡ater as W, with the same nesul-t. These findings wene intenpneted as indicative of a weaker, and so undetectable, sheep impness by companison with the cattle imp::ess.

4-43 Regressíon ømLyses for Stipa nitída-

No :regnessions were calculated on stipa data fon time 5, both paddocks. Incidence'then was less f¡an 2o%, and the numben of zeno scones Írer.e 1ike]y to introduce complications if analyses \^Iere attempted (see section 3-52).

(a) Density nefationships; tabl-e 4-3

D r b2" Sn =bonbl(Bt-B srr) '* ts) In the sheep paddock, significant pa::tial :regnessions (s,p.R. of D on (na -n) were found at times 3,7 and 8. The s.P.R. was positive at tirne 3, negative at othe:: times. Thene were no S.P.Rts of D on W. In the cattle paddoek, no significant

All equations as discussed in section 3-52 IAB!E 4.4 . IABIE tl. 3 REGRESSIOI COE¡TICIEXIS REGRESS ION COEFFICIEI{TS S?IPA S?IPA WTMA DENSIÎY RELAÎIONSHIPS NTTMA BIO}IASS RELATIONSHIPS

Paddock Ti¡ne bo brD b2(Bt-B) bgH R2 Paddock Time bO b1(Bt-B ) bzH R2 (t) (t) cattle 0 46. 4 2.38E-03 -4.3',t8-02 1 . I 3E-02 .576 g.2gàtt 2.46 2.88.1 . 138 q4 Sheep 3 20 34 28..t 1. 26 Cattle 1 -1. r+.388-03 -3. 86E-03 2. 96E-03 .735 3.72*, 0. 95 73.07*^* 0.29 0.7 0

Sheep 1 54500 -28.6 -4. 79 .096 CattLe 2 -3.81 3. 26E-03 6.75E-03 1.81E-03 .8L7 2.67t^ 0.65 76. 86* ^' 0.62 0. 8s Sheep I st700 -76.2 7. 86 . 156 CattIe 3 o.276 3.258-03 7.418-03 -6.998-04 .s22 3.45t* t 7.30 8.37.t, 0. 57 0. sl Cattle 4 ¡r.218-03 9. 98E-0|l -1.75E-04 .593 \_0'161 9.B8,ttt 0. s6 0.24 CattIe 6 3. 88 5,49E-0r+ -5.12E-04 -1.05E-03 .827 72.72*'t* 0. 07 0. 61 Cattl.e 7 1. 528-03 6. TABLE 4.5 -1.98 90E-03 9.59E-04 .692 17.92**. 0. 83 0. 21

REGRESSION COEFFICIENTS Cattle -1. 01 1.67E-03 -1.33E-02 I . 35E-0 3 .754 12 <^I** 1.20 1.51 STI2A NTTIÐA AVER'{GE SIZE RELATIONSHIPS

Sheep a -16.9 7.34E-03 7.528-02 6. 82E-03 .792 t,1 (Bt-B ) bzH R2 * â Paddock time bO 74. 0g* " "qt*I 0.63 (t) (t) Sheep 2 3. 66 3.83E-03 2.I28-O2 -1.14E-03 .692 77.641à1 1.43 0. 22 1. 61E-06 .165 CattIe 0 2.81E-06 -6.408-06 Sheep 3.63*r* 2. 37 3 -3.90 3.71E-03 5.75E-02 1.95E-03 .698 70.59*.4 2.08 0. 51 .184 Sheep 1 3. 65E-03 8.92E-06 9.85E-07 Sheep 4 -o.702 8.25E-03 1.45E-02 -1.12E-03 .747 3.8]tt',t'* 0. 91 73.37t*'t 1. 82 0. 39 .153 Sheep 3 1.79E-03 1.09E-05 -9.41E-08 Sheep 6 -1.83 I . 54E- 0r+ 6.99E-03 -6. 36E-0r{ .80r+ 3.48tà* 0. 20 16.23* t* 1. 30 n tt -1.068-06 7.998-07 .104 Sheep 7 3. 55 1.00E-03 1.29E-03 Sheep I 4.25E-03 -8.43E-04 .799 (D 2.571* 1. 51 75.451t* 0. 74 0. 33 (¡ Sheep I 62.9 4.488-03 -4. 058-02 3.75E-03 .683 70.054** 2.17 0. 1? 96. relationships wer:e found.

Cb ) Biomass ::e.l-ationshiP"; table 4-4

B nb2 Cs Sn =bO+brD Sn t -B^)+¡^w5nó

B on D was significant in atl cases. OnIy one S'P'R' of B on (ea - B) was found' in the sheep paddock at time l-' B on W was significant (positive) in the cattl-e paddock, time I.

BlD (s + brw Sn =bOfb, t -B,rr)

Fo:r cattle paddock data, a positive S.P.R. of B/D on (S was found at time O. Fon sheep paddoek data, positive t - B) s.P.R. rs for the same nelationship wene found at times I and 3' and a negative S.P.R. at time 8. No othen significant coefficients were found.

(d) Discussion.

only in th-e cattl-e paddock did a plant panameten show a significant nefationship with W, and this was an isol-ated instance.

Clea:ren demonstnation of piosphere s5rmmetry is required before any inferences can be dnawn concerning the nel-ative effects of sheep and cattfe, Thenefo::e it is concluded that neithen cattle non

sheep lrave affected the abundance of Stipa nitída to any detectable extent. The effect of the presence of ottren species on Sttpa appeaned more directly associated with- density on average size than with- 97. biomass, although- the times at wh-ich-::elationships were detected vanied witb- ttr-e panticula:n plant panameten. Th-e a:rgument of section 4-34 may h e invoked to account fon trre sign changes with time, obsenved fon r.elationstr-ips within the sheep paddock, but foundens in the case of th-e cattle paddoek, whe::e the relation between B and (ta - B) was negative at a tjme of pasture scarcity'

4-44 Regressíon anaLyses fon Enneapogon spp..

(a) Density r:elationships; table 4-6 DE. = bo + b, (ta - Br.) + brw S.P.Rts of D on (B- -B) wene found fon bothpaddocks as follows: in the cattle paddock for times O, and 4 to I incl-usive (a11 positive), and in the sheep pad.doek fon times 1 and 3 Cpositive), and times 7 and B (negative). D on W was significant (positive) in the cattfe paddock, for data fnom times 0r1r2 and 8; but was non-significant at a1l times in the sheep paddock'

(b) Biomass nelationships; table 4-7

B + bz (e Br") + br!Í Ea =bO+brD Ea t

B on D was significant in all cases. B on (B- - B) was significant only for sheep paddock data fi:om time 2. B on w was significant (positive) fon cattle paddock data fuon times 0 to 3 inclusive, time 5 and time 8. Fon sheep paddock data, B on W was significant (negative) at times 2 and 6. 98

TABLE 4.6

REGRESSION COEFFICENTS ENNEAPOGON SPP. DENSITY RELATIONSHIPS

2 Paddock Time bo b (s. -s) b I^T R t t 2 (t) (t)

Cattle o 106000 -36s 57.6 .202 9<** 2. 87** " Cattle I 94600 -251 43.2 .757 2.43 2.64** Cattle 2 59200 -342 52. 5 .241 2.41 3. B3** rÌ Cattle 4 185000 -60.4 12.9 .235 4.31*x* 0.90

Cattle 5 156000 -673 17.2 .252 4. 38* x rt L.35 Cattle 6 202000 -672 30. I .205 3.77*** 1". 49 Cattle 7 162000 -259 42.8 .165 2.76** 2.05 Cattle I 187000 -s89 59. 7 .256 3.72*,+* 2.77** Sheep I 25700 469 -3.11 .100 2.70x* 0.7L Sheep 3 11000 111 -2.83 .L47 3.38** 0.94 Sheep 7 395000 -743 -I9.2 .163 3.49*** 0. 68 Sheep I 385000 -421 1. 33 .401 6.68*** 0. 06 99

TABLE 4.7

REGRESSION COEFFICIENTS

ENNEAPOGON SPP, BIOMASS RELATIONSHTPS

Time b btD b (B -B) bsw R2 Paddock 0 2 t (t) (t) (t) .792 Cattle 0 -131 1 . 64E- 03 -0. 151 0.r32 7L.25x xx 0.94 5. 81*** Cattle t -150 1.43E-03 -0.0167 0.114 .740 9.97x** 0.73 5. 6Bx * tt .811 Cattle 2 -99. 3 1.118-03 0.2r1 0.0565 12.39xx* 1_.95 5.08*x* 0693 .607 Cattle 3 -35.4 6.228-04 5. 94E-04 0. 6.BB*rt* 0. 00 5. 60rt * lt .279 Cattle 4 -16. B 1.008-03 0. 347 0.048 4.37x*x 1. L7 7.80 22 364 Cattle 5 -107 9.28E-04 0. 0494 0.1 3.27*x* 0. L2 4.0Bxx* .410 Catt1e 6 50.9 7. 39E-04 -0.234 0.393 5. 15* x rr 1_.0L 7.60 Cattle 7 33. I 1. 86E-03 0.286 0;0916 .479 6.80**x 1.29 7.90 Catt1e I -tt4 1.64E-03 -0. 300 0.194 .750 9.07*/\* L. 76 5.64xx*

Sheep 1 1.19 8. 39E-03 0. 184 -0.0771 .665 10.02/t** 7.47 2. 57 Sheep 2 -]-3.2 3.81E-03 0. 0735 -0. 0420 .664 6.67*xx 8.02*xx 2.60** Sheep 3 -4.52 3.93E-03 0.0765 -6. 45E-03 .709 11. 46* /t* 0. 77 0.76 Sheep 4 -4. 5B 1. 59E-03 0 27 9 -0.0452 .545 8.34*xx 1 39 7.73

Sheep 5 25.3 9.65E-04 0. 0910 -0. 0367 .504 7, Bg*rtrt 0.44 1.58 Sheep 6 191 1.69E-03 -0,229 -0.135 .695 L0,g5xx* 0,63 3. 0Bx * Sheep 7 285 2.2+E-43 -0. 0371 -0.104 .695 10, 87x*x 0.70 2.17 Sheep I 109 3.578-03 -0. 331 0. 264 .521 5. ?Bx *x 0. B2 2. sB 100

TABLE 4.8

REGRESSION COEFFICIENTS

ENNEAPOGOÌ¡ SPP. AVERAGE SIZE RELATIONSHIPS

2 Time bo b_ (B -B) b It R Paddock It 2

CattLe 0 6.13E-04 4.82E-07 7.828-07 .377 0. 54 6.24*** Cattle 2 8.148-07 2.15E-06 6.00E-07 .264 1.62 4.69*xx Cattle 3 -1.428-02 1.73E-06 1.08E-06 .169 0. 59 3. 66** * Cattle 5 -2.92E-04 6.20E-06 9. 588-07 .139 1. 62 3. 02* * Cattle 7 -2.2]-8-03 3. 8sE- 06 4 . 61E-0 7 .127 2. B7** L. 54 8.61-E-07 9. 10E-07 .28+ CattIe 8 -8.64E-04 c t r*** 0. 68 ¿. Ja)

Sheep ! 5.218-03 1. 28E-05 -2 . 11E-0 6 ,253 4.13**x 2.08

Sheep 3 2.248-03 1.27F-05 - 7 . 608-07 .I22 2.77*x 1.. 80

Sheep 7 2. B6E-03 7.90E-06 -5.25E-07 .188 3.63*x* 1..81. Cc) Avenage size ne.l-ationships; table 4-8 B/Dn. = bo * ur_ Cta - BEu) + nrw

B/D on (Bt B) was significant, fon cattle paddock data fuom time 7, and fon sheep paddock data fnom times 1r3 and 7. B/D was significant (positive) fon cattle paddock data from times O'2'3'5 and 8; but no significant ::elationsbip was observed fon the sheep paddock.

(d) Discussion.

In bothpaddocks, a numben of significant nelationships between plant pa::amete::s and distance fnom waten welle detected.

From the national-e in section 3-4, th-ese ane considered stocking effects. In the cattle paddock, sueh nelationships we:re displayed by all three plant panametens. They became obscr¡r'ed duning on inmediately aften peniods of pastue g:rowth. rn all cases, negression coefficients wene positive. There was vaniation between panametens in the clarity with which they displayed the piosphene s5rnnnetry: nelationships with distance to water: wene found most frequently when using biomass as the dependent vaniable.

Pnobability values wene much lowen fon the S.P.Rfs of biomass and average size on distance to wate:r than fon the S.P.Rfs of density on distance to waten. This implied that tlre fonmen two parametens, since they showed the stocking impness with mone centainty, ltere more irnmediately affected by th-e cattle than density.

In the sheep paddocks, piosphene nelations welle detected only by biomass, æd the S.P.Rts wene negative. Decreasing 1-02. abundance of EnneapOgOn witlt incneasing distance fnom water Cand so, with- decreasj¡g stocking pt?essune) is sunprising, since the two species involved ane reported to decr:ease with incneasing stocking plressure (Beadle, 19.48). Howeven, piosphene relationships wer:e not detected as commonly as in the cattle case, but were not necessanily obscu:red by pasture gnowth (..g. B on !t at time 6).

The conclusions in negand to stocking effects wene thus: 1. Both cattle and sheep imposed piosphe::e s¡rmmetry

on the distni-bution of Ermeøpogon sPp. . 2. The sheep impness was the nevense of the cattle impness: abundance incneased with distance fnom waten under cattl-e, but decneased with distance

under sheep. 3. The cattle imp::ess was the more consistent, but

became undetectable duning or immediately following peniods of majon pastu::e gnowth. 4. The sheep impress was detected only intenmittently, but was detectable in one pe::iod of pastr::re gnowth. 5. The c1a:rity with which píosphene symmet::y was shown vanied with the panticulan plant parameten used' 6. The panametens which dernonstrated the symmetry most cleanly differed unden sheep- and cattle-stocking. The variable ttbiomass othen speeiesrt, on (Ba - B), was oniginally intended as an index of plant competition (section 3-52).

This appea:rs to be a misconception. The sign changes obsenved

ft-rn ttlis tenm in regressions using sheep paddock data suggest at least two pnocesses in operation, but the angument of section 4-34 103. cannot account fon the cattle case, where alf relationships involving this va:riabfe wene found positive. The latten case suggests that competition played no detectabJ-e pant in influenc-

íng ttr,e abundance of Enneapogo,t4 spp. in the cattle paddock. S.P.Rrs on biomass of othen species were most evident when density vras the dependent variable, for: both sets of data. Whether or not other speciesrbiomass was an index of competition, it is su:rpnising that its effect was most clea:rì-y displayed by density, the least quantifiable pa::ameter used.

4-45 Regressíon qna.LAses fot' Aristida contoï'ta.

(a) Density relationships; table 4.9

D B }I Ac =bOnbf(Bt- Ac )+u 2

No S.P.Rfs of D on (B- - B) wene detected. D on W was significant for cattfe paddock data fnom times O'1r7 and 8, and fon sheep

paddock data fuom times 2r3 and 8. B on W was positive in each case'

(b ) Biomass nel-ationships ; tabl-e 4.lO

B + blD *bz (s B )+¡ W Ac =bo Ac t Ac 3

B on D was significant in all cases. No S.P.Rrs of B on

(S. B) were detected. B on W was significant (positive) only t - for cattl-e paddock data flrom times 4 and 5. 104.

TABLE 4.9

REGRESS ION COEFFÏCIENTS AETSMDA CONTORT'A DENSITY RELATTONSHIPS

2 Paddock Tímê b b1(Bt-B ) bzw R 0 ft) (t)

Cattle 0 1040 -28.7 23.4 .1304 7. 34 3. L4t * Cattle t -6 56 -23.5 19. 9 .1344 1.1_0 3. 79* * Cattle 7 18800 -35.2 28.4 .1653 2.75 3.44*x* CattIe I 19700 -44.1 25.6 .0989 2. L9 2,58**

Sheep 2 -3180 1.74 3.61 .0950 0. 38 2.65x* Sheep 3 -1800 -o.32 2.30 .1043 0.07 2.79** Sheep I -6782 13.9 64. 4 .2728 L.1.6 4.57xrÍ*

* 105.

TABLE 4.10

REGRESSION COEFFICIENTS ARISTTDA CONTORTA BIOMASS RELATTONSHIPS

Paddock Time b b D b2(Bt-B ) bew R2 0 ! (t) (t) (t)

Cattle 0 -72.4 1.73E-03 +.628-03 9.45E-03 .898 21.76*** 0. 32 L.82 Cattle t -]-4.2 1.70E-03 t.768-02 9. 32E-03 .774 13.41*ttx 0.7 5 L. 28 Cattl-e 2 -1-0.2 1. 058-03 3.48E-02 4.15E-03 .793 74.07*** 1.. BB L.20 Cattle 3 -+.31 9.13E-04 1.48E-02 3. 11E-03 .837 75.98*x* L.69 2.00 Cattle 4 -7.70 7 . 9 3E-04 6. 56E-03 9.64E-03 .7 st 72.63*** 0. 64 3.58*x* Cattle 5 -3. 60 4. 84E-04 -9.33E-04 4. 50E-03 .77 4 13.7 4x* * 0. 1B 3. 00x x Cattle 6 -8. 34 1.85E-03 -t.7IE-02 8.45E-03 .793 14. íBx * * 0. 66 L.72 Cattle 7 55.4 1. 91E-03 7.098-02 -5. 178-02 .352 6.03lt** 7.6L 2.23 Cattle I 15.4 1. 37E-0 3 5.83E-03 -1.05E-02 .688 11.75*xx 0.29 1_.05

Sheep t _2.8I 6 . 15E-0 3 6 . 49E-0 3 9. 56E-04 .907 23. 86*x* 1 0E 0. 32 l+r Sheep 2 -2.34 3. 97E-03 2.05E-01 -7.63E-03 . 568 5.47*x* 2. 50 0. 89 Sheep 3 -4.37 1. 52E-03 2.978-02 1. 73E-03 .497 7.61*xrt 2. 30 L.05 Sheep 4 -0. 446 4. 71E-04 6 . 21E-0 3 4. B1E-04 .s28 8.17**x 0.97 0. 20

Sheep 5 3. 79 4. 21E-04 -8. 098-04 -1.2!E-03 .620 o oo**tl 0. L5 0. 71 Sheep 6 8. 00 1. 20E-03 -2,238-03 -2. 67E-03 642 10. 54* * * 0. 30 0. 55 Sheep 7 -1. 5B 1.62E-03 8.048-04 2. 56E-03 .904 24. 25* *x 0. 25 L.55 Sheep I 28.3 2.228-04 -1.45E-02 9 .24E-03 797 22. B9* *x 1. 4L 0. 91 loe

TABLE 4.tt

REGRESS ION COEF}'IcIENTS ARTS?TDA CONTOETA AVERAGE SIZE RXLATIONSHIPS

2 Paddock Time b b1 (Bt-B ) b W R 0 2 (t) (t)

Cattle 0 1.23E-05 1 . 368-0 7 7.018-07 .226 0.24 3.59x** Cattle t -1.61E-04 -1.28E-06 1.1+E-06 .27L 1. 36 4. 17x ** Cattl-e 2 -2.64E-04 5.19E-07 4.478-07 .239 0.74 3.54x** Cattle 3 -3.63E-04 1.35E-06 5. 81E-07 .273 L. 22 3. L4*x Cattle 4 -2.21E-04 7 . 548-0 7 5.218-07 .176 1. 26 3.36** Cattle 5 -2.298-05 -2.81E-07 3.238-07 .177 0.94 3.72x ** Cattle 6 -7.44E-04 -2.61E-06 9 . 53E-0 7 ,723 1.79 2. 72* *

Sheep 'il t -1.23E-03 3.518-06 8. 14E-07 .119 2.90*x 1..23 Sheep 3 -1.44E-03 1. 23E-05 5. 63E-07 .273 4. gltr * * L. 27 to7 .

Cc) Avenage size refationstrips3 table 4.11 B/oR. = hO t b, G- - BA") + nrw

B/D't on (e, - B) was signíficant (positive) fon sheep paddoek data fuom times I a¡d 3. B/D was significant (positive) for but no such nelat- cattle paddock data fnom times 0 to 6 inclusive ' ionship was evident fon sheep paddock data. fn the cattle case' the equation was less efficient when fitted to data fnom peniods of pastu:re growth (tfunes 4 and 6 ) than when fitted to data from no-

g:rowth pe::iods (times O to 3, and time 5 ) .

(d) Discussion.

In both paddocks, significant relationships between plant parameters and distance fi:om waten Ïrene observed. From the nationale in section 3-4 these ane considered stocking effects. In the cattfe paddock, the piosphe:re symnetny was detected by a1l- thnee plant panameters, but most fnequently by the avenage size variable, suggesting that this pa-nameten was the most immediately affected by cattle (c.f. Enneapogon). The ::elationships wene positive in each case. Unl-ike the Enneapogon analyses, symmetny was detected i4 growth peniods as wel-I as no-growth pe:riods, but the pattenn

was pantially obscuned, as evidenced by the fower efficiency of the

equation when fitted to data fnom g::owth peniods. In the sheep

paddock, piosphere sSrmmetry was detected only with density as the

dependent variable, and only in no-gnowth peniods. The nefation-

ships wene positive, as for the cattl-e case.

TLI-e conclusions witb- regand to stocking effects wene thus: 1. Both cattle and sheep imposed piosphene s¡rmmetry on the 108.

distnibution of Aristida eontorta,

2 The gener.al impr:ess was th,e same in both cases: abund-

ance incneased with- distance firom water. 3. The cattle imp::ess was the mone consistent, but became at least pa:ntially obscuned in peniods of majo:: pastu:ne

growth-.

4. The sheep impness was detectable only in no-gnowth peniods.

5 The cla:¡ity with which-.th-e piosphere s5rmmetny was shown varied with the panticulan plant pa:rameten used.

6 The parametens which demonstnated tlte s5rmmetry most clear:Iy diffened unden sheep- and cattle-stocking.

4-46 Regression anaLyses fot' Kochia astrotricVn aduLts.

(a) Density ::elatíonships; In both the sheep and cattte paddocks, the density of K. astro- tríeha nemained constant thnoughout the study (figure 4.11). Thene- fone K. astrotrieha density could not be negarded as an independent va:riable in the s¿tme sense as, fo:: example, density. Fur:then' any piosphere pattenn found in density distníbutions woul-d necessanily have had its onigins pnion to the study, and in the cattle paddock, prior to the introduction of cattle. The equation

Dx" bo + bt}l

was fitted, to detect any pnion impress. The negression I^Ias non- significant fon data fnorn both' paddocks, with R2 n.ao zero. 109.

(b ) Biomass nelationsÏtip* ; table 4.12

B bo + btD nbz Cs 8.. ) + b^w Ka Ka t KAJ

Positive S.P.Rts of B on D we:re obsenved th:roughout. A positive S.P.R. of B on (ta - B) was detected fon sheep paddock data fnom tirne 2. No othen significant relationships were found.

B/Dx_ = bo * bf (Bt - t*.) + b2}I

(B- (positive) only for B/D on t - B) was found significant cattle paddock data fnom time 5. S.P.Rts of B/D on !'l wene found only for" sheep paddock data, fuom times 2r3 and.7. The negression coefficients were positive.

(d) Discussion.

Fnom th_e density data, it can be concl-uded that no long ter"m stock impness on K. astrotrieha was present in eithen paddock. significant nelationships between ptant panametens and dist: ance fnom water wene only found for data fnom the sheep paddock.

Fnom the r.ationale of section 3-4, these relation-ships alre considered a stocking effect, specific to sheep. Piosph-ene s¡rmmetny was only detected with avenage size as the dependent vaniable, hence this was the panameten most immediately affected by the sheep.

S5rmmetny was obser:ved in both- glowtb- and no-gnowth peniods, but the equation was mone efficient when fitted to data fi:om the latter. Th-us, as befone, pastur:e growth obscur"ed ttr-e stocking relationsLips. 110. TABLE 4.12

REGRESSION COEFFICIENTS K. ASTROTRTCHA (EIUITS ) STOM¡SS RELATTONSHIPS

Paddock Time b b D b2 (Bt-B ) baw R2 0 1. (x) (t) (t)

Cattle 0 16. s 2.088-02 6. 75E-04 -5 . 19E-0 3 .725 13.05xx* 0.04 0. 79 Cattle ! 20.4 1.69E-02 -8.93E-03 -7 . 09E-0 3 .750 73.74**x 0. 61_ 1_.47 Cattl-e 2 11.9 7.13E-02 5.03E-03 -6. 93E-03 .776 14. 96xx x 0.35 2. 34

Cattle 3 5.27 t.3tE-02 -4. 688-03 2. 09E-04 .814 16. B7*** 0. 29 0. 07 Cattle 4 19. 6 2.OtE-02 -1.86E-02 -6.83E-03 .791 Ìt 151. 12** 1_. L6 L. 59

Cattle 5 9.48 1.438-02 -1. 92E-03 -1. 60E-03 .77s 15.06*xx 0. L6 0. 46

Cattle 6 13.0 1.27E-O2 -9 .92E-03 -3.70E-04 .694 71.90**x 0. 69 0. LL Cattle 7 13. 3 t.448-02 -4.04E-03 2.03E-03 .546 8.93**x 0. 41 0.38

Cattle B 1. ss 1. 09E-02 -t.7rE-02 9.47E-03 .683 11. 30rt xx 1_.52 2. 48

Sheep t 0.27I 4.94E-02 3. 53E-02 -7.65E-03 .617 g.g2x** . 1:83 0.76 Sheep 2 -9.66 1. 89E-02 1. 30E-01 6.46E-04 . 511 5.94*x* 6.70* ** 0. L0 Sheep 3 -]-7.7 1.448-02 4. 01E-02 8. 51E-03 .536 g.50xx* 2.00 2. 50 Sheep 4 -5.17 4 . 39E- 02 4. 94E-02 -1.278-03 .583 o E1**ls L. B4 0. 1_4 Sheep 5 -6.97 +.378-02 3. 38E-02 7.10E-03 .634 10.71*x* 1. L5 0. B6 Sheep 6 15.2 3.45E-02 -2. O7E-02 5.49E-03 . 507 7. BZx lt* 7. 44 0.60 Sheep 7 2.28 3. 34E-02 -2. 60E-03 6. 25E-03 .563 8.94*** 0. 23 0. 83 Sheep 8 17.3 +.748-02 -6. 93E-03 1. 088-02 .536 B. 43* lf x 0. B2 1. 09 111

TABLE 4.12a REGRESSION COEFFICIENTS .K. ASTRUTRICHA AVERAGE SIZE RELATIONSHÏPS

2 Time b b1 (Bt-B ) bzl R Paddock 0 (t) (t)

Cattle 5 1.00E-02 4.85E-05 -3.26E-06 .276 4. gg*** 7.72

Sheep 2 1.138-0? -4.44E-06 1.17E-05 .tt4 0. 35 2. B5** Sheep 3 4.63E-03 9.08E-06 9.518-06 .165 0. 58 3.57*x* Sheep 7 2.258-02 -4.668-07 1 .408-05 .100 0. 06 2.63*x

TABLE 4.13

REGRESSION COEFFICIENTS

K. ASTROTRICã.A.. REGRESSION OF JUVENILE ON ADULT DENSITIES, CATTLE PADDOCK

2 Time b b1Dx" R2 Time bo b1D,

0 .63 .146 . 195 5 .58 .181 .210 4.05*** 4.25*** t .44 .199 .298 6 .51 .L77 .200 5. 38* * rr 4.12*x* 2 .26 . 158 .300 7 54 .180 .222 5t. 40lr * * 4.47**lr 3 .55 .150 .!77 I .61 .!7 5 .tB7 3. 82*** 3.gs*** 4 .87 .180 .242 4. 66* rt * 772.

1?re conelusions v¡ith- regand to stocking effects wene thusi 1. Only sheep inposed piosphene symmetny on the dis* tnfüution of K, astnol;r'ícha-

2. Ábundance of th-is species incneased with distance

fuom water.

3. Ttre impness was intenrnittent, but was detectable i¡ both g::owth and no-g:rowth'per:iods. 4. The impness was onJ-y evídenced by the avenage size

of bushes. s.P,Rts of plant parametens on biomass of othen species we:re

observed twice, and we::e positive in both cases. Thus, as fon Aristída, it appears that the pnesence of othen species had littl-e effect on the biomass olr aver?age size of K. astrottieha.

4-47 Regression útaLAses fon KocLria astrotricVø iuueniLes.

Befone investigating piosphere relationships, the degnee of

dependence of juveni.l-e on adul-t ,K. astnotnicha densíties was sought, through the equation

D b br-DKt Kaj 0 *

In the cattle paddock, the reg::ession hlas significant (positive) at all times (tabte 4.1-3). rn the sheep paddock, no significant refationship was observed. Howeven, the latten situation may have anisen fnom the fow occumence of juveniles in the sheep paddock (see fig. 4.11). Inclusion of B_r as an index of abundance of

potential competitors, and of W as independent va::iabl-es did not 113. incnease R2 values significantly; nor ÌIere any new ::el-ationstr-ips obser.ved.

Thus, in th_e cattle paddock, juvenile density was dependent only on the density of adults, sh-owed no evidence of conpetition with othen species, or, fon that matten, with adults of its own species, and vlas not affected by stocking to any detectable extent. In the sheep paddock, none of the above factons appeared to affect juvenile densities. The ovenall diffenences in densities between paddocks, shown in figune 4.11, may stil1 have been a diffenential stocking effect, but without demonstnation of piosph-ere s5rmmetry there are no gnounds fon stating this with centainty. Therefone it nust be concluded that neithen sheep- nor cattle-stocking have affected the distnibution and abundance of K. astrol;v'ieha juveniles.

4-48 Regxession ana.LAses foz' Koehia pAnØnídatq,.

In both the sheep and cattle paddocks, the density of K. pavûn- ídata nemained constant thr:oughout. Stnictunes thus applied fon K. astrotricha (q.v. section 4-46 ). O*n on l,l was non-significant fon data fnom both paddocks, with R2 value nearô zeno in the sheep case, and O.O7 in the cattle case. Hence it was concluded that no significant long-tenm stock impness existed The standard biomass equation was fitted to the data, but the only significant nel-ationships detected we:re between biomass and density (table 4.14). No significant relationships were found fon average size. Hence it was concluded th-at; 1. Neittrs: slreep nor: cattle affected th-e abundance of

K. pATØnídnta. to a significant extent; and 114.

TABLE 4.14

REGRESSION COEFFICIENTS K. PYRAMIDATA BIOMASS RELAT]ONSHTPS

Paddock Time b br-D ur(n--n ) bew *2 0 ft) ft) ft)

Cattle 0 6.01 5.228-02 -+.25E-02 1. 34E-02 .632 9.45*** L.60 L.32 Cattle t 12.9 5. 85E-02 -5. 96E-02 1.36E-02 481 7.00xx* 1.29 0. 89 Catt.l-e 2 !0.7 3.978-02 - 5 . 19E-02 5.97E-03 .573 B.59*** 1. 27 0.70 Cattl-e 3 1.97 5.278-02 -6.7+E-02 1.248-02 .613 9.33x** 1_. 1_7 1_.25 Cattle 4 5.88 5.21E-02 1. 30E-02 -8 .58E-04 .716 12. 43* * x 0. 52 0 .J¿

Cattle 5 6.49 4.19E-02 -2.838-02 6.14E-03 .681 11.03*x* 7.33 0.95 Cattle 6 21.8 5.35E-02 -5. 34E-02 1. 10E-03 .750 1Ð "E*** 2. 07 0. 16 Cattle 7 14. 0 4. 89E-02 -1. B3E-02 2.91E-03 .743 72.65x*x L.77 0. 4B Cattle 8 10 .6 4. 56E- 02 -2.428-02 B. 30E-03 .657 10.06*,t* 1_.49 1_. 0L

Sheep t 2.7 3 0.154 -2.158-03 -t B 5E-03 . 909 25.77*rr* 0.72 J-. 13 Sheep 2 -1. 53 0.105 9. 55E-04 4. 53E-04 .892 23.47xxx 0. 29 0. 37

Sheep 3 -1. 63 0.1-02 7.59E-03 1. 31E-04 .844 18. B5* * x 0. B9 0. 09 Sheep 4 -r.23 0.197 3. B9E-03 -9.32E-04 .853 L9.65*** 0. 5L 0. 34

Sheep 5 1.68 o.152 -2 .57E-03 1. 02E-03 .866 20.63*** 0 .óo 0. 87 Sheep 6 4.5s 0.140 -5. B7E-03 -1.63E- 03 .8s3 19.69/1 *rt 1.7B 0. BL Sheep 7 5. 79 0. 143 -3.83E-03 -2.97E-03 .794 16.04x** L.03 1_.1_9 Sheep I 7 .76 0. 570 -4. 20E-03 -1.638-03 .815 17. 14'+ x * 1 0? 0. 64 115.

2, The abundance was not dependent on the oth-er" species . present. 449 GeneraL diseussion drl ?egoession qvtaL!:ses.

With, the exception of biomass equations with a density ter:m, the r:eg::ession equations wene inefficient in accounting fon the variation of ttre plant pa:rametens. This was only to be expected: the :regressions wene intended fo:: the detection of piosphene pattenns

supenirnposed on the intninsic vegetation patterns, and wene not pnoposed as models of the total system. Thenefone, such undenlying i¡rfluences as soil moistune, nutnient levels, insect gnazing and the nest, which largely detenmine pÌant abundance, were excluded fnom considenation. Much of the varia¡ce of the dependent var:iabl-es was not accounted fon, because such factors wene excl-uded.

Equations with biomass as the dependent vaniable and density as an

independent var"iab.l-e wene acco::dingly mor"e efficient, since density p::ovidedrin effect, an index of the backg::ound variables infl-uencing

species abundances. The efficiency of al-l- equations was least during on immediately aften peniods of pasture gnowth. This was considened to be due eithen to confusion of the short te:rm piosphene pattenn nesulting fi:om

the animal-s I lessen dependence on the waters at such times o:: to the obscuning of stock effects by the incnease in fonage. For instance, with an ovenall stocking nate of I sheep per 5ha, the

fonage consumed over 100 days would be equivalent to a 4Okg/ha (assuming tlre sheep to take 2kg (d.w. )/day -- Weston and Hogan, 1C73). If th-e pasture at the beginning of th-e 100-day peniod

had. a totat biomass of 100kg/ha and no gnowth occurned, the gnazíng 116. effect would b.: readily evident; as it was for th-e fÍrst few obsenvations. Howeve:r, an incr"ease in pasture levels of 3OOkg/ha (as obsenved in ttre sheep paddock between times 3 and 4) must tend to obscune the loss to stock. The p::oportionate inc::ease in backgr.ound noise Ci.e. the vaniation in the effects of othen envi:ronmental infl-uences) oven such a peniod would necessanily lessen the efficiency of equations designed to isofate the stocking influence.

In genenal, equations welre mor:e efficient when applied to data fuon the cattfe paddock than when appJ-ied to data fuom the sheep paddock. This might be taken as indicative of an impact heavien for cattle- than fon sheep-stocking, with the cattl-e piosphene pattenns being the more easil-y detected in consequence. This in tunn coul-d be r:elated to the gneaten dependence of cattle on the watens pnovided (see section l--62). However, in the pnesent caset the heavien impact of cattle may also have been due to (a) a heavien ove:raIl stocking r"ate in the cattle paddock than in the sheep paddock (section 4-2); (b) a concentration of cattle grazing in the sampled vegetation type (fig. 4.4) as opposed to a mone even spnead of sheep grazing; and (c) the two watens of the sheep paddock confusing and weakening the piosphe::e pattelrn centred on eithen one. Thus the cJ-arity of pattenns unde:r cattle-stocking nelative to that under sheep-stocking may have anisen from the special ci::cumstances of the stud.y; hence the diffenences'in clanity cannot be considered nep::esentative of the genoral case.

The nefationship between abundance of a par:ticul-a:: species

and the abundance of othen species was obscu:re. As alneady stated,

th-e tenm can not be consider:ed simply as an index of competition 717 . as was intended: significant regressions , wltet'e detected, llelîe always positive in the cattfe paddock case, and both positive and negative in th-e sheep paddock case, nathen tha.n consistently negative as would be expected wene the tenm a comPetition index. With- out independent evidence, th-ere is no point in speculating on the natune of the influences which the tenm was detecting. Nonetheless, th-e tenm necessanily nemained in the equations: praesence of othe:r species did inftuence the abundance of the individual species, and this infl-uence, whateve:r th-e cause, Ìrad to be taken into account to isolate th-e stocking effects.

The mai¡ objective of the negnession analyses llas the detection of piospher"e symmetry, with va:riations in this symmetny allowing objective companrison of the impacts of sheep and cattfe. Neithen he:rbivor"e imposed disce:rnible pattern on the distnibutions of Stipa nitídø and KocVtia pATarnidata. This may have been due to the low incidences of these species fon much of the pe:riod of study. Sheep-induced pattern in Stipa has been weff documented by othens

(".g. Osborn et al, 1931). The majon diffe::ences between the pattenns imposed by sheep and th-ose imposed by cattle r¡Iere: 1. The invense distr"ibution of Ewteapogon spp. abundance in the sheep case relative to the cattle case: and 2. The piosphene patts¡¡ for K. astrotTicha, obse::ved in the sheep case but not in the cattl-e case. Both cattl-e and sheep imposed simil-an s¡rmmetries on th-e distribution of Az,ístída eontorta, but the panametens wh-ich most consist- entl-y and cleanly displayed the s5rmmetny differed between the two

cases. 118.

Both- the diffenences and simi-tanities i¡ sheep- and cattle* imposed patterns ane explained i¡r pant by th-e dietany pneferences of th,e two animafs: the cattte have consumed only g:rasses, while the sheep have consumed both grass and sh-nub, as would be expected fnon the outline of pnefelrences in section l--63. This explan- ation is not entinely satisfactory, ho\^Iever. Enneøpogon avena,ceus (scored as pal?t of Enneapogon spp. ) is known to be p::efer::ed by both cattt-e (Chippendale, 1968; Low,l-973) and sheep (Leigh and Mulhan, 1965; Beadle, f94B). If gnazing by sheep and cattle was the cause of the piosphene patte::n, then an increase in abundance with decneasi¡g stocking pressutle should have been found unden both her"bivones.' Enneapogon abundance incneased with distance (i.e. with decreasing grazi:ng pnessulre) in the cattle paddock, but decneased with increasing distance in the sheep paddock.

G:razing as a direct i¡fluence cannot explain the l-atter nelationship.

Further", other mectranisms plloposed, such as tnampling on nutrient transfen, would be expected to apply equally in the cattle paddock. The detection of piosphere pattenn in the distrjòution of K, astrotricha average size is also inadequately explained by citing pnefenential g::azing selection as the cause of the pattenn. K, astrotrícha, although mone palatabl-e to sheep than most othen penennial chenopods (Jessup, 1951; Beadle' 1948), nonetheless cannot be considered mone highly p:referred than the grasses studied (as indicated by Jessup, 1951-; also see section .l--63 on sheep- gnazing tniaÌs involvíng other Kocfuia species). Sheep-induced patterns in no-gnowth- peniods Ce.g. times 2r3) may be explained hy the r:emoval of the mone pnefenned species by sheep, followed by 119. more intense gÌ,azing of tb-e shrub; but by ttr-e same token, such pattenns in K. altnotv'1:cho. sîze are not expeeted when th-e preferned species ar"e abundant, as at time 7. No:r can th'e patter:ns at such- times be consider"ed as a residue of th-e pattenns imposed in the no-g:rowth peniods: the latten Patterns welre completely obscu¡ed by the fi-::st pastune gnowtlr-. Ay,istida contorta,, showing sirnilan piosphere pattern unden both animal-s, provides a useful- basis fo:r comparing the mechanics of eattle and sheep grazing. In the catt.l-e case, piosphere Pattern was detected most comrnonly by the avenage size parameten, but in the sheep case by th-e density panameter. This illustrated the closer grazíng of sheep (discussed in section 2-3): the whole plant was taken at a time, hence thene v¡as no neduction in quadnat biomass olr average size independent of the reduction in density. on the othen hand, the cattle took only pa:rt of the plant: th.ene was reduction in biomass and size independent of the reduction ín density.

It was hoped that shont-term pattenns wpuld indicate the potential- futu::e tnends in the pastune, by showing the beginnings

of diver"gences between sheep and cattle paddocks. For instance,

welle the pattet?ns mone consistent, a long-term decnease in

Enneoptogon populations with cattle-grazing, and an incnease with

sheep-gnazing might be pnedicted. Howeverr in no case we:re the patterns evident th:roughout. In the cattfe case' any piosphene

sSfmmetny was obscuned during pastune gnowth, and in the sheep caset patte::ns were very inte::mittent. Thus alÌ the pattenns obser:ved wene considered eptremenal, dependent only on season, and so coufd not be 720. used as ttre basis fon pnedicting longe:-term outcomes of stocking. In summa::L tb conclusions fnom tlre foregoing analyses !^relle: 1. Neittren sheep non cattfe affected Stipa nítída and Koehi'a

PAaØnidatq. to a discer"nible extent. 2. Both sh-eep and catt.Ie irnposed sinitan piospher:e symmetnies

on the distr'íbution of At'istída contorta: but the Pal?am- etens which- displayed ttr-e patter:ns most clea:rly differed

between paddocks, as a result of g::azing habits differ"ing

between sheep and cattle.

3. Sheep and cattle imposed markedly diffenent piosphere symmetries on the distnibutions of Enneapogor¿ spP. and Kochia astrotr'íeha,

4. The hypothesis, that pr:efenential gnazing selection is the cause of stock-refated pattenn, accounted fon only par"t of the fonegoing nesults. 5. All the obsenved pattenns wene ephemer:al, genenally becom-

ing obscu.ned with pasture gnowttr-. 6. There was no evidence of the initiation of long tenm dive:rgences between paddocks as a nesult of the diffenential stocking.

4-5 ABUNDANCE OF MI}IOR SPECIES

The densities of minon species fon eaÔh observation ti¡ne ane plotted in Appendk 1, with data fuom open quadrats, both paddocks. Table 4.I5rcompiÌed fuom th-ese p1ots, demonstrates th-e pnepondenance of the species which gr:ow following late spning and sunnen nains.

The plots also demonstnate th,e eptr-emenal natune of many species, 121.

TABLE 4.15

SEASONAL RXSPONSES OF PASTURE SPECIES AT MT VICTOR

(a) Density increase folJ-owing summer nain. ArLstida Bassía SaLsola contorta patentùcuspis kaLí AtrípLeæ DactyLoctenium Sida Limbata raduLans cottugata AtripLeæ ConuoLuuLus CLíøtthus spongiosa erubeseerts foz,mosus Bassía Erodium Inagus bifLora eagnorwn qustv'aLiartus Bassía Kochia Tz,ibuLus Lmieuspís eæeauata terrestris Bassía Triraphis Lirnbata moLLís

(b) Density incnease following both late spning and summen nains.

Enneapogon Euphorbia Sída spp- fuwnnondii uirgata Ba.ssía Sida diaeøttha intricata

(c) Density incnease following spring rain. Stipa Bassia HeLiptez,wn ní.tida paladora fLoribundun AtripLeæ KocVtia Angianthus stipítata georgei pusiLLus Bassia MLnuría CaLocepltaLus d.ecuv,rens Leptopha LLa díttr+ehií Bassia obLiquicuspis

(d) Density inc::ease following winten nains.

Bassía Enagrostís Brachyscome dì,uarLeata ' dielsiÌ eiLiæis

TLre remainder of the species 1isted in Table 3.3 showed no discernjòl-e density response to nains. 722.

often considelîed perennial; pa:rticula::Ly Bassia species, considened by Black (1Sa3¡ to be biennial or sþor"t-tived per"ennials.

fn most cases, fluctuations in density sFrowed tlre same Pattenn b etween paddocks, although ovenal_l densities often diffened Þtween paddocks. However, centain species showed a continuous decl-ine in one on othen paddock, notwithstanding the abundant and continuous r:ains dur:ing the latten hal_f of observations. In the cattl-e case, the species were the ephemenals AtripLeæ Límbata, and Ez,agrostis díelsii. In the sheep case, they were the ephemenals Bassia diacøttha artd Bassia obliquicuspis.

The diffenences between paddocks of these species t trends

Ì.¡ere considened potentially indicative of divengence between paddocks, ::esulting fnom the diffe::ential- stocking. Símple linean negnessions of the speciest densities on distance fr-om waten wer:e calculated, but wene found non-signif-icant in atJ- cases. Inclus- ion of total biomass of majo:: pastune species as an independent vaniable a.lso nesul-ted in non-significant regnessions.

Any stock-induced pattenn pnesent may have been obscuned by the low incidence of the species (fnequency of occunnence l-ess than

2O9o), but, nonetheless, evidence of shont-tenm piosphene synmetny was necessany to a1low centain statement reganding long tenm divengence. Since no piosphene symmetry was evident, the diffenential- tnends between paddocks fon these species wene not considened to be a dinect nesu.l-t of tlre panticulan stocking negimes. Even wer:e the divengences neal, and not a clnnce effect, they wene not such as to Ímply a majon change in ttre chanacter of the ti^¡o 123. pastunes: the species inyolved cornpnised only a small pnopo:rtion of the total- flona, and contri-buted littfe to th-e total- standing cnop.

4-6 INFLUENCE ANALYSES

Figu_nes 4.I7,and 4.1-8 show associations at the .01 l-evel for the cattle and sheep paddocks ::espectively, and fon al-I tirnes. The nodes of intenaction did- not show the reinfoncement (cross-correl-- ation) usually r"equisite fon an Influence Analysis, and vanied so gneatly with time that individual analyses may not have been compan- abl-e in any case. The lack of reinforcement cou-l-d be ascni-bed to the small sample size (N=ZO) Uut the changes in associations must have :resulted from changes in r"elative abundances. fnfluence Analysis nonetheless should be able to detect a particul-ar infl-uence negandless of the panticula:r species pnesent' and used in the analysis, at any one time. Fon example, Banker and Lange (rgzo) :reponted stocking influences, detenmined by back-plots f:rom nodes composed of largely ephemenal species (e. g. Bassia species). Lange (rgzr) stated that fnfluence Analysis woul-d be viable in nonitoning changes in the stoeking inffuences with time. The implication is that the influences would stíl1 be detected, no matter which species were

observabl-e at each time. However, if the nodes of figune 4.lB wene used i¡r Infl-uence Analyses, despite the lack of reinforcement ' va:rious infl-uences would be confused; at time 2, species lrl5'17

and 20 for"m a single node, but at time 3, species l ancl 15 form pa::t of one node, and. species l-7 and 20 pant of a separate node. Fig. 4.17 & Fig.4.1B SPECIES ASSOCIATIONS IN THE CATTLE AND SHEEP PADDOCKS (SÞc

TIME I Nod. I

TIME 2 Ho¿. t @-æ Nod€ I o-@-@-@ Nodc 2

TIME 3 Ho¿e I @-.@-æ Node I o-@ o Nodc 2 Node 2 æ-@

TIME 4 Node I Node 1 @-@

Nodc 2 -@

TIME 5 Node I

Node 2 *

TIME 6 Node I @-@ Node 1 @-@ @

Node 2 Node 2

Node 3

TIME 7 Node t Node I æo@ @-æ-

TIME I Node t Node I

Node 2

Þos,l¡ve ¡¡egotr?. 01)P 001)P 3 l.lode Ä550ClATlON5 t !24.

The fi:rst case suggests one influenee in openation but the second suggests two in openation. In the present study, rnfluence AnaÌysis was intended only as a finst o:rden appnoach. Investigations using tt¡-is rnethod were abandoned aften the above conplieations became evident.

+_7 PRINCIPAL COMPONENTS ANALYSES

4-7l- Results

Few pnincipal components appea:red nelated to stocking pl?essure.

Fon cattl-e paddock data, the r:egnessions of site affinities fon the five vanimax components on distance f:rom waten were significant only fon components R3 at time t, R5 at time 2, R3 at time 3' R2 at time 6 and. R2 at time 8. Fon sheep paddock data, the negr:essions wene significant onÌy fo:r components RI at tine l-, R3 at time 2,

R5 at time 3, and. R5 at time 5. The efficiency of the negnession was low in all cases. The equations ane given in table 4.1-6. Since these components displayed piosphere symmetny in thein exPness- ion, they were conside:red functions of the stocking influence. Species affinities fon each of these components alre given in table 4.17. The other components may have nelated to var"ious environmental influences, ol? to vaganies of the data, but, since they pnessulle welre not fu::then showed no nefationship with stocking ' investigated.

Fon the stoeking components, intenpnetation pnoceeded as out- Ii¡ed i¡ section 3-53, with the species affinities being used to 125,

TABLE 4.16

REGRESSIONS OF SITE AFFINTTIES ON DTSTANCE FROM ÌTATER

(a) Cattle paddock

Time Va::imax b b f,l Significance R2 0 t component

t R3 .823 -ô.12E-04 .001 p .t7t ' 2 R5 -.943 7.888-04 .001 p .224

3 R3 .719 -5. 35E-04 .0t p .001 .130

6 R2 -.693 5.16E-04 .01 p .001 .!2t I R2 -. 678 5. 04E-04 .01 p .001 .tI6

(b) Sheep paddock

t R1 -. 523 4. 20E-04 .07 p .001 .094 2 R3 .512 -4. 10E-04 .01 p .001 .090

3 R5 .655 -5. 23E-04 .001 p .147 4 R5 .543 -4. 36E-04 .01 p .001 .701

,i 126 TA.EI,' 4. I ?

sPÈcI¡s ArrtNli'IEs (x100) rop s10c!{tNG co¡1poì,¡Et¡Ts

Ctl?TLt PrlDù)CX S¡TEEP PAÐDOCI( Time. I 236 ;3 lf Spccies. Compt' F3 P.5 P3 R2 F,2 Rl R3 R5 R5

1 Stipa 00 -11 42 _.16 ¡r 54 23 0r{ 18 nítída -0 - Ettneqogñ -4€ J9 -69 19 32 o7 -21 23 6lr gpP.

3 Artot¿da _lt8 04 -'t7 1l{ t, 27 o7 -14 -03 ænto?ta rt Roohia -30 -06 03 10 09 -05 77 00 27 oßtrotríoha

5 Koohia 79 -07 32 08 -31 66 06 01 15 pyrøtttdata 7 Atríplæ 00 t û , * û t á * Lít¡ùaèa I AtripLet I * û a * û i È -6{ spong'ioaa 72 Bæaía -07 -30 15 -27 o'l a i É t deaÌ.oMg

13 Ba¿ola -23 0q 03 -11 0q -81 52 30 10 diØútha llr Bøeía 77 -20 -15 18 1rr t È û t díwid¿ata 15 8æeía -15 05 t tl 15 -07 03 løíwpía -62 -66 !7 Bæaìa 61 09 o2 -15 01 15 -s8 1g o7 obliqvirepía rt t * 18 Bæaía 4l¡ I rt -27 15 t PatentìcÆPíe tt û t 19 Bdssía 09 t * 08 18 * patadæa. 20 Dusía 35 -1q 27 -78 02. 13 -38 -39 -08 ael¿rolaøoídea 2l Ðactvloceniw * t * * t è t ù l7 taduløte 22 Cqtnlwlw â t * -02 i á * û ñ etzbeeeaæ A il * 23 Dod.onæa 25 1¡l 39 -72 t rictozYga û i It t 23 Er4gtþâtí.o 11 * It â dlel.eií ù * 26 Erodiø t * a û 03 ølgnow * * 27 Euphæbla f t * -25 * * 39 iltnødií n * â 28 Eraohyecøe lll o2 -49 2tt eilíaria 30 Rodúa -06 52 0¡l 15 3B -01 75 10 eæû)dta A rt i 3r,t llínwía * ,i -26 -5-q Ieptophylla 37 rz.Ltuphía 03 * Ji 2\ -80 27 rcltíe 39 Saleola i 05 fr 16 kali ¡r0 si,da i * h -16 30 ærvt gata $1 Sída t2 A rt 05 03 * iì intrie¿a q6 Tpagw á n þ It 0g ùâ tøLíùtuÁ \'l Tribulw * * '13 IB teryeatPie rr8/c Cattle o'l -79 -19 a2 ît dung {0/s sheep 20 !'l 0lr -a1 dung ¡rg Kangarco I t l¡ -79 iì Í á 11 dúB ^

nlg€nvaluê. 1. 85 1. 16 1.38 2,2'l 2.13 1,65 1.30 1.10 | .32

Prcpontlon of .082 .078 .100 ,079 . 110 . 150 . 110 . 100 -067 total variance.

h nrecíea' ôrequenc¿' o:' ccqrrenec lein i0'" at thíø lire; 'l: I irl:rJl¡i¡ i 7n dnat'tãL¿. 127.

TABLE 4.18

SPEcTES WITH DENSITIES INCREASINE (+) OR DECREAST*E (-I }TITH TNCREASING DTSTANCE FROM WATER

CattLe paddoek Sheep paÃ.doek Species Tine Time t 236 I t 23 4

Stipa + nitida EVrneapogon +++ + spp. At istída + + eontorta AtrLpLeæ + spøngiosa Bassia diacøttha Ba^ssia ++ Lø.ict*spis Bassia + obLiquùeuspís Bassía patuttieuspis Bassía ++ scLeroLaenoídes Bnachyscome eiLí,aris Euphonbia dmnrunondLi Koehia + + eæcauata, Mínuria LeptopVtyLLa

Triraphi @ noLLis t2B. infen the vari¿tions in species densities with distance fnom water:. The nesults ar:e summa¡ized in tabl-e 4.18.

4-72 Discussíon.

Evidence of stocking influence vras intenmittent, as fon the neg:ression analyses of section 4-4. Stocking r:elationships were langety obscu::ed. in the second half of the perÍ-od of study. Although inferences wene therefone limited, as in the negnession analyses, the following points wene appar"ent.

1. Whene any species slrowed stnong affinity with trre stock- irrg influence at mone than one time, the nelationship

was consistent within paddocks (e.g. Stipa nLti'da and

- Ba.ssia pate.n,t.ieuspís, always decreasing in density with incneasing d.istance fnom water in the cattl-e paddock).

2. Ìfhene majon species wene involved, nefationships with the stocking influence as Í¡rfe:rred fnom P.C.Ars were coincident with ttre refationships detected by negnession analyses in section 4-4 (u'g' Enneøpogon spP' in the cattle paddock, infenned fnon the P.C.Ars to inc::ease in density with distance fr.om waten, and found to increase with distance in negnession analyses).

3. A netationship between stípa density and distance fnom

water was found in P.C.Ars, but. v'¡as not discennible in the previous negnession analyses. The fir"st point irnplied a consistently necurning pattenn, albeit short-tenm in nature. The second. point langely confinmed 129. that the pa:rticulal? components had been corr^ectly interpneted as nepresenting the stocking infl-uence. The thind undenl-ined the advantages of consider:irig the vegetation as a who.l-e, rathen than dealing with individual- species: nelationships between species t abundances and stocking pnessute, not evident when species wene considened in isol-ation, became evident when analysis dealt with all species simul-taneously.

Species I densities showed mar^kedly diffenent distributions between sheep and cattle piospher:es. Eithen piosphene pattenn in a speciesr distnibution was detected in only one paddock, on, if detected in both, the ::elationship with distance fr.om waten showed opposite signs in the two paddocks. Hence, fon the species showing stocking influence in tlre P.C.Ats, cattle and sheep had ma::kedly diffenent effects. In the cattle case, pnefenential gnazing selection of gnass species, with reduction of density nealr waten and a consequent incnease in density of less favouned species, accounts fon most of the pattenns obser"ved. The gnasses Enneapogon and Arístida incneased with decneasing grazing pnessul?e, while the less favouned forbs Bassia spp. decneased with decreasing gnazing pressune. (Hovreven, the grass Stípa would not be expected to behave in the same mannen as these fonbs). fn the sheep case, pnefe:rential g:razing of certain species cannot account for the observed pattenns. As befone, the piosphene pattenn in Ervteapogon is the neverse of the pattenn expected if pneferential- selection is the main cause of stock-nelated pattenn. Further, the chenopcd forbs Bassía diacantha, 130.

B, seLeroLaenoídes and the small shrub Koehia eæeauata ane considened to be mo:re or less equally palatable to sheep (e.g. Jessup, 1951); the differ"ences in their piosphene pattenns ane thus difficult to account fo:: in terms of differences in g::azing selection.

Veny few of the 55 pasture species scor:ed in quadnats showed any detectable nesponse to gnazing pnessure. F\:nther, pattenns wene only detected for: the most commonly occunning species, notwithstanding the high degnee of selectivity of stock (see the grazíng tnials neponted in section l--63, whene much of the stock intake was of minon eomponents of the pasture). This fr:rther suggests that some factor othen than dietany selection was the immediate cause of piosphene patterns.

As v¡ith the negression analyses, pattenns became obscur"ed in gnowth peniods. P::ediction of long-terrn divengence between paddocks was not possible because of this: the patte:rns had to be considened ephemenal-. Nonetheless, patterns which necu::ned in the course of the study, because of thein consistency, ane pned.icted to necur. elsewhene.

In summany, the main conclusions were:

1. Sheep and cattle imposed piosphene symmetny on few of the species encountened in the cou:rse of observations.

2. Sheep and cattle impresses diffened mankedly.

3. The hypothesis, that pneferential selection of species is the main cause of stock-nelated vegetation pattenn, was unable to account fon much of the obsenved pattern.

4. Piosphène s.mmetny was detected intenmittently, becoming

obscured duning or after- peniods of pastune g:rorvth. 131.

5 The obsenved patterns wel?e ephemenal-, giving no sound basis fon p::ediction of diver.gences in the vegetations

unden the two gnazíng negimes.

4-B ANALYSES OF PHOTOPOINT DATA (UTION STUDY)

[-BL The effeets of eattLe ød sheep grazing on pLant habits.

Plates 5, 6, 7 and B, taken at tines 1 and 3 in the cattfe and sheep paddocks nespectively, show cleanly the diffenences in the effect of g:razing on grass species . The cattle, overr this per"iod, acted as tt.Iawn-mowersfr: the visi-bl-e gl?ass standing cr?op was evenly neduced without much change ín density. Sheep on the other hand nemoved smafle:: individuals completely, while leaving entine the

:ranken fonage (pa:rticularly Stípa nítídn individuals). The same pattenn can also be seen in the photopoint sets of appendix 1.

These obsenvations conr-espond with the infenences drawn fr"om negnession analyses. In section 4-4 it was considened that sheep- gnazing most cl-osely affected the density of gnasses, but that cattle-gnazing nemoved a pnopontion of the biomass without causing a commensurate neduction in density. The photogr"aphs suppont th-is.

4-82 Brousing by cattle. Section 1-63 has stated that bnowse is unimpontant in , cattle diets, but may constitute a significant part of sheep diets. Trees in both paddocks displayed a well ma.nked bnowse l-ine maintained by stock, but leaf growth below this br"owse l-ine, within the neach of PLATE 5.

Photopoint 2, cattle paddock, 0.Smi north of Blacl< Hil-l Dam. 4/9/72

PLATE 6.

Photopoint 2, cattle paddock, 0.5mi north of Btack Hill- Dan. Note the even reduction of grasses by comparison with the standing

cnop in plate 5. t3/2/73

PLATE 7.

Photopoint 6, sheep paddoek, 0.6mi west of gate. 4/9/72

PLATE B.

Photopoint 6, sheep paddock, o.6mi west of gate. Note that l-ange individuals of Stipa nítidø (A) are stil-L standing, despite the marked

reduction in amount and numbe:: of grasses by companison with plate 7.

t3/ 2/ 73 .tfi:

la.

:'11l r.¡n. 732. stock, was very limited (see the exclosune photopoint, sheep paddock, in appendix I). Hence the tnee forage was considered to be of little significance in both cases, by compar"ison with the g:round pastune.

However:, an intenesting facet of cattle gnazing, not obsenved

in the sheep paddock, vras the almost total- destnuction by b:rowsing of sevenal- individuafs of EremophiLa duttonií, a shrub species grow-

ing to a maximum height of 3m. Photopoint 4, cattle paddock' included one of these shrubs in the foneground of the camena fiel-d (see appendix l-). This individuat was almobt totally dest::oyed by cattì-e in the period of observations, with foliage nemoved even' duning times of pastu:re abundance. Bnoken branches, but no leaves, wei:e found anound the plant at each obsenvation, hence the cattfe wene br-owsing, and not trampling the bush. In view of the dietany pnefenences of cattle, this was unusual. Furthen'- EremophiLa species in general-, and E. duttoníi and E. sturtii in panticulan, contain in the l-eaves large quantities of tunpenes, which nenden them highì-y unpalatable to any ungulate . @. stuv'tii is commonly

named rltunpentine busht', because of the obvious odoun of the

tur:penes). Gr.azing of such species by cattle was thus unexpeeted, but was observed both in the sample anea, elsewhene in the paddockt

and in othe:r cattl-e paddocks thr:oughout the study pe::iod. Some

saplings (to 3ocm tall-) of both species wer:e found in quad:rats, and

these r^rere canefully obsenved. Thene was no evidence of gnazing.

Hence there are no gnounds fo:: assuming that the g::azing of adufts of these species by cattle wil-I affect the maintenance of the population. 133.

4-83 Shrub mC gyass dgnøníes. Figunes 4.19 to 4.22 inclusive were compiled fnom the photognaphs, using the method of section 3-65. fn exclosunes in both padd.ocks, very little decnease in standing biomass of shrub on gnass was evident between tirnes from the stant of observations to time 3, although an ovenall d.ecnease coul-d be seen. In g:razed photopoints of the sheep paddock, a decline in g]^ass biomass was evident by time 2, and a decline in sh:rub biornass by time 3. The time lag between obsenvation of clean decrease in g:rass biomass and obsenvation of clean decnease in shnub biomass was intenpreted as showing incneased grazing of the shrub l. astnotrieha by sheep, consequent to the majon r^eduetion of mone pneferned grìass species. fn the cattle paddock, decreases in grass biomass wene evident fr"om the star.t of obsenvations to tirne 3, but decneases in shrub biomass wene langely undetectable between tirnes, despite an ovenall- decnease. This was intenpreted as demonstrating the exclusive gnazing of grasses by cattle. The pastune growth between times 3 and 4 was cleanly evident fon aIl- photopoints. Between times 4 and 5, exclosunes showed no visible dec:rease in eithe:: grlass or shrub biomass. In the paddocks' however, some slight decneases wene noted, mostly for gnass biomass.

Between tirnes 5 and 6, both decneases and incneases wene noted fon exclosu::es, but most photogr:aphs showed little difference. In paddocks, the pattenn was confused.

Between times 6 a¡d. 7, majon g:rowth of both grass and shru-b was evident, in exclosunes and in paddocks. At time 7, however, the g::ass langely obscuned the sh:rubs--a rar?e phenomenon which was Fig. 1.19 Fig. t*.20 RELATIVE CHANGES IN APMRENT BIOMASS OF RELATIVE CHANC€S IN APPARENT BIOMASS OF K. ASTROTRICHA IN THE CATTLE PADDOCK K.ASTROTRICHA IN THE SHEEP PADDOCK

eHofoPorN'r OATE PIIOTOPO NI DAfE PHOTOPOI N T OATE PHOTOPOTNI DAIE l:.È È sË F ñ F + È ts o= RÑ + * s o N cì ø o ?E = Blæk H,tt 0om m \ 4 0.3mr 0 .3mr V ts 05mi -\ A -/ þt 0m \ ^ 4 0'6n¡ V ^ a^ ^ ts 0.6m¡ V ^ a^ ^ z æ F 0 '9nt l0n, \ -7 U 5mr ^ U 0'9 mr \ ^ a^ ^ V ^ ^ 14ni \ \ ^ ^ 4 ðmt ^ 4 1- 2mt \ ^ a ^A z m V ^ ^ ^ G É U U r 2 Omr \ \ ^ ^ l'5mr ^ a ts 5m V ^ -/^ ^ I 2 d 9 E 2 5mi 4^ 4 U 1.Emr \ ^ a ^ ì 0m ^ A A^ O5m 4 F ^ dloik H,li Bore U ú 2.1 mt V \ ^ a a^ I Omr 4 z 2'1nt ^ a É a\ U I 5mr 4 7 A F 2'7mt \ ^ ^ U ul É 0 5mi \ 4 4 \ 2 Omr 4 f, I z 3' 0mr \ ^ l't I Cmr \ \ \ -\ ^ ^ ^ ^ o U r ole a\ U | 5mr \ V \ ^ \ ^A \ U (dotum) \ 4 ^ 4 4 2 0T' \ V V \ \ lst woler ^ ^ ^ / tz z U C.5mr G 2 5Ti \ ^ \ ^ \ I V ^ u ^ -7^ È 3 0n, A^ V ^ l¡l 2îd ^ U ViV A \ E ^ z ¡otâ \ \ \ a 4 É 3 5m, \ V .\ o ^ ^ o O) ^ -/ z 4 4 obvrous o¡Mous . Omr V ^ ^ \ rncreose V decreos€ 7 sl.rght slr ght ^ ^ ^' rncreose \ dære 09e

05vo5 úìcrecie V ;gJ:å:. '/^ s,gh.. \ eLght ncr€ose decíw Fig. 4.21 Fig tr.22 RELATIVE CHANGES IN APPARENT BIOMASS RELATIVE CI-IANGES IN APPARENT BIOMASS OF GRASSES IN THE CATTLE PADDOCK OF GRASSES IN THE SHEEP PADDOCK

PHOIOPCINf PIPfOPO NT DAfE PHO'TOPO NI DATE Pt-tof oPotNT OAIE

r: r È.- ñ È)s 3+ I I o I OO I r I ó{ ts { =N { ç ..\ Srock tl,li Dom V \ a Al4 0 5m \ 4 0 3mr V 03m V { { A A rDolm porn0 ui 0. 6rnr 0'6., 3'5nr V \ .0 tn \ ^ V ^ ^ i \ V ^ V A z -/ \ a G i . cùr \ V ^ ^ \ U 5 m \ A^ 0 9¡r \ \ ^ -\ Â ^ 0 9mr V ^ V A ^ U | .2mr 2m¡ t, G ì¿1r V V V ^ A^ $ m A -/ V V ^ a A ^ z 4 \ \ Â Y ^ A ^ U ! E A r J U ^ 1 .5mr ^ F 5m i 2 0¡t -\ \ V \ \ -\ \ \ a A^ Ø V \ A Â o U z 2 .5¡r \ V V ^ V ^ 1 '8 rnr \ V \ A^ a ^ A , 8n V ^ ^ . 05m V V -\ V \ \ z É ^ ttocl l'ill Bore \ -/^ A A^ V t1ul U 2.1mr \ V ^ ^ ^ 0m¡ \ V ^ ^ ^ \ ts \ 4 2.Lmt É1z ^ ^ ^ A^ 5m \ \ V ^ \ ^ I È 4 \ 2.7mr V ^ ^ U 0 5mt \ \ ú2 0m \ V ^ \ ^ ù ^ ì \ 3.0mr a^ -/^ ^ a l l-l 1.0nr ^ ^ ^ o 3 \ \ A { A \ ^ o 4 4 U 1.5nr Gote \ V \ \ \ d (Cclum) ^ 4 ..\ 2'C¡r \ V V \ Â ^ \ z lsl wqttr V ^ V Â ^ 5 ^ z E { U 2'5¡r V ^ I 0 '5m¡ V V -/ ^ \ \ ts ^ ^ { ^ É 4 3 !;¡¡ u¡ o V ^ v A^ G dql.er -\ \ A^ 7 ^ 7 t f \ ^ V ^ z 5¡r o E 3 V ^ \ J ^ ^ o \ o \ V ^ ^ obvrous obvrou s .i .Cmr ^ ^ J V \ I ncíecse V dlrght \ slroh'l / rncreos¿ S ¡.Z,.ut. 134. not anticipated. Observations of shrub biomass wene thenefone difficult, and the changes stated to occu:: nust be viewed with caution. In the cattle paddock, exclosu::es showed littl-e change in grass or shnub biomass in the long intenval- between times 7 and 8.

Howeve::, open photopoints showed a genenal decr:ease in g::ass biomass, though not in shnub biomass. As befone, this was considered the nesul-t of cattle gnazing the glîass nather than the shrubs. Ïhe sheep paddock pnesented a markedty differ:ent pattern. Changes ovel? this per:iod in exclosures were the same as obsenved in the cattfe paddock exclosunes. Howeven, the gr:asses in open photopoints, almost without exception, showed a manked inc::ease in biomass.

Sheep had been :removed fnom the paddoek three months pnior to the taking of final photogr:aphs, hence much of the incnease was due to grass pr:oduction in the absence of domestic gnazing. fnspection of the sheep paddock photognaphs in appendix I will show that the gnass standing cnops of open photopoints were not visibly different fuom the standing er.ops of exclosunes. Hence it appeans that the exclosures had neanly neached peak biomass by tirne 7, but the pad.dock could not attain this peak unden the pressune of stocking.

This suggests a maximum limit on biomass, in contnast to the statement in section 4-33 (q..t. ) ttrat increases in biomass did not appean limited by the biomass attained duning pnevious g::owth. 135.

4-84 Piosphere pattetws.

T?rene was tittle evidence of piosphene pattenn in the distribution of change il pastur:e, as detenmined fnom the photognaphs. Both the negression and Pnincipal Components analyses indicated that the piosphene pattern only accounted fon a sma11 par:t of the va-niance of biomass, hence such a r:esu.l-t was not entinely unexpected. Nonetheless, a patte::n was detected in the no-gnowth peniod between tines O and 3, fon grass biomass j¡ the cattle paddock. Photo- poiats within 2ni of water:, along the nontheast t::ack, showed some decrease in gr"ass biomass between tirnes, fnom tine 0 to tine 3.

Photopoints beyond 2mi showed no appa::ent change except between times 2 and 3. ThÌs was intenpneted as showing that (a) a gradient of grazing pressune existed between times 0 and. 2, extending to 2mi fuom waten; beyond which grazing vlas not evident' and (b) between times 2 and 3, this gnadient extended to the limits of sarnpling' as a nesutt of cattl-e fonaging fi:nthen in a peniod of seve:rely diminished pastune. This pr:ovides independent evidence that the changes in the piosphene pattern of total- biomass discussed in section +-42 (q.v. ) wene indícative of pr:og::essively mone distant gnazing by the cattle, and wene not chance events.

4_9 DTSCUSSION AND CONCLUSIONS

4-9I The reLatiue írnportance of stocl

Few species encountened in the counse of obsenvations displayed stocking effects. These effects were only obsenved 136.

inte:rmittently, and wh-en detected, accounted fon only a smal-f pnopontion of the total vaniation in the ::espective speciesr abundances. Therefore it is concluded that the stocking inffuence duning the study was minor by compar:ison with othen envi:ronmental influences. The above is not necessa:rily true fon the genenal case. The heavy nains dur"ing the study nesulted in abundant pastune g:row-th. Considening the stocking nate, it is not sunpnising that stocking

effects became obscuned in such pe::iods: the amount of pasture

r:emoved by stock would be negligib-te by comparison with the ovenall incr.ease, hence any piosphene patter^n still- extant woufd be

langely hidden by backgnound noise. Howeven, in a dnought, the stocking influence would probabty be the major: facton controlling the dist::ibutions of pastur"e species. This was pantly evidenced by data from the first thnee tr:ips. The negnession analyses were most efficient when applied to this data, hence the stocking in-

fluence was at its peak at these times. Al-I the piosphene patterns observed duning the study wene ephemeral, though usually necunning. This did not necessanily

nesult fuom the seanch for pattenns evol-ving in the shont-te:rm. The emphasis was placed on short-tenm Pattern detection, to avoid asenibing prior" sheep impness to the effects of cattle introduction.

New patterrrs welte expected to anise under. cattle, and it was a¡tic- ipated that the patterns would be maintained. That they wene

epLemenal in nature has an impontant beaning on the use of piosphene pattenn in monitor:ing stocking effects (discussed bel-ow). 137 .

4-92 Linean z,eLationsVrLps uitÍin th-e píosphere.

Linean ne.Iationships between plant panameter:s and distance fnom water welîe sought, following the evidence of such :rel-ationships neponted by Lange (fg6g) and Rogers and Lange (1971). Linean nel-ationships wene found to exist within the sampling limíts, unden both sheep and cattle stocking. Even whene the plotted data suggested non-l-inean nelationships, as in the distrilution of total biomass in the cattl-e paddock, linean functions nonetheless fitted the data with efficiencies equivafent to those of non- .Iinean functions suggested by the data cunve. Linearity of relationships, although ohserved, is difficul-t to nationalise. Pl-ant abundances proportional- to distance fnom waten implies that stocking pnessune is inversely pr:opontionaf to distance fnom water". In tenms of gnazing behavioun, this implies that stock gyaze out from water: at an even ::ate. (to expand: the standing cnop of pastu::e, P, is assumed inversely pnopo::tional- to the grazing pnessune, G. G:razing puessure is defined as pnopontional to the time avail-able for: gr:azing, over the a::ea avail-abl-e, T/A. Stock gnazj-ng out fuom waten at an even nate have avail-able at each unit of time .,2 2 an area lt (r+r- I -fr', whene i is the distance incnement. Expansion of this gives A = d,i2 + 2t,ø i-, on A pnopontionaf to n since d and i a:re constant. thus G is inve:rsely pnopontional to n, and P pnopontional- to r . ) However, the availabl-e evidence nega-r:ding patterns in gnazing behavioun indicates that the cumu.l-ative effect is not likeJ-y to be stnuctuned in this m¿mnel?. A trampling rathen tha¡ a gnazíng effect might therefone be pnoposed as the cause of the lineanity, since stock have to netunn to the water constantly 138 . regandless of grazing pattenns; but even so, pattenns of movement tend to be stnatified j¡ such a way as to r:ende:: such cumulative pattern impr"obable. Fon example, Schmidtts (fg69) wal-kens and non-walke::s would not be expected to impose linear patte:rns in the vegetation th::ough tnampling, because of the hetenogenous movement pattenn.

Nonethel-ess, this study has shown linean nel-ationships to exist in the cattle and sheep piosphenes, theneby pnoviding independent confir"mation of Langets (1969) observations. It nemains to be determined whethen these relationships nepr"esent the true situation, opr instead, app:roximations to more subtl-e functions of stockirg pllessune.

4-93 Possible canses of stoek-induced uegetation pattern,

As shown by the negnession and Pnincipal Components analyses, the species which decreased in abundance with incneasing stocking pnessure were gnasses under: cattle, and guasses, sh:rubs and fonbs unden sheep. These nesults could be expected, fnom the pneferences of the two herbivones, as aI::eady discussed at length. Howeven, such intenpnetation is complicated by the species obsenved to incnease in abundance with stocking pnessunet pantic- uIanly in the sheep case! many of these inc::eases cannot be explained by r.educed competition following the g::azing down of pnefenned species.

Othen potential detenminants of stock-r'elated vegetation pattenn ane distunbance and nut::ient t:ransfer (section I-8). 139.

However:, these factor.s welîe considened- to be of the same onde:: in the two paddocks. If ttrey wene the- detenminantsr anY diffenences in patter.n between paddocks strould have been diffenences in magnitude, nathen than the never?se situations which were found (e.g.

fon Enneøpogon spp.). Thus, neither pnefenential selection, nesulting in a decnease in p::efen::ed species and an attendant incnease in less p:referned species, nor tr:ampling, noll nutnient tnansfen can individually account for the vegetation patter"ns obse::ved. In the finst case' diffenences between paddocks would be expected; but the differences actualÌy found did not necessanily nelate to the animal-s t pr"efe:r- ences fon the species concerned. In the latten two cases' some differences might be expected between paddocks, because of the potential diffenences in stocking levels, but the revensed signs

between paddocks fon piosphere pattern could not be explained by the action of these factons.

* 4-94 On monitoring stock-induced changes.

Many, if not most, of the Austnalian reseanrch neponts on stoCk-induced vegetation tnend, in the nangelands, have been based

on surveys made at a sing]-e time, with pattenns in space being used as the base fon inference of tr.ends in time. Pnime examples ane the studies of Beadle (fg+g), Jessup (fSSf) and Newman and Condon (fgOg ), ir which long-tenm gnazing succession was infenred by comparison of site Suuveys, i^Iithout long-te:rm obse:rvation. Such 140. philosophy is used extensively in the remainder of th-is thesis. Banken and Lange Cfs0s) and Lange CfgZf) discussed th-e need to moniton t::ends by nepeated obsenvation, in such a way that pasture degnadation could be detected, and cornective action taken, befone penmanent damage ensued. The monitoring agent pnoposed was obser:vation of piosphene patterns at each time, using fnfluence Analysis which, of counse, deals simultaneously with a number of species. The national-e was that incipient degnadation could be detected by changes in the extent of piosphene pattenns, befone neaching the stage of damage to the penennial speciest populations.

T?re results of this chapten must cast doubt on the vatidity of such app::oaches, and indeed on the nemainden of the r:esearch r:eponted in this thesis, which is based on similar appnoaches. The piospher-e patterns, and afso the pattern of plant associations, wene found to be ephemeral in nature: piosphere pattern was only evident at ce:rtain tímes, and species intercornefations ehanged mankedly in the course of obse::vations. Hence the inference of long-term tnend which might be d:nawn fi:om obse:rvation of patterns in space at a single time wil-I be verSr g?eatly dependent on the conditions at the time of observation, and may not ind.icate the true long-te::m tnends at al-l .

The ephemenal natune of the pattelrns may be ascribed in pant to the ephemer:al species showing piosphe::e symmetny in thein dis- b:ibutions, but maly of these species ane the very species considered to be advance indicator"s of pasture degnadation. Fon example, the various Bassía species ar?e generally neponted to increase with 747. grazing prôessure, and to demonstr"ate degnadation, by Ratcliffe

(fggO), Crocken and Skewes C1941), P,eadle (fS+A), Jessup (fgSf)'

Moone (rgss¡), Ça:r:rodus et a1 c1964) and Barker and Lange G970). Stipa nítid.a is reponted to decr"ease unde:: stocking Cosbonn et al, 1931), as is Ertneapogo¿ (Beadle, 1948). The nesul-ts described in this chapten strow that such species cannot be reganded as neliabl-e i¡rdicators of th-e leve] of stocking Pressune; the stock-induced patterns appean intermittently, and so detection of the stoeking infl-uence is highJ-y dependent on the time of survey' Thus great difficulty can be anticipated in tb-e inte::pnetation of changes in patteun with- the view to monitoring pastr:re trends'

The pr"oblems in detecting incipient pastune degnadation thus remain unsolved. Once the degrradation becomes appar:ent in ttr.e populations of the long-lived penennial pastune plants ' near- permanent damage has afneady occunned (see section f.5). How- ever, the shorten-Iived species which might be expected to show the eanly puoglress of degradation, befone the damage is serious, must be considered unreliabl-e indicatons because of the inter- mittent manner? in which they display stocking effects'

4-g5 The short-term irnpact of cattle intnoduction'

Thepr:imePurposeofthestudywasthed.etenminationofthe effects of cattl-e introduction on the vegetation of former: sheep 142. pastune, in re.ation to the effects of continued sheep stocking. The individual stocking effects have been discussed in detail- in ttre cou:rse of analyses, and the detailed concfusions can be summal- ized as fol-lows. 1. Ttre stocking effects of the two h-erbivones, particulanly as evidenced by the existence of piosphene symmetnies, diffened with regand to ttr,e species concenned, the nection of the nel-ationships, and the pJ-ant panametens most immediately affected by stocking. 2. Stocking effects we¡e most easily detected in the distnib-

ution of common species, which, howeven, compnised only a small per"centage of the total- ffor"a.

3. A1l- stocking effects obsenved, fon both herbivo::es' wel?e

ephemenal in nature; becoming obscured on disappeaning entineJ-y folJ-owing ne-g::owth of pastu::es. The thind conclusion is pe::haps the nost significant: it implies that the short-term pattenns cannot be construed as indicative of long-tenm divengence between vegetations stocked with sheep and vegetations stocked with cattle. If this wene not so, the short- te::m effects would imply consider.able divergence in pasture comPos- ition, because of the ma¡ifest differ"ences between the herbivonesf effects on the most common,species. The effect of the nains of l_973-74 must be consídered once more. As alneady stated, the stocking effects rnust be, and ane, most pl?onounced dur:ing dr"ought. The natu¡e of the :rainq, and the consequent f::equency of pastune gr:owth, would l-ead to the delay of any J-ong-tenm divengence by mininising the effect of stock. This undenlines the difficulty 143. of mounting any investigation in the arid nangeÌands: the r¡eathe:r in the shont-tenm cannot be pr"edicted with ce::tainty, and the pnedictive value of the ::esu.lts wil-l be highly influenced by the weathen pattenns prevailing duning obsenvations. This was

accepted at the outset of the study, but the nainfall was expected to be eithen absent, or intenmittent, in keeping with past histony. In eithen case, results of gneater pnedictive vafue than those actual]y gained wer:e anticipated. The heavy and continuous rains of 1973-74 were without panatlel in a centuny of neeondings, and the pnedictive vafue of data gained. in such an

atypical period $Ias necessarily low. Nonetheless, on the basis of the data gained fuom the shont-

tenm companison of sheep and cattle stocking, the fina-l- conclusions, on the effeet of cattle introduction to fonmen sheep pasture, must be as fol-l-ows: l-. The int::oduction of cattle to fonmen sheep pastu::es nesults in the formation of shont-term vegetation pattenns, significantly different fi:om those imposed

unCe:: sheep-stocking. 2. There is no evidence that the diffenences in impact - wil-l result in long-ter:m divergence between pastr::res switched to cattle and pastunes sti1I unden sheep' CHAPTER 5

LONG TERM CHANGE IN VEGETATION GRAZED BY CATTLE

5:1 INTRODUCTION

This chapten describes two sunveys, which wene unde¡taken to determÍne the long-te:rm tnend in vegetation grazed by cattle. One sunvey companed differences in species abundances between virgin and cattfe-gnazed areas, and the other rnade the companison betlveen cattl-e- and sheep-gnazed areas. The effects on abundanceS of such factons as soil vaniation were minimised by careful selection of sample aneas. Diffel?ences in abundance were considened to be evidence of long-tenm pasture dÍvengence between stocking negimes. The sunveys were not nepeated. The basic assumptions, fon infenr:ing long-tenm change, were; 1. The status of that pant of the plant community, obsen- vable at a particulan time, would neflect the status of

the whol-e community. 2. The status of the community unde:: vanious grazing negimes could be dinectly infenned by compar"isons of plant abund'-

ances between negimes; pnovided that diffenences

r"esulting fnom disparities in landfonm, soil type, oniginal - vegetation on shont-tenm weathen effect wene finst eliminated. A caneful matching of site landscapes minimised the effect of landfonrn, soil type arrd oniginal vegetation on diffenences in plant abundance between sample areas. Any nesidual dispa:rities in these featunes were taken into consideration, befone stocking effects wene d44- 145. infer:red. Diffellences in weather sample aneas wene minimised, by selection of aneas as close to each other as possible, and by simulta¡eous sampling.

Sampling aneas $rere smal-l (550ha), but the sampling was objective, with high inter:naI neplication (ca l-10 quadnats/sample anea). Species abundances between sample alreas were companed by subjeetive appnaisaf of mapped data. 0n1y those differences

ímmediately obvious vùelre considered funthen. The choice of sites was limited, by (a) the fer^r stations, within the dog fence, which had cannied significant cattle hs:ds for at feast ten yeans, (b) the veny few which- had nestricted cattle to set areas, and (c) ttre nequinement for matched l-and- scapes. Sites fon the vingin/cattle companison we::e sefected on Pennatty station, and sites fon the cattle/sheep comPanison on

Roxby Downs station (locations: fig. f.1); in the sand/clay systems mentioned in chapten 2.

5_2 METHODS

5-2I SeLection of sanpLe aï'ea's. - In selecting aneas with matched landscapes, soils and land- fonms could be dinectly companed, but tþ-e oniginal vegetation had to be infer"ned. Iþee distnibutions were used as the indicatons of the original vegetation. Matune tnees nespond much more slowly to the irnpact of domestic gnazing than do shor"ter-fived pastune species. The exact lifespans of tnee 146. species arte no- known, but tb-ey are over a centu:ry for such species as CaLLttris eolLumelafl:s Clange' l-965) a¡rd Aeaeía aneuï'a. (Pneece, 197Iarb). Hence, although understoneys may have been completely nodified by stocking since settl-ement, the tree stnatum still- pnovides an index of the original vegetation. The::e are some limitations, howeven. Although matune trees are little affected by stock, seedlÍllgs and small saplings may be completely eliminated--see Hal-l et al- (1964). Funthen, neduction of tree populations by p::evious logging may lead to inconnect infenences. The initial match was by gnound neconnaissance, aided by aenial photographs. The validity of the match was checked in detail, in the counse of the sr:lveY.

5-22 Patønetez's used in sønPLing.

Density counts (discussed in section 3-3) wene attempted fon at1 pasture plants encounte:red, dead or a.Iive; although density

was unsatisfactony fon certain species, individuals of which coul-d not be clearly defined. Fon instance, Myt'iocephaLus stuartíi was pnesent in the vingin/cattle survey, but as dnied and fall-en sticks

and seedheads: in this case, incidence only was necorded. grew fr:om undengnound Eragrostís dustt'q.Lasiea, also p:resent ' rhizomes: in this instance, fnequency was estimated, fnom ten t 2x;-;mz quadnats at each sample point. Fnequency has sevene statistical limitations in situations such, as the Pnesent one, where adequate ::eplication is not possÍ}Ie cGnieg-smittr-, 1964).

However", as Goodall- (1970) suggests that it is a valid measurle' if considered in its own night nather than as an index of othen !47 . measunes, frequcncy was employed in this investigation'

5-23 SønpLing system.

20 x L 5m2 quadnats wene used th'oughout. Such quad:rats we::e used successfully fon density counts in similar situations by

Banken and Lange (1970). The me:rits of ttre shape have alneady been discussed (section 3-3). The edge effects noted in that section wene not appanent in this case. The same in/out rules we:re used. In each survey , II2 quadrats wene laid in a rectangulan gTid' dimensions l-8oo x 32OOm2 Czooo x 35OO yd2)' Dams wer:e the datum points, set on on nea:: a shont side of the grid' As with the shont-term study of the pnevious chaptens, this anea of greatest tnaffic was expected to show changes wrought by stock most clea:rly' (f90g), Rogens Regulan nadial sampling, as used by Barken and Lange

and Lange (rgzr), and :ra¡dom sampling wel?e considened, but nejected on the following g::ounds. 1. flith mapping as a main aim, an onthogonal grid was the most efficient system, since it gave a consistent scale oven the whole area (Vlítliams, 1971)' 2.Radialsarnplingcentned.onwaterswouldhaveovensanrpled .clayflatsinwhich-watenswe:resited,nelativetosand

areas of equal concern' S.Onlyovertdiffe::encesinabundanceandpatternwel?e soughtrandstatisticaltneatmentsnequiringestimates of pnecision we::e not contemplated'

The sampling gnid was staggered; as the dune nidges had a more oll 148. less ::egular" pattenn, it was necessany to avoid bias which migtr-t nesuLt fnom a negular, nepetitive soil patte::n. The gnids ane show-n in figure 5 .1.

5-24 Data presentation ø¿d anaLysis.

Density data was displayed in map form. Comparisons welre dnawn on the basis of obvious differences in occurnences and/on abundances between sample areas. The maps ane pnesented in appendix 2. As an adjunct to the density maps, associations based on species

incidence welle cal-cufated, and are Plîesented as constelfation

diagnams.

5-25 Criteria for inferr'íng Long-tertn ehnnge.

The vi:rgin/cattle sunvey, on Per:natty station, was nun in dr:ought, but the cattle/sheep compa:risonron Roxby Downs station,

was made in a peniod of gnowth, following nains in fate summer and autumn, 1973 (ta¡te 5.1). The method of inference differed

accondingJ-y between surveys. In the first case' a diffe:rence in

abundance of a penennial speeies between sample alleas was considered evidence fon a long term change. However, lower abundances of an

ephemer"al species in the gnazed anea may have :resulted fnorn short-

te::m gnazing since last growth, or fi:om a long-tenm dec::ease in the population under gnazing. No inference was dnawn, since the

two possibilities cou.l-d not be distinguished within the context of the sunvey. On the oth-en trand, greater abundance of an ephemenaÌ i¡ the grazed ¿Lrea, irnespective of shont-terrn gr:azing effectst was consider:ed to indicate a long-tenm j¡rcnease in the PoPr:fation Fí9. 5..| SAMPL!ft{G GREDS FOR VIRGIN/CATTLE AND CATTLE/SHEEP SURVEYS

VIRGIN/CATTLE SURVEY (Pernotty) /,

VRGIN AREA Emu Blulf Dom

Avenue A Dqm

CATTLE AREA

White Dqm

CATTLE/SHEEP SURVEY N (Roxby Downs)

CATTLE AREA Todd's Dom

SHEEP AREA

Sundoy Hole Oom ¡ 500m TABLE 5.1 1+9.

RAINFALL IN THE PERNATTY-ROXBY DOI{NS DISTRICT

(a) Rai¡fall in mm at Pennatty and Roxby Downs for the six months pneceding sunveys.

Pennatty Ro>

1971 Nov 4t 1972 Nov 4

Dec 8 Dec 0

1972 Jan 73 1973 Jan 12

Feb 6 Feb 97

Mar: 0 Mar 8

APn 0 APn 16

(b) Rainfall in the three months pneceding the for aneas in the vicinity of the Roxby Downs survey

Roxby Pennatty Andamooka lloomena Downs

1973 Feb 97 ]-2l- 99 l-2I

Man I 12 11 l3

Apn 16 18 41 I

Data cour:tesy of Buneau of Meteonology, Adelaide 150.

unden grazing.

The cattle/sheep companison had a sirnplen basis. Ì{itlrout the benchrnark of a vilgin site, speeies abundances w.ere expressed only

in tenms of r:elative diffenences between sheep and cattle ar?eas. All differences wene tneated as long-term nather. than short-te::m in

origin, since tlre survey was made in a gnowth period. chapten 4 pnovides adequate evidence that shor t-tenm gnazing effects a:re obscuned in such tirnes. Majon differences between matched pastu¡es hence may be considened long-tenm, pr"ovided that both pastu:res have neceived equival-ent nain. The nains immediately pneceding the cattle/sheep survey r{ene widespnead and consistent (ta¡te 5.f), the station managen r"eported that both areas had received equivalent amounts, and there were no obvious differences in the stage of gnowth between areas (plates g, j.0 ).

5-3 RESULTS: VIRGIN/CATTLE COMPARISON.

5-3I Stockíng histotg of sønpLe dr1ea.s.

The sample alleas , in cattle-g:razed and in vingin pastune, r^/erle located as shown in figur"e 5.2. The oniginal- Pennatty homestead was established in l-866, nean the gnazed anea sampled. Log fences wene built shontly aften, containing both the frood flat of per:natty cneek and the surnoundi¡g sandhills. Sheep wene the main domestic stock, but a smal-l- catt]-e hend (ca 20 head) was maintained. Seve::e pasture.degnadation occur:red, particuranly in the fir.st two decades of this century. Erosion hras aggnavated by peniodic fl-ooding.

Dykes wene buil-t acnoss the cr"eek flats in 1g26, to ressen fl_ood PLATE 9.

Nean Toddrs Dam, Roxby Downs station. New growth of ephemenal species forming a spallse gnor:nd cover. Tnees ane ACAciA' aneLna (for:egnound) and CaLLitrLs coLwneLLarLe (extrerne l-eft)' t7/5/73

PLATE 10. ephemeral Nean sunday lIole Dam, Roxby Downs station. New g:rowth of species, largely grasses, fonmíng a sparse gnound covell' Tnees a:re Heterodenárum oLeaefoLiwn. fl/5/73 .*${''. i:+1, ' , , \*rl I ---¡+

:{¡q-.-, ¡ . F'9. 5. 2

LOCATION OF SURVEYS ON PERNATTY AND ROXBY DOWNS STATONS

(b) (o) PERNATTY (VIRGIN/CATTLE) SURVEY ROXSY OOIINS (CATTLETSHEEP) suRvEy

+s . 0 CAlILE Cd60l@ 0om

Em L, TORRENS 0om PLAT EAU

+c tê S*+ c

Dqn C Oom C Dom Ç , S

c PERNÂlfYar.. o H5 r , , I ¡ À _..^i Nøth Gos . - >\ N O@ ,'+ n--"'S Augustq .Ayo +5 t^ - LdkC \r Molcnor Rrdqq -.'- t,------

r ROIBY i_ xsOOIVNS tt, ,.. + Woterpoint (S lø shecp. C for catttc, ?r---" t Homesteods or lulg lo Woom6o ---- Ma,it roods ----- Olhtr trocks trqrr,.dliu, +---+-Dog tmc" lOtm 151 .

damage, and have been maintairied since. These hal-ted t1r-e waten

enosion, and the pastune of the fl-ats negenenatec to some extent.

The cattle her:d was incneased gnadually, fnom 50 head in l_930 to

200 head by 1950. Her.d size has since fr-uctuated between 2oo and

500 head, with the mode at 300 head. By 1950, the pe:rnatty creek and the su::nounding dunes had been turned over entir:ery to cattle,

as a semi-open range. The nangets limits r^/er1e, in rg72, yertacowie outstation and Bal-dockts dam. Cattl-e tended to move fnom water to

watern, with- camps in sandhilJ-s up to ZóOOr fi:om the creek: maximum

penetnation into th-e sandhil-.J-s was ca 30OOn. Whil_e sheep wene

nesponsibl-e fon the initial inpact and nesultant change in veget-

ation, cattle wene expected to have imposed patterns of thei:: o¡m dur:ing the twenty years of continuous and exclusive gnazing.

The El-izabeth C::eek area was al-so l-eased. in the l-860 t s , but sho:rtage of waten pnevented its utilisation. The nean-vingin pant sampled may have been g:razed, by sheep, for a shont per"iod in the

late 19th centu::y, but thene nemains no evidence of damage. Emu

Bluff dam, used as the sampling g:rid datum point, r^ras sunk in

1958; but has only held waten íntermittently. Fr"om l-960 to

1972, stock (pnobably cattle, but no certain:recond) were nun fnom

it, fon shont periods totalÌing seven months at the most. Pnion to 1958, the alrea was ungnazed in living memol?y. The soil lichen

cnust was intact, and shrubs and tnees bor:e no evidence of bnowsing.

5-32 Descníption of natched Lætdscapes.

Figune 5.3 shows the soil distributions fon both vingin and cattle*gnazed aneas. None of the rnapping units a:re completely Fig.5.3 SOIL LAh¡DSCAPES IN THE VIRGIN/CATTLE SURVEY

CATTLE-GRAZED SAMPLE AREA

+ \ \ I s \ I I I \ I I I \ ì 0 I I I 5 1 I t/ \ \¡ I \ i i'\ \ I I \ t, I j\. I t ¡\ I ¡ q\ \ I '! \ I \ ,0

9 a--

^/ + H 250m

VIRGIN SAMPLE AREA

N () I * ô / @

to< 9 \

\ 5 \ 7 s\ I I ..s I I 0 s / *.-I r- _.-./ \ \q

250m

Dom + Sonds

Soil sompling cores 6 \'-...- Stoney loblelond Dune rrdges Cloys Creek flots Ctoypon G 752,

homogenous: maj¡r units contained smal-l pockets of othen types.

The pnincipal pnofile forms (p.p.fts; Nonttrcote, 1971_) fon the

mapping t¡¡rits, as obsenved fnom augeï, cores, were;

1. UeI.22 for sands--coarse ned. sand.s with some accumulation

of onganic matter in the sur:face lOcm, but no othen pedologic onganisation; not cohenent;

2. D::r.33 fon crays--duprex soirs, with a loam over"laying heavy clay, A ho::izon with a sunface crust, a sporadie_

aIly bJ-eached A, honizon, inci:easing arkal-inity down the

pnofile, calcaneous in the B honizon, with gypsum

incl-usions appeaning ; 3- Drl-.13 fon stoney tableland soil-s--as fon D:rr.33 excepting the absence of an A, honizon and gypsum incfusions, but with sunface pebbles and stones in the pr"ofi1e.

This concuns with the mapping in the Atlas of Austr.al-ian soil-s

(1960). l¡trhil-e the p.p.Frs were the sarne fon both surveys, thene

existed detailed diffenences. The A ho:rizon in the cattl_e sample was much deepen than in the vingin sample, possibly due to silt deposition behind dykes. Both ho::izons in the cattl-e sample contained a lowen pnopor:tion of soluble salts than was found in the vingin sample. Diffenences ane summanized, in tabl_e 5.2. No such diffenences were found between areas fo:: stoney tableland soils or: sands.

Both sample ar:eas embr"aced equivalent topognaphies, vj.z. , areas of stabilised sand nidges, as manked on figune 5.3, extensive clay flats containing rnajo:: dr.ainage channers, and some nises of stoney tableland associated with the neighbou:ring Lake Tonnens and. Arcoona plateaux. 153.

TABLE 5.2

SALT CONTENT AND HORIZON DEPTH OP CLAY FLAT SOILS.

Fnom five cores in each sample a.rea, nanked i¡ orden.

% sol¡bl-e salts (t:5 soil;waten) Depth of A honizon in cm. A tror. B hon.

vingin cattle virgin cattle vr_l?grn cattle

3.0 I.7 4.5 2.O 10 100

1.4 0.O2 3.s 0.8 10 60

o.7 O.O2' 3.0 0.4 7 40

o.2 0.02 2.s o.2s 5 25

o. 05 0.o2 1.1 0. 25 5 25

!È 154.

Figu::e 5.4 çhows the tnee distribution in each sunvey. As

with th.e soils, th-ere was little diJfer.ence between the two sample aneas fon tr"ees of th-e sands and stoney tablelands. The clay flats in the cattl-e area bone more tnees, in par:ticular EucaLyptus

cønaLduLensis, pnobably as a result of the lowen salinity by comparison with the clay fl-ats in the vingin sunvey. The match in l-andscape was thus not complete. Nonetheless,

it was adequate fon compa::ison between pastures of the sand and stoney tabl-el-and soifs, but comparisons between pastr:res of the clay flats had to al-low for soil differences.

Quadnats laid on the soil- types, fon vir"gin and cattle-grazed

aneas respectivel-yr 67 and 73 on sands,37 and 48 on clays, and I and 11 on stoney tableland soÍl-s.

5-33 tlbtrtdanee of pasttne species.

The distnibutions of 39 species othen than trees o:r tall shrubs

wene mapped. Table 5.3 summanizes the comparisons of abundance 3 between sunveys.:'s Abundances r^rene companed fon the particulan soil

types; since most species wene nestricted to one on othen type, ovenall compa:risons between sample areas would have been biased by the diffenences in pnopontion of quadnats on the types. Ten species we::e found r^rith no overt diffenences in abundance between sample areas,

l-4 wene less abundant in the cattl-e ar?ea (including one completely

Maps ane p::esented in appendix 2. Fig.5.4 TREE DISTRIBUTIONS IN THE VIRGIN/CATTLE SURVEY

CATTLE-GRAZED SAMPLE AREA

\¡ a a I ¡ Y v ¡ ¡'\ ¡ . lY j:1.{ ¡ ¡ '¡ ¡ ¡, ¡ a I

a\ _¡-ôA_ ¡ô \J Â -a-- -a--a------A a Õ ::,::: '<:S ,.:li'.1 - .-o-'- -r-- _

250m

VIRGIN SAMPLE AREA

¡ c cc c c Âc^' c c /r ¡ '¡\c t ¡' ,la c'. c a t I ):-L c L ¡' À ¡\- Jlt i ^' ¡ \x L

¡ ¡.r t¡ \o .tÁ¡ ìi.l '¡ I ' ¡ I t ¡ \. v\ ,L

¡ a t

I \¡ \l I a J

250m Acocio oneuro Heterodendrum oleifol¡um a Dyke

Acocio l¡nophyllo x Colt¡tns columellqils /l Orornoge lrne

Acocio sowdenrr Prtlosporum t Swomp

Cosuonno crislolo c Myoporum plotycorpum m Acocro lrguloto - Dodoneo vtscoso Acocro vrctorroe A Euc comoldulensrs a 155 lAnLr 5,3

VIP,CIN,/CAI'TLI: ColtPAF.ISollr SPllclnS A¡U:l¡Allcl;s IIJ ltll CR^ZED ^Rf,A R¡LATIVE TO ABU}¡¡ATJCLS I¡I TIif VIRTI]}i AF-I:A,

Species Perenni¡l Yain soil Abr¡ndance I OP tyPe ln grazed Ephercral a!ea

Atriplet apøgiosa E C Iays +

Bacsia bmchgptem I] Clavs t

---- díaoøtha E Clays +

---- ,lívayicat¿ tì CIays +

---- püq7om I Clays +

Enneqogq aùilacew f Llb i quitous +

EragrcatíB aßtm Laaica P clays +

---- diel¿íí E Clays t

Plagioeetw refrætm Sands + o1 Sieynùríø sp. E Sands +

Itagw utruIiøua E CIays +

Xotthin spinoaw E Clays +

Aríatlda ffitorta Sands

Atrtpbz Lítribata E Clays and silds ---- tseaÌcø"ía P Clays

Czo tal æia dí e e'i. tí fto m P Sands lbohia aatpotricha P CIay6

---- georve¿ P Clays

Ilyriooeplnlre etuøtíi E Sande

IJitrcria BoWañ P Clays

Pottulaø oleracea E clays and sands

Slda vlngata P sandõ

SaLaoLa kalí E Clays and sands

Paelryæmía tauíe P Clays

Abutlln otocú?ø E Sand s (+)

Aircot qøiÞifiàm P Clays (-)

Phyatlathw fwruohríí P Clays (-) Baeaín obl,íqvirepía E Stoney t/land clays 0

Drachyaarrc eíLíæíe E Yb iquitous o

CitnLLüB tøa¿us E sands o Eqlnrbím ù¡zrnnãíi I Sands o

Kæhìa pywíÅota P ClaYs o

---- tüfr¿osa P Clays 0 CarpobrcLß eqrílacemLe P Clays 0

Pøicm effusø f, Ubiqultous 0

Î:titqåís íollía E S¿nds 0 zygochlæ paadoæ Sañds o

a Abundance ratlngs: + nore äbwdånt in grazed relðtive to vlrgln. - Ìess àbundãnt ín grazed nelative to vlrgln. ( ) difference is mrginal. o no diffe¡ence betveen ¡reas. 156. absent), and 13 species were more abundant in the catt.l-e anea

(including two not pr.esent in th-e vingin aï-ea). Two species showed an appal?ent shiFt Ín pr:efenr"ed soil type Þtween sample alleas.

0f the species l-ess abundant in the gnazed anea, five wene ephemenals, fon which no infenences of long-te:rm decnease could be drawn. The nemai¡rden wene perennial-s, two of which showed onJ-y ma:rginal differences, and se wene not consids:ed. Of the nest,

K. georgei was completely absent fiom the cattle-g:razed ar.ea;

Nitraría schobeni, Atz'ipLeæ uesica?ia anð. ct,otalaria dissitifLora were rlar?e by compa:rison with the vir:gin ar?ea; Paehycoz,nia tenuís was absent fnom the sands of the cattle area; and both Kochia astz'otz'icha and sidn uirgata lnad incidences simil-an between areas, but lowen densities in the cattre anea. N. schobez+ is a characten plant of saline loams on clays (Specht, Ig72), and its lowen abundance was ascnj-bed to the lowen salinity of such soil-s in the cattl_e anea. Such an a:rgument did not apply to the othen elay-soil species . K. astz,otri.cha, K. georgei and .4. uesicaria tol-e::ate a rlange of salinities gneaten than that found in the sulrvey

(Jessup, 1951), and a:re usual-Iy J-ow in abundance in Nitz,aria-dominated vegetations in any case (Specht' 1972, App.V). Non was the sl^Iamp present in the cattl-e area extensive enough to account for their lower abundances. C. dissitifLora and S. uirgata grew on sands, witlr- no edaphi.c diffenences between sample at?eas . P. tentris is usually found on saline l-oams and clays (Jessupr 195f), and its occunnence on sands in the virgin anea was unusual-. However: , its absence fnom sands in th-e cattle area, when pnesent i¡ ttre othen, !57 . also vras not a nesult of edaph-ic differences but :rathen of stocking. Thusr' ttre l-essen a.bundances of ./.. Þesicanta, C, diss[tíflora,

K: geoz,geí, K, astrotz,[c[La, P. tenuís and ,9. uùrgata in the gnazeò, area, by compa:rison with the vingin area, r^rene also considened í¡dicative of a long-tenm decr:ease in thein populations, nesulting fuom stocking.

Of the thinteen species with a greate:: abundance in the cattl-e ar?ea, onlry Ez,agrostis austnalasiea was pe::ennial_. A chanacte:: plant of swamps (specht , rg72), its abundance was pnobabty due to ephemeral swamps fonmed by the dyke within the sample area.

The ephene::als AttípLeæ spongíosa and Inagus austraLiøtus were positively conrelated with it. These three species, and Abul;ilon otocarpwn, only ma::gina1J-y mol?e abundant, hrere not consider.ed funther, Bassia panadoæa and the alien invaden Xanthium spinosun were widesp::ead in the cattle-gnazed anea, but completely absent fnom the vir"gin area; Bassia bz,achyptera, B. diaeantha, B. diuar- ieata, Enneaptogon a'Denaceus, and Ez.agrostis dieLsíi showed greaten a-bundances in the cattle al?ea. These species were more abundant, not only on clay fl-ats whene the cause may have been the soil diffenences, but also on clay pockets isolated fr"om the f1ats, and on' sands. PLagiosetum nefractwn and of Sisymbnium sp., nestricted to sands, welre more fnequent and mone abundant in the cattle anea, despite the sinil-anity of the sands between sample aneas. Thus, the gneater abundances of Bassí,a bnachyptera, B. diacontha, B. diuay- icata, B. pa.r,adoæa, Enneapogon qlenqceus, Enagrostis dieLsií, PLagiosetum refractum, of Sisyrnbriwn sp., and Xanthíum spínoswn in 158. the grazed ar"ea, by cornpar"ison with th-e virgin area, were consido:ed indicative of l-ong-term decneases in th-e speciest populations, nesulting fuom stocking.

AtnípLeæ inflata and Babbagía aeroptera sTrcwed appanent diffenences between sampre areas in soil habitat. rn.the vírgin area, the former was found on alI.soils, with maximum densities on sands; and the l-attenrwhile g:rowing nainly on clay soirs, was also found on sands. rn the cattl-e a:rea, .4. inflata was less fnequent and fan less abundant on sands by compa::ison with clay soiJ_s, and B. aeroptera was not found on sands at all-. Possible explanations are that (a) these species wene mone abundant on clays than on sands in the cattle at?ea, by compa:rison with the vingin anea, because of the less saline and mor:e loamy soils; and (b ) their absence fi:om the sands was a nesult of sho:rt-term gr"azing. Funther investjø- ation was not made. No inferences wene dnawn neganding the long ter"m effect of stock on these species.

5-34 .Assocíatíon of pasture species.

An association analysis fon each treatment divided the species into two main gnoups, positively associated within groups, and negatively associated between groups. The gnoups col?nesponded to sand habitat and clay soil habitat respectivefy (fig. 5.5).

Main g:roups in the cattl-e pasture wene not as cÌearly defined as those in the vir"gin pastune: ttris may have been due to both gnazing and soil diffenences. FIGUR.E 5. 5

PLANT ASSOCIATIONS IN THE VIRGIT'i/CNTTIN SURVEY

KEY

1. AtnipLeæ Límbata 23, DactyLocteniwn YadttLans 2. A. infLata 24. AtrLpLeæ sPongiosa 3. A. oesicaria 25.Ábchía tomentosa 4. Pachyeov'nía tenuis 2 6. PaLgì,osetum refractum 6. Bassia decuY'Yens 27. cf Sisymbtiun B. Babbagia acroPtera 28. Dod.orøea uiscosd. g. EupLnrbia drwwnon'dií 29. Bassia brachYPtera 10. Sida, petroPhíLa 30. EnchyLaena tomentosa LL. PLryaLLanthus fuernrohz'Li 3L. PortuLaca oLeracea 72. Brachyscome ciLiaris 33. Panieun effuswn L3, Aizoon quaÅrífídun 34 Bassia diaeantha L4. SaLsoLa kaLi 35. B. paradoæa 16. MyriaeePhaLus stuartií 36. B. &Luaricata L9. Eragrostis dieLsií 38. EnneaPogon auenaeeua 20, At'istídn, contorta 39. Eragrostis austY'aLasica 27. Iriraphís moLLis 40. SoLarwm sP. + 22. Iragus austtaLiaruls 52. Tragus austraLianus * Dacty L o ct eníum r aduL ott's

* Confused identification in certain quadrats ' -i l-_t I

9 (q) VIRGIN AREA

2 N¡@ -1. , I I -t-¡'- -

I t 't I I I

(b) CATTLE-AREA

0l )p).00f '001)p Positive Negot¡ve

ASSOCIATIONS 159.

5-4 DISCUSSION: VIRGIN/CATTLE COMPARISON

The decrease of the per.ennials ,4. uesicaxía, C. dissítífLona,

K. astrotricha, K. geoz,gei and P. 1;enuís was a like1y nesuft of the heavy sheep stocking and subsequent enosion pnion to the intnoduction of cattle, as discussed in chapten 1. Nonetheless, the peniod under. cattl-e was conside::ed adequate for: significant negeneration of at least some of these populations. (Fon example, negenenation of Atz,ipLeæ species on the Koonamone Vegetation

Reser:ve was weff advanced aften an equivalent peniod, despite sever:e inte::vening dnought--Hal-l- et al-, (fgO+). Seed sources fon the l-atten three species wene found within the sample anea, and sources fon the former two were pnesent upstneam, along Per:natty C::eek. The peniodic flooding of this watercounse must have bnought seeds into the sample ar^ea at several inter"val-s after the int:roduction of cattle. At the Ieast, the cattle have maintained the oniginal sheep impness on these penennials by suppnessing regenenation, if not neducing the surviving populations fur"ther. For the ephemeral species encountened in the Pe::natty survey' seve::al genenations would have elapsed since 1950. Chapten 4 indicated that short-te:rm stocking effects on such species ane langely obscuned in gror+th peniods. A series of ephemenal growth peniods in the absence of grazing must obscr.re, if not oblitenate, the longen-term stocking effects. Populations of the grass Stípa nitida, on the Koonamo::e Vegetation Resenve, showed mar"ked differences in abundance between populations within the ungnazed Resenve and neighbouning gnazed paddocks, although the Reserve had 160.

only been in existence fon slx year:s (Osbonn et alr. 1931--the autlrors considened Stípa to be perennial, but the fluctuations in its populations , ::epo:rted Ìn ctr-apten 4, show it to be ephemenal ) . Sfunilanly, a senies of ephemer.af g:rowth peniods, following a change in gnazíng negime, must obscu¡e the effects of the eanlien, and show most clea:rly the effects of the faten negime. Hence, the infenned incneases in epheme:ral populations were regarded as due only to cattle. The gener:al tr-ends Ímposed unden cattfe gnazing, as indicated Ey th-e nesul-ts of the compar.ison, alle:

Ì. The negene:ration of some perennial pastu:re species is

suppnessed; 2. Incneases in abundance, due solely to cattle gnazing,

are found only for ephemer"als.

This is the same pattenn ::eponted to develop unden sheep grazing

Gection 1-52). The particuLat species involved al?e neporlted as neacting in the same mannen to sheep grazing: A. uesicaria,

K, astrotric\ta and K. georgei ane repolrted as decr.easer:s (Ratc1iffe,

1936; C::ocker and Skewes, J-941; Beadle, 1948; Jessup, 1951; Canrodus et aI, 1965), and Bassía (afl- the above species),

Babbagia acroptera., and the invader X, spÌ,noswn ane :reponted. as incneasens under sheep stocking (above ::eferences; and Leigh and Multram, J-965). The only conflict between the pnesent infe:rned tnends unden cattle and the r eponted tnends unden sheep is the i¡cnease of Enneapogon auenaeeL:s, a species decneasing unden sneep (Beadle, 1948). The increase also conflicts with other reponts on the effect of cattl-e, Chapten 4 indicates that the 161 . species decr:eases with incneasing cattle stocking; and Low

(fgZe) has shown its impontance in cattte diet in the Alice Spnings anea. Hence, under- pnoÌonged cattfe grazing, the species could be expected to decnease r"athen than to incr:ease. Pnefenential g:razing selection of plant species is often consider"ed the majon cause of stock-induced vegetation pattern

(section 1-8), resulting in a dectine of pnefenr-ed species, and an incnease of less prefenr^ed species due to lowened competition.

Cattle pnefenences show some similanities in genenal, but manked diffenences in detail, by comparison with sheep prefenences (section 1-63). Hence, tb-ere should a::ise manked diffenences in the effects of the two her"bivores. In the pnesent case, the infenned long-tenm effects of cattle an:e closely similar to the nepor.ted long-tenm effects of sheep. If the animalsr pr"efenences ane the main causes of pattenn, this is a most unlikely resu.l-t. Altennative determinants ane dístunbance and nutnient transpont.

The latten plays a minon ro1e, because of the 1ow stocking nates: excepting perhaps in the immediate vicinity of the waten (section

1-B), the trsacr:ifice zoner'. Mechanical distu,r"bance of soils and plants is pnoposed here, as the majon cause of long-tenm vegetation patter.n and change nelated to stocking. ff this is conr:ect, most of the pattenn shoul-d be independent of dietar"y prefer"ences of the panticula:: herbivor.e. This can expJ-ain the close simil-anity between the obse::ved cattle effects and the repo:rted sheep effects, despite the animaÌsr different diets.

The special situation at Roxby Downs was selected to test this 162. hypothesis in part. The comparison was drawn between cattle- and sheep-grîazed aneas, with equivalent stocking (in the managerfs estimation). The pnedicted nesult was no diffenence in vegetation between alreas, attributable to diffe::entiat dietany selection of sheep and cattfe.

5-5 RESULTS: CATTLE/SHEEP COMPARISON

5-5I Stockíng historA of sønple ateas.

Ihe sample areas, in cattle-grazed and in sheep-gnazed pastunes, were focated as shown in figur:e 5.2. The nonthernmost part of the present Roxby Downs station, including the cattl-e- grazed sample allea, was stocked fi:om 1B7O to ca l-930 unden a sepanate l-ease. Sheep h¡eue run inÍtially, with a gradual tnans- ition to cattle. By 1945, the nonthennmost paddocks were g::azed solely by cattle. Degradation of pastu:res fol-lowed settfement' and continued at l-east until- 196I. At this time, the lease came unde¡. the p::esent management, and pasture negenellation was attemp-

ted (B.G. Lay, pens. comm. rf' has infonmed me that the penennial pastures in 1971 wene stil-l in wonse condition than in l-949).

From 1945 to 1973, cattle numbens va:ried between l-00 and 700 head'

with above 2OO head in most yea:rs. The highest continuous level

was 5OO head, between 1945 and 1951. Fnon 1966 to 1973, numbers had incneased to 600 head. As on Pennatty, cattle vlene nun on

:"' M.Sa. Thesis, Botany Dept., Univensity of Adel-aide Q972). Lay nepeated Jessupts (rgsr) assessment of pastu:re conditi-on on this station. 163.

semi-open nange, with no maintenance of the oniginal subdivis-

ionar sheep fencing. Howeve:r, thene was l-ess movement between

wate::sr pnobably because of the widen spacing. The managen

neponted that hends tended to g:raze fnom a panticulan r^rater. The datum point fon the cattle-gnazed sampl_e ar"eas was Toddrs d.am,

sunk in l-945. The piosphere centned on this dam was induced

solely by cattle, with no p::ion sheep impness, as the nea:rest watens befo::e 1945 were too distant fon sheep to graze the anea.

The southenly par"t of the station, including the sheep-gnazed sample alrea, was gl:azed by sheep onty, fr"om 1g7O. pastune degradation occunned, as in the nonthennmost parts. Fr^om l_946 to 1973, sheep numbens varied between grooo and 25rooo head., with the mode at 15rooo. Highest continuous levels wene 23-24rooo head between 1953 and t-957. Nunbers had declined from 22rooo in

1969 to 1Ì,000 in 1973, to permit majon expansion of the cattle hend.

The datum point fon the sheep-grazed sample area was sunday Hore d.am, sunk l-920 as a smal-l catch, enlar"ged in l-945 and 1961. Stock numbens on this dam p::ion to l-96f wene not known, but the pnesent management has nun 600 head, with littl_e va::iation, since 1g6-ì_, and considened this equíval-ent to thein cattl-e-stocking nates.

5-52 Descz,iption of matched Landscøpes.

Figr:re 5.6 shows the soil distnibutions fon bo-th sample areas. Mapping units al?e more homogenous than in the vingin/cattle companison. The P.P.Fts fon the mapping units, as obsenved fnom Fig. 5.6 SOIL LANÐSCAPES IN THE CATTLE / SHEEF SURVEY

CATTLE-GRAZED SAMPLE AREA

I I 2 I -( I / ,L I /i ! I , I I /lt/: + i I / I N L 4 I I 4' I 6 I \ tir' / \ t I I L_ í I I \ l I /1 I I I I I I I \ I t/ tl / I I I I it I I I ¡t

250m

SHEEP-GRAZED SAMPLE AREA

ls I I I 2 I / I 6 1 I \) I I I I l / I L I I I I 3 /\ + I \ / 'ì I I Ir I I tr, / I I I

250m Dom + Sonds

Soil sompling cores J Cloys 0une ridges \'--.__ Fence t-¡-r clqypon Fig. 5.7 TREE DISTRIBUTIONIS IN THE CATTLE /SHEEP SURVEY

CATTLE-GRAZED SAMPLE AREA

a. ¡ ¡v. ¡ ù t I .rl l¡ ¡ Í^ ¡l rJ¡ ¡Y h ¡ t ¡ Y¡ ly ¡ h ¡ v:l:: -l ¡ ¡ ¡ I t I ;:"'+ att ¡ ¡ ¡ rJ lo ¡ t I .I¡v t ¡ l ¡ I ijvr ¡Y Y¡Y ¡ ¡ O ::1Y I ¡.. ¡ Ljjj I iv' I ¡ ¡ I ^¡ I Y¡ ¡ ¡ I t .1a: ¡o ¡ ô ¡ ¡ h c ¡ ¡¡ Y¡ Y¡ _t ¡ "¡ålo v ¡¡¡ :i. ii o il I l lY ¡ ¡ I ¿ ¡ it {,:.1. ¡ ::::j: I t^^ ¡ AY a t ¡ ¡ ¡ ¡ i:.t ^ ¡ t ¡ t t ,nY \ ¡ ¡ I ;'r'i' ,:Y.i' ,.a a m m m ¡ ^ 250m

SHEËP-GRAZED SAMPLE AREA

F-----{ 250m

Heterodendrum oterlollum a Cossro nemophrlo c

Cottitr¡s columelloris À Hokeo leucoptero h

Mulgo Acocro lrgulato- Oodorieo vr>Los(J

Acocio sowdenil fn Fence x--f-x 164.

augen cones, were: . 1. Uc1.22 for sands (descr^ibed in section 5-32); 2. Gn3.12 fon clay soits--gnadational- soil-s with a sandy loam Aa honizon, no Arhorízon; a ned clay B honizon, calca:reous with a smooth-ped fabnic. The P.P.Fts wene the same fon both aneas, and no differences in othen pa:rametens \¡relre found (ta¡te 5.4 c.f. table 5.2).

TABLE 5.4

SALT CONTENT AND HOR]ZON DEPTH OF CLAY SOILS From foun cotles in each sample area, nanked in onden

% solubte salts (1:5 soit:waten) Depth of A honizon in cm.

A hon. B ho::. cattl-e sheep cattl-e sheep cattle sheep

.1 1.0 .5 1.0 40 45

.1 .06 .4 1.0 40 40

.05 .03 .4 .1 35 20

.04 .02 .2 .1 30 20

Both a::eas had simil_ar topogr:aphies. The dominati_ng land- fonms wer:e low sandnidges, as befone, interspersed with clay subsoiled fl-ats. No rnajon dnainage channel-s r¡rene pnesent, but instead thene .v¡ere small-, ill-defined channer-s dnaining into claypans within the sandrÍdges.

Figur.e 5.7 shows the tr.ee distnibutions fon each a:rea. Aeaeia aneut'a. and Á. LinophyLLø wene mapped as one unit, rrmuJ-garr. l'[Lr-ile these most common species coufd be distinguished on the 165. gr-ound, they were indistinguishable fnom aenial photographs. Hence thein distr.ibution could not be intenpolated between the gnound tnansects with confidence. There wene slight differ^ences in tree distnibutions, wítln CalLitz,is colLumeLaris (native cyp::ess) mone continuous in the cattle a::ea than in the sheep arîea, and, mulga uiee uersa; but diffenences wer:e not so gneat as to cast doubt on the compa::ability of the two vegetations.

The match was mone compl-ete ttr-an in th-e pnevious compar?ison.

On the basis of the soils, topognaphies and tnee distr"ibutions, the vegetations of b oth sand and clay soils r^relre nega:nded as companable between sample ¿Llreas. The number of quadr:ats laid on the two soil types lrere, for. the cattle and sheep surveys nespectively, 79 and 72 on sands, with 33 and 35 on cJ-ay soils.

5-53 ,Abtmdance of pastuz,e species.

The distr"ibutions of 43 species othe:r than tnees on tall shnubs wene mapped. Table 5.5 summanizes the comparisons of abundances between sunveys. Abundances wene companed as descríb- ed in section 5-33. 26 species showed either manginal on no differences in abundance between sample aneas, 1I species we:re mone abundant in the cattle-grazed area, and 6 species less abundant in the cattle al?ea. Only two of the speeies showing diffenences wene penennial. Following the nationale of section 5-25, these diffenences were ascniöed to long-term changes in the populations caused by stock, with d-iffenent outcomes fon sheep and cattle stocking. I

IArLE 5.5

CÀITLE/SHEEP CoãPARISoñ: SPEeI$ IX Îl{E CållLE ^rur{Ðl\lrcEs ARE RELÀîM 10 ¡\BtllDÂllCES IX lHE SHEEP ^RFA.

Spec:es PeÈnnial l.laln soil ¡òúdecê i Speclcs PcæEia¡ llðlD sol.I Áb@dücet or tYPe in cattle o! type ID cattl.ê ¡.phæral aæa EpheÈrðl åFô

iñt¿catd E c1âys 0 Àízø q'aøìtifìdø P Clays 9@.ía ob|ì4uíaapic E Clays 0 .iltripler tallata E SaDds -- pøadoaa E Clavs o ?ata lü¡{û .l:- ea í t íf Lont P Sands ------prteil¡eratís L CIay s (- ) t)ro Ìiw ey¡now Ubiquitous i a¿lctoiae\ol:,les I Cl avs i) ?L.taulùt:¿ oIèPã.'èa Clàys + ---- Frcehysøe ei!íarío f, CLÀys li¿l¿ol¿ k¿li sànCs t ciídthw t a¡nds ') .l'i.q L,jftrdld sànds lawsß

)actyl,1e ! êa i,ø rø¡tl (aa t Clavs (+) Iri¡!¡¡6 ¿.irPcsr¡:s 5êncs lragrectio eriopoal E :¿nds í) îri¡o¡r.n .'o i iijt'rie L sånd s EwhotL'íø i¡wandíí I lrb iauitous i.:g g oph l, ! t! i f C lays -u m at iacø Ko chi a ¿s t r i.:'- rí L.rÃ P Cla,'s

e1;,¡ttle'ìr P ? (rðr.e ) )tri¡ l.tt at'argícwa CIàys ---- ?hg allørL'"hø wobtií .4rièti,lÃ cr4¿orad tlbiouitous :t¿ !La!íoÐ.a1n reþ¿¿èø lìands t- ) -'1ir¿¿ills;a¿frr S¿nds !'tí;otÆ (hoüarus I (ra¡e) ::rú e,41ù gùn Ite¡k¡êå4 tlbiquitous !-! Ía carruge Lu ll Ltn,i:i (- ) i'NgrosLie ¿;cisii. E C Iavs of :ìíeyd,ri.n sg:, :ìnnJs :,1¿ !cr8 lloi!Et ach¿uo sand s :Yagilg Lvtnti.ûus I: : ùnds

Abut¿ioa olocaÞú iqn¿s G)

Atrípiez lìabata Ublquitous 0 û Abudante ¡¡tings: + nore lun<.i¿nt ün('-er c¡rt.Le-qrèzint reI¿tive ---- aeoíctía clays 0 to un,ler she€i'-r',r¿zinf . - lr.ss c,uod¿nr under cJttle-{ratinf relative Sabbagía æPopcem Clèys 0 . to un,lor sheec-gra:ing,

'¿røêi,lptetd ( L.l insI . 9daôi.d C.làys 0 ) i flerence is mrr8 cll a) fo ¿ i f l'er.ence l,et!êen ùeai. ctr ---- d¿ewre¿e C.:.avs o ---- d.ìæøtha ublquitous (+) ---- litøècata Clays (r) FIGURE 5.8

PLANT ASSOCIATIONS IN T}E CATTLE/SIIEEP SURVEY

KEY

2. AtrLpLeæ spongíosa 23. Ar'ístida contorta 3. A. uesicaz"ia 24. Iz,í,raphis moLLís 4, Pacltycornía tenuis 25. Inagus an¡straLianus 5. I{ochia astnotz,íeVø 26. PLagiosetum refractum B. Bassia íntz"Lcata 27. Enneapogon cyLindrLcus L0. B. diuaz"icata 28. AbutiLon otocarpum 1L. B. brachyptera 29. Sidn uirgata L2. B. paradnæa 31-. Aizoon quadrLfidwn L4. B. decwrens 32. PottuLaca oLev,acea 1,5. B. diacantha, 33. SaLsoLa kaLi L6. B. patenticuspis 34. Etq'lnrbía dtwmnondLi LB. ZygopfryLLwn aurantiacum 36, Ero&ium cygnorwn 20. DactyLoeteníum raduLans 40. Good.enia eyeLoptena* 2L. Ermeapogon aÐeytaceus 45. Babbagia acroptexa 22, Enagrostis &Lelsii 46. CLianthus fonnosus

*ldentification doubtful . :A 3

'ì-.; Í,- 20

29 8-11 ¡ tl , tl / i--

l0 16 It

15 (o) CATTLE AREA

r2 7 12

10 1t, 7 riì 11

I t- 2 tu ¡ /_t_ I t 2 2 t- - 5o -r'- t t t_ -l I 16 @ I a

5 .,\

L I 2 (b) SHãEP AREA

20 L6 .15

0l)p 001)P Posrtrve Negoltve

ASSOCIATIONS 767 .

5-54 .Associatic,n of pasttue species,

An association analysis for each treatment ciivided the species, as befone, into two g:roups, one corltesponding to sand, the othe:r to clay soil- habitats (f:.g. 5.8). A point of note is the patte:rn of associations for PLagíoseLum refz'actum. As will be shown in the next chapten, this species is a colonissr of disturbed on unstable

soils. ft was negatively associated with both the main groups, suggesting that the species of these groups a:re decneasens where thene is undue distr.l¡bance.

5-6 DISCUSSION: CATTLE/SHEEP COMPARISON.

Some infonmation exists in the l-itenattu"e concenning the

relative impor"tance to sheep and cattle of Aristída contorta,

Atz.ipLeæ spongiosa and Enneapogon qvenaeeus (a11 l-ess in the cattle area) , Erodium eAgnorm, SaLsoLa kali and Tripogon LoLäformis (¡ottL mone in the cattle area). Enneapogon lnas { alneady been discussed. A. contorta, although neganded by pastonalists as poon fodden, has been shown by Low (1973) to contnibute significantty to cattte diets in the AÌice Spnings anea. Jessup (fSSf) neganded it as unpalatabfe to sheep.

Chapten 4 showed this species to be affected in the same way by both sheep and cattle oven a short term. 0n the basis of the two papens cited, diffenential gnazing sefection might be invoked as the cause of the diffenences between sunveys, but this conflicts with the results of chapter 4. 0f the othen species, all but 168.

A. spongiosa ar?e r?eg¿Lrded as highl-y prefer"r:ed by both sheep and cattle (Et,odíun and SaLsoLa, Leigh and Mul-harn, 1965, for sheepr' Chippendale, l-968, fon cattle; Ir'ípogon, Leigh and Mulham' 1965' fon sheep, Low, 1973, for: cattle). While there ane no relative studies, these henbs and gnasses should be of mo::e impontance to cattle (section I-63), and hence should be fess abundant unden cattle than unden sheep if dietany selection is the main cause of pattenn. A. spongiosa., an incneaser unden sheep ("'g' Cnocken and

Skewes, 1941), i= not known to be gnazed by cattle' Little infonmation exists conce:rning the othen species which

showed differences between the sheep-gTazed and cattle-gnazed

sample areas, and so the fotlowing is langely speculation' The review in section l--63 has shown that, on th-e whole, cattle diets contain a g:reaten pnopo::tion of henbs and gnasses than do sheep diets: the latten animals eat more fonbs and sh:rubs. Hence, if dietany selection detenmines the effects of stock on rangeland vegetation,ahigherabundanceofforbsandshr.ubs,and'alowerr

abundance of he:rbs and gnasses, would be expected unden cattle grazhng, neJ-ative to sheep grazing; in the absence of diffe::ences discussed, in baekg:round environment. other than species al-ready fess two herbs (CitnuLLus Løntus anð. PtíLotus poLystaeheus) wene abundant, but foun herb s (AtTipLeæ ínfLata, Pot'tuLa.ea oLera'cea' TtibuLus tercesttis andzygophyLLwn atæqntíacwn) were mone abundant

nnden cattle, T\^ro sh:rubs and one forb (Aizoon qtnfuífidun' cz,otaLæia dissitifLoz,a anò. sída oingatd l^Iere mor?e abundant under cattle. There is good agreement between the results of the vir:gin/ cattle and cattle/sheep compar"isons, as outlined in table 5'6' TABLE 5.6 169.

SPECIES SHOWING CHANGE UNDER CATTLE- AND SHEEP:STOCKING

Species Inferred Reponted Relative change under change under a-bundanee cattle sheep unden cattle (Pennatty ) (Roxby Doms )

AtnipLeæ D Dl- 0 uesieoyia ,2 ,3

CnotaLæia D Dl + dissitífLona

Kochia D Dr12,a 0 astxotricha

T 0 Bassia rt 2 3 bnaehyptena , r rr12 0 rz I tu o díuayicata ', I I1r2r3r4 0 panadoæa

Enneapogon T Dz a)enaeeus Eragrostis I ï o dieLsii I PLagiosetum I r?s o refracturn

Sí,dA D ? + uirgata of Sísgnibrium sp, I rt 0

Refer.ences: t Jessup (1951) 2 Beadle (fg+e) 3 Cbocken and Skewes (1941-) 4 Banker and Lange (1970) 5 Murnay (193 ) ]-70

The tabl-e lists increasing and decr.easing species under" cattle as infe::r ed f-nc.rm the fir.s L companison, the species I neported- behaviou:r unden sheep, and thein r"el-ative abundances in the second compa:rison. It can be seen that, in general, the same species increase on decr:ease in abundance under" both sheep and cattle stocking; only 3 of the 12 indicaton species showed diffenences in abundance between cattle- and sheep-stocking. These obsenvations str ongly suggest that the concl-usions of section 5-4 a¡e cornect.

The fonegoing nesul-ts indicate that only a small part of the diffenences between the vegetations, unden the two grazÌ:ng negimes, can be adequately explained by differ.entíal- dietany selection. Because of this, it cannot be said that ther.e is fir"m evidence fon mechanical distu:rbance as the cause of stock-nel-ated vegetation change, even though the reaction of the majority of plant species would suggest this. The results may thenefone be construed as showing both pnocesses in oper:ation, but to va:rying degnees.

Inter"pneting the cause of obsenved stock-induced change, by indirect observation and neference to the l-iteratune, has p::oved inadequate. In hindsight, a bnoader appnoach was nequined. As wel-l as ::esultant pattenn in plant populations, the ::eseanch should have dealt with the detaíI of stock action--gnazíng behaviou:: and disturbance. A full study of gnazing behavioilrl \^ras beyond the scope of the resear.ch, but the nelationship between distur"bance a¡d plant distributions was investigated, and is neponted in the following chapten. 171,.

5_7 SUMMARY AND CONCLUSIONS

The virgin/cattle conrpalison indicated long-tern trends in pastune unden cattle, simila:r in outline to tnends, neponted in the litenatu.ne, in pastu:re under: sheep: cattle-stocking on

Pe:rnatty station has been accompanied by a decl-ine in abundance of perennial species and an incr:ease :l-n abr:ndance of ephemenal species. The cause of such trend is unl-ikely to be the prefenential- sel-ection of centain species by cattle: cattfe do not consume shrubs to a great extent (section 1-63) and, whil-e no specific infor"mation exists, al?e unlikely to graze ctr-enopod shnubs in par.ticuì-a:r, because of these speciesr salt content

(section l--62 ) .

The particulan species found to decnease and incnease under. cattle stocking on Pennatty a::e reponted o almost without exception, to neact in the same manner unden sheep-stocking. Pneferen- tial grazing selection by cattle cannot be the cause of the cbanges, since, although thene ane some gene::aì- simil-arities, there ane major" diffenences in detail between the preferences of sheep and cattle (section l--63). Dietany selection is said to cause change in gnazed vegetation, by nemoval- of the mone pnefer:red species, r^/ith a nesultant increase in less prefenned species due to the lowered competition (e.g. Leigh and Nob1e, 1969). If this is so, then cattle and sheep should remove different species, because of thein diffening pneferences, and the species benefitting fuom the lowered eompetition shoufd not be the same in both cases.

lulechanical- distutbance, unl-ike dietar"y pr.efenence, is common to both animals. SoiI distr:rbance, for instance, will have a 172. centain effect on a plant community, langely independent of the distu:rbing agent. Fon example, breakup of a l-j-chen cirust will nesult in a l-ower: nitnogen input (see Rogens et al , 1966); erosion of a disturbed soil- wil--l- remove much of the plant nutrients (Chanley and Cowl-ing, 1968). In both cases, the changed nutrient status wil-l- favoun the establishnent on incnease of species better" adapted to l-ow nutr.ient l-evel-s, and the decrease of species less well adapted, Mechanical disturbance, independent of the particufar he:rbivone, is ther:efore pnoposed as the main deter-minant of stock- :rel-ated vegetation change, to account fon the similarity of tnends in vegetation, stocked with catt.l-e on Pernatty, to tnends in vegetation stocked with sheep elsewhere. The pantial'test by the compa:rison of equivalently stocked sheep a¡d cattle pastures, on Roxby Downs, yielded no concl-usive evidence for on against the hypothesis. In this lattelr comparison, the cattl-e and sheep pastures displayed some diffe:rences which could nesuft from cattl-e and sheep gr"azing diffening species. 0n the othen hand, such diffelrences '.¡ene found for very few species: most showed no detectabl-e diffenences in abundances between grazíng negimes, as would be expected fo:: equivalently stockèd paddocks, with- a more on l-ess equivalent level of distr.rbance, wene the hypothesis cornect. Nonetheless, alf,hough the nesufts of the vingin/cattle comparison on Pernatty can only be explained by pnoposing disturbance as the cause of stock-nel-ated vegetation change, the validity of the hypothesis remains uncentain, and mo::e detail-ed testing is required. CHAPTER 6

RET,A TT ONSHTPS ENS P ECIES ABI]NÐANCES AND SOIL DISTURBANCE

6-1 TNTRODUCTION

In the previous chapter, mechanical- distunbance of soils and plants by stock r^ras p?oposed as the main deter"minant of long-term trend in gnazed pastunes. The investigations neponted in this chapten pr"ovide a pa:rtial test of this hypothesis. The effects of vaniations in soil stability on species abundances were obsenved in a virgin situation. From these observations' the effects of soil disturbances were readily inferned. fn the light of this inforrnation, the data of the pnevious studies r^¡ere lre-examined, to determine the extent to which an incnease in soil distunbancet nesulting from stocking, might account fon the obsenved changes in species abundances. A significant nurnben of ttre species encountered on the

Pernatty ancl Roxby Downs studies occun al-so in the Simpson Desert (l_ocation: fig. 6.1). Except on the margins, this sandridge desert has never" been grazed by domestic stock. The soils have local- vaniations in stability: these variations' collrel-ated with topognaphy, follow a regul-an pattenn, alle marked and clearly defined' and ane well documented. Hence pattenns in local abundances of species can be simply rel-ated to the stability of the soil- habit- ats, without the complication of stockíng effects; and the out-

come of soil disturbance is readily inferred. Although neither tb-e Dese::tts l-andfonms nor the stnuctr¡:re of its vegetations are dj-rectly comparable with those of the study 4nÕ F¡E a 6 a 1 LOCATION OF S¡MPSON DESERT STUDY SITES

oLD 1: I .-'-* Wotercourse NÏ | i'. t ' lPoeppet Corner \l aRDSvr-tE ----- Trock ì._-:';!:r'---,;:r. The Frnke + Mt Dore @ o -4-.-: Roilwoy HS @ w--- c .+ { L Thomos '|,-' Atton fr' o Ø Dorns H. S,ot! + Homesteod -i' I t I J(i I O Study site )x Dolhousie I à'iii' I I .|00 Sprrngs I N $\ I km N or FiE"t '..1 Peero Peero l:,Mt Soroh Pootowno 'i, \=s Loke "t N.T OLD 'ç- The ñ L, -{. Mocumbo .. .a ir Alrce H. S. 5çrrngs l,c Todmorden ,,'n HS I r360E ÞrmÞss . Cho.lotte ADATTA ì.ó I Wqlers. ..,., *ì....G w c a o,.: h it,. Oodnqdqllo. .,.: ; óu f;ì \6 S. A, NSW Loke Eyre 4 \ \ . Broko 5L \ Pt Aúgusto Hrll ',.uo P o c k" Þp ¡i. ì t7+. sites already discussed, tlre sur:face soils, and particularly the sands, are closely similar. Therefone, a species found to occun on unstable surface soifs in the Desert should indicate unstable surface soils if found in th-e other study areas. Further' an incr.ease in such a species I abundance with stocking can be most simply explained as the resul-t of soil- distu:rbance.

Data wene gathened dur:ing a napid tnavense of the souther"n

Simpson Desert in July, l-972 (see fig. 6.1). The Desent was crossed in dr:ought, hence only 35 species wene sufficientJ-y abundant for inclusion in analyses. Table 6.t l-ists these species, and gives their" occurnences in the other study aneas.

6-2 DESCRIPTION OF T}IE DESERT

6-2I CLimate

The Simpson Desent incl-udes the rnost anid pants of the contin- ent (Gentilli, Ig72). Most of the availab]e c]imatic data has been recorded on the mangins of the Desert. The rainfal-l- is veny low and var^iabl-e: the mean annual nainfalls at Oodnadatta (S.e. ) and Condil-l-o Downs (S.e. ), oven a 4g-year" pe:riod, were 147mm and ll6nm nespectively, with annuaf fall-s nanging f::om 25mm to 34Omm (Cnocker, t946); and at Bi::dsvill-e (Qld), the mean annual nainfal-Ì oven a lO-yean peniod was 1l6mm, vtith a nange fuorn 50mm to

33Omm (Boy1and, l-970). Rainfall- has a pronounced seasonality' with most falls in summen (Cnockenr l-946). Solan :radiation

Jevel-s ar:e high (fig. 1.8), and the mean annual tempenatunes on the mangins exceed ZOoC (Crocker, 1946). 175

TABLE 6. 1

OCCURFTIICE ]F SPiCIT]S IN Sl14PSOì'I L)LSI.Ii'I' AND CTIIi'],I :]?UTJY ARIAS

(a) Species in Bel-t Transect analyses-

Desez't Petmatt 7oæblt [¡esert Pcwtaltlt ])oxb:,t

! Tt4odia 72 |remophiLa + + basedouii LongifoLia' 2 rJoodenia + t3 Atripleæ + + eycLoptera Limbata 3 At'istida + + !4 Plagiosetun + + eontorta refraetun 4 Abutilon + + 75 Zygochloa + + otoeæpum paradoæa. 5 Phyallønthua + + 76 ?R.utídosís fuernrohrii h.cLichta soirlcs? 6 Enneapogon + + t7 PtiLotus aueTiaceus LatifoLius 7 LeschenauLtia 18 ilgriocephalus + diuaricata stuavtii I Euphorbiun 19 Ero.gt ostis + + uheeLeri CieLsii 9 Bobbagia + + 20 Enneapogon + acroptera eyLindricus I0 Bassia + + 2t Ptilotus + parad.oæa poLystaeh.eus tt + + 22 TribuLus ---fficøttha histrir

fb ) Additional- species in densit,y studies

Desert Pernattu .3@L Desert Pernattv .I9gþ!

Aeacia Eragrostís + cønbagei Lrniflora Frankenia dietyophLeba sP. PortuLaca + + mu?Taaana. oLeraceq. Cnotalaz,ia Rhagodia + + ctmninghønii spinescens + + So.LsoLa + + dissitifLora kali DícnastyLís Sida + + dov,attií uirgata Dodonaea + + Tragus + + viscosa ausl;ralianus

Eragrostis Tr.íchodesnn + + er4opoda zeyLøticum 776.

6-22 Physiographg and soiLs,

The principal featune of tlre Desent is its sandnidge system.

The sandridges ane pa:rallel, t::ending nonthwest-southeast, and appar:ently windformed (-!ladigan, 1936). They vany between 8m and 3Om in height, wittr- a separation of 13Om-550m betr¡reen nidges, and may exceed 3OOkr in lengttr (t"taidigan, f936). They are usually unstable at the crests, but the l-owen slopes and intendune ffats are stabi.l-ised by spar:se vegetation CCnocker', 1946; Boyland, 1970).

The eastern slopes of ridges (slipslopes ) are steeper', and often less stable, than the westenn slopes (backslopes)' The soil- of the ridge crests and slopes is a fine, red quantz sand (type ucf.22 or ucl-.23). The cnests have slightly ]anger gfains than the slopes, but the::e are no cthen soil differences between c::ests and slopes, apart fi:om soil stabiÌity (Cnocker, 1946)'

The interdune flats are eithe:: ned clayey sands (l,itc¡rfield' 1962) orr in nestnicted. aneas, clay soils with an ovenlying sandy loam (Cnocker,1946).

6-23 Vegetation

- The vegetátion of the Desent is pr"ima::ily a hummock gnass- tand (Spednt, 1972-- see discussion, section 1- 51, on the main vegetation divisions of S.A. ). A checklist of the known flona is given in Symon (1969). The vegetation was first desc::ibed by cnocken (rg+o), as a Tt'io&La basedouü-Spinifeæ paÎadofiustc

ttedaphic comp]-exrr, comprising five well-defined associations'

Those of impontance hene are th-e ubiquitous Spinifeü paraáorus

¡¡ Now Zy Loa paz'adoæa !77 association of th,e mobile cnests, the Tt'iodia basedol'ïii association of the stable lower slopes and sandy interdune conridors, and the

Acacia camba,gei association of ctay intendune con::idons in the easte::n Desert. The l-ast is a l-ow open woodland. Simil-an associations r^rere defined by Boyland (1970 ). Wiedemann (tglZ) , unl-ike the pr.evious authons, applied objective techniques to the description of the vegetation. He found that, although pa:rticulan soil- habitats could be clear"ly defined, the vegetation varied contínuously: the definition of distinct plant associations was not real-istic. Analyses in this chapter support this view. Connelations between species abundances and topography ale obvious fnom these reports, but all authors agree that the co:rne1- ation is indinect: species distributions ane pr:imanily determined by the stability of the substnata. Soil- features themsefves alîe dir:ectly cornelated r^rith topography, hence topog:raphy pnovides a convenient index of soil- conditions. Nonetheless, al-though top-

ography pey se may play some pant in contnolling species distnibutions, it is a minon factor by companison with soil- features (Crocken, 1946; Boyland, 1970; Wiedernann, 1972).

6-24 Domestíc gz'azing in the Desert.

On the mangins of the Desent, cattle alle nun whene underground

hraten is accessible, o:r whene stneams exist. Within the Desent,

very deep bores alre nequir"ed to tap the Great Artesia¡¡ Basin. The-,

expense of such bones, and the difficulties of access' are the majon reasons fon the absence of stock within the Desent, nathen than any intni¡sic featur"e of the pastunes. Where deep bo:res have been 1.7 B. sunk with-in the Desert, cattle are nun. An example is Pur"ni Bore (fig. 6.1), which. was sunk in 196f by a petnol-eum explor:ation tearn in the wester:n Desert: it Ïras since been stocked with cattle.

6.3 COì4PARABIL]TY OF THE DESERT AND OTHER STUDY AREAS

Fr"om the fonegoing descriptions, the physiography and vegetation of the Desert diffens gneatly firom the study al?eas on Pennatty and Roxby Downs stations; panticula:r1y in the size and nelative instabil-ity of the Desent dunes, and the pnepondenance of hummock gr^asslands in ttre Desert by companison with the talf shrubl-ands or" low woodlands of the other aneas. Nonetheless, the sand soifs are dinectly compar:able, and many species ane common to both aneas. Thenefone this study concentnates on the distributions of those species in common, occr:lning on sands. The Deser-t pnovides an excell-ent situation fon observing in isolation the effect of soil stability on the speciest populations, with resufts directly applicabfe to the sand ar:eas of Pernatty and

Roxby Downs stations. On the Dese¡t d.unes' the::e are no significant soil- diffenences other than in soil stabifity; differ:ences in stability are clearly discer"nible; and ehanges in stability can be simply indexed by changes in topogr?aPhy.

6_4 METHODS

6-4I Fot obseruatíon of specíes íneidences acnoss dune profíLes.

At each camp on the tr"averse, except site 6, a stnaight belt

transect was nun across the r"idges, including at least one complete 179. topog:raphical cycle. contiguotts 4m long by 2m wide quad::ats wer"e laid, and incidence of all species necorded. Inffuence AnaÌysís (section 3-51-) was applied, using th-ose species with fnequency of occunuence gneater" than lO%. Only associations at on below the .OI pnobability level were considered. Influence Ratings (IRts) were backptotted on a topogr"aphic pnofile. Fþom this backplot' the effects of soil stability on the incidence of the species involved were l?eadilY infenned.

6-42 For obseruation of speeies densíties aeToss dtme pt'ofiLes.

At site 6, la:rge and welt defined dunes, with mobil-e crests and vaniety of i4terdune areas' wel?e selected fo:: detenmination of nel-ationships between species abundances and soil parametens. Both sand and clay soils wene present. The profile was divided into four topogr"aphic divisions: dune cnests (unstabl-e sand), slipslope (semi-stable sand), backsJ-ope (stable sand) and fl-at. The soil of the ffat was a texture-contnast tyPe, sandy loam over: a 'clay subsoil-: this flat supponted a fow open woodfand of Acaeia cønbagei, unl-ike the hummock gnassland of the sand aneas. The topognaphíc profile is given in figune 6.2 and shows the relative extent of each division.

I^Iel?e run along Three panallel tnansects, at 4OOm intervals ' this pr:ofile. A totar of sixty-six 2oxl-m2 quadrats wene laid, with each quadr:at set at r.ight angles to the tnansect. Slipslopes and cnests weue sanpled. mone densely, relative to thei:r antea, than the other divisions: tLis allowed sufficient neplication fon 18 0. compar?ison with oth-er divisions. Quadrats wene negularly spaced along tnansects with-in divisions. ALL ínuestigations at site 6 uere eorr'¿ed out iointLy uith Dt,, Susan Barker, *

6-5 RESULTS

6-sI InfLuenee AnaLyses on speeies ineidences.

Fon belt transect data, the nodes of association are given in figune 6.3 and the fnfl-uence Analysis backplots in figune 6.4. In backplots, quad:rat scores (IRts) ane negulanly spaced fon clanity: since the quadnats followed the gnound contour"s, the accompanying pnofiles ane slightly distorted.

"' Twenty-two species wene inc]uded in the analyses, but not al-l- were encountered at each site. However, the numben of combinations found wene sufficient to determine the pattenn of associations which would anise wene all species encountened in one sunvey.

On this basis, the species list was divided into two main gnoups, species 1-l-3, and species 14-22. Within groups, speciest assoc- iatíons wer"e posi.tive, between groups negative. Fon consistency of pnesentation, the high IRrs, derived for any node' welre awarded to quadrats containing species of the second gnoup. Fon instance, if species 1114 and 15 fo::med a node, with species 1 negativeÌy associated with the othen two, the maximum fR of 3 was awarded to

¿t Research, Sch-ool- of Biological Sciences, Austi:alian National Univensity. Pnesent Address: National Parks and Wil-dl-ife Senvice, Adelaide. Fig. 6.2

PROFILE OF TRANSECTS AT SITE 6

Acqciq combogei o o oooo o oo o o o

backslope ctoy f lqt stip- bockslope S crest slope crest stipstope Fig.6.3 NODAL STRUCTURE FOR SP€CIES AS NUMEEREO IN IABLE 6.1

SITE I (Púß lq., Node I c.olrdl R r2tas 5 +++++ 0

s¡lE 2 (vrgh) o--@ Nod. t ¡R5t4 l-+ o +-

Nodc 2

tRr36 I5 16 r? rc ++++ 0 ++ +

Nodo 3

IF t rlt t0 ?1 4 -+t+ 0 +---

SITE 3 llryì, æ--æ--e--@ tlodc t IR 9þllr¿É22 6 ---+++ 0 t++---

ETE (llshl ' l{odc I

tn 56 ló ll z2

3 + + 0 ++ +

sllE 5 (Yke¡ñt t{odG I

tR I t4 ls 22 {++ 0+

sllE 7 (vkorñ) @@ Node I tR ! ñnt!É¡2 ASSOC¡AT¡ONS 6 ---++ + +++-- Þoa{ú |l.9otiw

Node 2

0r, p o0l, p lÊ 216

0 Fig.6.4 INFLUENCE RATINGS FOR BELT TRANSECTS

Nod. t 1 212233¿53272?222222 l!l4l¿¿5(21

SITE 1 (Oro..dl fPurni Bore

Nodc I oooooooo 222 2 22 z 22 270200202220 I 3 I 30 1 I Nodc 2 ll 1 I t t I o r lool o2o?266tG 535 3331 oo!222 3 I I 21000 1031 I ll Nod" 3 2 22 o22ota2aal 2 ooo 2o o 0000 220020 z 1'22 31

SITE 2

aloy

NOdC I 23 ¿4a556355512L4451 332323 l2

SITE 3

cloy

Nodr I ¡35¿¿55(54555 ¿545 (522 410 00 I 4¿5555

SITE 4

Node t I I ll lt t 1t7taa! 32 33 2l oooooooooooooooooo oooooo0 0000000o10000

Acqcio comboge¡ SITE 5

c loy 21 I 221 I 3r s 36 6 5 t666 666 66151! 22 2 3 2322 oOO 2223 l l l Oì '1010032 todG I gc 2 I 2? 221l2lll r r 2 21 2

Eucolyplus mrcrotheco SITE 7

c oy #

Fig. 6. s

SPECIES ABUNDANCES ON ZONES AT SITE 6

GROUP I SPECIES GROUP 3 SPECIES

t5 l5 15 Single record on ftols

10 for Bobbogio acroptero 5 Erogrostis eriopodo t- FSSCBS FSSCBS F55CBS FSSCBS Erogrostis É. Crotolorio Crotolorio FSSCBS loniftoro porodoxo FSSCBS FSSCBS o cunninghomii d¡ss¡t¡ltoro Tributus histrix Zygochtoo Fronkenio sp Bossio dtocontho Atriplex limboto Trogus oustrolionus f, o Ø f- t- (r z O GROUP 2 SPECIES J :l o- FSSCBS FSSCBS FSSCBS FSSCBS a Acocio dictyophtebo Acocio murroyono Dodonoeo v¡scoso Ptitotus totifotius U) .ç zF i- 30 J o- U) r!z ._c o ?5 f- FSSCBSFSSCBSFSSCBS tl'l FSS CBS z FSSCBS FSSCBSFSSCBS Swoinsono rigido Goodenio cycloptero . Ptogiosetum Trichodesmo IJ FSSC85 refroctum zeytonicum o Arisirdo contorto Enneopogon ovenoceus Erogrostrs dielsri Portuloco oleroceo Min¡mot record for Dicrostylrs doron¡i on bockslope

Mrnrmot records for

20 Abutiton otocorpum on ltot, bcckslope F Flot wrth A. combogei Ctoy soil N=14 Rhogodio sprnescens on llot. bockslope SS Strpstope Sond soit N=12 BS Bockstope Sond soil N=22 FSSCBS F55CBS C Crest Sond soil N=18 .* One Solsoto kolr Srdo virgoto occurrence only tB1-. quadnats r^rith Þpecies 14 and 15 pnesent, but species I absent. similar:ly, if species 1r2 and 14 for"ned a node, wittr- species J-4 negatively associated with- thre other two, the maximum fR was awanded to quadnats with- species 14 pr"esent, and the otlre:: two absent. At vingin sites (sites 2 to 7), the high IRrs wene most consistently found on and around dune cr^ests, the mobile sand al?ea. Therefore, incidences of species of the second gnouP wel?e gneatest in such areas, and incidences of the fi::st group least in such areas. IRfs vanied elsewhere along the pnofiles, with values displaying the infl-uence of sand or: clay substnates. The vaniation was continuous, as woul-d be expected from the studies of Wieclemann (1972)' and not disjunct as would be expected fnom the othen autho::sr descriptions. At site 1 (Purni bone, stocked with cattle), high natings wer:e found on both nobite crests and the rnar:gin between dune and fl-at.

6-52 Density distributions aLong topogz'aphíc profiLes.

Figune 6.5 dispÌays the aveuage density, on each topognaph- ical division, for" the species encountered at site 6. These species for-rned three main groups: those restricted to the clay ftat, those restr:icted to sands, and those pnesent thnoughout.

On the sands, the only species nestnicted to a partieular divis-

ion wene Zygochloa paz'adoæa, Ct'otaLæía dissítífLora' and PtiLotus LatifoLius, on th-e cr.ests; and the uncommon Aeacía

mw?aAünq and Diø,astyLis doranii, on slipsJ-opes and backslopes 182. respectively. Nonetheless, the abundance of otlrer. species yar:ied considerably fnom division to division, and so with vaniations in soil stabiJ-ity. In panticular, on the mobil-e cr:ests, Atistída contorta., AbutiLon otocalpum and Enneapogon qÐena.ceus

showed. lowest abundances, oll ürene comPletely absent' and

PLagíosetum nefz'actum reached its g:reatest abundance. This pattern collresponds with- that shown by belt tnansects. Thene

wer-e only two dispa-:rities between the ::esults at site 6 and the inferences fi:om the belt transects. The latter indicated

Goodenia cycLopterø to be a stabl-e-soil species, and Eragrostis dieLsií an unstabte-soil species, the reverse of the actual findings at site 6.

6-6 D]SCUSSION

6-6I ?he effect of soiL disLttrbørtce.

Fon virgin sites, both appnoaches demonstrated that the

species found on sand.s coul-d. be g:rouped acco::ding to the stability of thein soil- habitats. Thus, if sands are distu::bed, a decnease in the stable-sand species and an incr-ease in the unstable-sanc1 species should occull at the site of disturbance. Examples of this

wene found in the Deser"t. The at?ea ðnound Pu::ni bone (site 1) was stocked with cattle:figure 6.4 shows the position of the bore nelative to the bel-t b:ansect. Cattle at this bone wer"e obsen-

ved to move mainly along the intendune corridons, only crossing

dunes by means of the track usecL fon the vehicfe t::aver:se. This

tnack was SOOm distant from ttre transect. The backplot of site 1 18 3. rRr s showed th-e highest r:atings to be not only on mobile crests , but also on the ma:rgin of the flat, where cattle in th-eir movements had distr::rbed the sand. A funther e5ample is distr:r'bance caused by noadmaking. The unimpr"oved ea:rth tnack used for the vehicle traverse was built in 1961 by the explonation team mentioned ear.lier, and has since been used by tour"ist par:ties. The nemoval- of vegetation in th-e path of the t::ack, and in par:ticula:r the perennial hummock-grass Tt:iodia, nesufted in wind erosion, as would be expected. fnom the studies of Manshal-l (1973). AJ-though nelativel_y infrequent, the tounist excr:r'sions have langely pl?ev- ented r"egeneration. Although the nesultant deflation I^Ias most obvious where the tnack cut through crests, it was also found on pneviously stable sand ffats. Such aieas were being colonised by plants nonmally restnicted to cnests, panticulanly Plagiosetwn refractum. Table 6.2 l-ists the species corrlesponding to stabl-e and unstable sand habitats, which on the above basis, could be expected to decnease or incnease folJ-owing soil distunba¡ce'

6-62 Compa.yison uyLth resuLts ft'om Pernatty otú Roæby Doums swueAs.

Tabl_e 6. 3 gives the species, considered on the for:egoing evidence to change in abundance with disturbance; the changes in theilr populations under: stocking infe::ned from the vingin/cattle companison (Pernatty); and th,ei:: :relative abundances in the

cattle/sh-eep comparison (Roxby Downs ). If mechanical- distr:-nbance by stock is th-e cause of 184,

TABLE 6.2

SPECIES EXPECTED TO INCREASE OR DECREASE FOLLOVTTNG SAND DISTURBANCE

Incr:easens Decneasens Evidence fnon Plagiosetwn AbutiLon nefractum otocarpwn

Zygochloa Av"istida Belt transect paradoæa contotta and density data Enneapogon aüenaceus PortuLaca oLev,acea

Mynioeep haLus PhyaLLant?rus stuapti U fuernrohz'íi BeIt tnansect data only PtíLotus polystacheus

Cz,otalaría SalsoLa dLssitífLora kd.L1, Density data only Sida uirgata 185. TABLE 6.3

SPEC]ESI REACTION TO DISTURBANCE IN THE DESERT AND TO STOCKING ELSEWHERE

Species Change Infenned Abundance following change under unden cattle disturbance catt]-e nef. to sheep ( simpson ) (Pe:rnatty ) (noxUy Downs )

CrotaLayia I o + dïssítífLora MyrioeephaLus I stunrtii PLagiosetun I I o refnaetum PtiLotus ï poLystaelteus ZygocVtLoa I 0 patadoæa

AbutiLon D 0 otoearpum Atistida D eontorta Enneapogon aDenaceus D I Portulaca D + oLeracea PhyaLLøtthus D (D) 0 fuernnohríi SaLsoLa D + kaLi Sida D D + uirgata

D decnease. in abundance I incnease ,l' Ephemenal , no infer"ences drawn r'c:t "¡"utt1 + mone abundant ) cattfe/sheep comPan:-son less abundant) 0 no change, olr no diffenence. 186. stock-rel-ated ,regetation change, the species considered to dec- ïìease o:: Íncr.ease witt¡- distunbance íJì tlrc Dese:rt woul-d be expected to show similar change unden stocking; and such change could be inferned fnom the results of the vingin/cat-Ele sunvey' Many of the species rârrled ane ephemeral , and no inference was dnawn concerning them (see previous chapter')' For the rest' pLagiosetum rtefraetum and sid.a uirgata, respectively inc:reasing

and decreasing species with disturbance in the Desert, showed similan changes with stocking. PhyaLLanthus fuernt'o?rcii' a

dec:rease:r with distu¡rbance in the Desent, showed only a manginal

decr:ease with stocking. However, ct,otaLaTiA dissitífLoTa and the neverse Ewteapogon ü)end.ceas showed changes with stocking ' ofthoseexpectedifincreaseddisturbanceduetostockwasthe cause of the changes in their" abundances' ZygochLoa paradoæa' an incneaser with distr.urbance in the Desent, showed no appa:rent

change with stocking. Inthecattl-e/sheepcomPar"isonrthreeoutcomescouldbeexpec- and ted we:re disturbance th,e main cause of change. If the cattle

sheep distu::bed the sands to the same extent, no difference

between paddocks woufd be expected in the abundances of the species of table 6.3, since thd species woul-d react in the same

mannelr in both paddocks. If cattl-e-stoc1

bances would be refatively more abundant, and those decr:easing r^rith distunbance would be :refatively less abundant in the cattl-e

paddock by companison with the sheep paddock. If sheep-stocking !87.

caused the more distu:rbance, the lleverse situation would be expected. Tab1e 6.3 displays no sucÏ¡- consistency'

6-63 ConcLusions.

Once more, thore is no conc'l-usive evidence for on against th-e distu:rbance hypothesis. Centain species , the neactions of which to soil- distunbance in a virgin situation have been noted, have ::eacted to stocking as would be expected wene the hypothesis cor:nect: othen species have not' Funthen, the results of the cattle/sheep sul?vey have not shown the consistent pattenn exPec- ted f::om the hypothesis. Nonetheless, the hypothesis cannot be dismissed with ce::tainty. The:re still o.*.ìrr" the p:roblem of the cfose simil-anity between the trends obser"ved under cattfe and the tnends nepo::ted unde:: sheep, inexplicable in tenms of dietanysel-ectionrand-mostsimplyaccountedfonbythedistur- bancê hypothesis. Distunbance as a potential majon cause of stock-refated vegetation change is wonthy of mone detaited investigation than was possibie in the counse of resear"ch for this thesis, date' and of mone emphasis in othen reseanch than has been given to CHAPTER 7

sot'fE EFFECTS OF CATTLE 0N PERENNIAL CHENOPODS

7_I INTRODUCTION

Two minor studies, investigating the long-term effect of cattle on perennial chenopods, ar:e descr"ibed in this chapten. The finst deatt with- changes in populations of AtrípLeæ Ðesicav'ia anò Koehia astrotríeha ín 45 yea:rs; and the second concenned possible negeneration of AtripLeæ and Kochiø species under cattle.

7_2 POPULATTON CP,ANGE IN ¿. VESICARIA AND 1(. ASTROTRTCHA

7-2I Methods Osborn et al (fgSZ) used belt transects, r:adiating fuom waten points, to estimate the density of A Oesicariq,: these provided the basis fon their" pioneer:ing PaPer on sheep gnazing effects. The oniginal data sheets a:re stored in the Botany Department, University' of Adel-aide. Apart from the Koonamone Vegetation Resenve necords, they nepr"esent the only eanly r-eseanch' in a,nid South Austnalia, capabJ-e of dinect nepetition. Of the many run, three tnansects were l-aid in virgin pasture on Frome Downs station (J-ocality: fig. 1.1), in 1929; immediately ' pnior" to stock introduction. They we::e mentioned, but not analysed, in the 1932 pape:r. Two we::e::un fr^om a newly sunk dam, which could not be nelocated due to lack of station ::ecords. The longest tnansect, of over 2OOOm, was run from Davies (Davists) dæ, sunk at th-e same time. This transect was nel-ocated and repeated (rig.7.t). In 1929, the tr.ansect was laid on a stnict compass bearing,

-188:- Fig. 7.1

LOCATION OF DA.VIËS DAN4 TRANSËCT

LAKE FROME

4 t Þ Btuff * .r2 \ Dom ii \ tl' \ Tronsect DAVIES AE DAM / ,, , I \/t-* FROME '\\'--',, DOWN5, -----o, H5,,, --+ ,' Bitteroo - - tl \ -¿)- t. Hut /.t .. //t/ t t'' I ; ( i + WoterPoint I ',, -----Mqit trocks I ', -- --- Other trocks I ,""'-)+Dog fence I ti 5mi NOTE: Dog fence turns eost 5km 5 km further south

I L L I TORRE N S ) FR OM L I o F rome I Downs I Ll',

Pt Augusto ¡ Yun to 189.

aided by sigh-ting posts, wittr- distance measu:red in long Paces (f .Sn). Ëveny bush actuall_v crossed by tlre obsenver Ìras scored; fon species (listed as Atr"LpLeæ or Kochía, i.e. .4. Þesicaria and K. astnotl,ichø) , ineight, diameten, vigor.ul as indexed by the state of the foliage, plresence or: not of a basal soil- mound ' and juveniÌe on adult. Cnitenia fon detenmining juveniles have not been pneserved. The tnansect could be reganded as a strip of quadr.ats, size one long pace by one shoulden width. The tnansect was nepeated in Manch 1974 r fnom the or"iginal stanting point, but allowing for" magnetic variation over the inten- vening peniod (zon). Distance was estimated by rangefinden and sighting posts rathen than paces. I^lhile the :relocated tnansect coul-d not be expected to coincide exactly with the oniginal, nonetheless it was considened adequately close to penmit dinect comparison of nesults.

7-22 Stocking histony of study q?ea.

Although no complete records exist, sheep rnay have been nun fnom

Davies dam fon the initial ten to fifteen yealrs. Cattle wel?e cer"t-

ainJ-y nun fi:om it after" 1946, when the State dog fence was consofid- ated afong the southenn bourrdary, and the l-ast sheep flocks were removed. Howeven, cattle have always been the main enterpnise

on Fr:ome Downs, and so the possi-bility of the t::anseet being grazed only by cattl-e remains. Past and present stocking level-s on the darn could not be determined, due to both- absence of necords, and the open l?ange 19 0.

system. Thene lrel?e aPProximately l-00 head on the dan at th,e tj¡le of survey, and tFre manager considened this a usual levef.

7-23 ResuLts.

The l-929 tnansect was nun in dnought, but the l-974 tnansect in an exceptional g::owth per:iod (see ehapten 4)' Thus foliation as a measune of vigoun was not comparable between times. The large numben of defoliated bushes neponted in l-929 wene not necessanily dead (Osbonn et al , 1932).

Tabl_e 7.1 summan:izes the fíndings. The clearest change is the a.l-most total- el-imination of ,4. Uesicalia and the deerease in density of K. astz'otricha. Figr::re 7.2 emphasises this.

K. aphyLLa was not mentioned at all in the oniginal data, non in th_e accompanying notes. As the notes wene compr.eh-ensive, the species must then have been a veny mino:r component of the vegetation.

The reduction in síze of K. astrotv'icha is also cfear.

Fig¡r:e 7.3 shows the :relative firequency distribution of diameters, fon both times. These suggest a possible change in age stnuctt::re, but this is r:ejected on the g:rounds that (a) ttLe pnopontion of dead ptants nemains the same, and (¡ ) in 1974 , many of the smal-l-er'

cahopies wene growing fnom extensive and o1d ::ootstocks, indicating a pnuning down bY stock. TABLE 7.1 191.

DAVIES DAM TRANSECT RESULTS

Ca) Uumben of bushes in üransect.

1929 1974

Total busbes r245 119 A. uesicaria, 946 4 K. astv,otrieha 299 BO K, pyranidata I K. aphyLla ar¡ 26 Total Koehia 299 t_I5 K. astrotrieha 6 5 (j uvenile ) *The only catego::y on l-929 data sheets was K. Astnotrícha

(b) Mean height a¡rd diameter of foliated bushes with associated standand e::nons.

1929 19 74 K. astrotríeha

I Height sz.s l rr.o 26.4! r2.2 (cn) Diameter 4s.6 I 16.7 30:6 t l-8.s (cm)

l_929 L9 74

K, astrotz'ícLta Dead 26 9 % totar 9 I1 Defoliated 135 5 % total 45 6 Fig. 7 .2 BUSH DISTRIBUTION ALONG THE DAVIES DAM TRANSECT

TOTAL BUSH

I I -t

I I t- I - -L - ¡ I oC, I E I c) I I E I lt a I a d) I I I I

I r-l

1000 æ00 Dislonce (m)

KOCHIA ASTROTRICHA

I rt, 30 -o I E ) I c I .c I ln l I I lD I

I I I

1000 2000 Dislo nce (m)

1929rBP ---- 1974 t r F ---= Fig. 7 .3 LARGEST DIAMETER OF K. ASTROTRICHA BUSHES on Dqvies Dcrm tronsect

le23

20 n=299

l0 o ã-

L) a- IJ af, lrl É, tL 30 trJ tr 1974 J n=80 lrJ 20 E.

O O O o o r,o (^o o + o c\¡ cf) \1 c- @ I o I I o I I I I I I I I r o o (\¡ (Ð { tJO @ Þ- @ O) D¡AMETER SIZE CLASS (Diometer s in mm ) 192.

7.24 Discussion.

Thene ar:e two possible conclusions, dependent on the past histony. l-. If cattl-e only have been nun, then they have brough-t about the rnajor destruction of shrub populations as described unden sheep (chapters I, 5). 2. If sheep were nun first, and were the oniginal cause of destnuction, then the catt.l-e have maintained this

impness for" th-irty years. In eith-er case, the impact of cattle on these species cannot diffen mar"kedly firom that of sheep.

7-3 A CROSS-FENCE COMPARISON OF BUSH DENSITY: CATTLE VS UNGRAZED

Cattl-e were introduced on Wabnicool-a station (locality:

Fig 1.1) in 1958. The small paddock shown in figune 7.4 was stocked continuousJ-y with cattle f:rom this time, the cattÌe being nestricted by strengthened fences (a nane example--see section 2-32). P::ion overstoclcing with sheep eliminated the oniginal AtripLeæ pasture and nesulted in an ephemeral grassland.

No necord exists of the sheep numbens. Cattle number vanied fnom 20 to 60 head (heifers) with the rnode at 30 head. As the paddock was only SOOLu (z^i2) in anea, this::epresented a heavie:r stocking pen unit area than situations reponted in the nest of the thesis, but a lighter" stocking in te:rms of number per: wate::. 193.

The paddo-k adjoining the western fence was par-t of anothe¡' station, and sheep were l?un in it, Howeve:r, shccp nanely pene- tr"ated so far, due pantly to the distance from the nearest waten

(Bkm), and partty to a barnien of -l-ow hills. ' The vegetation

showed l-ittl.e sign of gnazing: the soif sunface vlas undisturbed, the penennial- chenopods were tall- and spindly, and tree canopies extended to the gnound. (fne latter". two features indicated the absence of br.owsing).

A ::ectangulan staggered g:rid, dimensions l-OOOxBOOm2, was

sited over" this fence. 80 quadrats wene laid, dimensions 1.5x20m2. Density of adu1ts and juveniles wene scored, for: AtrípLeæ stipitata, A. uesicaîia., and .K. georgei, the majon

chenopods pr"esent. Juveniles wene defined as singJ-e stenrned

individuals unden lOcm in height. Figures 7.5,7,6,7,7 give

the distr ibutions of adul-ts fo:r these species , and figur:e 7 . 8 the distr"ibution of A. stipitata juveniles, the most commonì-y found.

Fnom these distr"ibutions, there appeared no significant

:re-establ-ishment of populations aclross the fence, despite the

seed source on the ung::azed side, the sJ-ope of the ground which would aid water tnanspont of seeds, the reported unpalatability of A. stipitata (..g. Ba:rken and Lange, l-970) and the nepor:ted regenen-

ative ability of the same species (uat:- et al, 1964). Fu:rther, the survey was made in May 1974, aften a year' of above avelrage nainfall (chapter: 4). At least some seedlings could be expected

to appear" in th-e gnazed paddock at such a time, if negene::ation was irr progress. 4 Fig. 7.4 LOCATION OF CROSS-FENCE COMPARISON AT WABR¡COCLA

or SAMPLE AR EA .To Koonomore * Bore

I \.\l'l Ltr'. Bore fì**'-.'- '1oo,-n I I I - .,.r¡( ..'r:ììi.ì'.*' coolo \ HS )

+a N .¡t -7 4^"

1 Zkm .r¡'\'''r'9-- TRACK ¡ Woroonee ."..' ..¡:' STATION Hitl T ,Teetutpo .,,,r.,.t,'t' BOUNDARY r*lJ.t'l't :ìr7¡11'I FENCELINE .li\, I HS

Sompte grid

a a o a a a a a

a a a a a a a o

a a a I o a a a ...... ""1 UNGRAZED Boundor . . ' -r -t-i-:--'--:-t - Fence CATTLE o a o a a a a a a

I a a a a a a a a

a a a a a o a a

a a a a a a a ground Trock slope

250m N Fig. 7 .7

O ..Oo OO ooo o K. CfORCf,l .OO. o . OO o o o a Oo ao o .O

Density ¡l a Density i v !t -v 1 plont/qraùot - a a O a plorìl/srockot a O 3 ptontrlqrodto, o I a. a a Ou o 3 ptonts/qucûot a o 5

Fig. 7.8 OO. o!. o a A. STIPITATA oa oO.O. O. o (JUVENILES) o .ooo O !o .o ooO oO oo .- Density - Density -v a I ptont/quodrot

O 3 ptonts/quoorot O 1 ptont/qudrot o, O 3 prontslqrodot Ou

L- __l 19 4.

Ilence, cattJ-e appean to have maintaj¡red, fo:r fifteen yeanst the slr-eep impness on these species.

7-4 CONCLUSIONS.

The above studies were neither extensive non detailed.

However:, they dealt with species which, by all accounts, would rank ver:y l-ow on a scale of cattl-e pneference (see section 1-63)' Yet cattle were found to induce changes in their populations, compa:rable with those neported for sheep, o:: to maintain pnion sheep effects

The :results thus pnovide independent support fon the contention of the preceding two chapters, that the J-ong term effect of cattle is not significantly diffenent fuom that of sheep' I\:nthen, they would not be expected if preferential selection of fodder species wene the cause of long terTn change. CIIAPTER 8

GENTBA!__-DIEçUSEfAN

8_]- INTRODUCTION.

The object of this thesis was to outline the ímmediate irnpact and the pr"obable J-ong-term effect, on anid Pastul:es, of the introduction of cattl-e onto fonmer sheeplands; and to detenmine, if possible, the mechanisms involved. This was attempted by the observation of a small number: of repr"esentative situations. Detail-ed discussions and conclusions have been given in the counse of nepontsr and ane summarízed as folfows: l_. The immediate irnpact of cattle is the formation of new pattenns in the vegetation' These pattenns are only

obse:rved intermittently, but al?e necurnent' 2. ltre long-ter-m outcome of cattle introduction is closely similar" to the long-ter:m outcome of sheep gnazing. Either the fonmer" sheep impress is maintained, on the cattle,ifintnoducedtoformenlyung:razedpastures' bning about changes in the vegetation as reponted to occtll' under sheeP stocking. 3. The cause of the stock-nelated patterns and changes in vegetationr.emainsuncentain,butitappeansth-atgrazi:ng per se is a minor facton except in the short-term' Before discussing the wider implications, both ecological and pastonal, two senious conflicts in the findings must be nesolved'

These ane (a) the val-id-ity of meth-ods used in the inference of long-

tenm clrange, h view of the demonstnated shor"t-term var:iation; and (b ) tne clispar:ity between the shont-term effect and long-te:rm

outcome of cattle int::oduction.

-19 5_ 196.

B.-2 THE VALIDITY OF INFERENCES REGARDING LONG.TERM PASTURE CHANGE.

The infenences of long-term pastu,re change due to stocking, as given in chapter- 5 and the latten pant of chapten 7, were based on data coll-ected at a single point in time. To necapitulate: the densities of pastu:re species wene comPalred between sitàs with diff- e::ential stocking histories. A diffenence between sites in density of a species was interpneted as showing a long-term change in the status of the speciesr population, with al-fowance being made fon the effect of background envinonmental infl-uences and gnazing in the short-term. In chapter 4, shont-te::m effects wene shown to fluctuate so gneatly as to cast sevene doubt on the validity of inferences of change in time dnawn fnom pattenns in space: the panticula:: infenences woul-d depencl to a great extent on the conditions pe::taining at the time of data coll-ection. Thus the accuracy of the infenences d::awn in chapters 5 and 7 depend to a g::eat extent on whethen adequate al-lowance !¡as made fon the shont-term stocking effects. As regards the long-lived penennials rpa::ticulanly the salt- and bl-ue-bushes, any significant diffenence in density relating to diffe::ences between stocking treatments will- have long-term consequences. The species a::e sfovr to negenerate, and until they do' thene remains the potential for permanent damage through enosion' Therefore a lowe:: density of penennial-s in one stocking tr:eatment relative to a¡othen must be considened evidence for: divelgences between the two, if not long-tenm in origin, then at l-east long- tenm in consequences. Thene is thus littIe doubt as to the 1-97 . val_idity of the inferences concerning ttr-e perennial species. . As negands ephemenals, th-e short-ts:m effects wene effect- ively rninimised, but this was due to fortunate (though not fontuitous)circumstances. These were the avail-abil-ity of (a) a virgin pasture as a basis fon the finst compa::iso¡ (vingin/ cattl-e, Pernatty station); and (b) a gnowth peniod in which to d::aw the second companison (cattl,e/sheep, Roxby Downs station). vingin aneas, suitable for comparison with grazed pastunesr al?e undenstandably tlare, aften a centuny of settlement; and compar-

ison between grazing treatments cannot always nely on convenient gnowth peniods. In the first comparison, sho:rt-tenm stocking effects were minimised by disr:egarding any ephemenal species with a lowen abundance in the gtazeð alrea, and only considering those wittr higtrer abundances in the grazed area, despite pr"obably short- In te:rm reduction in densities fnom the previous gnowth peniod' the second compar.ison, the gnowth peniod meant that shont-term stocking effects we::e obscured on absent--as shown in chapter 4--

and any nesidual- differences between tneatments must th-enefone

have neflected on a long-tenm divengence' The infenence of long-tenm change can thus be considered valid,

because of the special cilrcumstances. Nonetheless, since such ci::-

cumstances alre not always avaj'fable, the general wanning in section 4-94r on the intenpretation of changes in time fnom patterns ín sPace, still aPPlies' 198 .

B_3 DISPAR]TY BETWEEN SHORT-TERM AND LONG-TERM OUTCOMES.

Although the irnmediate inpact of cattl-e is the fonmation of patter"ns in vegetation diffening f::om those imposed by sheep, the long-ter-m outcomes ane closety similan. Such a dispanity is not easily resofved, since the tong-tenm outcome cannot be completêJ-y independent of short-term effects. Two possible explana-tions aLe offer-ed. l-. The eventual- outcome of stocking is not the resul-t of cumulative sho:rt-tenm changes in tlr-e vegetation; but is :]athe:: the result of occasional- catastrophic changes, occu::ning over a shont period, but with J-ong-tenm

consequences.

2. The eventuaf outcome of stocking is the cumulative resuft of short--Eerm changes which, though predominantly

ephemenal in natune, nonethefess have a nesidual effeet; but the seasonal va:riation ovell a short obsenvation peniod is such a.s to obscure the direction of the dr"ift

in Pastune condition. The two ane not necessarily mutually exclusive'

Few of the reponts on gnazing succession in th,e southern Australian r.angelands give a time scale : on the whole, changes have been lar-gely infen:ed, as in this thesis, fnom companison of different stockíng r.egimes, rather than by dinect observation of succession oven a long peniod. Hence the::e is little evidence as to which of the above plrocesses is in operation. Indirect evidence does exist for the first. The histony of gnazing on 199 . the Austr^alian rangelands is punctuated by pe::iodic catastroptr-es (section f-3). Pasture degradation folfowing settlement was not a g:radual pnocess. rn panticuJ-a:r, many of the perennial pastunes we::e destnoyed oven a shont pe::iod; duníng the lB90ts in western New South Inlales, (Anon' l-9Ol-), tSSO-t935 in South Australia (Ratctiffe, 1936), to quote only two instances' The r:esult of this damage--erosion, loss of dnought fodde:: lreserves' changes in nutnient status--have been detailed in chapten 1. Thus, although oecu:fing ovelr a very short peniod, this destruction had consequences still evident, and not refated to su'bsequent ' non-catastr"ophic short-tenrn gnazing effects. Nobl-e (lgzs) has also pointed out that tnend in pastr:r'es is not necessarily a gnadual pnocess. Penny (fgZO) considened that, following the eanlier. catastr:ophes, the rangeJ-and pastunes a:re at pnesent in a relativel-y stable state, a]though indívidual outcomes have been cletermined by the seve::ity of the past erashes. l[illiams (1968 ) neached simil-an concl-usions in studies on vegetation changes in the semi-a::id Rivenina. Eithen possi-bility can account for the dispa:rity repor^ted in this thesis. In the first caser the short-term obsenvatj-ons were made during a peniod of exceptional pastu::e g:rowth. stocking effects, because of the light overafl stocking' welfe necessarily minfunal, and catastr"ophic changes could not be expected to occun unden such conditions. In the second case, the observation peniod

u¡as too short, and the pastune fluctuations too gneat, to allow detection of incipient long-tenm changes nesulting fuom the cattle introduction. 200.

B-4 PTOSPFIERE PATTERN

. The vegetation patter:ns of the str-eep and cattl-e piosphelres wene Ínvestigated, pr:imanily to isolate stocking effects from the influences of backgr^ound environmental variabfes. Howeven, the findings have implications extending beyond the immediate p::oblems under consider.ation.

Banken and Lange (fSOg , 1970 ) and Roger.s and Lange (fgZl ) were pnimanily concerned with the demonsf¡ation and gene:raÌ delineation of piosphene pattenn in vegetations stock,ed with sheep. Foi: this purpose, they used nadial sampling g:rids, designed to amplify any existing nadial patte:rn, and mapping as the main fo¡m of data pnesentation. These methods were adequate to d.isplay the genenal fonm of piosphere patter-n, viz., the existence of continuous gradients in vegetation and the radial nature of the patter:ns. 0n1y minon attempts wene made to determine, in mo:re detail, the fonm of plant/d istanc e r"el-at ionships .

The p::esent study has demonstnated two significant points.

The finst is that an unbiased sampling system will detect piosphene pattenn'as neadi.l-y as the above sampling. The second is that the plant/distance relationships can be appnoximated by linear^ functions. The seanch fon finear r^elationships was proposed on the basis of Langers (fgOg) obser.vations, of J-inea¡ refationships between sheep

dung d.:rops and dístance fr"om water:, and on evidence in Rogers and Lange (l-gZf) tnat lichen fnequency also shows a linean nelationship. The existence of linean gnadients in the pasture have now been anpfy demonstr"ated. Nonetheless, funthen investigations into these

nelationships a::e nequired. All- recent- studies, and the present one, 20L, have dealt with the piosphere pattern wef.I within the normal nange of stock. The :resutts cannot be rationaÌly extrapolated to the límits of the piospher.e: changes in plant pa:rameter-s with distance must appr.oach zer"o towa:rds the limits. Funther", ì-inea:r nefationships ane not easily accounted fon fnom reports of stock behaviour'.

Tlrese r"eponts indicate that the distr^ibution of grazing p::essune is not such as to give ::ise to linea:: pattenn in vegetation. Thus the detail of the variation nemains unknown, although the vegetation studies have demonstrated the existence of the patterns and. have shown how these can be appnoximated. Until the detail has been deter.mined, the direct nelationship between the patter-ns and the actions of the gnazing animal- cannot be conside::ed in depth. The pnesent study has demonstnated the existence of piosphene pattern caused by cattle. This was exPected, but had not been p:reviously documented in the Austnalian nangelands. The cattfe piospher^ shows the same basic outline as the sheep piospher"e \ 'continuous gradients in the vegetation and radial pattenn centred on the waterpoint. The differences between the sheep and cattle piospher.es are differences in detait--the par:ticular speeies affected--rathen than diffelrences in pr:inciple.

8-5 THE CAUSE OF STOCK_INDUCED VEGETATION CHANGE

Thnee potential factons in vegetation change nesulting fi:om stocking were discussed in chapten t, viz., the gnazing of pnefenned species, tnampling ol? genenal distunbance in the counse

of gnazingr ancl nutrient transfer over the pastune conrelated with 202.

stock ¡ovefiìeot. The fast factor v¡as not considered to play a major overa_L]:rol_e in causation of change, because of the fow animal:allea ratio, and the r"esultant rninimal- tuansfer (see section 1-B). The finst two have been discussed at length th::oughout the thesis.

fn summar:y, the pr.efenential sefection hypothesis is that changes in vegetation nesult fi:om the grazing down of prefenned species, and the consequen-L increase in fess pneferned species due to the 1essened plant competition. Hence patterns and tnends

obsernved shoufd neflect the dietany prefenences of the par"ticulan her.bívore involved. The qbservations neported in chapten 4 l-ent

some suppor"t to this. The findings indicated that the cattfe welre l.Iere consuming some grass consuming rnainly grass, wher:eas the sheep ' but also fonbs and shnubs. This conresponds wi'Eh- the genenal differences in pnefer:ence- between sheep and cattle, as neviewed in section 1-63. Nonetheless, the biypothesis I^Ias not convincingly valiclated. In particulan, there was evidence of shor"t-te:rm stock-induced pattenn in the distr"ibu-bion of Ewteapogon spp. , the nevelîse of the pattenn expected fi:om neponts of sheep preference. Furthen, Koehia astrotv"Leha showed short-te::m stocking impress

unden sheep at a time when more highly pneferned species were abundant. The review out]ined -the high degr:ee of sefectivity displayed by gnazing stock, and sheep in p¿nticular-. Foi: example,

Leigh and Mul-ham (1966a) r"eponted that species comprising only a

small- pnopontion of the total- pasture were majon contnibutors to the diet. Howeven, of the 55 speeies dealt with in the chapten 4 studies, only those most commonly occurning showed the influence of 203. stocking. If dietany selection wel?e the cause of pattenn, less common species slrould also have clenonstnated the s'i;ocking influence.

,Atthough the pnesent findings may arguably have anisen fuom the par:ticular: statistical meth-ods employed, Barken and Lange (1969) also found that the piosplrer-e pattenns wel?e only dispJ-ayed by a few of the more common species, despite thei:: use of sampling systems and analyses designed to amplify any existing pattenn.

The outcome of loñg-te:rm cattfe stocking, reponted in the fi:rst par:t of chapten 5, was completely inexp]--icable in ter"ms of dietany prefer.ences. The same species wer:e found to increase o' decr:ease und.er cattle-stocking as under sheep-stocking. The :results of chapten 7 ane afso inexplicable in such tenms; the cattle had and thus main- suppressed the :regeneration of penennial chenopods ' tained the oniginal sheep impness, despite the l-ow prefenence fon shnubs in genenal, and pnobably fon chenopods in palticula:: (section 1-63' see also chapter' 2). Mechanical distunbance was therefone proposed as the pnime causal factor, since the distunbance of soils and plants by stock would necessanily have an effect independent of the dietany pnefenences. Soil distunba¡ce was the feature considere<1 in depth. The

compa:rison of equivalently stocked sheep aDd cattle pastures, with a more on less equivalent levef of distunbance, indicated l-ittl-e diffenences between the long-te::m outcome of the two stockíng negimes, as was expected fnoni the hypothesis. Howeven, the findings d.id not completely rule out the possibility of divergences due to diffenences in gnazíng p:references. Refationships between plant distni-butions and soi'l- stability r¡Iene examined in more detail 204

(chapten 6), and tÌrc data from the precedí-ng chapten ne-examined, but again the changes unden stocking did not coincide pnecisely with changes predicted to folÌow soil- dis-tu::bances. Hence, although the resul-ts indicate th-at the pneferential se-l-ection h5rpothesis is rmsatisfactony, the valÍdity of th-e distunbance h-ypothesis remains equivocal. Nonetheless, the findings on the wh-ol-e concull with Heady I s (fg0+) statement, that

ttvegetation changes seem to be cornelated more with the intensity of r.ange stocking and use than with p:refe::ence, notwithstanding a widespr:ead belief that changes are due in lange measune to sefectionrt.

B-6 PASTORAL IMPLICATIONS

The investigations of this thesis deal-t with nepnesentative situations. The panticulan stations concer"ned had stocking histories conmon to most in the South Aust:ralian a::id sheepÌands, and wene stocked at a rnedian leve.I fon their" par:ticuJ-an districts.

The pastunes examined wene typicaÌ of those most likely to be switched to cattl-e, both in their stnuctune and composition. The outl-ine of nesults th-erefone can be considered indicative of the likely effect of cattl-e intr.oduction el-sewhere, although the detail will vary in accordance v¡ith the particular- ci-rcumstances.

The i¡vestigation did not consi.den the question of optimum r:ates. Rather", the outcome unden tlre usual r"ates set by the nange manager?s was sought, wheth-er" or: not these repr.esented the optimum. 205.

It must be realised th-at stocking levels ane set on the basis of o

Ttre managerrs assessment wil-l- 'bhus determine the outcomes fon much of the foneseeable future. In this explorator"y study, the:refone, only the consequences of the most usual assessments have been considered. As regar:ds the r"esul-tant condition of the pastune, the findings imply that neithen advantages nor dísadvantages will accrue in the long term fnom the replacement of sheep-stocking by cattle-stocking; the end r"esult wil-l- be the same as woul-d occun un

did not show any ch-ange j¡r ovenall perelniaf cover, alth-ough th-e dominant species wene appa:rently sJ-owly being ::eplaced with oth-en perennials (the panticulan instance llas' Koe'tía sedtfoLía, sJ-owly

decreasing, and K. pyramüdata, slowly increasing)' Centain of the statíons d.iscussed in this thesis have largely succeeded in :regenerating the per"ennial pastures, by a combination of spelling

and mechanica.l- treatments. These are, howevell' unusuaf situations' on the whole, the find.ings point to a gnadual attrition of the vegetation resources, which, because of the tong J-ife span of the species involved, mây not be evident except oven a very long period'

Thene appeans to be a short-tenm benefit of the switch to cattl-e, but this is equivocal. It depends langety on the mechanics of cattl-e-grazing and the pattern of seasons. In chapter 4, the

first r"egr-owth of rnajor fodden grasses was rePollted to be fuom old butts. Evidence was presented indicating that the cattl-e tended to.Ieave mo:re butts than did sheep; in the cattl-e caser

a component of biomass showed changes relating to stocking without

commensunate changes in density, but in the sheep case there was no change in biomass independent of the change in density. Hence the negrowth of grasses may be fasten initially under cattfe- stocking than under sheep-stocking, since there is not the time lag in production nequired fo:r germination and gnowth to an edibl-e size in the cattle case. However, ttr-is is a doubtful benefit; tÌrc¡e is pr.obably betten potential for production in a new rather th.an a senescent indivi.dual. Further, such- a benefit would only appfy for" seasons sucla as l-971--1974, whore pe::iod's without rain ane 207. too shor.t fo:: the death and bneakdovm of butts r"enaining f:rom the last gnowth, as woufd occulr in drougbt.

B_7 CONCLUSIONS

Such conside::ations aside, the formal- conclusions of this investigation ane as follows:- 1. The immediate impact of cattle on for"mer sheeplands is the fo:rmation of new patter^n in the vegetation. This

patterrr shows piosphere symmetny, similan in its basic

chanactenistics to slrmmetny in the sheep piosphene. The detail is, on the whol-e, diffenent to the detail unde:: sheep; in the species affected, the parameter"s affected. and the direction of ::elationships. 2. Ttre long-tenm outcome of cattle intnoduction is closely

similar to the ou-bcome of continued sheep-stocking. Per-ennial species decnease in abundance, theil? lregeneration is suppnessed, and ephemenals incnease in abundance. The particulan species invofved at?e repolrted to behave in the

same mannen unden sheep-stocking. 3. The cause of stock-nelated vegetation pattern and cbalge nemains uncertain. Preferential grazing selection of plant species can account fon some of the sho:rt-term patterns observed, but not fo:: the long-tenm changes. Distunbance, independent of the particula-n herbivore, caì

account fon much of th-e long-tenm ch-anges, but some changes 208.

and pattelrns stil-I nemain unaccounta-ble. Nonetlreless,

. Rrhateve:r factons a:re involved, they appear largely indep- . endent of the pa:rticulan henbivone. Conce::ning the natur:e of the piospher"e, two furthe:: conclusions can be added.

4. The piosphene concept, oniginally de::ived fuom obsenvations of sheep-stocking effects, can be val-idly extended to the

cattle case.

5. In both sheep and cattle cases, piosphere effects can onJ-y b obsenved in the distnibutions of a small proportion of pastur-e species. Implicit in the fir"st two concl-usions is the deg::adation of the pastunes in the South Austnaíian anid sheeplands. The deg:radation which can occun under. sheep has been well neponted. The intnoduction of cattl-e thus has the same potential. It is thenefore vital to deter"mine the basic factons, apparently common to both henbivones, which bring this about.

The piospher"e concept pnovides an excel-lent basis for the study of stock/plant intenactions. Nonethel-ess, it is evident that only

its bnoadest outlines a.re at pnesent known. Two points that would bean close investigation ane (a) the natune of the distnibution of gnadients within the vegetation, and (b) autecological studies of the species which, consistently show piosphene pattenn. Studies dealing with the fonmen poì-nt alre necessary to detenmine the true natune of the distribution of stocking pnessure within the piosphene. Without this knowledge, examination of the potential causes of 209. pattenrì must rîemain difficult. Studies dealing r¡ith- the fattsr point sb_oul-d considar, i-n close detail, changes ih soil st::ucture and nutr"ient status, as it appeans that these featu::es may be of g::eaten importance than gnazíng per se.

The economics of pastonal enterpnise arre at pnesent in a state of flux. Nonethefess, g::azing in the rangelands is likely to continue in the for.eseeabl-e future, particulanly as the demand fon animal pnotein grows. Both this a¡d past studies indicate that the gnazing by domestic ungulates, of plants which evolved in the absence of such henbivones, is nesulting in a genenal degnadation' In the past, this degradation has sometimes been veny napid. At present, the situation appears nefatively stable, Yet the indic- ations ar.e that, on the whole, a slow attnition of the vegetation resoullce is inevitaÌ,le under pnesent management systems. Inves- tigations ar:e requined, not only aimed at docurnenting the situation as this one has been, but also to determine the root causes of this decline, and to d.evise viable altennatives to the pnesent systems' B]BLIOGRAPFIY

ALBERTSON, F.W., RIEGEL, 4., and LAUNCHBAUGH' J.L. (fgSZ). Ecology of dr:ought cycles and g:razing intensity on gr:asslands of the centnal Gr.eat Pfains. EcoL. Monogz'. 27, 27-44'

ALEXANDER, G .I., and CARRAILL, R.M. (1973). The beef cattfe industry' pp143-170 in The Pastov'aL Industz'ies of At,r"stralia, G' Alexanden and 0.8 Wrll-r-ams (eds). Sydney Univ. Press, SYdneY.

ANDERSON, D.J. (ls71). Pattenn in desert Per"ennials. J. EeoL. 59, 555-560.

ANON. (1901) Royal Commission fon the New South I'lal-es Govt. Reponted in cArN (1962).

ARES, J.0., an¿ LEON, R.J.C. (L972). An ecological assessment of the influence of gr:ázing on plant community str"ucture. J. EeoL. 60, 333-

AUSTIN, M.P. (1968). Pattenn in a Zerna eTecta' dominated community' J. EcoL. 56, I97-2LB- (1971). Rol-e of regnession analysis in plant ecology' Pxoc EcoL. Soc. Aust. 6' 63-75.

and GRIEG-SI'1ITH, P. (1968). The appl-ication of quantitative me thods to vegetation surveY IT. Some methodological p:roblems of data fnom rain forest. J. Eco L. 56, 827-844.

BARKER, S. (1970). Quondong Station, South Australia: a field context fár applied nangeÌand research. TTaTts. RorJ, Soc. ,S. Aust._ 94, l-79-l-90.

., and LANGE, R.T. (1969). Effects of modenate sheep stocking on plant populatíons of a Bfackoak-Bluebush association. Aust. J. Bot. L7, 527-537.

., and LANGE, R.T. (1970). Population ecology of Att'ipLeæ sheep stocking. pp105-120 in The BioLoga of AtnipLeæ, --,undenR. Jones (ed. ). CSIRO Div. Pl-ant Industry' l-and BARNARD, A. (1969). Aspects of the economic histony of the.anid paátor"a1 industry. pp219-228 ín Arid Lands of Austt'aLia, n.O. Slatyer: and R.A. Penry (eds). Aust. National- Univ. Press, Canbenra.

BEADLE, N.c.l,i. (fg+e). The Vegetatíon øtd Pastt'tv'es of llestev'n IIsa South Wales. N.S.W. Dept of Consenvation and Govt Printer, Sydney. (1952). Studies in halophytes I. The germination of the seed and establishment of the seedlings of five species of Atr,í.pLeæ in Austnal-ia. EcoLogY 33, 49=62. -2I0- a1 I

BEADLE, N.C.W. (fgSS). Some aspects of ecological- r:esearch in semi- arid Australia. pp452-460 in Mongraphiae BioLogíeae vrrl: Biogeogrophy md EcoLogy in Austy,aLia. Dr" !f. Junk, Den Haag.

., and TCHAN, Y.T. (1955). Nitrogen economy in semi-arid plant commun ities I. The envir"onment and genenal- considenations. Pz,oc. Linn. Soc. N.S.W. 89,273-286. (1e70 BENNETT, D., MORLEY, F.H.I^I., CLARK, K.W., and DUDZINSKI, M'L' ). The effect of gnazing cattle and sheep together: ' Aust' J. ErptaL Agnie. Aním. Husbøtdry L0, 694-709.

BENNETT, J.8., KENNEY, F.R., and CHAPLINE, W.R. (1938). The problem: subhumid areas. pp68-76 in USDA Yeæbook of Agt'icuLtt'tz'e L938. U.S. Govt Pninting Office, Washington.

BLACK, J.M. (1943-1957). Elora of South AusttaLia. vols Î-IV. Govt Pninten, Adelaide.

BLEAK, 4.T., FRISHKNECHT, N.C., PLUMMER, 4.P., and ECKERT, R.E. (rgos). Pnoblems in antificial and natu::al ne-vegetation of the anid Shadscale vegetation zone of Utah and Nevada. J. Range Møtage. B,59-65.

BONSMA, J.C. (1949). Breeding cattfe for incneased adaptabiJ-ity to tnopical and subtropical envinonments. J. Agríc, Sci. 39, 2O4-22L'

BORDEAU, p.F. (1953). A test of nandom vensus systematic eco!-ögieaI sampling. Eeology 34, L¡99-5rr.

BOWES, K.R. (1968). Land settlement in south Austnalia 1857-1890. Libnanies Boand of South Austratia' Adelaide.

BOX, T.tr'I. , and PERRY, R.A. (1971). Rangeland management in Australia. J. Range Møtage. 24,167-17I. BOYLAND, ( ). Ecological and fl-oristic studies in the Sirnpson Desent National- PaÏk, south western Queensland. Pt'oc. Roa . Soc. ?Ld 82, 1-16. BROWN,G.D.'andHUTCHINSOI'I'J.C.D.(1973).Climateandanimalprroduc- tion. pp336-370 in The Pasl;oraL IndustrLes of Australia, G. Alex- anden and o.B. williams (eds). sydney univ. Pness, sydney.

BURBIDGE, N.T. (1945). Genmination studies of Austnalian Chenopodiaceae with special neference to the conditions necessalry for regeneration I. AtripLeæ Desicor¿a. Hewa::d. Trans. Roa. Soc. S. Aust. 69'73-8+.

CAIN, N. (f962). Conrpanies and squatting in the I'lestern Division of New Sout Wat-es, ]ASO-]SOS. ppa53-a56 in The SirnpLe FLeece: Studies in the Australiøt WooL industtg, A. Barnand (ed). Aust. National Univ. and Melbounne Univ. Pnesses, Melbou:rne'

CAMPBELL, N.4., and ARNOLD, G.W. (1973). The visuaf assessment of pu=i,-" yíel-d. Aust. J. EæptaL Agníe. Anim, Husbandry L3, 263-267.

CARRODUS, B.B., SPECHT, R.L., and JACKMAN, M.E. (1965)' The vegetation of Koonamore station, South Austnalia' Tt'øts' RoA' Soe. ,S. Aust. ' 89 , 4l-- 57 . 272.

CHARLEY, J.L., and cowLING, S.VI. (1968). Changes in soil nutrient status r:esulting fnom overgrazjlng, and their consequences in plant communities of semi-arid areas. Proc. EeoL. Soc. Aust. 3, 2B-38.

CHIPPENDALE, c.M. (rgOe). A study of the díet of cattLe ín CentraL Austxalia as deterwLned bg n)tnen sønpLes. N.T. Administration, Primany lndustries Branch, Tech. Papen f.

CLARY, W.p., and PEARSON, H.A. (1969). Cattl-e preferences for forage species in r¡or.thern Arizona. J. Røtge Møtage. 22, 114-117.

CONDON, R.W., and KNOI,,ILES, G.H. (fgSZ). Saltbushes. J. SoiL Consenuat. Seruiee N.S.W. B, l-49-157.

CONDON, R.W., NEWMAN,J.C., and CUNNINGHAM, G.M. (fS0g). Soil erosion and pasture degeneration in Centnaf Austrafia I. Soif erosion and degeneration of pastunes and topfeeds. J. SoiL Conset'U. 25, 47-92.

COOK, C.W., HARRIS, L.E., and YOUNG, M.C. (fgOZ). Bota¡ical and nutni- tive content of diets of cattfe and sheep unden single and conrnon use on mountain nange. J. Anim. Sci. 26' 7169--7174.

CROCKER, R.L. (t946). The Simpson Desert expedition l-939. Scientific nepo::t no. 8. The soil-s and vegetation of the simpson Desert and its borders. Ttøts. RoA. Soc. S. Aust. 70' 235-258'

., and SKEI^IES, H.R. (1941). The pnincipal soil and vegeta- ----tTõñ r"efationships on Yudnapinna station, South Australia. Tr'æts. EoA, Soc. ,9. Aust. 65, 44-60.

DIXON, S. (fegZ). The effects of settl-ement and pastonal occupation in Austr-alia upon the indigenous vegetation. TnøLs. RoA. Soc. S. Aust, L5., t95-206.

EICFILER, Hj. (1965). Supplement to J. M. Bl-ackts Flora of South Austra- Lía, Znd Edition. Govt. Pninter, Adelaide. cn GALT, H.D., THEURER, B. EHRENREICH, J.H. HALE, W.H., aad MARTIN, (fgOg). Botanical-' compositTon of 't-he' diet of steens grazing a desert grassJ-and range. J. Røtge Møtage. 22' I+-L9. cENTILLT, J. (ed) (1971).. CLímates of AustraLia and Na¿ ZeaLavñ. VoI. l-3 of llot'Ld suï'ÐeA of cLimatoLogy. Amsterdan. (1972) , Attstz'aLian elímate pattezns. Nelson , Melbourne.

GIBBS, W.J. (1969). Meteonology and climatology. pp33-54 Ln tutid Lartds of Austz'aLia, R.O. Slatye:r and R.A. Penr^y (eds). Aust. National Univ. Pness, Canberua.

GOODALL, D.W. (1952). Some considenations in the use of point quadnats fon the analysis of vegetation. Aust. J. Scí. Res. Ser'. B 5, 1--41' (1970). Statistical- pJ-ant ecology. Ann. Reu- EcoL. Syst" L, 99-!24 21.3..

GRIEG-SMITH, P. (1964). Quarltitatiue PLøtt EeoLogg. ::2nd edition. Buttenwo::ths, London.

, AUSTTN, M.P., and I^IHITMORE, T.C. (rS0Z). the appJ-ication of quant:'-tative methods to vegetation survey I. Assoc iation analysis and pr.incipal component or"dination of nainfonest data. J. EcoL.55,483-503.

and CHADtrIICK, M.J. (rg0S). Data on pattern within plant communaties III. Acacia - Cappæis sêmi-succulent scnub in the Sudan. J, EcoL. 53, 465-

HALL, E.A.A., SPECHT, R.L., and EARDLEY, C.M. (1964). Regene:ration of the vegetation on the Koonamone Vegetation Resenve 1926-1962. Aust. J. Bot. L2, 205-264.

HARKER, K.tr{. (1960). Defaecating habits of a herd of Zebu cattl-e. ITop. Agric. (Tv,ínidød) 37, l-93-200.

HARRTNGTON, G.N., and PRATCHETT, D. (rSZS). Cattle diet on Ankole ::angeland at diffenent seasons. Tt'op. Agrie. Irtnídnd) 50, 2tI-279.

HEADY, H.F. (fe6+). Pal-atability of henbage and animal preference. J, Røtge Manage. L7, 76-82.

. (1966). fnfl-uence of grazing on the composition of Themeda - -ffidîø grassland, East Afnica. J. EcoL. 54, 705-727. HEATHCOTE, R.L. (1965). Back. of Bourke: a stuáy of Land øpprai.saL artd settl,ement i.n semi-ayid AustraLia. Mel-bounne Univ- Pness, Me-Ibounne. (1969 ). Land tenr:::e systems: past and pnesent. ppJ-85- 208 in Lø¿ds of AustnaLia, R.o. SJ.atyen and R.A. Perr-y (eds). Aust. National Univ. Press, Canbenna.

HERBEL, C.H., ARES, F.N., and NELSQN, A.B. (1967). Grazing distr"i- bution pattenns of Hereford and Santa Gentr.udis on a southern New Mexico range. J. Rønge lúanage, 20, 296-298. HILDER, E.J. (1964). The distnibution of plant nutrients by sheep at pasture. Pt'oc. Aust Soc. Aním. Prod. 5r 247-248.

HUTCHINGS, S.S., and SCHMAUTZ, J.E. (1969). A field test of the nelative weight estimate method fon determining he:rbage pnoduction. J. Range \tløtage. 22, 408-t+71.

IVIMEY-COOK, R.8., and PROCTOR, M.C.F. (1967). Facton analysis of data fnom an East Devon heath: a comparison of pninciple component and notated solutions. J. EcoL. 55, 405-419.

JACKSON, E.A. (1958). A study of soiLs and some aspects of thehydro- LoW of Yudnapinna station, South Austz'aLia. CSIRO Soi-l-s and Land Use se::ies, No. 24.

JESSUP, R.W. (1951). The soil-s, geology and vegetation of northwestern South Australia. Trans. RoA, Soc. S. Aust. 74, tB9-273. 214.

,IONES, R. (ed) (1970). The bíoLogA of AtnipÌex. CSfRO Div. of Pl-ant ïndustry , Canber"ra.

KERSHAVI, K.A. (1968). Classification and o:rdination of Nigenian savannah ' vegetation. J. EcoL 56, 467-482. . (rgzg). Quøttitatiue and dynøníe plant ecoLory, 2nd edn. Edwand ArnoJ-d, London.

KNOI'/LES, G.H. (1951). Some westenn pastune species--Bl-uebush, cotton- bush and othe:r pe:rennial ehenopods. J. SotL Consev'uat. Sez'. N.S.rn. 7, !80-798.

LANGE, R.T. (1965). Growth ring cha:racteristics in an arid zone conifer. CaLLítnis eolwneLLarzls nea¡ i^loomera. TTarts. RoA. Soc. ,9. Aust. 89, 133-137.

. (1967). Nitnogen, sodium and potassium in foÌiage firom some anid and tempenate zone shrubs. Aust. J. tsioL. Scí. 20, !029-1032. . (f,g0e). Infl-uence analysis in vegetation. Aust, J. Bot. L6, 555-564. . (1969). The piosphere: sheep tnack and dung patter:ns. y' Rnage Møtage. 22 , 396-400. . (1971). fnfluence analysis and presc::iptive management of rangeland vegetations. Proc. Eco L Soc. Aust. 6, l-53-l-58.

LAYCOCK, I,.i.A., BUCHANAN, H. , and KRUEGER, I^/.C. (1972). Three methods of detenmining diet, utilization and trampling damage on sheep r?anges. J. Range I'|øtage. 25, 352-356.

LEIGH, J.H., and MULHAM, I,I.E. (1.965). PastoraT pLøtts of the Riuerine plain. Jacananda., Melòounne.

and (fg00a). Sel-ection of diet by sheep grazing semi-a::id pastur es on the Riverine plain I. A bladden saltbush (Atr"tpleæ uesicav,ia)--cotton bush (Kochi,a aphyLLa) cornmurrity. Aust. J. Erptal Agz,ic. Anim. Husbøtdry 6, 460-467.

., and . (1966b). Selection of diet by sheep gnazing semi-arid pastunes on the Riverine plain II. A cotton bush (Kochia aphy LLd --grasslan d (S t¿pa-Danthonia caespitosa) community. Aust. i. ErptaT, Agríc. Aním. Husband.ry 6, +68-474. ., and (1967). Sefestion of diet by sheep gnazíng semi-alrid pastur"es on the Rivenine plain III. A bladder saftbush (AtripLeæ uesícarLa)--pigface Oisphymq. austraLe) community. Aust. J. ErptaL Agrie. Anim. Husbøtdry 7, 42I-425.

., and NOBLE, J.C. (1969). Vegetation resources. pp73-92 in Az"Ld Lætds of Austz'aLia, R.o. Slatyen and R.A. Penry, eds. Aust. Nationai- Univ. Pness, Canberna. 2!5.

LEIGH, J.H., ItrILSON, 4.D., and I'{ULHAM, V¡.E. (fgOe). A study of menino sheep grazing on cotton bush (Kochía aphgLLù--grassland (stipa uariabilís-Døtthonia caespítosa) communtiy on the Riverine pì-ain. Aust. J. Agric. Res. 79, 947-961-. lirCHrrnlD, W.H. (1962). SoiLs of the Aliee Sprtngs anea. CSIRO Aust. Land Res. Sen. No. 6.

LOCK, J.M. (t}lZ). The effects of hippopotamus gnazing on gnassfands. J. EcoL 60, 445-467.

LOW, W.A. ¡972). Community preference,by fnee nanging shortho:rrrs in the Alice Spnings distr"ict. Aust. Soc. Anin. Prod. 9, . (1973). Feeding interactions of r:ed kanga:roos and cattle in an anid ecosystem. ( e papelr given at III Íio::J-d Confenence on Animal Pnoduction, at l,lelbourne, l-973. )

LYNCH, J.J. (fgOZ). Ranging behavioun of sheep in lar:ge aI'eas. Pz'oc. EeoL, Soc. Aust. 2, 167-169.

(1973). Menino sheep: some factors affecting their distnibution in veny large paddocks. (Pnoceedings, conference on Behavioun of Ungulates and its Re.l-ations to ManagementrCalgany, f971). Publ-. Intennat. Union for Conservation of Natune, Mor:ges, Suisse.

., and ALEXANDER, G. (1973). Animal- behavíour and the pastonal industries. pp371-400 in The pastoz'aL industries of AustraLia, G, Alexander and O.B. t¡lilliams (eds). Sydney Univ. Fress, Sydney.

LYON, L.J. (fgOe). Evaluation of density sampling methods in a shnub community. J. Range l,Iøtage. 2L, 16-20.

MACFARLANE, I,ü.V. (1964). Temestrial- animals in dny environments : ungulates . Am. Han&ook PVLysioL. 4, 509- . (fg6g). Pnotein fnom the wasteland: water and the @a1 ecolgy of ruminants. Aust. J. Sci. 3L' 20-30. ,, MORRIS, R.J.H., and HOIIAR-D, B. (1956). Ttre waten economy of tnopical men ino sheep . NatuYe 1-7 B , 304-305 . , HOWARD, B, and SIEBERT, B.D. (1967). Waten metabo- l-ism of mer^ino and Border Leicesten sheep g::azing sal-tbush. Aust. J Agni.c. Res. 18, 9+7-958.

McINTYRX, G.A. (fgSa). Estimation of plant density using line transects. J. EcoL. 41, 319-330.

MACONOCHIE, J.R. , and NELSON, D.J, OglZ). Regener.ation of vegetation in Central Austnalia. Pastoral Reuiea L972' ppJ-6*17.

MADTGAN, C.T. (1936). The Australian sandridge deserts. Geog. Reu. 26, 205-227.

MANNETJE, L.tt., and HAYDOCK, K.P. (l-963). The dny weight rank rnethod fon the botanical analysis of pasture. J, Bz,it. GxassLd Soc. L8, 268-215. 21.6.

MARSHALL, J.K. (1973). Dnought, land use and soil enosion. pp55-77 in The enuiz,onmental, eeonomic øtd soeial signíftcance of dnougltt, J.V. Lovet'- (ed). Angus and Robentson, Sydney.

MEAD, R. (1971). A note on the use and misuse of regr"ession models in ecology. J.EcoL. 59, 215-219.

I'[EIGS, P. (1953). World dist:iibutions of anid and semi-a¡id homoclimates. pp203-209 in Reuieus of vesearch on avid zone hyfuology. UNESCO' Panis.

MOORE, C.Ìtr.E. (fgS3a). The vegetation of the south-east Riveni4a, New South Wales f. The climax communities. Aust. J. Bot. L,485-547. . (f953b). The vegetation of the south east RiveninarNew South l¡lal-es fI. The disclimax communities. Aust. J. Bot. 1) 548-571.

MOORE, R.M. (1959). Ecological obse::vations on pJ-ant communities g:razed by sheep in Australia. pp50o-513 ín lúonogz'aphiae BioLogicae VILI: Biogeographa ørd Eeology in Australia. Dn !'1. Junk, Den Haag.

MORLEY, F.H.I/tr., BENNETT', D., and CLARK, K.l^1. (1964). The estimation of patsune yield in lar ge g:razing experiments. FíeLd Sta. Rec. Dit, PLant Industtg CSLR) (AusÐ 3,43-47.

MURRAY, B.J. (fg3t), A study of the vegetation of the Lake Tornens . plateau, South Austnalia. Tr'øts. RoA. Soc. S. Aust. 55; 97-772.

MUTOOROO PASTORAL COMPANY (1951), rhß lh¿tooroo pastoraL corrpany after 50 geans. Adventiser Pninting Office, Adelaide.

NEI,IMAN, D. M. R. (fgOg ). The chemical composition, digestibi]-ity and intake of some native pastu::e species in Central- Australia duning winter. Aust. J. ErptaL AgrLc. Anim. Husbandng 9, 599-602.

NEI¡IMAN, J.C., and CONDON, R.W. (l-SOg). Land use and present condition. pp105-132 in Az"id Lqnds of AustraLia, R.o. slatyer and R.A. Perny (eds). Aust. National Univ. Press, Canbenna.

NOBLE, I.R. (fgZt+). T::end in rangeland vegetation. Pape:: given at the U.S.-Austnalia Rangelands Workshop at Aliee Springs, L974. (fn pl?ess' Society fo:r Range Management, Denven.)

NORTHCOTE, K.H. (fgOS), A faetunL key for the recogn¿t¿on of Aust?alion' soíLs. CSIRO Div. Soils Div. Repor:t 2/65. (rgzr). A factuaL key foz'the reeognítion of AustraLiø¿ soiLs, 3nd edition. Re l-tim Technical- Publications, Adelaide.

NOy-MEIR, I. (1971). Mul-tivariate analysis of the semi-arid vegetation in south-easter"n Aystralia: nodal ondination by component a]ìalysis. Pz,oe. EcoL. Soe. Aust. 6' 159-193. , (fSZ+¡. Iful-tivaniate analysis of the semi-arid vegetation in south-eastenn Australia II. Vegetation catenae and envinon- rnental gradients. Aust, J. Bot. 22' t1'5-t40. x7.

OGDEN, J.c., BROI,IN, R.A., and SALESKY, N. (1973). Gnazing by the Echinoid Dí.adema antiLLatwn Philipi: for"mation-of halos anound West Indian patch neefs. Science 182 (4113), 775-717.

ORLOCI, L. (1966). Geometric modefs in ecology I. The theony and application of some ordination methods. J. EeoL. 54, 793-275.

OSBORN, T.G.B. (f925). On the ecology of the vegetation of arid Australia I. Introduction and general descríption of the Koonamone reserve. Tt'øts. RoA. Soc. ^9. At'tst. 49, 290-296,

. , ll00D, J. G. , and PALTRTDGE, T.B. (l-931). 0n the'aut- ecology of 5típa nitída : a study of a fodder grass in anid Austnalia. Pv,oc. Linn. Soe. N.S.l,l, 56,299-324.

., and (rggz). on the gnowth and reaction to grazr_ng of the penennial saltbush Atz,íp\eæ uesica.?iun. Proc. Linn. Soc. N.S.l,/. 47, 377- ) ') and (1935). on the cl-imate and veget ation of the Koonamo::e Vegetation Resenve to l-931 . Pz,oc. Linn. Soc. N.S.l,,/. 60, 392-427.

PATTEN, B.C. Q972). A simulation of the shontgîass prainie ecosystem. SímuLation L9, t77-t86.

PECHANEC, J.F., and PTCKFORD' G.D' (rgsz)' A weight estimate method for the detenmination of nange on pastune production. ./. An. SOe. Agron. 29o 894-904.

and STEI,trART, G. (1940). Sagebr:ustr--gnass nange sampling studies: size and structr:r'e of the s ampling unit. J. An. Soc. Agron. 32, 669-682.

PERRY, R.A. (1967). The need fon r"angefatds neseanch in Ar:stnal-ia. Proe. EeoL. Soe. Aust. 2, 1-l-4. (1969). Rangelands resear?ch and extension. pp291-302 in Az,id Land,s of AustraLia, R.O. Slatyen and R.A. Penny (eds). Aust. Nationaf Univ. Pness, Canberna. . (fgZO). Ecological guidelines fon resounce management. Proe. EeoL. Soe. Aust. 5, !-tt.

PIDGEON, I.M., and ASHBY, E. (1940). Statistical analysis of negenenation following protection from grrazing. Pz'oe, Línn. Soe. N.S.W. 65, 123-r+3.

rIELOU , E. C. ( 1969 ) . Introduetion to ÌrlathematícaL Ecology . John Wiley and Sons , New Yor"k.

PREECE, P.B. (fg7la). Contnibutions to the biol-ogy of Mulga f. Flowening. Aust. J. Bot. L9r 2.t-38. . (1971b). Contributions to the biology of Mulga fI Germination. Aust. J. Bot. 19, 39-49.

RATCLTFFE , I'. N, (1936). SoiL drift ín the arid pastoral av'eas of South AustraLia, CSIRO parnphlet No. 64. 218.

RATCLIFI-E, F.N. (rg+z). FLying Foæ and Dxiftíng Søtd. Angus and Rober"tson , Sydney. RXID, G.K.R. (fgOe). Economic aspects of the sheep indust:ry in the pastoral zorre. A, Reu. AgrLc. Econ. 21, 36-49.

RICHARDSON, N.A. (1925). The Píoneers of the Northllest of South AustnaL¿a. T?romas and Co., Adetaide. Facsimil-e edition 1969, S. Aust. Libnanies board, Adelaide.

ROBARDS, G.E., LEIGH, J.H,, and MULHAM, I,l.E. (fe6Z). Sefection of diet by sheep gnazing seni-a::id pastures of the Rivenine pJ-ain IV. A grassland (nøtthonia caespitosa) community. Aust. J. EæptaL Agrie. Anim. HusbqnfuV 7, 426-433.

ROBEL, R.J., BRIGGS, J.N., DAYTON, 4.D., and HULBERT, L.C. (1970). Rel-ationships between visual obstnuction measurements and weight of gr.assland vegetation. J. Rætge Møtage. 23, 295-297,

ROBERTS, S.H. (fgzt+). Hístoz,g of Austz,aLiwt, Løtd settLement. Macmiltan Mel-bour-ne ; reissued l-968.

ROGERS, R.W., and LANGE, R.T. (1971). Lichen populations ou arid soil- cnusts anound sheep watening places in South Austr.alia.)íkos 22, 93-100.

LANGE, R.T., and NICHOLAS, D.J.D. (1966). Nitnogen fixa- t lon l-ichens of a::id soil crusts. Nature 209 (50L8)r 96-97. scHMrDT, DT (1969). Behauiour of cattLe ín a hot drg cLimate, Unpub. M. Run. Sc. Ttresis, University of New England, A:rmidale, N.S.W..

SCHMIDT-NIELSEN, K. (1964). Desez,t AnímaLs. Oxfond.

SEARS, P.D. (1949). Soil, fertil-ity and pasture g:rowth. ppl03-125 in Sheepfarrning ArwtuaL, Massey Agnicultunal Cotl-ege, New Zeal-and.

SHOI,IN, L.M., MILLER, R.F.,and BRANSON, F.A. C1969). Sagebrush con- ve::sion to gr"assland as affected by pnecipitation, soil and cultural pnactices. J. Range Møtage. 22, 303-311.

SOKAL, R., and ROHLF, F.J. (1969). Biometrg. Fneeman, San Fnancisco.

SPECHT, R.L. Q972). The Vegetation of SouttL Australia (znd edition). Govt Printen, Adelaide. SQUIRXS, V.R. , and HIIIDLEY, N.L. (1970). Paddock size and l-ocation of watening points as factor"s in the drought survival of sheep on the centr"a.l- Rivenine plain. WooL TechnoL. øtd Sheep Breeding L7,49-54.

., and V'IILSON, A.D. (1971). Distance between food and water suppJ-y and its effect on drinking fnequency, food and water intake of Merino and Bonden Leicesten sheep. Aust. J. Agr'íc. Res. 22, 283-290.

SYMON¡ D.E, (1369). A checklist of flowening plants of the Simpson Desert and its immediate environs- Australia. Tz'qns, RoU. Soc. D Aust. 93, 17-38. 2J9.

THETFORD, P.O., PIEPER, R.D., and NELSON, A.B. (fgzr). Botanical and chemicaf composition of cattl-e and sheep diets on Pinyon- Juniper" gnassland range. J. Rø'tge Møtage. 24, 425-431.

THOMPSON, O. (1971). The WooL Ctisís: a nationaL szßueA of the Austv'a- Liart uool industry. The Austnal-ian, Sydney.

THREADGILL, B. (1922). Sout?t AtætnaLiott Land EæpLoz.ation. Govt Printer., Adelaide; fon the Public Libraries Board, S.4.,

TRUMBLE, H.C. (fg+S). The ecological refations of pastur"es in South Austral-ia. J. Brit. GrassLd Soc. 4, 135-159

., and WOODROFFE, K. (fgS+). fnfluence of cl-imatic factors on reaction of desent shrubs ot grazíng by sheep. ppt29-147 in Biology of Hot and CoLd Deserts, S.L. Cloudsley-Thompson (ed). fnst BioI. Symposium, 1954.

VALENTINE, K.A. (fS+Z). Dístance firom waten as a facton in g:razing capacity of r.angeJ-and. J. Forest. 45,749-754.

VAN DYNE, G.M., and HEADY, H.F. (1965a). Botanical composition of sheep and cattle diet on a mature annual- nange. HiLgardiar' 36, 465-492.

., and (1965b ). Dietar"y chemical composition of sheep and cattl-e grazíng in common on a dr"y annual nange . J. Røtge Møtage. LB, 78-86.

VIALKER, B.H. (l-970). An evafuation of eight methods of botanical analysis on gnassJ-ands in Rhodesia. J. AppL. EcoL. 21403-416.

WARING, E.J.C. (1969). Some economic aspects of the pastonal industry in Australia. pp229-257 in Ar,id Lands of AustraLia, R.0. SJ-atyer and R.A. Penr"y (eds). Aust. National Univ. P::ess, Canbenra.

l[EfR, W.C., and TORELL, D.T. (1959). Setective grazing by sheep as shown .1 by a comparison of the chernical composition of range and pastu::e forage obtained by hand ctipping and that col-l-ected by esophageal fistulated sheep. J. Aním. Sci. 18, 64I-649.

WESTON, E.J., and MOIR, K.W. (f969). Gnazing preferences of sheep and nutritÍve val-ue of plant components in a Mitchell Gnass association - in nonth-westenn Queensland. QLd. J. Agric. Anim. Sci. 26, 639-650.

v,IESTON, R.H., and HOGAN, J.P. (1973), Nutrition of herbage-fed ruminants. pp233-268 in The Pastov,q.L Industv"ies of Australia, G. Alexander and O.B. lfilliams (eds). Sydney Univ. Pr"ess, Sydney.

VITEDEMANN, A.M. (L972). vegetation studies in the simpson Desert, Nonthenn Ter.nitony. Aust, J. Bot. L9, 99-12+.

I.lfGHT, J.R. (1967). The sampling unit and its effect on saltbush yield estimates. J. Range Møtage. 20, 323-325. 220.

VIILLIAMS, O.B. (1956). Studies in the ecology of the Riverine plain II. Plant-soiÌ nefationships in three semi- anid gnassl-ands. Aust. J. Agric. Res. 7' 127-139. . (1962). The Rivenina and its pastonal industry 1860- I 9. pp411-434 in The SirnpLe fleece: Studies in the AttstraLiøt WooL Indt¿stW, A. Barnard (ed. ). Aust. National Univ. and Mefbounne Univ. Pness, Melbourne. . (1968). That uneasy state between animal and plant in ---Elmanfpulated situation. Ptoc. EeoL. Soc. Aust. 3, 1-67-774.

!üILLIAMS, W.T. (1971). Stnategy a¡d tactics in the acquisition of ecologicat data. Pt'oc. EcoL. Soc. Aust- 6, 57-62. i,lILSONr A.D. (fsoo). The val-ue of AtripLeæ (saltbush) and Kochia (bluebush) species as food fon sheep, Aust. J. Agríe. Res. 17, 147-1s3. , and LEIGH, J.H. (1969). Use of water" intakes in estimation of saftbush intake by gnazing sheep. FieLd Sta. Rec. Dit). PLmt Indz,ætrg CSIR) (Aust.) B, 1-8. ., IfiIGH, J.H., and I,IULHAM, I,¡.E. (1969). Study of merino sheep grazing a btadder" sal-tbush (AtTipLeæ uesicoria)--cotton bus}: (Kochia aphyLLd community on the Rivenine plain. Aust. - J. Agrie. Res. 20, t+23-7136. WILSON, J.W., and LEIGH, J.H. (1964). Vegetation pattenns on an unusual gilgai soil in New South Wal-es, J. EeoL. 52, 379-389.

WIMBUSH, D.J. , BARROI^I, M.D. and COSTIN, A.B. (fg6Z). Colou:: stereo- photognaphy for" the measu:rement of vegetation. EcoLogg 48, 150-152.

IIOOD, J.G. (fgg0). Regeneration of the vegetation on the Koonamor:e Vegetation Resenve. Incns. RoA. Soc. S. Aust. 60, 96-111. (1937). Vegetatíon of South Australia. Govt. Pninten, Adel-aide.

YARRANTON, G.A. (1971). Mathematical nepr"esentations and modefs in plant ecolory: response to a note by R. Mead. ./. EcoL, 59, 22t-224. APPENDIX 1.

DATA FOR CHAPTERS 3 AND 4

(a) Density of pasture species.

(b) fxampl-es of photopoint series. Appendix 1.

The fol-l-owing gr"aphs show the mean densities of pasture species encountened in the Mt Vícton study. Species are arranged by fieJ-d numben as fol-l-ows.

L, Stipa nítída 2. Enneapogo4 spp. 3. Ar-'i,stida contorta 4. Kochía astrotr"Lcha 5. K. pyranid.ata (Resul-ts fon the above given in text) 6. Abutilon otoeatpum 3L, Kochía georgei 7. AttipLeæ Línbata 32. K. sedifoLía B. A. sporqiosa 33. K. tomentosa 9. A. stipitata 34. Mínuria LeptophyLLa L0, A. uesicaz"La 35. MaLacoeera, tr,ícov.'ne LL. Bassia bifLora 36. PtíLotus obouatus L2. B. decurz.ens 37. fuiraphís moLLis 1-3. B. diacantha 38. Rhago&La spineseens uar deltophylLa L4, B. &Loav,icata 39. SaLsoLa kali 15, B. Lanicuspis 40. Sídn conrugata 1-6. B, Límbata 42. S, intricata L7, B. obLíquieuspis 43, S. uiz'gata LB. B, patenticuspís 44. Helipl;ez'wn florLbundum L9. B. paradoæa 45. CLianthus forTnosus 20. B. seleroLaenoídes 46. Tnagus atætralianus 2L. DaetyLocteniwn 47. TribuLus terrestyLs TaduLøts 22. ConuoLuuLus erubescens 50. BuLbirte semibarbata 23. Dodpnaea microzyga 57. Angianthus pusiLLus 24. EnchyLaena tomentosa 52. CaLoeephaLus dittrichii 25. Eragrostis dieLsii 53, EremophiLa duttoníí 26. Erodiun cAgnorvm 54. Enagrostis Laníflora 27. Euphoz,bia drwmnon&Li 55. Chenopodiwn desertor'um 28, Brachyscome ciLiarLs 56. Bassia brachyptera 29. Kochía aphyLLa 57. Sida cornugata uar angn¿stifoLía 30. K. eæcansata DENSIIY plonts (S=10xt quodrot. L=20x? o ó o

ó{

c) {I (o t- ot rn

@ D { o b o g o f o o (')x

5 Àt o (Þ \¡ o) (.ll (n t (¡ l- (¡ E

DENSITY /quodrot (S=l0xl L=20x2 oO o

ó{

ln t0 :r m m È T' d B 0 cl P o x(a

Fq ¡\) o (o @ \¡ o) (n (^ UI (, t-' r- tJr

È CATTLE PADDOCK SHEEP PADDOCK

13s

g o 13s I' DL o a f f cr C' (\¡ (\¡ ox x ô¡ (\¡o I I J J ó 14S -.o 14S E o D f o q qf oX ox il I vl tn

17 DL 15s et o E fss f õ ø c,) o o c d ,¡ c ¡ ó o. o. õF- F z 16S õ lJJ z 16s o oul

Jqn JU¡ Ju¡ ¡¡ ¡¡l 1gt1 1972 r973 Ín2 1973 CATTLE PADDOCK SI.EEP PADDOCK

17s 17s d D g o ! q d & s¡ 18S 18S x (\a o x õ¿ I ôto J I J o E ct o E gl ú q) 19S ox ll 06 o út I 04 vl ! 19s I ó E d o E f æ d ú 20s qa 20s o o c c L d o- l0 o t- õ F. z ùD l¡J z o ot¡.¡ 2rs 2rS

¡¡ ¡¡l ¡¡l tu 1972 tm r!¡?¿ 1972 973 t974 CATTLE PADDOCK SHEEP PADDOCK 225 225 o1 02 o2

õ ct ! 23 1 D 23 1 d ol o qf f 02 q0 c\¡ (\¡ ox x (\ (\¡o il 2t, nol present il 21 L J J oL d p E o o ) l (t v x 25S 25S o ox ll il ú ú) o d o E d cr ú) u) o o c C ó õ o- o- t- t- 26S õ zõ z 26S tdô oul

10 c8 06 a1 5 02

Ju APT Ju Jul Jul Oct Apr Jul 0ct Jon aPr Jul tg't2 1973 1974 1972 :373 1974 PADDOCK CATTLE PADDOCK SHEEP 275 275

28S 28S

U d ò rc ct d = f cr q s,l 29 1 1 N 29 x x o o ô¿ c\¿ lt il J J æS d 30s e E õ é q) e x x o o t¡ tl (n t¡, 3rL 31 not Present 0 o rl d E E 32 1 d 32 not present d a qa ø 33 1 o 33 1 o c C d I o- À 34S f- 34S l- u't 6z z UJ oUJ o

Jon ¡¡l 0Ét Jôn a9t ¡l O

35S

04 35 not present

36 1 é 36 1 d L E E ct o g) ql 37S è¡ 37S ô¿x ()x ô¡ No I tl J J d , 38L é 38 1 D rit D ú o ) q) o x x o r0 39S o 39S il il Ø ú)

E 0¿ õ D ó a2 õ o qJ C tt I 40s 40s c 3 d è t- t- ñ ûz 125 z 12S l¡J oUJ o

r0 43S 43S

apr Jst JU apr Jr¡ Ju Oct .h¡ 'r97¿ r9æ 1972 rc,3 1972 'g,a CATTLE PADDOCK SHEEP PADDOCK

145 145 o ó D E o o q) e present (\¡ 45 not ô.t 4ss x ox ô¿o (\ I il J J ó d E o d) o ú 45S e X 45S o 9 il I Ø ('r) o e o E o o f, ql q t 175 ul c tr 175 o o o- ô f- l- õ 50 not present zõ 50s z l¡J otlJ o 5rs 51 mt present

Jut Oct apr ¡¡ ¡Jl Jon Apr ¡¡ aPr Jul 1972 1973 ß7a 1972 lfrr3 r974 CATTLE PADDOCK SI-IEEP PADÐOCK

ct o ! E E o 52S c,j 52S úl c, (\¡ N x ox o .ìl 1l J J é Þ qI x ox 53 1 o 53 1 il l¡ vt ln

ég E D E ct o f q q (n c o o- 5/. S 5/. S 0 t- l- õ zõ z UJ g o 55S 55 not present

s6s s6s

57S 57 not present

¡l Apr Jsl }j oct Jorl aP. 1972 1973 gt4 1gt2 r9æ r97¿ Appendix 1. 'Ihe following a:re exanples of the photopoin'E series discussed, in chaptens 3 and 4. In each series, only one member: of the steneo pair is shown. The photopoints ane:

CATTLE

PADDOCK

Photopoint 4: 1.4mi north of Bl-ack HiIl Dam. Koehía astno-

tv,icha (l-ow bushes), Enneãpogon species (grasses) and '(Iange Ez,emophila duttoniL bush). The fence in the left hand colrnen is excl-osure 2.

Photopoint 2Oz 2.5mi nontheast of Black HiIl- Dam. Kochia

astrotrLcln (low bushes ) , Enneapogon species (l) , Av'istid.a, eontorta (B) an¿ Stípa nítidn @).

Exclosure 2: 1.4mi nonth of Black Hill Dam. Kochia astrotz"Lcha

(Iow bushes), Enneapogon species (g:rasses).

SI-IEEP

PADDOCK

Photopoint 1: Emu Leg Gate. Kochia astrotricha (tow bushes), Enneapogon species (gnasses), Heterodenfu'um oLeaefoLiwn

(tnee, centne ) , Acacia anel,tra, (tr"ee, fa:r night ) . Note gr.owth of Stipa between !3/2/74 and !/B/14.

Photopoint 10: 1.Bmi west of gate. IØchia astyotnicha (Iow bushes)

Ertneqogon species (short gnasses), Stipa nitida (tatt grass).

Excl.osu-ne 1: 0.1mi east of gate. Kochía astTotrLcha (fow bushes)

Enneøpogore species (A), Stipa nitída G). The tnee in the fan connen is a HeterodertÅ.rum.

Relative changes in apparent biomass fon these photopoints ane given in figunes 4.19 to 4.22 incl-usive. cP4 7-7 ,72

;."*r ¡t,O

CP/+ 1'9'72

cPA 1 .10 ,72 -- ''.::.--'-::å--:._.- '- - ^- - -j=----t ,- ,. ðÌ, 5 t:.' cP 4 2B ,11.7 2

al.- -ff -¿.4. ,

cP4 13,2,73

-r..¡ +t

CP/+ 1,11,73 r,\:-r i ¿ar - n , sfì6: -r?, ¡Ê*

.jâ I ¿ ;r.!-, .ir* ¡L. cPA 13.2.74 i6 . -.'l *'. r"_. : .¡r.;a,.Þ¡,úrùî:{s rT1.þ,r11t -.', *iÈÊ,<¡r¡à-.1i.:¡¡¡l-

qii'.. h .$ii;,t;' '-'.Ìr.r, ,', rå,, ,;ë

?,¿,:,"-.

.l , ù.!

cP20 28,t,.73 cP 20 u,.7 .73

l",l \i'

.1)!r|tr_¡æ: Í I;f {*- "$- ç. cP20.1 3,,. 8.1 0z d) I *'t.

,t.. 11

tÐpf:: CP znd N 4-9 '72

tátìÌ

rJETrätËlÌ,:: CP2ndN 1.10.72

¡.. {t; $f

.i r !:t 'r... -

,2'.73

,Lå-

F,+

,{F ': f:trl :.s; .,. ndN 2/, ,7,73 æËklT3'r?: 1 .rr "-

f r..

1.11.73 . i1+..* CP2ndN t" t- \ "':i l ,'þ t .i - ¿1¡r_å- '*s-. ' " i¡r-,- rÉ5--;",;'

CP 2nd N 13.2.74 CP Znd 1,9,74 y ! {,.ej(*_ N sP 1 4-9.72 lil ,l

5P1 5,10.72 SP 1 t: : ,fr'¡) 2841.72 .l I I (' ',:.

'.! //

,t, l.t

.lt

.\, a i-. {'. "ï

SP1 13.2,73 t .lì

.{:

r*/\ i

+, ,.1 SP1 ,ì 28 ,/r,T 3 I I SP1 1 24 t7 ,73

r\17 SP1 1 .11.

I SP1 I 13 '2'T+ ,1,9, I l,dS s¡ tlt + ¿i;tr1ytþ îd,(# i:.

j. ';- ¡- i.'

sP10 13.2 'r ì¡ 'Ur ..'l. ¡ "ò

-=E

sP 10 28.4.73

- tlr,¿ ' .l/: Þf,,'19

P10 13

\ .jtr

l¡r.:!.,rr 73 "t,SP¡J

I ,;J.. í, ú

¡". .1¡.: r t¡t Þ. { +).- {- v- ra ,t"

:t *'Åtr' ¡r"

'?l¡¿¡ ,J*ltl "t ;t

..r,,i'lù"¡, ,. i: ".,

1,7 3 ' l'1.,

' - - li .

E 13.2.7.

*,i il ,r"É i' SP lstE 1'8.74 .Ê?Àits APPENDIX 2.

DENSITY DATA FOR CHAPTER 5.

(a) Distnibutions of species in the vir"gin/cattle sunvey.

(b) Distnibutions of species in the cattle/sheep sunvey. Appendix 2.

(a) Distributions of species in the vingin/cattle survey

The foll-owing ane maps of speciesr abundances in the stud.y areas on Pennatty station. Speciesr maps are anranged in three sets with diffening scales. The finst set comp:rise those species fon which fnequency was estimated: the scale numbens nefer to the nurnber:, out t of ten, of 2xLm' quadnats in which the species oecu::r:ed. The nemaining sets are density maps: two scales rôrere necessany becâuse of the wide vaniations in abu¡dance. Species ane anranged alphabetical-ly within sets. Density counts ar:e given as plants per: quad:rat. Bqssio brochyptero Bossio intricqtq

o o I o o o oo vrRGrN -6õm VIRGIN 250m h o I o o o 3 o o o t o o o 5 o o O oo o oooo 7 o oo o o oo oo oo !ooo o CATTLE e- 'n E-om 250m Erogrostis oustrolosicq Nitroriq schoberi

o a o fo ooo fo vtRGtN în vrRGrN 250m ^ -2som ! o o

o ? o d o o o o oa o 5 o o o oo 7 oo ooo o CATTLE eA- CATTLE .n 256,

-250m Portulocq olerqceq Zygochloo porodoxo

o OO O o o oo o o vtRGtN vrRGrN 1r, în 250m

o o o o o o o ? o o o o 5 o o o o OD o O O 7 a o CATTLE .n CATTLE .¡- 25õm ZSO.' Abutilon otocorpum Aizoon quodrifidum

a o o o

a a a

o a

a a

VI RGI N vtRGrN 250m în 75õm

a ¡O o a

a a 1 5 o a o 10 o 20 o a o 50

100 CATTLE Ê^_ o cAîì'_i .a- " 250m ,so. Aristidq contortcl Atriplex inf loto

a ) o a oooo o. a a O.o a a a.O o a o d. o fo oo or O. fo Or a o ) o OO a) Ooo .Oo ..4 o o a oooooo ooo o a o Ooo O n a I a o o '4. vlRGrN zsom vrR GtN 1n ^ 250m

a O a

a U Oo .a o o .O oO 1 Oo oof å 5 oo o o l0 oo. o o o o 20 )o oOoo O. oo o 50 O .a 100 oo a a c ^tÌr o CATTLE 'n " 250m 75om F

Atriplex timbqto Atriplex spongaoscl

o a

a a

o oo o o oO a o o a. a o o o o a a o vlRGrN 1^ vtRGtN în 2Eõ-m 25Onr

l o 5 o a o o 10 o 20 o 50 .O a o )o o O o oOO. 100 o ooo o €ñ CATTLE n o CATT LE - 25C"''l 250m I

Atriptex vesrcclno Bobbogio ocroPtero aof o a o o o a o o a a DO a DO Oo Oo OO a a o?. a o o c o a a a o o o o a o

o o o .f e a o a o vlRGlN f^ 'l RGIN A ' ==^-lSun' -250m o o a a o a

,1 o o 5 o a o 10 c 4.. O. o 20 a a a a 50 o a.

100 o a o a io' o €; .Alflç O CATTLE '' uF L! ^ lSum - 250¡r - á

Bqssio divqricqto Bossio diocontho

a oo o

VIRGIN în vrRGIN A 75Om -250m

a o o o o 1 o å 5 a o a oo o o 10 a O a o ?0 o o o 50 oo o a I a o a o o a o a 100 o a o O r^TTl F o CATTLE 'n 25om !H¡,LL n - 250m I

Bossiq obtiquicuspis Bossiq porodoxo o

vtRGtN în vrRGrN ,f,, zsom -z:om o o o oo a 1 å 5 a a o 10 o 20 O oo o ooa 50 o o D. o o Oo 100 a o a oo CATTLE ." a F o zso. ^-Tt 250m UU9¿ urog¿ U g j-rlr-vl ---à 3]IIV]

a a a 00r a o o aaaa a 0ç o a a a a aaoaaa a 0z o a a a a 0l o a a a a I I ? a

utOg¿ u.r09z N tCU l^ u¿ Nle Ut^ "/ o a a a

a aaao a

a a a a a

a a a

a a a a a aa

a aa a a a a a

o a aaaa a snrÐlDltnbeo snlorqod.ro3 srJDrlrf, Ðr,lJof,sÁqlDJg Citrullus lonqtus Crotoloriq dissitifloro

o D o

a o a o o o

o a a o

o o o vtRGtN în vìRGIN 1" 250m -250m

a 1 å 5 a o 10 o 20 a o 50

100

.n o CATTLE 'n 25om CATTLE 25Om a o

o f o u o c A) o c o o ol o E o o- O E ro o o c! t.r) o c\¡ o a c a l. c ., l¡J s-c o LiJ) z a l- (D a t- É; ()

-1. P R 3 3 .ooooo ! ! a a o a O a o a a

o a a o o c o o oa ! J a ñf- t o a 3o Ëe o o o O o o d O a o a a a cÞ a e =o o o o 3: co o o o 0) a a a E ô O õq O .o. LO a c\ loIE o o o ln oo ã8(J .l- o lôl o o c o o o oå I o --c :o o o t! t-) a O a z :: F_ O a o a () o a É- o o I ¡ Erogrostis dielsii Euphorbio drummondii

o oo a o o o o a DO a o oo o OO a o o O. Oo !o oooo o a t a 4.. a a o oo a o a o o oO o vlRGlN 1^ v IRGìN A zmn, " ===-l5um

o D oOo o O I o O oO' ''o o o o OO. o 1 DO o 5 oO oo o ra ooo o o 10 a O D a Oo o .Oo OO o 20 o o o 50 o o o o a 100 ç o o CATTLE 'n CATTL E o 250m n 250m Kochiq ctstrotrichq Kochio george¡ o a a

I o a o a a

v r RGrN vtRGtN 1" 1" 25om 250m

o 1 å q o o 10 o 20 a o 50

r00 CATïLE .n o CATTLE .. 25Om zso, Kochio pyromidoto Kochio tomentoso

a a.

a a

o a o o a o

a vrRGrN în vlRGlN 1n ZEõm

a 1 o å 5 a c t0 o a o 20 a a a o 50

100

1- CATTLE n o CAITLE 1r 25õm ZO; ur09¿ ur u 09¿ -!à 3-l11vf l-lLLvl a O Oo 00r a o a o o 0ç a a o o. o a a o oaO 0z o a a 0r c a I o I a

trr09¿ ItrOç¿ t1 NtCUt^ "l Nte UtA roo a Ooorf o a o a o .a a a a a a

O. o ) a a a D O o oO o a a o ooO o a a oo a a O o a a a - ol aoaa a o a Oo Oo ' a O a a rO o o o o a a a a a a

srnuÐl DtuJof,Áqf,Dd [uDnls snlDqdÐf,otJ^hl Ponicum ef fusum Phyltonthus f uernrohiii o a o o a a O. a a O C a o o t t o a o o o a a o .o Oo a oO o vlRGri.r vrRGtN în A 25Om -250m

a o o O , a a o o a o a 1 o o 5 a å a o O o 10 a o a o 20 ¡OO o a o 50 Oo o a o

o 100 CATTLE -n o CATTLE .¡- 25Om ZSO. Sido virgoto cf. Sisymbrium (DEAD MATER IAL ) ¡ o o a o o o o o o a

a I o o o o o a lo a o o a a o o a vrRGlN vrRGrN 2Sõm h 250m ^

a a o o a O. aa a a a o Oo oo o a a c ,l oOo O a a O 5 a aa E c 10 o Oao o 20 o o o o o o 50

'100 a CATTLE F;- o CATTLE €n ' 250m 250m Trirophis mollis Xqnthium spinosum

a o a

o a o vrRGtN 1" -Sõm vtRGtN în 250m

O o o a

a t 'I a å 5 c 10 o o o )Ct o o 50 o o 100 CATILE .n 250m o CATTLE " 250m Appendix 2.

(b) Dist::ibutions of species in the cattle/sheep survey.

The following ane maps of speciesr abundances in the study areas' on Roxby Downs station. Speciesr density maps are an::anged in two sets because of the wide llange of abundances. Species ane annanged. alphabetícatly within sets. In all cases, counts are given as plants pen quadnat. Aristido contorto Enneopogon ovenoceus o o o t oo oJt oo o o OO o oOo o o o O o o I o o o o o o ooo o o o o o o oo o o o Oo oo oo I o o oo a oa) oO o DO o Oo oooaooo CATTLE 1_ cATTLE |n r¡ 25om 250m

o 1-50 51- o r00 101- o 200 201 - o 500

) 500

SHEEP în sHEEP fn o 250m 250m Goodenio cYcloPtero Trogus oustrolionus G. Pinnotifido ooo o O O Oo oOo o o o o oo oo OO oo o o oo o o a oo oo o o oo OO Oo O rao o ooo CATTLE t_ CATTLE ln r¡ 25om -250m t rttt o ! o o o ooo o oo O o o 1- 50 o oo ooo o 51- o r00 o 101- o o 200 oo oo o 201 - o Oo oo o o s00 o oo fo )500 o oooo SHEEP În SHEEP n o 250m 75õm Abutilon otocorPum Aizoon quodrifidum

o a O. o o oo o o a o Oo a o oooO a o I o o o ..O o oo ¿ a a OoO CATTLE t_ cATTLE În r¡ 25om

-250m a O ! o a o o a a Oo 1 o å 5 o o o 10 20 o o o a Ooo o o o o oo o o 50 a. o a ooo 100 o a o o SHEEP 1n SHEEP 250m 250m Atriptex infloto Atriptex timboto

a o o OO. ' o oOo o ' o o oo. o o o oo o o a ooo a a Oo o oo aa a a o Oo

a o CATTLE 1_ cATTLE 1n r¡ 25om

-250m a

1 O o 5 o 10 o 20 o 50

100

o SHEEP 1n SHEEP ln 250m 250m Atriplex spongioso Atriplex ves¡cor¡o

o o ooo a a a

a o

CATTLE În CATTLE 1_r¡ 250m 25om

o o a o o a o 1 o 5 o o 10 o 20 DO O o 50 o o ö r00 o o SHEEP SHEEP n În 75om 250m t

Bobbogio ocroptero Bossio brochypterq o o o o

a D

cATTLE 1n CATTLE 1_ 250m rn 25om

a o a o I o 5 o

o 10 a o 20 o 50 a o 100 SHEEP SHEEP 1; 1n 250m 250m t

Bossio decurrens Bossiq diocqnthq o o a o o o a o o a o a o o a a o a o o O a o o a o a o o o f O a o o aa a OOo o o CATTLE CATTLE 1_ 250m rn 25om

a O o o o o

1 a o o a o 5 o 10 o o o a o o ?o o o o o lo o a o 50 a a o o 100 a o o o SHEEP SHEEP î. ln 250m 250m ú

Bqssio divqricqto Bossiq intricoto o o o o o o a ) a o o a o o a o o DO o O.o o o

cATTLE ln f5õm CATTLE 1 l¡ 250m o ! t)o

a o 1 å 5 a o l0 o o 20 a o o 50 a I o o 100 o O. o SHEEP T o SHEEP 250m În 250m n

Bqssiq obliquicuspis Bossiq porodoxo

o a I

I a

a o cATTLE ln CATTLE 250m În 250m

o a o o a 1 å 5 o o 10 o o 20

a o 50 a o a

100 o

srEEP o I în SHEEP In 250m 250m Bqsslo potenticuspis Bqssiq sclerolqeno¡des o o1, I t o o o ao a

a o cATTLE |n CATTLE 1_ 2 50rr rn 25om !

a a o 1 å 5 a o o o o 10 o o ool o 20

o 50 o o o o 100 o sHEEP fn SHEEP în 250m 250m Citrullus lonotus Ctionthus formosus a o

a a

a o a

o a a

.O o o CATTLE ln CATTLE 250m 1n 250m

o a o O. o ) a o 1 o o 5 o a o 10 o o 20 a a o .o a 50 a o a aaa o o ..O o o o 100 SHEEP SHEEP T în 250m 250m Crotolorio dissitifloro Erect Prostrqte o

o a o a o o D a a a O o o o o oo o oo o o a o o a a cATTLE |n CATTLE 1_ 250m rn 25om o a o

1 å 5 o 10 o 20 o 50 o a o r00 sHEEP fn SHEEP În 250m 250m Doctyloctenium rqdulqns Enneopogon cylindricus o o o a a o ö o a o o o o o a CATTLE 1n CATTLE 1 25õ- rn 25om - U

a 1 o o 5 o o o l0 a o o o o 20 o o Oo o Io o 50 o o o o o o OO'ot 100 o SHEEP o T SHEEP În 250m 250m Erogrostis dielsii Erogrostis eriopodo o o o a o o o o

o o a a o cATTLE 1n CATTLE 1_ 250m rn 25om

I O 5 a o 10 o 20 o o 50 a o a o a 100

o SHEEP SHEE;) În 250m 250m Erodium cygnorum Euphorbio drummondii

a t o o a a ! ! Ooo o o o o o o o a o oo ! oOoO o o a a Oo o oo o oo a. o . . OO. o o Oo O .OO o fo oo o ooo O. O. a a oOooO O' oo CATTLE CATTLE 1_ în zsom r¡ 25om O.. oo a oo o oo a o 1 o o oo å 5 t o O o l0 a O Oo 20 o a oo O o OO. o o o 50 o o O ot Ooooo oo O.O r00 o' Oo a o o oO o o SHEEP SHEEP 1n 250m 250m Kochio qstrotricho Kochio erioclqdo o o

oa a o

CATTLE CATTLE 1_ În r¡ 25om -250m a

a oo a I o 5 a o o 10 o 20 a a O o 50 o o 100

o SHEEP SHEEP În 250m 250m Plogiosetum refroctum Portuloco olerqceo a o ! a Oo o o o o t o o o o I o o o oo a o oo o ) o o o opO ,a o o CATTLE În CAÏTLE 1_ 250m rn 25om

a o o !o a 1 O o c å 5 o o o 10 a o oO o 20 50 ¡ O o aa o o 100 o .a O a oa o SHEEP SHEEP n În 250m 250m Ptilotus potystqchyus Ptilotus obvotus o ! a o a a o o a

a o Oo'a a o

o a a o o o oo Oo. o of o I o o cATTLE |n CATTLE 1_ 250m rn 25om

o 1 å 5 o a o 10 o o 20 o 50 o o 100 a a a o a o a sHEEP fn SHEEP În 250m 250m L

Sidq corrugoto Solsolq kqli vctr. ougustifolio a o o Oo a o o o o o o O Ooo a D. o o o o o oo o o o o o o D a o o a o o a o oo. o a a oOoO o CATTLE CATTLE 250m în 250m o o I a a o o o o o 1 o o 5 ao o o o 10 aa a a o o 20 o o 50 o o o o o 100 o a o O sHEEP o fn SHEEP în 250m 250m virgoto Sido cf. Sisymbrium (DEAD MAïERTAL)

a Oo oo O O o oo o oo ofoo o a oOo I .a.O oo oo a o oO a).aa oo o o of . oo. ' oa o oofO of o CATTLE În CATTLE 1 250m rn 25om - O O. o o a ! o o o o o o a o o a a I å 5 o .O o 10 oo o o o 20 o oO a a o o 50 oQ o o a o o oo o f oo o 100 SHEEP SHEEP T 1n 250m 250m Solonum petrophilum Tribulus terristris

a o o o o o oo o. o o o o ooo o o a OoO o ö

a CATTLE o oo. o n o zSom CATTLE 1_ rn 250m

a o o a a 1 a å 5 o 10 a 20 a o o o o o a a o 50 o o o

100 SHEEP 1; o SHEEP 250m ln 250m Tripogon toliiformis Zygophyllum ourqntiocum o a o oo o a o e Oo o o a o oo Oo a o a 'oo o CATTLE 1^ CATTLE 1_ 250m r¡ 25om o o o oo o o Ooo o ,l o Oo o å 5 o o o o 10 o a o 20 oo o o oo o 50 a o o o o O o oo 100 o o o o SHEEP sHEEP fn în 250m 250m